United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park. NC 27711
EPA-453/R-98-011
May 1999
Air
& EPA
NATIONAL EMISSION STANDARDS FOR
HAZARDOUS AIR POLLUTANTS (NESHAP)
FOR PESTICIDE ACTIVE INGREDIENT
PRODUCTION:
SUMMARY OF PUBLIC COMMENTS AND
RESPONSES
-------
EPA-453/R-98-011
National Emission Standards for Hazardous Air Pollutants
(NESHAP) For Pesticide Active Ingredient Production
Background Information for Promulgated Standards--
Summary of Public Comments and Responses
Emission Standards Division
U. S. Environmental Protection Agency
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
May 1999
-------
DISCLAIMER
This report is issued by the Office of Air Quality Planning and Standards, U. S.
Environmental Protection Agency. Mention of trade names and/or commercial products is not
intended to constitute endorsement or recommendation for use. Copies of this report are
available free of charge to Federal employees, current contractors and grantees, and non-profit
organizations-as supplies permit-from the Library Services Office (MD-35), U. S.
Environmental Protection Agency, Research Triangle Park, North Carolina 277U
(919) 541-2777) or, for a nominal fee, from the National Technical Information Service,
5285 Port Royal Road, Springfield, Virginia 22161 (703) 487-4650).
n
-------
ENVIRONMENTAL PROTECTION AGENCY
National Emission Standards for Hazardous Air Pollutants for
Pesticide Active Ingredient Production--
Summary of Public Comments and Responses
1. The final National Emission Standards for Hazardous Air Pollutants (NESHAP) will
regulate emissions of hazardous air pollutants from pesticide active ingredient
production. Only those operations that are part of major sources under section 112(d) of
the Clean Air Act as amended in 1990 will be regulated.
2. For additional information contact:
Mr. Lalit C. Banker
Organic Chemicals Group (MD-13)
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
Telephone: (919) 541-5420
3. Paper copies of this document may be obtained from:
National Technical Information Service (NTIS)
5285 Port Royal Road
Springfield, VA 22161
Telephone: (703) 487-4650
U. S. EPA Library Services Office (MD-35)
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
Telephone: (919) 541-2777
4. Electronic copies of this document may be obtained from the EPA Technology Transfer
Network (TTN) over the internet by going to either of the following addresses:
http://www.epa.gov/ttn/oarpg/t3bid.html
http://www.epa.gov/ttn/uatw/pest/pestpg.html
111
-------
IV
-------
TABLE OF CONTENTS
LIST OF ACRONYMS AND ABBREVIATIONS FOR UNITS OF MEASURE xi
1.0 SUMMARY 1-1
1.1 SUMMARY OF CHANGES SINCE PROPOSAL 1-1
1.2 SUMMARY OF IMPACTS OF PROMULGATED REGULATIONS .... 1-4
2.0 OVERVIEW OF PUBLIC COMMENTS 2-1
2.1 ORGANIZATION OF COMMENT SUMMARIES 2-1
3.0 APPLICABILITY 3-1
3.1 SELECTION OF SOURCE CATEGORY 3-1
3.2 DEFINITION OF AFFECTED SOURCE 3-9
3.3 APPLICABILITY TO PRODUCTION OF INTERMEDIATES 3-18
3.4 RECOVERY DEVICES 3-20
3.5 EXEMPTIONS FROM THE RULE 3-21
3.6 NEW AFFECTED SOURCE 3-25
3.7 STARTUP, SHUTDOWN, AND MALFUNCTION 3-27
3.8 OVERLAP WITH OTHER REGULATIONS 3-29
3.9 STORAGE TANK ASSIGNMENT 3-30
4.0 COMPLIANCE DATES 4-1
4.1 EXISTING SOURCES 4-1
4.2 NEW SOURCES 4-1
5.0 DEFINITIONS 5-1
5.1 NEW TERMS DEFINED IN THE FINAL RULE 5-1
5.2 REVISED DEFINITIONS THAT DIFFER FROM COMMENTERS
SUGGESTIONS 5-2
5.3 REVISED DEFINITIONS CONSISTENT WITH COMMENTERS
SUGGESTIONS 5-9
5.4 DEFINITIONS UNCHANGED SINCE PROPOSAL 5-11
6.0 STANDARDS-GENERAL 6-1
6.1 HEAT EXCHANGERS 6-1
6.2 OUTLET CONCENTRATION STANDARDS 6-2
6.3 COMPLIANCE OPTIONS 6-4
6.4 ROUTINE MAINTENANCE 6-5
6.5 MACT AND MACT FLOOR 6-6
6.6 CLARIFICATION 6-8
-------
TABLE OF CONTENTS (continued)
7.0 STANDARDS-PROCESS VENTS 7.!
7.1 MACT AND MACT FLOOR 7-1
7.2 APPLICABILITY CUTOFFS 7.4
7.3 HCL STANDARDS 7-7
7.4 COMPLIANCE AVERAGING PERIOD 7-9
7.5 OVERLAP WITH OTHER STANDARDS 7-10
7.6 CLARIFICATIONS 7-10
7.7 REQUIREMENTS FOR SURGE CONTROL VESSELS AND
BOTTOMS RECEIVERS 7-11
8.0 STANDARDS-STORAGE TANKS 8-1
8.1 STORAGE TANK MACT FLOOR 8-1
8.2 STORAGE TANK STANDARD 8-3
8.2.1 Applicability Cutoffs 8-3
8.2.2 Control Level of the Standard 8-4
8.2.3 Format of the Standard 8-5
8.3 COMPLIANCE AVERAGING PERIOD 8-5
8.4 STORAGE TANK DEFINITION 8-6
8.5 OVERLAP WITH OTHER STANDARDS 8-7
8.6 CLARIFICATIONS 8-7
9.0 STANDARDS FOR EQUIPMENT LEAKS 9-1
9.1 LEVEL OF THE STANDARD 9-1
9.2 CROSS-REFERENCING THE CONSOLIDATED AIR RULE 9-8
10.0 STANDARDS FOR WASTEWATER 10-1
10.1 GENERAL 10-1
10.2 AFFECTED WASTEWATER 10-1
10.3 TREATMENT OPTIONS 10-5
10.4 STANDARDS FOR NEW SOURCES 10-6
10.5 OVERLAP WITH OTHER REGULATIONS 10-8
10.6 CLARIFICATION OF CROSS-REFERENCES 10-9
11.0 STANDARDS FOR BAG DUMPS AND PRODUCT DRYERS 11-1
11.1 JUSTIFICATION FOR THE STANDARDS 11-1
11.2 LEVEL OF THE STANDARD 11-3
VI
-------
TABLE OF CONTENTS (continued)
Page
12.0 POLLUTION PREVENTION STANDARD 12-1
12.1 FORMAT OF THE STANDARD 12-1
12.2 POLLUTION PREVENTION FOR REACTANTS AND
GENERATED HAP 12-4
12.3 COMPLIANCE PROCEDURES 12-6
12.4 RECORDKEEPING 12-7
12.5 BASELINE YEAR 12-7
13.0 EMISSION AVERAGING 13-1
13.1 COMPLEXITY OF METHODOLOGY 13-1
13.2 RESTRICTIONS ON EMISSIONS AVERAGING 13-1
14.0 TESTING PROVISIONS 14-1
14.1 TESTING CONDITIONS 14-1
14.2 TEST PLAN 14-6
14.3 EMISSION PROFILE 14-6
14.4 TEST DURATION 14-7
14.5 CLARIFICATIONS 14-7
15.0 INITIAL COMPLIANCE DEMONSTRATIONS 15-1
15.1 GENERAL 15-1
15.2 INITIAL COMPLIANCE WITH PROCESS VENT STANDARDS 15-6
15.2.1 Emission Estimation Calculations 15-6
15.2.2 Compliance with Outlet TOC Limit 15-10
15.2.3 Exemption from Performance Testing 15-11
15.2.4 Provisions for Flares 15-14
15.2.5 Provisions for Condensers 15-14
15.2.6 Clarifications 15-16
15.3 INITIAL COMPLIANCE WITH STORAGE TANK STANDARDS 15-17
15.4 INITIAL COMPLIANCE WITH WASTEWATER STANDARDS 15-19
15.5 PLANNED ROUTINE MAINTENANCE 15-19
vn
-------
TABLE OF CONTENTS (continued)
16.0 MONITORING 16_j
16.1 IMPACT OF MONITORING REQUIREMENTS ON CONTROL
LEVELS 16.!
16.2 ESTABLISHING PARAMETER LEVELS 16-3
16.3 MONITORING FOR SPECIFIC CONTROL DEVICES 16-7
16.3.1 Bag Dumps and Product Dryers 16-7
16.3.2 Flares 16-9
16.4 MONITORING FREQUENCY 16-10
16.5 QUALITY CONTROL PROGRAM AND PERFORMANCE
EVALUATIONS OF CONTINUOUS MONITORING SYSTEMS 16-11
16.6 MONITORING DURING STARTUP, SHUTDOWN, AND
MALFUNCTION 16-12
16.7 MONITORING FOR STORAGE VESSEL CONTROLS 16-13
16.8 DATA AVAILABILITY REQUIREMENTS 16-15
16.9 VIOLATIONS 16-15
16.10 MISCELLANEOUS 16-19
16.11 EDITORIAL CLARIFICATIONS 16-21
17.0 RECORDKEEPING 17-1
17.1 RECORDKEEPING BURDEN 17-1
17.2 RECORDS OF GROUP 1 DETERMINATIONS 17-2
17.3 RECORDS OF TURNOVERS 17-3
17.4 RECORDS OF PROCESS VENT EMISSIONS 17-3
17.5 RECORDS OF WASTEWATER CHARACTERISTICS 17-4
17.6 RECORDS OF BAG LEAK DETECTION ALARMS 17-5
17.7 LOCATION OF RECORDS 17-5
17.8 OVERLAP WITH RCRA RECORDKEEPING REQUIREMENTS 17-6
17.9 CLARIFICATIONS 17-6
18.0 REPORTING 18-1
18.1 PRECOMPUANCE REPORT 18-1
18.2 PERIODIC REPORTS 18-2
18.3 NOTIFICATION OF COMPLIANCE STATUS REPORT 18-3
18.4 STARTUP, SHUTDOWN, AND MALFUNCTION REPORTS 18-3
18.5 NOTIFICATION OF PROCESS CHANGE 18-4
18.6 EQUIPMENT LEAK REPORTS 18-5
18.7 RECORDS OF REPORTS 18-5
18.8 OVERLAP WITH RCRA REPORTING REQUIREMENTS 18-6
Vlll
-------
TABLE OF CONTENTS (continued)
Page
19.0 IMPACTS 19-1
19.1 ENVIRONMENTAL IMPACTS 19-1
19.2 COST IMPACTS 19-1
19.3 ECONOMIC IMPACTS 19-3
20.0 MISCELLANEOUS 20-1
20.1 STANDARDS FOR ENDOCRINE DISRUPTORS 20-1
20.2 RISK-BASED STANDARD FOR HC1 20-1
20.3 RELATIONSHIP BETWEEN THIS RULE AND OTHER RULES 20-2
20.4 RELATIONSHIP BETWEEN THE NOTIFICATION OF
COMPLIANCE STATUS REPORT AND TITLE V PERMITS 20-4
20.5 OMB REVIEW 20-4
20.6 EDITORIAL CLARIFICATIONS 20-4
20.7 COMMENT PERIOD 20-5
20.8 SUPPORTIVE COMMENTS 20-5
LIST OF TABLES
2-1. LIST OF COMMENTERS ON THE PROPOSED NESHAP
FOR THE PAI INDUSTRY 2-2
20-1. EDITORIAL COMMENTS AND RESPONSES 20-6
IX
-------
LIST OF ACRONYMS AND ABBREVIATIONS FOR UNITS OF MEASURE
ACRONYMS
ACPA
Act
ACT
Administrator
Agency
ALR
ANSI
APCD
BACT
BIF
CAA
CAM
CAR
CEM
CEMS
CFR
C12
CMPU
CMS
C02
CTG
CWA
EG
EPA
FBCA
FDA
FID
FIFRA
Fr
FR
GMP
H2
HAP
American Crop Protection Association
Clean Air Act
Alternative Control Techniques
EPA Administrator
EPA
average leak rate
American National Standards Institute
air pollution control device(s)
best available control technology
boilers and industrial furnaces
Clean Air Act
compliance assurance monitoring
Consolidated Air Rule
continuous emissions monitor
continuous emissions monitoring system(s)
Code of Federal Regulations
chlorine
chemical manufacturing process unit
continuous monitoring systems
carbon dioxide
Control Techniques Guideline
Clean Water Act
emission guidelines
U. S. Environmental Protection Agency
fixed-bed carbon adsorber
Food and Drug Administration
Flame lonization Detector
Federal Insecticide Fungicide and Rodenticide Act
fraction removed
Federal Register
Good Manufacturing Practice
hydrogen
hazardous air pollutant(s)
XI
-------
HC1 hydrogen chloride
HCN hydrogen cyanide
HON Hazardous Organic NESHAP
EFR internal floating roof
LDAR leak detection and repair
MACT maximum available control technology
MON Miscellaneous Organic NESHAP
N2 nitrogen
NAAQS National Ambient Air Quality Standards
NAICS North American Industrial Classification System
NESHAP national emission standards for hazardous air pollutants
NO2 nitrogen dioxide
NOCS notification of compliance status
NOX nitrogen oxides
NPDES National Pollutant Discharge Elimination System
NSM new source MACT
NSPS new source performance standards
NSR new source review
O2 oxygen
OAQPS Office of Air Quality Planning and Standards
OECA Office of Enforcement and Compliance Asurance
OMB Office of Management and Budget
OSHA Occupational Safety and Health Administration
OVA Organic Vapor Analyzer
P&R Polymers and Resins
P2 pollution prevention
PAI pesticide active ingredient
PEC purchased equipment costs
PhRMA Pharmaceutical Research and Manufacturers of America
PM paniculate matter
PMPV Pharmaceutical Manufacturing Process Unit
POD point of determination
POTW Publicly - Owned Treatment Works
PSD prevention of significant deterioration
QA/QC quality assurance/quality control
QIP Quality Improvement Program
xn
-------
R&D research and development
RACT Reasonably Available Control Technology
RCRA Resource Conservation and Recovery Act
SARA/TRI Superfund Amendments and Reauthorization Act/Toxic Release Inventory
SIC Standard Industrial Classification
SO2 sulfur dioxide
SOCMI Synthetic Organic Chemical Manufacturing Industry
SOX sulfur oxides
TOC total organic compounds
TRE total resource effectiveness
TRI toxic release inventory
TSCA Toxic Substances Control Act
USDA United States Department of Agriculture
VOC volatile organic compounds
xni
-------
ABBREVIATIONS FOR UNITS OF MEASURE
Btu
C°
dscf
dscfm
dscm
oF
ft3
gal
gr
hr
K
kg/yr
km
Kpa
L
Uyr
Ib
m3
min
mg
Mg
MMm3
MW
MW-hr/yr
ng
ppm
ppmdv
ppmv
ppmw
psia
scmm
ton/yr
Mg
jim
wk
British thermal unit
degrees Celsius
dry standard cubic foot (@ 14.7 psia, 68 °F)
dry standard cubic foot per minute (@ 14.7 psia, 68°F)
dry standard cubic meter (@ 14.7 psia, 68 °F)
degrees Fahrenheit
cubic feet
gallon
grains
hour
degrees Kelvin
kilograms per year
kilometer
kilopascals
liter
liter per year
pound
cubic meter
minute
milligrams (10~3 grams)
megagram (106 grams)
million cubic meters
megawatt
megawatt-hours per year
nanogram (10~9 grams)
parts per million
parts per million by dry volume
parts per million by volume
parts per million by weight
pounds per square inch actual
standard cubic meters per minute
ton per year
microgram (10"6 grams)
microns
week
year
XIV
-------
1.0 SUMMARY
On November 10, 1997, the U. S. Environmental Protection Agency (EPA) proposed
national emission standards for hazardous air pollutants (NESHAP) for pesticide active
ingredient (PAI) production (63 FR 60565) under authority of Section 112 of the Clean Air Act
(Act). Public comments were received from 34 sources consisting mainly of PAI manufacturers,
trade associations, and other interested parties.
All of the comments that were submitted and the responses to these comments are
summarized in this document. This summary is the basis for the revisions made to the standards
between proposal and promulgation.
1.1 SUMMARY OF CHANGES SINCE PROPOSAL
Applicability
Eliminated applicability of inorganic PAI's.
Introduced concept of "process units" and "PAI process units" analogous to the
definitions in other regulations.
Introduced concept of "process unit groups" and "primary product determinations" for
defining applicability of PAI process units that are constructed from nondedicated equipment that
are also used to produce non-PAI products.
Added provisions regarding overlap with other standards.
Process vent standards
Added 20 parts per million by volume (ppmv) option for HC1/C12 halide emissions.
Changed control requirement for HC1/C12 at new sources from 99.9 percent to 99 percent.
Added an alternative standard that requires control to 20 ppmv as TOC and 20 ppmv as
HC1/C12, and compliance is demonstrated using CEM's.
1-1
-------
Storage vessel standards
Changed MACT floor and MACT cutoff from pounds per year (Ib/yr) format to vapor
pressure cutoff.
New source cutoff changed from 1 Ib/yr and 7,000 gallons (gaJ) to 16.5 kilopascals (kPa)
and 10,000 gal, and control requirement changed from 98 to 95 percent.
Clarified that allowance for planned routine maintenance of control devices applies only
to control devices used for storage vessels.
Final rule allows owner or operator to designate storage vessels as Group 1 storage
vessels.
Added alternative standard option like the one for process vents.
Wastewater standards
Changed definition of wastewater to exclude maintenance wastewater streams with low
emission potential.
Added exception to the cross-referenced provisions in the Hazardous Organic NESHAP
(HON) to allow compliance only with outlet total organic compounds (TOC), not both outlet
TOC and total organic hazardous air pollutant(s) (HAP).
Equipment leaks
Changed requirements to be consistent with provisions in the proposed Consolidated Air
Rule (CAR) instead of subpart H; the provisions have been included in the final rule.
Pollution prevention
Changed basis for baseline factor from 1 year of data to 3 years.
Added equation to calculate required reduction in volatile organic compounds (VOC)
factor.
Added requirement to submit a demonstration summary describing methods to be used to
measure and record consumption and production.
Test methods and initial compliance
Revised requirement to correct outlet concentrations to 3 percent oxygen only when using
a combustion device and supplemental gases are combined with the vent stream. Also added
requirement to correct outlet concentrations when supplemental gases are combined with vent
streams that are controlled in non-combustion devices.
1-2
-------
Deleted option to demonstrate compliance for a condenser based on test results; final rule
requires compliance based on temperature measurement and calculation of the emissions. As a
result, also deleted peak-case conditions based on maximum heat removal capacity.
Deleted option to test under representative peak-case conditions.
Expanded definition of peak-case conditions to include most challenging conditions other
than maximum HAP load. Also expanded definition of the emissions profile to include
equipment limitations as well as actual emissions.
Added optional equations to calculate emissions from heating and depressurization.
Changed the cutoff for calculating incremental values for emissions from heating from
50 degrees Kelvin (K) below the boiling point of the material in the vessel to 10 K below the
boiling point.
Added equations for calculating emissions from gas evolution and air drying.
Added equations for calculating controlled emissions from condensers when uncontrolled
emissions are generated from one of the seven types of emission episodes for which equations
are specified in the rule
Added provisions for demonstrating initial compliance when using a flare.
Monitoring
Added monitoring parameters for catalytic incinerators.
Changed monitoring provisions for bag leak detectors by allowing adjustments after
initial settings under certain conditions and deleting the requirement to develop a Quality
Improvement Program (QIP).
Clarified that averaging periods for batch unit operations may be over a "block" period
not to exceed the duration of the operation or series of operations as well as over 24 hours.
Reporting
Changed submittal date of Precompliance report from 12 months to 6 months before the
compliance date.
Added procedures for submitting notification when the owner or operator wants to
change anything that was (or would have been) submitted for approval in the Precompliance
report.
1-3
-------
Miscellaneous
Numerous editorial changes and changes in definitions to improve readability and clarify
intent, especially in the monitoring, recordkeeping, and reporting sections.
1.2 SUMMARY OF IMPACTS OF PROMULGATED REGULATIONS
The final standards will reduce nationwide emissions of HAP from the production of
PAI's by 2,500 megagrams per year (Mg/yr) (2,800 tons per year [ton/yr]), or 65 percent
compared to baseline emissions that would result in the absence of the standards. Wastewater
generated from water scrubbers used to control hydrogen chloride (HC1) emissions is expected to
increase by an estimated 10.8 million liters per year. No increase in solid waste is expected.
Energy usage is expected to increase by an estimated 4,880 x 109 British thermal units (Btu) per
year. The total annual costs for control at existing and new sources are estimated to be
approximately $37.5 million and $5.16 million, respectively (June 1995 dollars). The economic
impact analysis shows that the market price will increase by less than 3 percent, and market
output will decrease by less than 3 percent. No plant closures are expected from compliance with
the final standaidi.
1-4
-------
2.0 OVERVIEW OF PUBLIC COMMENTS
The public comment period following the November 10, 1997 Federal Register notice
(proposed rule) lasted from November 10, 1997 to January 9, 1998. However, a number of
commenters requested that the comment period be extended, and EPA did grant an extension of
the comment period until February 9, 1998. Late comments received after February 9, 1998 were
also accepted. A total of 34 letters commenting on the proposed rule were submitted and these
comments have been placed in the docket for this rulemaking (Docket A-95-20) under
categories FV-D (received on or before February 9, 1998) and FV-G (received after February 9,
1998). Table 2-1 presents a listing of all persons submitting written comments on the proposed
rule, their affiliations, and the recorded docket item number assigned to their correspondence.
In some instances, commenters supported their comments by referencing comments
submitted by other comments. This was particularly true of the comments submitted by the
American Crop Protection Association (ACPA) which were referenced several times by their
member companies. Commenters IV-D-20, IV-D-22, FV-D-25, and IV-D-29 incorporated the
comments of the ACPA (comment letter IV-D-16) by reference.
2.1 ORGANIZATION OF COMMENT SUMMARIES
Chapters 3.0 through 20.0 present a summary of the comments on the proposed rule along
with EPA responses. The comments are grouped by subject areas, and the organization of topics
is similar to the organization of the preamble to the final rule.
2-1
-------
TABLE 2-1. LIST OF COMMENTERS ON THE PROPOSED NESHAP
FOR THE PAI INDUSTRY
Item No.
IV-D-01
IV-D-02
IV-D-03
IV-D-04
IV-D-05
W-D-06
IV-D-07
IV-D-08
IV-D-09
rv-D-io
IV-D-11
IV-D-12
IV-D-13
IV-D-14
IV-D-15
IV-D-16
IV-D-17
IV-D-18
W-D-19
IV-D-20
IV-D-21
IV-D-22
IV-D-23
IV-D-24
F/-D-25
IV-D-26
IV-D-27
W-D-28
IV-D-29
IV-G-01
IV-G-02
IV-G-03
FV-G-04
IV-G-05
IV-G-06
Commenter and Affiliation
D.Gustafson, The Dow Chemical Company
L. Hott, Novartis Crop Protection, Inc.
W. Dickerson, BASF
T. Gilding, American Crop Protection Association
A. Stungys, Tomen Agio, Inc.
A. Deshmukh, Occidental Chemical Corporation
V. Jones Clorox Company
T. Gilding, American Crop Protection Association
J. Cooper, AlliedSignal, Inc.
B. Higgins, STAPPA/ALAPCO
R. Godbole, Phelps Dodge Corp.
D. Ailor, Coke and Coal Chemicals Institute
W. Adams, Kennecott Utah Copper Corp.
R. Phelps, Eastman Chemical Company
A. Stungys, TomenAgro, Inc.
T. Gilding, American Crop Protection Association
M. Wax. Institute of Clean Air Companies
D. McKinnon, Manufacturers of Emission Controls Association
K. Parameswaran, ASARCO, Inc.
R. DiMenna Rohm and Haas, Co.
J. Dumelow, DuPont Agricultural Products
B. Raff. Novartis Crop Protection, Inc.
C. Wysong, Engelhard Corp.
R. Pauline, Chemical Specialties Manufacturers Association
C. Keffer, Monsanto Company
S. Tirey, Chemical Manufacturers Association
A. Dawson, American Cyanarmd Co.
L. Swaim, The Dow Chemical Company
J. Dege, Dupont SHE Excellence Center
D. Powers, Merck & Co., In.c
R. Smerko, The Chlorine Institute
N. Carlson, Elf Atochem
A. Deshmukh, Occidental Chemical Corporation
R. Randolph, Missouri Department of Natural Resources
C Wvsong Engelhard Corporation
2-2
-------
3.0 APPLICABILITY
3.1 SELECTION OF SOURCE CATEGORY
Status at proposal: Under the proposed rule, the source category consisted of the
production of all pesticide active ingredients that are used to produce an insecticide, herbicide, or
fungicide pesticide end-use product. The preamble to the proposed rule explained how the
source category was developed, and requested comment on the scope of the source category. The
procedure used to develop the source category is summarized in the remainder of this
introduction.
The initial list Oi categories of major and area sources included 10 source categories in
the agricultural chemicals industry group. In June 1996, butadiene furfural cotrimer was moved
from the polymers and resins industry group to the agricultural chemicals industry group (61 FR
28197). In the notice of proposed rulemaking, EPA made the following additional changes:
(1) all manufacturers of active ingredients within the meaning of FEFRA section 2(a) that are
used in herbicide, insecticide, or fungicide pesticide end-use products were added to the
agricultural chemicals industry group; (2) the individual initial and new source categories in the
agricultural chemicals industry group were combined into a single source category; and (3) the
new source category was named "pesticide active ingredient production."
Comment: Thirteen commenters addressed the issue of the source category definition.
Four commenters discussed the scope of the source category in general terms:
commenters TV-D-09 and TV-D-12 opposed the expansion, commenter TV-D-28 could not tell
from the available information whether or not it was appropriate, and commenter TV-D-20
supports the scope of the applicability and the definition of PAL Eight commenters (TV-D-07,
TV-D-11, IV-D-13, FV-D-19, rV-D-23, IV-D-24, IV-G-02. and TV-G-04) requested exemptions
for specific processes or classes of processes that would be subject to the proposed rule; the
3-1
-------
primary concern of these commenters was that the proposed rule would apply to inorganic PAJ
production processes, many of which, the commenters contend, are significantly different than
organic PAI production processes. One commenter (IV-D-15) stated that the Captan process
(one of the original source categories) should not be combined with other PAI processes because
it is significantly different from other PAI processes. Details of the comments are described in
the remainder of this section.
Commenters IV-D-09 and IV-D-12 opposed the expansion of the source category because
(1) the processes subject to the proposed rule are dissimilar in that some are from naturally
occurring materials, but others are synthetically produced and (2) many of the processes produce
compounds that are not primarily used in insecticide, herbicide, or fungicide products.
Commenter IV-D-09 is also concerned that the change has been largely undetected or
misunderstood by the regulated community because the proposal did not identify either the
number of processes that would be covered or examples of the processes. Without reviewing the
processes to ensure that process operation, emission characteristics, control device applicability,
and costs are similar, the commenters contend that the proposed regulation is arbitrary and
capricious, is inconsistent with the Act and EPA's procedures for developing maximum available
control technology (MACT) standards, and defeats the purpose for creating industry categories.
Both commenters suggested that EPA change the definition of PAI to mean any synthetically
produced material because the Basis and Purpose document only describes processes to
synthetically produce chemicals. Commenter IV-D-09 noted that this approach would be
consistent with the definition of intermediate, which is a compound produced in a chemical
reaction and that is further processed in one or more additional chemical reactions to produce a
PAI. Both commenters also indicated that previous rulemakings recognize this distinction (e.g.,
the HON, the new source performance standards (NSPS) for equipment leaks, and the Effluent
Guidelines forPesticide Chemicals all focus on processes to synthetically produce compounds;
whereas the Effluent Guidelines for Gum and Wood Chemicals target production of compounds
from natural sources). Commenter IV-D-12 also noted that the Effluent Guidelines are explicit
and unambiguous in that each process subject to the guidelines is specified.
One commenter (IV-D-28) stated that the Federal Register notice itself did not contain
enough information about the similarities and differences of the processes that would be covered;
3-2
-------
therefore, the commenter could not determine whether it is appropriate to combine the source
categories. However, this commenter stated that, for this rulemaking, EPA should not further
expand the source category beyond the processes covered by the proposed rule because owners
and operators of other processes may not have read the preamble closely enough to determine
that EPA is thinking about adding other processes. Also, this commenter contended that the
necessary data to evaluate such processes could not be provided in the comment period. This
commenter also specifically supported the exclusion of antimicrobials, rodenticides, and biocides
from the definition of PAL
Commenter IV-D-24 believes the proposed rule is too broad because it does not
distinguish between antimicrobial PAI's and other types of PAI's. The commenter requested that
EPA exempt formulators of antimicrobial PAI's for the following reasons: (1) production,
pollution prevention, and treatment practices for antimicrobial PA1 processes are different from
other PAI processes; (2) the antimicrobial PAI's have low toxicity, biodegrade rapidly, and have
undergone extensive safety testing; (3) EPA recognized the unique nature of sodium hypochlorite
and similar chemicals by excluding them from the Effluent Guidelines for Formulating,
Packaging, and Repackaging Facilities in the Pesticide Chemical Category; and (4) formulators
of antimicrobial products have no air emissions.
Commenter FV-G-02 requested that EPA exempt chlorine and sodium hypochlorite
production and packaging plants from the rule because chlorine is already highly regulated. The
commenter cited Occupational Safety and Health Administration (OSHA) regulations, Superfund
Amendments and Reauthorization Act/Toxic Release Inventory (SARA/TRI) reporting
requirements, EPA's pending risk management program requirements for facilities with more
than 2,500 pounds (Ib) of chlorine onsite, and the pending MACT standards for chlor-alkali
production. Another commenter (F/-D-07) also requested that EPA exempt sodium hypochlorite
production because the product: (1) does not present an unreasonable risk to the environment;
(2) is not emitted to the air in the manufacturing process; (3) ionizes, dissolves rapidly in water,
rapidly degrades, and does not volatilize; (4) the production, pollution prevention, and treatment
practices differ from these practices for other PAI's, which EPA recognized by excluding it from
the Effluent Guidelines for the Pesticide Chemicals Category.
3-3
-------
Commenter IV-D-23 requested that EPA exempt kaolin (aluminum silicate) production
because: (1) the process, which consists primarily of mechanical action and treatment with
inorganic chemicals, is very different from the flow diagram in the Basis and Purpose document;
(2) the commenter's production facility is not a major source of HAP emissions; (3) kaolin is not
a typical pesticide in that it repels pests rather than killing them; and (4) kaolin is nontoxic (i.e., it
has an unlimited tolerance level).
Commenters IV-D-11, IV-D-13, and IV-D-19 believe sulfuric acid production should be
exempt from the rule. All of the commenters operate copper smelters and produce sulfuric acid
from the sulfur dioxide in the smelter offgas. These processes would be subject to the proposed
rule because the sulfuric acid is registered as a herbicide, but only small amounts of it are sold for
this use. (The commenters noted that other sulfuric acid producers also would be subject to the
proposed rule, although their sulfur dioxide feedgas is produced by burning sulfur.)
Commenter IV-D-13 explained that the smelter offgas contains various metals and metal
compounds, some of which are HAP. These HAP are removed by paniculate controls prior to
the catalytic oxidation of the sulfur dioxide to sulfur trioxide. The commenter cited EPA's final
summary report for the primary copper smelters NESHAP, which stated that the participate
controls result in HAP metal control efficiencies of greater than 99.9 percent, and, in most cases,
the combined emissions of all HAP metals in acid plant tail gases are less than 0.1 ton/yr.
Commenter IV-D-13 also cited the following reasons to support an exemption: (1) the process
does not use or generate organic HAP or HC1, which are the major HAP according to the
preamble to the proposed rule; (2) sulfuric acid plants are an effective control for metal HAP
emissions from smelters; (3) regulating the process would provide no environmental benefit but
would impose a significant burden (e.g., to demonstrate that equipment does not emit HAP);
(4) EPA lacks authority under section 112(c) of the Act to apply NESHAP to sulfuric acid plants
because the plants are not major sources and the de mini mis emissions would not present a threat
of adverse effects warranting regulation of area sources. < 5) most sulfuric acid manufacturers do
not think of the process as a PAI process, which may lead to noncompliance due to ignorance;
and (6) the production process, emissions, and controls differ from the original 11 processes.
Commenter IV-D-19 provided similar reasons: (1) the preamble to the proposed rule indicates
the rule is designed for organics and HC1, but no mention is made of other commodity
3-4
-------
inorganics; (2) no information in the docket supports a PAINESHAP for sulfunc acid plants;
and (3) the final NESHAP for copper and lead smelter furnaces will indicate that sulfuric acid
plants are MACT. Furthermore, both commenter IV-D-13 and commenter IV-D-19 indicated
that sulfuric acid plants are not included on the list of source categories, and commenter IV-D-13
believes that if it were listed at all, it logically would be listed among the categories of inorganic
chemicals, not with agricultural chemicals. Commenter IV-D-11 concurred with the comments
submitted by commenter IV-D-13.
Commenter IV-D-11 requested that copper sulfate production be exempted from the rule.
The commenter explained that copper sulfate is a byproduct from liquors used at copper
refineries and rod mills, and it is registered as a herbicide. To support the request for an
exemption, the commenter used arguments similar to those used to support the request for an
exemption of sulfunc acid production: (1) the process is not a major source of HAP (but it may
be co-located at a site that is a major source); (2) the process emits no more than trace amounts of
organic HAP and HC1, the major HAP emitted from the source category according to the
preamble to the propose i ruie; (3) the rule would not achieve an environmental benefit but would
impose a significant compliance burden on owners and operators to show they are below cutoffs;
(4) owners and operators do not think of their processes primarily as PAI processes, which is
likely to lead to noncompliance due to ignorance; and (5) this process and organic PAI processes
do not have similar process operations, emission characteristics, control device applicability and
cost, and opportunities for pollution prevention. As an alternative to adding an exemption for
this process, the commenter suggested that EPA could limit the applicability of the rule to the
manufacturing of organic PAI's.
Commenter IV-D-15 believes the Captan process (one of the initial source categories)
should not be combined with other PAI processes because it has properties and process
requirements that make it different from other PAI processes. The commenter contends that the
impacts of combining the processes in one source category were not considered when drafting
the proposed rule, and the commenter suggested that EPA study the impact of grouping these
processes. According to the commenter, some of the differences are: (1) the process vent flow
rate for production of the intermediate is much lower than the process vent flow rate for the
active ingredient production, which leads to differences in the complexity and cost of the control
3-5
-------
devices; (2) the Captan process has both volatile organic HAP and participate HAP emissions;
(3) the cost to control carbon disulfide emissions from the active ingredient production would be
much higher than the modeled costs (see section 19.2 for additional discussion of the cost
impacts).
Commenter IV-G-04 stated that EPA should exempt chromic acid and sodium
bichromate processes from the rule because these processes were part of the Chrome Chemicals
source category, which EPA delisted.
Commenter IV-G-05 supports the addition of PAI's such as organochlorides and
organophosphates because these compounds are bioaccumulative, persistent, and acutely toxic.
Response: In response to the comments, EPA reexamincd the scope of the source
category and determined that the proposed rule included some processes that are not similar to
the others. For the final rule, changes were made to narrow the scope of the source category; in
addition, for processes that remain in the source category, changes have been made to exempt
some processes and to clarify requirements for others. These changes are: (1) a statement has
been added to specify that the provisions of the rule apply only to PAI process units that
"process, use, or produce HAP;" (2) the definition of PAI has been changed to mean any organic
material that is an active ingredient within the meaning of FIFRA section 2(a); and (3) a
statement has been added to specify that the rule does not apply to the production of ethylene
(processes subject to the HON are also exempted, as they were in the proposed rule).. Finally,
EPA decided not to limit the source category only to production of compounds by chemical
synthesis. Each of these decisions is discussed in more detail later in this response. However,
this discussion first summarizes EPA's rationale for expanding the source category list to include
PAI's other than those on the initial source category list, and for aggregating them all together in
a single source category; this rationale was presented in the preamble to the proposed rule.
From a survey of a subset of the industry, EPA determined that: (1) production of
compounds on the initial source category list as well as many other compounds have a number c •:
similarities, including process equipment, emission characteristics, control applicability, and
control costs; (2) many of the additional compounds also are produced at facilities that are major
sources; and (3) the initial compounds and the additional compounds are PAI's that are used in
herbicides, insecticides, and fungicides. Because common control techniques can be applied,
3-6
-------
EPA concluded that developing separate standards for each PAI is not warranted and that it is
technically feasible to regulate emissions from a variety of PAI processes by a single set of
emission standards. Variability in the quantity of HAP produced by different processes was
addressed by incorporating applicability thresholds in the standards (section VI.D describes
changes that have been made to the thresholds for storage vessels in the final rule). At the
facilities that produce more than one PAI, often in the same equipment or same "pool" of
equipment, compliance would be facilitated by having only a single set of emission standards.
Finally, it would be more efficient and less costly for EPA to develop only a single set of
emission standards, and compliance and enforcement activities by regulatory authorities would
be more efficient and less costly.
The provision that the rule applies only to PAI process units that "process, use, or
produce HAP" has been added to the final rule because EPA did not intend for owners and
operators to demonstrate compliance for processes that do not meet this condition. Note,
however, that this provision does not automatically exempt process units that do not emit HAP;
for emission points in such process units, an owner or operator must demonstrate that emissions
are less than the applicability thresholds.
The EPA decided to exclude production of inorganic compounds from the source
category because (1) inorganic PAI's comprise only a small percentage of the total PAI
production, (2) many of the inorganic PAI production processes do not use or emit HAP, (3) data
are unavailable on the use, emissions, and control of HAP compounds other than organics and
HC1, (4) some of the inorganic PAI's are included in other active or delisted source categories,
and (5) most of the inorganic PAI's are used primarily for non pesticidal purposes. In this
context, "organic" means any compound that contains carbon and hydrogen with or without other
elements. Based on a review of pesticide registration data in 1996, less than 10 percent of the
PAI's in pesticide products that are registered as insecticides, herbicides, or fungicides are
inorganic compounds. Inorganic compounds comprise a similar percentage by weight; based on
1993 consumption data, the top 25 compounds account for nearly half of the total PAI
production, and the two inorganic compounds in the group (sulfur and copper hydroxide) account
for less than 10 percent of the total.
3-7
-------
Of the inorganic PAI processes, only those producing HC1, chlorine, and compounds
containing arsenic and chromium are known to use and emit HAP. Both HC1 and chlorine
production processes have been specifically exempted from the final rule because they are the
subject of other MACT standards that are under development. Chromium-based compounds are
part of the delisted chrome chemicals source category and thus, EPA agrees with the commenter
that they should not also be part of the PAI source category. Data on the existing control levels
for arsenic-based compounds are unavailable. In the absence of such data, EPA has decided that
production of such compounds should not be part of the PAI source category, but they may be
investigated at a later date.
The commenters cited examples of some inorganic compounds that are primarily used for
nonpesticidal purposes. The EPA believes there are other inorganic compounds that could be
added to this list of compounds used only in minor amounts as pesticides. Conversely, most of
the organic compounds are specifically designed as PAI's. Exceptions include ethylene, which
has been specifically exempted in the final rule because it is the subject of a MACT standard that
is under development, and several compounds covered by the HON such as acrolein, ethylene
oxide, napthalene, and propylene glycol.
Production of organic PAI compounds that are derived from natural materials is retained
in the source category. Natural materials used as PAI's fall into one of two categories. One
category includes herbs, tobacco dust, dried blood, chitin, putrescent whole egg solids, pyrethrum
flowers, cinnamon, sawdust, and ground sesame plant. These compounds are simply harvested
or collected and the only processing involves mechanical action. None of these compounds is a
HAP. As a result, these processes are not subject to the final rule because the production
processes do not process, use, or produce HAP. The second category includes compounds like
turpentine that are extracted from natural materials. Extraction processes are not exempted from
the final rule because they tend to use large amounts of solvent and have a high potential for
emissions. Emissions from extraction processes tend to be more concentrated than emissions
from many of the operations in chemical synthesis processes, and they tend to be larger scale
operations than extraction operations that are part of a chemical synthesis process. These
characteristics make control of extraction processes more cost effective than control of many
chemical synthesis processes. However, because the final rule includes a primary use criterion
3-8
-------
for determining applicability (see the response to comment 2 in section 3.2), extraction processes
are only subject to the final rule if the product is primarily used as a PAI.
The EPA disagrees with the assertion by one commenter that the Captan® process (and
the associated intermediate process) should be considered separately from other PAI processes.
The intermediate appears to be an integral intermediate and thus, would be subject to the rule as a
separate process. Although the flow rates of the intermediate and Captan® process vent streams
differ, the flow rates and other process vent stream characteristics for both processes are well
within the range of characteristics for process vent streams at other surveyed PAI facilities.
These differences were accounted for in EPA's impact analysis by using different models to
represent the two processes.
In addition, the paniculate emissions from product dryerss also are considered to be a
separate type of emission point like process vents or storage vessels. The fact that this plant is
the only one of the MACT floor facilities to have HAP emissions from product dryers is not
considered a significantly unique characteristic. It is analogous to the fact that some of the other
plants have HAP storage tank emissions or wastewater discharges and are subject to the specific
standards for these emission points, where other plants are not.
Finally, EPA believes the cost analysis is correct. Carbon disulfide can be controlled with
many of the same control devices that are used to control other organic HAP. If incinerated, the
resulting sulfur oxides (SO2) emissions can be controlled using scrubbers comparable to those
used to control HC1 emissions. A detailed discussion of the cost analysis is presented in section
19.2. Therefore, EPA believes the Captan process is not sufficiently different from other PAI
processes to warrant development of a subcategory or a separate source category.
3.2 DEFINITION OF AFFECTED SOURCE
Status at proposal: Section 63.1360(a) of the proposed rule defines the affected source as
the facility-wide collection of process vents, storage tanks, waste management units, heat
exchange systems, cooling towers, equipment identified in § 63.149, and equipment components
(pumps, compressors, agitators, pressure release devices, sampling connection systems, open-
ended valves or lines, valves, connectors, and instrumentation systems) in PAI manufacturing
operations at a major source of HAP emissions. The EPA received several comments on the
affected source. The comments focused on the following issues: (1) limiting applicability to
3-9
-------
processes where the primary product is a PAI, (2) limiting applicability to processes where the
product is primarily used as a PAI, and (3) clarifying the definition of affected source.
Clarification of the definition of the affected source is addressed in responses to comments in this
section; changes to the definitions of terms used in the definition of the affected source are
discussed in chapter 5.
Comment 1: Commenter IV-G-03 requested clarification of the definition of the affected
source to ensure that only those processes and equipment associated with the processes that
produce a PAI are affected. The commenter believes the clarification is necessary because the
phrase "facility-wide" in the definition could be interpreted to mean that all emission points,
whether or not they are associated with a PAI process, are subject to the rule.
Response: Although the definition of the affected source in the proposed rule was
described in a very long sentence, the EPA believes that it clearly indicated that "facility-wide"
applied specifically to process equipment that are used to produce PAI's. However, to improve
clarity and reduce redundancy, the affected source in the final rule is defined to be the "PAI
process units" that process, use, or produce HAP, and that are located at a plant site that is a
major source. The final rule also indicates that an affected source includes waste management
units, heat exchange systems, and cooling towers that are associated with the PAI process units.
As noted in the response to comment 4 in this section, most of the language in the proposed
definition of the affected source has been shifted to the definition of the new term "PAI process
unit" in the final rule.
Comment 2: Two commenters (IV-D-14 and IV-D-28) urged EPA to specify, as in other
MACT standards, that a process or process unit is subject to the rule only if its primary product is
a PAI. Commenter FV-D-14 defined primary product as the one with the greatest annual design
capacity on a mass basis. According to this commenter, a primary product determination is
needed because much process equipment that is only occasionally used for PAI production could
be subject to the proposed rule, and the exemption for equipment operating in HAP service for
less than 300 hours per year (hr/yr) is likely too low to exempt many processes. Commenter
IV-D-28 stated that in cases where a single process unit makes more than one product, the
concept of "primary product" helps the manufacturer determine which rules apply. This
commenter cited three situations where a primary product determination is important. The first is
3-10
-------
in contract manufacturing, where the process unit is reconfigured as needed to switch from one
product to another (the frequency of reconfiguration may be variable-months, a year, or even
longer). If a PAI is produced for only a short time during the year, the commenter believes it
would not be appropriate for the process unit to be subject to the PAI rule. The commenter also
noted that a simple way to define applicability is to specify that if a process unit stops making a
PAI, the PAI rule no longer applies. The second situation involves processes that make two or
more products at the same time (one is the main reason the process was built; the others are
incidental, beneficial by-products). By identifying the primary product, the process unit would
be regulated together with other process units that produce that product under a specific,
appropriate MACT standard. The third situation is when a facility makes a change in the process
unit that is intended to be permanent. The commenter could not find any provision in the
proposed rule that would allow such a process unit to be exempt from the rule if they stop
making a PAI.
Another commenter interpreted the proposed rule to mean that the rule would apply
whenever a PAI is produced. If a facility uses non-dedicated equipment, the commenter realized
that this could mean that other rules would apply when the equipment was reconfigured to
produce a different product (e.g., the proposed Pharmaceuticals rule used the same language).
The commenter believes that complying with two standards for the same equipment would be
confusing. Therefore, the commenter suggested that the PAI rule apply only when 50 percent or
more of the annual production from the equipment is a PAI, or that EPA allow a facility to
comply only with the most stringent rule that would apply to the equipment, regardless of the
configuration or the product being produced.
Response: In response to the comments EPA evaluated several options for including a
primary product determination. The analysis considered two types of situations. The first
situation consists of processing equipment that produces only one PAI, produces different PAI's
at different times, or simultaneously produces coproducts (one of which is a PAI). In each of
these, PAI is produced to some degree with each operation of the equipment. The second
situation involves processing equipment that produces different products periodically, and some
of the products are not PAI's.
3-11
-------
For the first situation EPA determined that a primary product determination is not needed.
This conclusion is obvious for equipment that only produces PAI's because no other rule could
apply, provided that compounds subject to the HON are exempted from this regulation. The
analysis is more complicated if a PAI is produced as a byproduct or is produced in minor
quantities relative to some other product of the process. The EPA is not aware of any such
situations. However, if such processes exist, they may already be subject to the HON, in which
case they are exempted under § 63.1360(d) of the final rule. The only other standard that might
apply to such a process in the future is the Miscellaneous Organic NESHAP (MON). The MON
will cover a wide variety of compounds in many different industries. Thus, EPA believes that a
process unit producing a PAI, even if the PAI is not the primary product, has more in common
with other PAI process units than with process units that will be subject to the MON. Therefore,
EPA also believes it is more appropriate to regulate all such process units under the PAI rule
rather than the MON.
The EPA considered four options for defining the applicability of the rule to equipment
periodically used to produce chemicals other than PAI's. The first option is no change from
proposal (i.e., no primary product determination). The second option is to include all equipment
used to produce different products in a "process unit group," and always comply with the
regulation that applies to the primary product for the group, regardless of what product is being
produced. The third option is to define the applicability of the rule based on the primary product
of the process unit. The fourth option is similar to the second option, except that the applicable
rule for the process unit could, under certain circumstances, be a rule other than the one for the
primary product of the group.
Under option 1, there is no primary product determination, and the standards apply to
each PAI process unit. Equipment used to produce multiple products is part of a PAI process
unit whenever it is producing a PAI. This option was rejected because, as the commenters noted,
it has the undesirable effect of requiring an owner or operator to comply with a different
regulation each time the feedstock changes or the equipment is reconfigured to make a different
type of product.
The second option is to lump all nondedicated equipment into one or more "process unit
groups" and require the owner or operator to comply with the rule that applies to the primary
3-12
-------
product within the group. A variation on this option would be to require compliance at all times
with the most stringent regulation that would apply to any of the individual process units within
the group. This option was rejected because the promulgated Pharmaceuticals standard does not
include a provision that would allow an owner or operator to elect to comply with the final PAI
rule when a pharmaceutical is produced in a process unit group that has a PAI for the primary
product. The variation also was rejected because it would be difficult to implement; the most
stringent regulation would vary depending on the mix of different types of emission points at a
given facility and could require mixing and matching different requirements from different rules
that apply to the various emission points.
The third option would specify that the rule apply only if the primary product of the
process unit is a PAI. This option was rejected because it does not solve the problem of
equipment being subject to multiple regulations. A process unit is defined only by the product it
makes. If the raw materials are changed or the equipment is reconfigured to make a different
product, the result is a different process unit. An exemption for a process unit when it no longer
produces a PAI would be meaningless because, by definition, a change in product creates a
different process unit. In other words, it is not possible to make a permanent change in the
primary product of a process unit because the change creates a different process unit. The
process unit is based on the product being produced, not a specific collection of equipment.
The fourth option, like the second option, includes the concept of process unit groups.
Under this option, the owner or operator may elect to comply with another existing MACT
standard for any PAI process unit(s) if the primary product of the process unit group is subject to
the other standard on the promulgation date of the PAI rule or the date of startup of the process
unit group, whichever is later. Thus, PAI process units within a group, even if the PAI is not the
primary product for the group, are subject to this standard unless and until the process unit group
is subject to another MACT standard that covers the primary product of the group. This option
also allows the owner or operator to elect to comply with the Pharmaceuticals standard for any
PAI process unit(s) if any of the products produced in the process unit group are subject to the
Pharmaceuticals standard. Thus, pharmaceutical manufacturing process units within a group that
are covered by the Pharmaceuticals MACT may comply with those standards even if a PAI is the
primary product of the group. This provision is included because the pharmaceuticals rule does
3-13
-------
not have a provision that would allow an owner or operator to comply with the PAI rule while
producing a pharmaceutical product when the primary product of the group is a PAI. However,
two provisions in the Pharmaceuticals rule are not applicable when producing a PAI. First, the
process vent emission limit of 0.15 Mg/yr in the PAI rule applies instead of the 2,000 Ib/yr limit
in the pharmaceuticals rule because the 2,000 Ib/yr cutoff would not be consistent with the
MACT floor for PAI process vents. Second, the owner or operator of a new source that will
produce PATs as well as pharmaceuticals must comply with all of the requirements regarding
application for approval of construction or reconstruction in § 63.5 of the General Provisions; the
exclusions in § 63.1259(a)(5) of the pharmaceuticals rule do not apply. Again, EPA believes this
change is necessary to avoid disparate treatment of PAI producers.
Under option 4, the primary product of a group is defined as the product (e.g., a PAI,
pharmaceutical, HON chemical, or currently unregulated chemical) with the highest estimated
operating time or total production rate for the 5 years after the compliance date for the PAI rule
or after startup of the process unit group, whichever is later. The owner or operator proposes the
number of groups and the boundaries of each group based on site-specific operation, but a group
may only include equipment that is or may be used with equipment that is used to produce a PAI
(i.e., some equipment must overlap between the PAI process unit and some other process unit for
all equipment in both process units to be part of the same group). This option was selected
because it simplifies compliance by allowing an owner or operator to comply with only one
regulation for a process unit group. It accomplishes this goal without sacrificing emission
reductions because the requirements of the rules are similar. It also does not require that an
existing regulation be amended.
Clarification of the 300 hour exemption that one of the commenter's cited is also needed.
This exemption does not apply to process units; it applies to pumps, valves, and other
"equipment" that are subject to the equipment leak provisions in subpart H. This provision has
been clarified in the final rule.
Comment 3: Cornmenter IV-D-09 found the definition of the affected source to be
confusing ana .T>biguous regarding substances that are only occasionally produced for PAI use
or are not nom ;sed as PAI's. For example, many chemicals that are registered active
ingredients alst many other uses. According to the commenter, the rule should clarify that
3-14
-------
production of a chemical exclusively for nonpesticidal purposes is not subject to the rule. The
commenter also noted that for some chemicals that are registered PAI's only a small percentage
of the total quantity produced is sold for use as a PAL Therefore, the commenter stated that the
rule should be clarified to restrict applicability to production of chemicals that are primarily
intended to be used as PAI's. Another commenter (IV-D-12) also believes the rule should not
apply to processes producing compounds that are not primarily used as PAI's.
Response: Since proposal, EPA has evaluated four options for determining applicability
of process units that produce a product for use both as a PAI and other purposes. Option 1 is to
require no primary use determination (i.e., no change from proposal). Option 2 is to list in the
rule compounds that are registered as PAI's but that would not be subject to the rule based on
determinations that their primary use nationwide is not as a PAI. Option 3 is to require site-
specific determinations of primary use. Option 4 is to list in the rule all PAI's that are subject to
the rule.
Option 1 would encompass the most process units and would therefore achieve the
greatest environmental benefit. The EPA rejected this option, however, because it could result in
inequitable regulatory treatment of a given type of process unit. For example, one facility might
produce a compound for multiple purposes, including a small amount for use as a PAI, but other
facilities produce the same compound exclusively for other purposes. Under this option, only the
facility producing a small amount of the compound for use as a PAI would be subject to the rule
even though otherwise identical to the other facility.
Under option 4, a list of PAI's subject to the regulation would be included in the
regulation. Compounds for which the primary use is the collective non-PAI purposes would be
excluded from the list. This option was rejected because it would not accommodate changes in
the industry. This is a dynamic industry with new compounds being developed and registered as
PAI's every year. Between 1984 and 1995, the industry added an average of 14 new compounds
per year (although not all of these new compounds would meet the definition of organic PAI
subject to regulation under this rule). As a result, updating the list every year would be
impractical. Another disadvantage to this option is that EPA's pesticide reregistration process is
not yet complete. Presumably compounds with incomplete evaluations would be included on the
3-15
-------
list. The list then would have to be amended periodically to delete compounds whose
registrations are canceled.
Option 2 was rejected because, like option 4, it would not automatically accommodate
changes in the industry; the rule might have to be amended periodically to exempt new
compounds that are primarily used for non-PAI purposes. Another concern with option 2 is that
it would be difficult to ensure that the list is accurate and complete.
The final rule adopts option 3, which requires site-specific determinations of primary use.
This option was selected for several reasons. First, this approach is likely to result in a given
process being subject to only one, most appropriate regulation because EPA is not aware of any
compounds for which the primary use is as a PAI for one facility but not others. Furthermore,
EPA does not expect the primary use at a given facility to vary. However, if the primary use
changes to non-PAI purposes, the PAI final rule will still apply to the process unit (based on
EPA's "once-in, always-in" policy); if the primary use changes to a PAI, the PAI final rule will
apply only if the process unit is not already subject to the HON. A second advantage of this
option is that it automatically accommodates new compounds that are developed in the future,
and existing compounds that are found to have a pesticidal application. A third advantage is that
minimal additional recordkeeping and reporting is required. Manufacturers are required under
FIFRA to record and report the annual production of each PAI that they produce; today's final
rule requires that they also record and report the total production to demonstrate that the
compound is produced primarily for non-PAI purposes. Finally, the pharmaceuticals rule
provides a recent precedent for including a primary use provision.
The final rule incorporates the primary use concept in the definition of PAI process unit.
Specifically, a process unit is considered to be a PAI process unit if more than 50 percent of the
material produced is used as a PAI or integral intermediate. The primary use is determined based
on the projected annual production from the process unit in the three years after the effective date
of this rule or startup, whichever is later.
Comment 4: Commenter IV-D-28 believes the definition of affected source needs to be
revised to include not only the emission points, but also the process unit and emission control
technologies. The commenter recognizes that the definition in the proposed rule is similar to the
definitions in other MACT standards, but the commenter has recently realized that it is too
3-16
-------
narrow. For example, in determining whether changes constitute "reconstruction," the changes
must cost more than half as much as building a new similar affected source. However, under the
proposed rule, the affected source includes only process vents, not the reactors, distillation units,
or other process equipment of which the vent is a part. Similarly, it includes valves and
connectors on process piping, but not the piping itself. The commenter also contended that the
cost of installing emission controls is a legitimate part of the cost of building a new affected
source, but to consider that cost in the reconstruction analysis, emission control technologies
must be included in the definition of the affected source.
Response: Reconstruction, as defined in the General Provisions, includes: "the
replacement of components of an affected .. . source to such an extent that... the fixed capital
cost of the new components exceeds 50 percent of the fixed capital cost that would be required to
construct a comparable new source ..."
The EPA agrees with the commenter that the equipment and piping within a process are
components of an affected source that should be considered in the fixed capital cost analysis for
determining whether changes constitute reconstruction. For this and other reasons, the final rule
includes the term "PAI process unit," which is defined as the process, any associated storage
tanks, equipment identified in § 63.149, piping and ducts, and components such as pumps,
compressors, agitators, pressure relief devices, sampling connection systems, open-ended valves
or lines, valves, connectors, and instrumentation systems that are used in the production of a PAI
or integral intermediate.
The EPA disagrees with the commenter's assertion that control devices should be a
component of an affected source for the purposes of determining reconstruction costs. The
preamble to the General Provisions cites EPA's policy on this issue, which was originally stated
in the preamble to a December 16,1975 regulation that deals with modification, notification, and
reconstruction requirements under 40 Code of Federal Regulations (CFR) part 60. That
preamble states: "Costs associated with the purchase and installation of air pollution control
equipment (e.g., baghouses, electrostatic precipitators, scrubbers, etc.) are not considered in
estimating the fixed capital cost of a comparable entirely new facility unless that control
equipment is required as part of the process (e.g., product recovery)."
3-17
-------
Comment 5: Commenter IV-D-28 supports the statement on page 60571 of the preamble
that the rule only applies to facilities that use, produce, or emit HAP.
Response: The EPA does not know where the commenter found this statement; it is not
on page 60571, nor is it anywhere else in the preamble to the proposed rule. However, EPA
agrees with the commenter that the rule should not apply to facilities, or process units within
facilities, that do not use or produce HAP. Process units that use or produce HAP but do not
emit it are not automatically excluded because EPA believes a facility must demonstrate that the
HAP is not emitted or that the emissions are below the appropriate applicability cutoffs.
Therefore, the applicability section of the final rule has been revised to specify that the rule
applies to PAI process units that process, use, or produce HAP.
Comment 6: Commenter IV-D-24 expressed concern that the proposed 25,000 pound
production threshold could lead to unequal treatment of companies with one large facility as
opposed to several small facilities.
Response: The Clean Air Act requires that EPA develop standards for major sources
(i.e., stationary source located within a contiguous area and under common control that emits or
has the potential to emit considering controls, in the aggregate, 10 tons/yr or more of any HAP or
25 tons/yr or more of any combination of HAP). The EPA interprets this definition to mean that
all emissions from the plant site are to be included when determining whether a source is a major
source. Small facilities (i.e., area sources) are to be regulated only if the emissions from those
facilities are determined to pose a threat of adverse effects to human health or the environment or
if otherwise listed under section 112(c).
3.3 APPLICABILITY TO PRODUCTION OF INTERMEDIATES
Status at proposal: Under § 63.1360(a) of the proposed rule, PAI manufacturing
operations would include manufacturing of any intermediate that is integral to a PAI production
process and for which more than 50 percent of the annual production of the intermediate is used
in the onsite production of PAI's. An integral intermediate process was defined as a process
manufacturing an intermediate that is used in the onsite production of PAI's and is not removed
to storage before being used to produce the PAI(s). An intermediate was defined as a compound
produced in a chemical reaction that is further processed or modified in one or more additional
chemical reactions to produce a PAI. Section 63.1360(h) of the proposed rule also would allow
3-18
-------
an owner or operator to elect to include other intermediate processes in the affected source. The
other intermediates included (1) integral intermediates for which less than 50 percent of the
intermediate is used in the onsite production of PAI's, and (2) isolated intermediates (i.e.,
intermediates that are removed to storage before being used in the onsite production of PAI's).
Comment: Two commenters (IV-D-16 and IV-D-25) addressed the issue of including the
production of intermediates in the affected source. Commenter IV-D-25 supported the provision
that would allow an owner or operator the option of including production of isolated
intermediates under the PAI rule because the commenter produces intermediates that would be
subject to the Miscellaneous Organic NESHAP, but with this option the commenter may avoid
the complexity of complying with two NESHAP. Commenter IV-D-16 recommended
clarifications to §§ 63.1360(a) and (h) of the proposed rule. To improve the clarity of
§ 63.1360(a), the commenter recommended replacing the phrase "that is integral to a PAI
production process" with "that meets the definition of integral intermediate process." The
commenter also recommended replacing the word "and" in § 63.1360(h) with "or" to clarify that
either or both of the two listed types of intermediates manufacturing operations may be included;
according to the commenter, the language in the proposed rule could be interpreted to mean that
the option may be used only if both types are included. Another commenter (IV-D-28) stated
that the term "isolated intermediate" should not be used because it has a different meaning under
Toxic Substances Control Act (TSCA), and different definitions for a single term will cause
confusion. Another commenter (IV-D-15) stated that the rule needs to include a definition for
"storage" to clarify which intermediate processes are integral.
Response: The intent of the proposed rule was to consider each integral intermediate
process to be a separate process within the affected source, and to allow the owner or operator to
elect to include any other intermediate process in the affected source. To improve the clarity of
these provisions, EPA made several changes in the final rule. The first change was to include the
production of integral intermediates in the definition of the new term "PAI process unit," as
described in section 3.2 of this chapter. This change clan ties that production of each integral
intermediate is a separate process unit. The second change was to delete the term "isolated
intermediate" to eliminate possible confusion with the term as it is defined under TSCA. The
impact of this change was minimal because the term was only used in the proposed rule to
3-19
-------
describe intermediates that are not integral intermediates. The third change was to replace the
term "integral intermediate process" with the term "integral intermediate" and change the
definition to mean an intermediate for which 50 percent or more of the annual production is used
in the on-site production of one or more PAI's and is not stored before being used in the
production of another integral intermediate or the PAI(s). For the purposes of this definition, an
intermediate is stored if it is discharged to a storage vessel and at least one of the following
conditions is met: (1) the processing equipment that discharges to the vessel is shutdown before
the processing equipment that withdraws from the vessel is started up; (2) on average, the
material is stored in the vessel for at least 30 days before being used to make a PAI; or (3) the
processing equipment that discharges to the vessel is located in a separate building or processing
area of the plant than the processing equipment that uses material from the vessel as a feedstock,
and control equipment is not shared by the two processing areas. Processes that satisfy any of
these conditions are considered to be significantly distinct and separate. The fourth change was
to clarify the provisions allowing the owner or operator to elect to include any intermediate
process in the affected source. The final rule specifies that an owner or operator may elect to
designate production of any intermediate that does not meet the definition of integral
intermediate (and is not otherwise exempted) as a PAI process unit in the affected source. See
chapter 7 for a discussion of integral intermediates in the development of the MACT floor.
3.4 RECOVERY DEVICES
Comment: Commenter IV-G-03 requested that EPA clarify the applicability of
equipment that is used for multiple processes when the recovered material from a PAI process is
used in a non-PAI process. The commenter noted that § 63.1360(f)(5) provides some guidance
for storage tanks, but this guidance needs to be expanded to cover other pieces of equipment.
Response: The term recovery device in the proposed rule had the same meaning as in the
HON, but it should have been used only in conjunction with the wastewater provisions. The
MACT floor for process vents is based on the concept that certain condensers are part of the
process (i.e., process condensers) and any other add-on devices are considered to be control
devices; the concept of recovery devices as in the HON does not apply to process vents. For the
final rule, the term recovery device has been revised to include only devices used with water
streams, and to specify that equipment based on gravity separation may be a recovery device only
3-20
-------
if all of the inlet streams are two-phase liquid streams. The material recovered in a recovery
device may be used in any process, including non-PAI processes.
3.5 EXEMPTIONS FROM THE RULE
Status at proposal: The proposed rule exempted research and development facilities,
emission points subject to the HON, emission points subject to other MACT standards, a variety
of water discharges, and equipment (i.e., flanges, valves, etc.) in organic HAP service for less
than 300 hr/yr.
Comment 1: Two commenters requested additional exemptions in § 63.1360(d) for
emission streams with certain characteristics. Commenter IV-D-27 suggested the following
should be excluded from the rule: processes where the sum of uncontrolled organic HAP
emissions is < 10,000 Ib/yr; process vents with concentrations <50 ppmv total organic HAP
concentration or <0.005 standard cubic meters per minute (scmm) flow; and storage tanks with
uncontrolled organic HAP emissions <500 Ib/yr. Commenter IV-G-03 requested an overall
exemption for PAJ processes where the total organic HAP emissions are less than 500 Ib/yr.
Commenter FV-G-03 suted that the justification for the exemption is that it meets the intent of
the CAA in that (1) it is protective of human health and the environment, and (2) all PAI
processes achieve a level of control already being achieved by the better controlled and lower
emitting processes. The commenter believes that this overall exemption would be consistent
with the intent of allowing exemptions for individual types of emission points associated with
production of a given product, and the sum of these individual exemptions could conceivably
even exceed the suggested overall cutoff of 500 Ib/yr. The commenter also believes an overall
exemption would give potentially affected facilities flexibility to reduce emissions in the most
cost effective way.
Response: The EPA decided not to exempt processes and storage tanks based on
emission stream characteristics. However, based on similar comments regarding the MACT
floor, the applicability cutoffs for storage vessels were revised for the final rule; see section 8.2
for a discussion of the changes. In addition, the definition of process vent was revised to exclude
streams that are undiluted, uncontrolled, and contain less than 20 ppmv of HAP; see the
responses to comment 8 in section 5.2 and comment 1 in section 7.2 for additional information.
3-21
-------
An overall cutoff was not included in the proposed rule, and also is not included in the
final rule because it is either inconsistent with the MACT floor or would require a restructuring
of the MACT floor with no clear benefit. Moreover, EPA believes the emissions averaging and
pollution prevention provisions in the rule provide comparable flexibility. It is not clear whether
the commenter wants the exemption to replace the cutoffs for individual types of emission points
(i.e., "individual cutoffs'') or to be in addition to them. If the overall cutoff were to be in addition
to the individual cutoffs it would be inconsistent with the MACT floor because a facility would
only take advantage of it if it were less stringent than the sum of the individual cutoffs, which are
based on the floor. An overall cutoff could not simply replace the individual cutoffs because the
MACT floor consists of individual cutoffs plus control efficiencies for emission points that
exceed the cutoff. Thus, the entire MACT floor would have to be reevaluated to determine an
overall control efficiency to go with it.
However, the EPA agrees with the commenter that flexibility in compliance is important.
This is the reason emissions averaging and pollution prevention provisions were included in the
proposed rule, and are retained in the final rule. The emissions averaging provisions allow an
owner or operator to determine the total required emission reduction for a group of emission
points, excluding equipment leak emissions, and to control these emission points in any way that
achieves the total required reduction. The pollution prevention provisions allow an owner or
operator to take credit for substantial reductions in HAP consumption as an alternative to using
add-on control devices to achieve the specified emission reductions.
Comment 2: Commenter IV-D-28 requested an exemption for equipment that does not
handle process fluids (e.g., heat exchange or refrigeration systems). The commenter noted that
previous MACT standards have distinguished production equipment, which handle process
fluids, and other equipment. The commenter suggested the following language based on
§ 63.160(e) of the HON: "Except as provided in this subpart, lines and equipment not containing
process fluids are not subject to the provisions of this subpart. Utilities and other nonprocess
lines, such as heating and cooling systems which do not combine their materials with those in the
processes they serve, are not considered to be part of a process unit."
Response: The final rule contains requirements for heat exchange systems, which can
include equipment that does not handle process fluids. The EPA has required some type of
3-22
-------
monitoring of noncontact cooling water because heat exchangers are known to leak in the
industry, and could contribute to significant emissions if left uncorrected. For this reason, the
final rule, as well as many other MACT standards, including the HON, and the Pharmaceuticals
NESHAP, contain requirements for monitoring of cooling towers under the heat exchange
systems requirements. Regarding the referenced language from subpart H of part 63, EPA
agrees with the commenter that, unless otherwise stated in any subpart that references subpart H,
lines and equipment not containing process fluids are not subject to the provisions of subpart H,
as the language in § 63.160(e) states. The language of subpart H, however, was meant to exclude
only components in lines that do not contain process fluids from the Leak Detection and Repair
Provisions of subpart H. The proposed rule, by referencing the provisions of subpart H,
including § 63.160(e), also contained this provision.
Comment 3: Section 63.1360(d)(2) of the proposed rule states that the provisions of the
PAI rule would not apply to emission points in PAI manufacturing operations that meet the
applicability requirements under subparts F, G, H, and I of the 40 CFR part 63.
Commenter IV-D-28 supports the concept that one rule takes precedence, but believes there are
two deficiencies with the proposed statement. First, the commenter interprets the statement to
mean that, in order to be exempt from the PAI rule, an emission point would have to be subject
to all four subparts at the same time. However, the commenter contends that few, if any,
emission points would meet this condition. Second, the commenter believes it is not useful to
have an exemption that applies only to specific "emission points" within a process unit because
the PAI rule and the four specified subparts deal with different subsets of HAP. As a result, the
commenter believes it is possible that, within a process, unit that is subject to one or more of the
specified subparts, the PAI rule may apply to some emission points that are not, individually,
subject to the other subparts. This would mean that two different regulations would apply within
the same process unit, which the commenter contends would be too complex.
To make the exemption achievable and useful, the commenter suggested two changes.
First, revise § 63.1360(d)(2) to state that emission points in PAI manufacturing operations that
meet the requirements under subparts F, G, and H are not subject to any provisions of the PAI
rule, and emission points that meet the applicability requirements under subparts H and I are not
subject the the requirements in the PAI rule that are related to equipment leaks. Second, add
3-23
-------
language similar in concept to that in § 63.160(c) of the HON that would allow a facility to elect
to comply with either another applicable MACT standard (e.g., the HON) or the PAI rule for all
relevant emission points associated with a process unit.
Response: The EPA agrees with the commenter that clarification of the proposed
provision is needed. The EPA did not mean that all four subparts must be applicable
simultaneously for the exemption to be allowed. Further, EPA agrees with the commenter that
all of the emission points in a given process unit should be part of only one source category for
the purposes of applying MACT standards. Therefore, this provision was revised in the final rule
to specify that emission points within a process unit that is subject to subpart F are not subject to
the final rule. References to subparts G and H were deleted because these subparts are cross-
referenced from subpart F. The proposed exemption for emission points subject to subpart I was
replaced with a provision that allows the owner or operator the option of continuing to comply
with the provisions of subpart I or switching to compliance with the equipment leak provisions in
the final rule (see section 3.8 for more information). These changes are consistent with the
commenter's first suggestion.
Although some of the compounds subject to the HON are registered as PAI's, and others
are intermediates in the production of PAI's, EPA believes the HON should take precedence for
any process that is subject to the HON because such a process is likely to be used predominantly
to produce chemicals in bulk for various uses. Even if the production at a particular facility is
captively used only to produce PAI's, the HON should take precedence because the process is the
same as that at other facilities that are producing the same chemical for other uses. Therefore,
EPA disagrees with the commenter's suggestion that a facility be allowed to elect whether to
comply with the HON or the PAI final rule. Overlap with other MACT standards may occur for
some intermediates and nondedicated equipment; the applicability in these situations is discussed
in sections 3.2 and 3.3 of this document.
Comment 4: Section 63.1360(d)(3) of the proposed rule would exempt emission points
in PAI manufacturing operations that meet the applicability criteria under any other existing
MACT standard. Commenter FV-G-01 suggested expanding this exemption to allow a facility to
comply with a more stringent rule in lieu of the PAI rule. For example, the commenter described
a situation in which nondedicated equipment may be used to manufacture pharmaceuticals part of
3-24
-------
the year and a PAI for the remainder of the year. Under the proposed rules for the PAI and
Pharmaceuticals source categories, the equipment would be subject to the PAI rule when making
a PAI and to the pharmaceuticals rule when making pharmaceutical products. The commenter
believes complying with two standards for the same equipment train would be confusing. As an
alternative to broadening the exemption, the commenter suggested that the rule apply only when
50 percent or more of the annual production from the equipment is a PAI (see comment No. 1 in
section 3.2 of this document).
Response: The EPA decided to delete this provision because: (1) it is redundant with the
provision described in comment 3 above for processes subject to the HON; (2) other than the
HON, there are no existing MACT standards that would apply to PAI production; (3) inorganic
compounds that may be registered as PAI's (e.g., sodium dichromate, chromic acid, hydrogen
chloride [HC1], chlorine [C12], hydrogen cyanide [HCN]) are part of source categories that have
been deleted or for which standards have not yet been promulgated, and production of inorganic
compounds has been deleted from the PAI source category); and (4) changes to the applicability
of intermediates in the final rule mean it would be unnecessary for production of intermediates
(see section 3.3). The commenter's concern about nondedicated equipment being subject to
multiple rules is addressed in the response to comment 1 in section 3.2.
Comment 5: Commenter FV-D-28 suggested editorial clarifications to the language used
in §§ 63.1360(d)(2) and (3). Specifically, the commenter suggested using the phrase "within an
affected source subject to" instead of the phrases "that meet the applicability requirements" and
"that meet the applicability criteria."
Response: In an effort to improve clarity, the final rule uses language similar to that
suggested by the commenter.
3.6 NEW AFFECTED SOURCE
Status at proposal: Section 63.1360(g)(2) of the proposed rule states that an addition of
PAI manufacturing operations at an existing plant site would be subject to the requirements for a
new source if the addition has the potential to emit 10 tons/yr or more of any HAP or 25 tons/yr
or more of any combination of HAP, unless the Administrator establishes a lesser quantity at a
plant that currently is an affected source. New source requirements also would apply to a new
3-25
-------
plant site on which construction started after November 10, 1997 or a reconstructed source for
which reconstruction started after November 10, 1997.
Comment 1: Commenter IV-G-03 suggested that § 63.1360(g)(2) be reworded by
inclusion of the phrase "by regulation" after the word "Administrator" so that it does not
arbitrarily extend the Administrator's decisions of establishing lesser quantities without specific
guidance or regulations.
Response: The phrase "unless the Administrator establishes a lesser quantity" is
redundant with section 112(a)(l) of the Act. Because of this redundancy it is not included in the
final rule.
Comment 2: Commenters FV-D-21 and FV-D-29 requested confirmation of their
interpretation of § 63.1360(g)(2) to mean that a source with minor actual emissions but major
potential to emit could elect to accept a Federally enforceable "synthetic minor" operating permit
with an emission limit below the 10 and 25 tons/yr cutoffs, and thereby avoid the new source
requirements for process vents, storage tanks, and wastewater. Another commenter (IV-D-28)
supports the approach whereby an addition is subject to new source standards only if it meets the
10 or 25 tons/yr emission level by itself.
Response: The new affected source provisions have been revised for the final rule. As
noted in the responses to comment 1 in section 3.2 and comment 1 in this section, the term "PAI
manufacturing operations" and the phrase "unless the Administrator establishes a lesser quantity"
are not used in the final rule. The EPA also is concerned that the term "addition" may be
ambiguous. The intent was that the requirements apply only to additions consisting of
equipment dedicated to the production of a single PAI. Therefore, the final rule specifies that
new source requirements apply to an affected source for which construction or reconstruction
commenced after November 10, 1997, or to any single PAI process unit that meets the following
conditions: (1) it is not part of a process unit group, (2) construction commenced after
November 10, 1997, and (3) it has the potential to emit 10 tons/yr of any one HAP or 25 tons/yr
of combined HAP. Under this definition, if an owner or operator elects to accept Federally
enforceable conditions that limit the potential to emit for a single PAI process unit that is added
to an existing facility to levels below these thresholds, the PAI process unit would be subject to
existing source standards, not new source standards.
3-26
-------
3.7 STARTUP, SHUTDOWN, AND MALFUNCTION
Status at proposal: For batch processes, § 63.1360(e) of the proposed rule would require
an owner or operator to comply with the provisions in the rule during periods of startup and
shutdown, and periods of malfunction would be regulated according to § 63.6 of the General
Provisions. For continuous processes, § 63.1360(e) of the proposed rule specifies that only
§ 63.6 of the General Provisions would apply during periods of startup, shutdown, and
malfunction.
Comment: Three commenters addressed the startup, shutdown, and malfunction
provisions in § 63.1360(e) of the proposed rule. Commenter FV-D-28 agrees that routine startups
and shutdowns between batches should be covered by the rule, but stated that it should not apply
during other startups and shutdowns because normal emission control techniques may be
inappropriate or ineffective during those times. According to the commenter, some of the other
situations include (1) initial startup of a process unit, (2) startup after a malfunction or an
extended period of nonoperation, and (3) shutdowns due to a malfunction. The commenter
explained that during initial startup, control devices and monitoring systems need to undergo
"shakedown" and debugging, and may need time to reach their full efficiency. After an extended
downtime, process equipment also will need time to get back to normal operating conditions, and
control devices will need to reach operating temperatures or equilibrium. Although the
commenter understands that the proposed rule would not apply during malfunctions, the
requirements during a shutdown associated with the malfunction were not clear.
Commenter IV-D-28 also stated that the final PAIMACT standards should not-
incorporate § 63.6(e) of the General Provisions for four reasons. First, the requirement in
§ 63.6(e)(3)(i)(A) to minimize emissions "at least to the levels required by all relevant standards"
is ambiguous. Second, the General Provisions do not address shutdowns of compliance
equipment such as control devices. Third, the General Provisions do not address startups,
shutdowns, and malfunctions that affect only a portion of the process. Fourth, the General
Provisions do not say how to deal with periods of nonoperation. To address these concerns, the
commenter suggested revising §§ 63.1360(e)(l) and (2) to refer to a new § 63.1360(e)(3) in the
rule instead of § 63.6 of the General Provisions. For the new § 63.1360(e)(3), the commenter
recommended using language similar to that in the HON. The commenter also noted that the
3-27
-------
discussion of § 63.6 in Table 1 (the General Provisions applicability table) would need to change
as in the HON.
Commenter IV-D-29 recommended that EPA consider revising § 63.1360(e)(l) of the
proposed rule to allow batch processes with air pollution control equipment to comply with the
startup, shutdown, and malfunction requirements in 40 CFR 63.6(e). The commenter explained
that operating practices for controls used with batch processes are the same as those for controls
used with continuous processes; for both types of processes, operators verify that all control
equipment is on-line and functioning properly to minimize emissions at all times (consistent with
§ 63.6(e)(l)(i) of the General Provisions). Furthermore, the commenter stated that maintenance
and corrective actions after a malfunction of a control device are the same for both batch and
continuous processes. Therefore, the commenter recommended that EPA consider revising the
rule to include the following language: "For batch processes with air pollution control
equipment, startup, shutdown, and malfunction shall be regulated according to section 63.6 of
subpart A of this part. For batch processes without air pollution control equipment, the
provisions of this subpart shall apply during startup and shutdown, and periods of malfunction
shall be regulated according to § 63.6 of subpart A of this part."
Response: The EPA has reconsidered the applicability of the rule during periods of
startup and shutdown and determined that the requirements of the rule should not be applied
under certain situations for batch processes as well as for continuous processes. For batch
processes, these situations include initial startups of new or reconstructed processes, and
shutdowns that are not part of intended operation (e.g.. for maintenance, replacement of
equipment, or other repair as a result of a malfunction). These are times when the operators may
be unfamiliar with the equipment operation or it may not be possible to follow standard operating
procedures. However, setting the equipment in operation after maintenance (including
maintenance to correct a malfunction), after switching to a product that has been produced in the
past, or for each batch during a campaign are all routine, normal operating conditions that should
result in the same emissions profile. Similarly, the cessation of operation at the end of a
campaign, between batches, or for planned, preventive maintenance are all normal operations
with the same emissions profile. Conversely, for continuous processes, startup and shutdown for
any reason results in operation under conditions different from the normal steady-state operation.
3-28
-------
To account for these differences between batch and continuous processes, the final rule provides
definitions for startup and shutdown that differ from the definitions in the General Provisions.
Specifically, the following definitions have been added to the rule:
Startup means the setting in operation of a continuous PAI process unit for any purpose,
the first time a new or reconstructed batch PAI process unit begins production, or, for new
equipment added to any PAI process unit, including equipment used to comply with this subpart,
the first time the equipment is put into operation. For batch process units, startup does not apply
to the first time the equipment is put into operation at the start of a campaign to produce a
product that has been produced in the past, after a shutdown, or when the equipment is put into
operation as part of a batch within a campaign. As used in § 63.1363, startup means the setting
in operation of a piece of equipment or a control device that is subject to this subpart.
Shutdown means the cessation of operation of a continuous PAI process unit or any
equipment within the PAI process unit, including equipment required or used to comply with this
subpart, for any purpose. Shutdown also means the cessation of a batch PAI process unit, any
equipment within the PAI process unit, or any individual piece of equipment required or used to
comply with this pan or for emptying and degassing storage vesselsas the result of a
malfunction. Shutdown does not apply to cessation of a batch PAI process unit at the end of a
campaign, for routine maintenance, for rinsing or washing of equipment between batches, or
other routine operations.
The EPA has also clarified in the final rule that the provisions can apply to processing
equipment, as well as control, monitoring, and recordkeeping equipment. Additionally, in
response to the commenter's concerns regarding ambiguity of the general provisions, EPA has
replaced the reference to the general provisions with language from the HON that specifically
clarifies applicability of provisions during startup, shutdown, and malfunction events.
3.8 OVERLAP WITH OTHER REGULATIONS
Comment: Three commenters stated that, in addition to the exemptions in § 63.1360(d)
of the proposed rule, the rule must also address overlaps with other regulations.
Commenter IV-D-28 identified overlaps with NSPS under part 60, NESHAP under part 61, and
RCRA equipment leaks requirements that are not addressed. Commenters FV-D-21 and IV-D-29
went a step further and identified specific NSPS and NESHAP with overlapping provisions for
3-29
-------
Group 1 PAI vents (i.e., NSPS subparts Kb, m, NNN. and RRR; and NESHAP subparts BB, FF,
and G). All three commenters suggested revising the rule to include language similar to that in
§ 63.110 of the HON for provisions dealing with process vents, storage tanks, and wastewater.
Commenter IV-D-28 also suggested borrowing language from §§ 63.160(b) through (d) to
address overlapping provisions that deal with equipment leaks.
Response: The EPA agrees with the commenters that the rule must address overlap with
other regulations. The final rule includes language similar to that in § 63.110 of the HON for
overlap with NSPS requirements for storage vessels in subpart Kb of 40 CFR part 60 and RCRA
requirements in 40 CFR parts 260 through 272. The EPA also added a provision specifying that
an owner or operator subject to both this rule and the equipment leak requirements in subpart I of
40 CFR part 63 may elect to comply with the requirements of either rule.
The requirements in NSPS subparts EH, NNN, and RRR apply to individual vents,
whereas the process vent standards in the final rule apply to the sum of all process vents within a
process. As a result, a facility generally must comply with both the final rule and any applicable
NSPS. One exception is provided in the final rule. If an owner or operator elects to reduce
emissions from a process vent by 98 percent (or implement an equivalent control option), then
the owner or operator is required to comply only with the provisions of the final rule.
The final rule does not address overlap with NESHAP in 40 CFR part 61. Subparts BB
and FF regulate emissions from benzene production, which, because it is subject to the HON, is
not subject to the PAI rule. Subpart G is reserved and also is not covered in § 63.110 of the
HON.
3.9 STORAGE TANK ASSIGNMENT
Status at proposal: Section 63.1360(0 of the proposed rule specifies procedures to
determine if storage tanks, including storage tanks in a tank farm, are part of the affected source.
Under the proposed rule, a storage tank would be part of the affected source if the majority of its
throughput is associated with PAI processes. If the use varies from year to year, the
determination would be based on the utilization that occurred during the year preceding
publication of the proposed rule (i.e., November 10, 1997). A storage tank in a tank farm would
not be part of the affected source if all of its throughput went through intervening tanks between
it and the applicable processes. The proposed rule also specifies that if there is a significant
3-30
-------
change in the use of the storage tank that the owner or operator would be required to reevaluate
the applicability of the rule for that tank.
Comment: Commenter IV-D-28 addressed two aspects of the storage tank applicability
determinations. First, the commenter stated that for storage tanks in a tank farm, the provisions
in § 63.1360(f)(3)(ii) of the proposed rule do not indicate whether material from or to intervening
tanks should be excluded or included in the "predominant use" calculation. The commenter
recommended that the language from § 63.100(g)(3) of the HON be used without modification
because those provisions are clear and complete; it would also make the PAI rule more consistent
with the HON. Second, the commenter stated that a facility should not be required to reevaluate
the applicability every time the predominant use may have changed. The commenter believes
that frequently changing storage tank assignments would make the compliance program difficult
to administer, and that the reevaluation is unwarranted because changes in usage between process
< units are accounted for in the initial assignment determination.
Response: The EPA agrees with the commenter that the rule must indicate how material
from or to intervening tanks is to be used in the predominant use calculation. Language from
§ 63.100(g)(3) of the HON was not copied verbatim into the proposed rule because a storage tank
was only assigned to the PAI manufacturing operations in general; it was not necessary to assign
it to a specific PAI process unit. However, because the concept of process units was added to the
final rule for other reasons, the language from § 63.100(g)(3) was also added to the final rule.
The final rule retains the requirement to reevaluate the applicability assignment if the storage
vessel begins receiving material from (or sending material to) another process unit because these
are significant changes that are not accounted for in the initial evaluation. For example, if a tank
is no longer associated with a given process, it should no longer be assigned to the process unit
for that process.
3-31
-------
4.0 COMPLIANCE DATES
4.1 EXISTING SOURCES
Comment: Four commenters addressed the issue of the compliance date for existing
sources. Commenter IV-D-15 stated that EPA should delay implementation of the rule to allow
facilities to comply with the HON and then reassess their potential to emit.
Commenters FV-D-16 and IV-D-28 supported the proposed compliance date of 3 years after
promulgation. Commenter IV-D-17 urged EPA to seek expeditious promulgation,
implementation, and enforcement of the proposed limits; this commenter believes a compliance
date sooner than the statutory maximum of 3 years after promulgation might be appropriate and
is certainly achievable.
Response: The compliance date cannot be more than 3 years after promulgation because,
as one commenter noted, 3 years is the maximum time allowed under the Act. However,
section 112(i)(3)(B) of the Act and § 63.6(i) of the General Provisions allow an owner or
operator to request an extension of compliance of up to 1 additional year if the additional period
is necessary for the installation of controls. To make this provision more visible, it also has been
stated in the final rule. The EPA does not believe commenter FV-D-17 presented a compelling
argument for a compliance date sooner than 3 years after promulgation. Thus, EPA continues to
believe that 3 years is a reasonable time period, and appreciates the support of two of the
commenters in this regard.
4.2 NEW SOURCES
Comment 1: Commenter IV-D-28 believes the proposed provision that would require
new sources to be in compliance upon startup should be revised to require compliance by initial
startup or the promulgation date of the rule, whichever is later.
4-1
-------
Response: A provision requiring that new sources be in compliance by initial startup or
the promulgation date, whichever is later, is consistent with other MACT standards and has been
added to the final rule.
Comment 2: Commenters FV-D-21 and IV-D-29 suggested adding a provision that would
extend the compliance date for new sources that commence construction after proposal but
before promulgation if the final standard is different from the proposed standard and the owner or
operator complies with the standard as proposed during the 3-year period immediately after the
effective date.
Response: The language suggested by the commenter is similar to the provisions in
sections 112(i)(2) of the Act and § 63.6(b)(3) of the General Provisions (except these provisions
refer to final standards that are "more stringent," as opposed to merely "different," than the
proposed standards). The EPA assumes the commenters are relying on these authorities for their
requests. Because the final rule is not more stringent than the proposed standards, there is no
need to include this provision in the final rule.
Comment 3: Commenters IV-D-21 and IV-D-29 believe EPA should either allow new
sources a period of up to 6 months to complete any required testing after startup, or change the
definition of startup to stipulate that startup is not complete until all required performance testing
is complete and that this testing must be completed no later than 6 months after steady state
production for continuous processes or until 6 months after a successful batch production run has
been completed.
Response: The EPA does not believe that the compliance date needs to be changed to
accommodate required emissions testing. Under the proposed rule, an owner or operator would
be required to submit the Notification of Compliance Status report no later than 150 days after
the compliance date (i.e., startup for a new source). This requirement is consistent with other
MACT standards (e.g., the HON, Polymers and Resins [P&R] I, and P&R IV), and it is nearly
the requested 6 months after the compliance date. Furthermore, much of the work (e.g., the
emissions profile) may be completed before the compliance date. The amount of time needed to
reach steady state production or to complete a successful batch production run should not be
greater in this industry than in other chemical production industries. Therefore, the final rule
4-2
-------
retains the provision to submit the Notification of Compliance Status report no later than
150 days after the compliance date.
4-3
-------
5.0 DEFINITIONS
5.1 NEW TERMS DEFINED IN THE FINAL RULE
Comment 1: Commenter IV-D-28 requested that EPA provide a definition for Flame
lonization Detector (FID).
Response: In the final rule, FID is defined to mean "a device in which the measured
change in conductivity of a standard flame (usually hydrogen) due to the insertion of another gas
or vapor is used to detect the gas or vapor."
Comment 2: Commenters IV-D-28 and IV-G-03 requested that EPA provide a definition
forTOC.
Response: In the final rule, TOC is defined to mean "those compounds measured
according to the procedures of Method 18 or Method 25A of 40 CFR part 60, appendix A."
Comment 3: Commenter IV-D-28 requested that EPA add definitions for startup and
shutdown consistent with the HON.
Response: Definitions for these terms in the final rule are presented in the response to
comments on startup, shutdown, and malfunctions in section 3.7 of this document.
Comment 4: A number of commenters (IV-D-14, IV-D-21, IV-D-28, IV-D-29, and
IV-G-03) requested that the rule define both process wastewater and maintenance wastewater.
Response: The final rule includes definitions for both of these terms that are consistent
with the definitions in the HON. The terms were added because the requirements in the final rule
for maintenance wastewater differ from the requirements for process wastewater. See
section 10.2 for a discussion of the changes to the maintenance wastewater provisions.
Comment 5: Commenters IV-D-14, IV-D16. and IV-D-28 requested that the final
regulation include a definition of "bag dumps."
5-1
-------
Response: The final rule defines bag dumps to be equipment into which bags or other
containers containing a powdered, granular, or other solid feedstock material are emptied. To be
complete, the final rule also defines product dryer as equipment that is used to remove moisture
or other liquid from granular, powdered, or other solid PAI or integral intermediate products
prior to storage, formulation, shipment, or other uses.
Comment 6: Commenter IV-D-15 stated that the rule needs to include a definition for
"storage" to clarify which intermediate process are integral or non-integral.
Response: The definition of "storage" as it relates to intermediates is described in the
response to comments in section 3.3 of this document.
Comment 7: Commenter IV-D-28 requested that EPA define "organic HAP" by listing
covered compounds because compounds like phosgene and cyanide compounds do not have
universal interpretations.
Response: The final rule defines organic HAP as "those HAP listed in section 112(b) of
the Act that are measured according to the procedures of Method 18 or Method 25 A, 40 CFR
part 60, appendix A, and exist in the vapor phase at ambient conditions."
5.2 REVISED DEFINITIONS THAT DIFFER FROM COMMENTERS SUGGESTIONS
Comment 1: Commenter IV-D-28 requested that EPA change the definition of "Group 1
storage tank" to match the HON because control of 240 Ib/yr or 1 Ib/yr is unlikely to be cost
effective, and consistency with the HON reduces regulatory burden.
Response: The responses to comments in sections 8.1 and 8.2 of this document describe
changes to the MACT floor and standard for storage vessels. As a result of these changes to the
floor, the definition of "Group 1 storage vessel" for the final rule has been changed to be
consistent with the HON, except that the vapor pressure cutoffs differ. For existing sources, the
vapor pressure cutoff is 3.45 kPa for storage vessels with a capacity greater than or equal to 75
cubic meters (m3). For new sources, the cutoff is 16.5 kPa for storage vessels with a capacity
greater than or equal to 38 m3 and less than 75 m3, and 3.45 kPa for storage vessels with a
capacity greater than or equal to 75 m .
Comment 2: Commenters FV-D-21 and IV-D-29 stated that the final rule should include
a definition of impurity because one of the exemptions is tor "vessels and equipment storing
and/or handling material that contain no organic HAP and/or organic HAP as an impurity only."
5-2
-------
The commenters suggested the following wording: "A substance that is produced coincidentally
with the primary product, or is contained in the final product as a contaminant serving no useful
purpose, or is present in a raw material. An impurity does not serve a useful purpose in the
production or use of the primary product and is not isolated."
Response: The EPA agrees with the commenter that a definition of the term "impurity" is
needed to clarify the use of this term in the definition of storage vessel. The final rule defines
impurity as "a substance that is produced coincidentally with the product(s), or is present in a raw
material. An impurity does not serve a useful purpose in the production or use of the product(s)
and is not isolated." The exemption cited by the commenter also has been revised in the final
rule by deleting the words "and equipment" and the words "and/or handling." Removing these
words clarifies the original intent of the statement, which was to exempt certain storage vessels.
Comment 3: Commenter IV-G-03 stated that the definition of "wastewater" needs
clarification with regard to the second criterion, which is that it must be water that "is discarded
from PAI manufacturing operations at a major source." The commenter stated that this implies
that the "source" in question that must meet the definition of "major" is the vessel, tank, process
equipment.
Response: In the final rule this criterion has been changed to be water that "is discarded
from a PAI process unit that is at an affected source." The term "PAI manufacturing operations"
was replaced with the term "PAI process unit" to clarify that the wastewater is discarded from
process equipment or storage vessels and because, as discussed in the responses to comment 7 in
this section and to comments 1 and 2 in section 3.3, the term "PAI manufacturing operations" has
been deleted from the final rule. The phrase "at a major source" was replaced with the phrase
"that is at an affected source" to clarify that the discharge is only classified as wastewater if it is
from a PAI process unit that processes, uses, or produces HAP and is located at a major source.
The definition of wastewater was also changed in other ways as part of the changes in the
provisions for maintenance wastewater; these changes are described in section 10.2 of this
document.
Comment 4: The proposed rule states that to be a process condenser, the primary
purpose of the condenser must be to recover material as an integral part of a unit operation, and
the condenser must support a vapor to liquid phase change for periods of source equipment
5-3
-------
operation that are above the boiling or bubble point of substances. Several examples of process
condensers were also provided. Commenter IV-G-01 recommended changing the proposed
definition because the requirement that source operation be above the boiling or bubble point
temperature means that a given condenser will sometimes be a process condenser, and sometimes
it will be an air pollution control device. The commenter believes this will cause confusion for
the facility, and tracking the changes from one classification to the other will be burdensome.
Therefore, the commenter recommended that the definition be changed to mean "a condenser
whose primary purpose is to recover material as an integral part of the unit operation, or is
essential to effectively run the process." Furthermore, the commenter believes the facility, which
is more familiar with the process than anyone else, is in the best position to determine whether it
should be considered a process condenser or an air pollution control device.
Commenter IV-D-28 requested that the definition of "process condenser" be clarified
because the term "integral" is not defined, the term "unit operation" is used incorrectly (see
comment 7 in section 5.4), it is not clear what the term "support" means, it is not clear what the
difference is between the boiling and bubble point, the term "substance" should be limited to
HAP, and the examples are circular because they use the term "process condenser."
Response: The EPA agrees with the commenter that a given condenser may be classified
as a process condenser at some times and as an air pollution control device at other times.
However, EPA believes the conditions for both types of situations are clearly defined in the final
rule and that this approach is preferable over potentially arbitrary classifications based on its
typical use over all uses. Because one of the formats of the process vent standard requires that a
reduction from uncontrolled emissions be applied across all vents within a process, EPA is
concerned about the opportunity for crediting reductions achieved by condensing boiling streams
on other sources in the process (i.e., overestimating the control by calling the condenser an air
pollution control device for all inlet streams).
Although the basic concept of the definition is unchanged in the final rule, a number of
editorial changes were made to clarify the meaning. For the final rule, process condenser is
defined to mean "a condenser whose primary purpose is to recover material and is an integral
part of a unit operation. The condenser also must cause a vapor-to-liquid phase change for
periods during which the temperature of liquid in the process equipment is at or above its boiling
5-4
-------
or bubble point. Examples of process condensers include distillation condensers, reflux
condensers, and condensers used in stripping or flashing operations. In a series of condensers, all
condensers up to and including the first condenser with an exit gas temperature below the boiling
or bubble point of the liquid are considered to be process condensers. All condensers in line
prior to the vacuum source are included in this definition." The definition retains the term
"bubble point" because a liquid mixture does not boil at only a single temperature. For a mixture
of two or more components, the bubble point is the temperature at which the first bubbles of
vapor form in the liquid. In the examples, the process condenser is considered "an integral part
of the unit operation" because process could not be operated without the condenser.
Comment 5: Commenter FV-G-01 recommended that the definition of "pesticide active
ingredient manufacturing operations" should not include waste management units. The
commenter stated that in other standards the waste management units are not part of the
applicable process units and thus are not subject to the standard but are instead used to comply
with the standard. The commenter also concluded that the proposed definition could be
interpreted to require compliance with the new source standards at an existing waste management
unit simply because a new and major PAI manufacturing operation has been built that will
contribute wastewater to the unit.
Response: As discussed in the response to comments in section 3.2 of this document, the
term "PAI manufacturing operations" has been deleted from the final rule. The definition of the
affected source has also been revised, and a new term, "PAI process unit," has been added. The
PAI process unit is comparable to the chemical manufacturing process unit (CMPU) in the HON,
and it does not include the waste management units. However, waste management units are part
of the affected source, as in the HON and other standards. Finally, the commenter's conclusion
regarding the application of new source requirements is correct. If a new PAI process unit that
meets the requirements for new source applicability (see section 3.6 of this document for a
discussion of changes in these requirements), then the waste management units associated with
that new PAI process unit would have to meet the requirements for new sources. The practical
impact of this requirement, however, is expected to be minimal because the requirements for new
sources and existing sources are identical except when the HAP load to the waste management
unit exceeds 2,100 Mg/yr. Based on survey data from the industry, no single existing PAI
5-5
-------
process unit discharges wastewater with such a high load (and only one facility discharges
wastewater containing that much HAP).
Comment 6: Commenter IV-D-09 stated that the definition of "pesticide active
ingredient" must be revised to be consistent with the basis and purpose document. The
commenter asserted that PAI's derived from naturally occurring substances are not "synthetically
manufactured" or "produced in a chemical reaction" and should not be regulated by the proposed
NESHAP.
Response: Processes that derive PAI's from naturally occurring substances most likely
will be subject to the rule if they are extracted with HAP solvents. Such extraction is comparable
to extraction used in production of synthetic organic PAI's. Thus, the production of PAI's
derived from naturally occurring substances has not been deleted from the definition of PAI in
the final rule. Only sources that use a HAP are regulated; therefore, many producers of PAI's
from naturally occurring compounds are not regulated. However, as noted in the response to
comments in sections 3.1 and 3.2 of this document, production of inorganic compounds has been
deleted from the affected source, and production of compounds that are primarily used for non-
pesticidal purposes is not subject to the final rule.
Comment 7: Commenter IV-D-28 believes EPA should replace the term "process" with
the term "process unit," and then add a new term called "PAI process unit." The commenter
finds the proposed definition of "process" unclear for several reasons. First, it forces a subtle,
non-obvious departure from common usage for the term; by defining it as equipment that is used
to produce PAI's, it means equipment used to produce other products cannot be called a process.
Second, it does not require a process to consist of more than one unit operation; the commenter
noted that the HON was amended to clarify that a CMPU consists of two or more unit operations
(the commenter also objected to changing the definition of unit operation from the definition in
the HON; see comment 7 in section 5.4 of this chapter). Third, inconsistencies among
regulations are likely to lead to problems in establishing and implementing compliance programs
at facilities with processes subject to different regulations. To resolve some of these concerns,
the commenter believes a generic term, "process unit," should be added with a definition that
encompasses all of the different types of process units in other MACT standards. The
commenter also would add the new term "PAI process unit" to define a specific type of process
5-6
-------
unit that is subject to the PAIMACT standard. The commenter supports the exclusion of
formulation from the definition of "process" (or "PAI process unit").
Response: Many of the changes suggested by the commenter have been incorporated in
the definitions in the final rule. The term "process" has been retained, but the definition has been
changed to refer to the production of any product, not just PAI's. The generic term "process
unit" has been added to parallel the term CMPU in the HON, and the term "PAI process unit" has
been added and defined as a process unit that is used to produce a PAI or integral intermediate.
The definition also describes other features of a PAI process unit based on language in the
proposed definition of "process," but it does not specify that a PAI process unit must consist of at
least two unit operations. See the response to comments in section 3.2 of this document for more
information about the changes in these definitions as they relate to the definition of the affected
source.
Comment 8: In the proposed rule, a process vent was defined as "a vent from a unit
operation through which a HAP-containing gas stream is, or has the potential to be, released to
the atmosphere." Examples of process vents and emission points that would not be process vents
were also included in the definition. The proposed rule defined a product dryer vent as "a vent
from an atmospheric dryer through which a gas stream containing gaseous organic HAP,
paniculate matter HAP, or both is, or has the potential to be, released to the atmosphere.
Commenter FV-D-28 requested that EPA modify these definitions to include a de minimis HAP
concentration cutoff (because the proposed definition would apply even to controlled sources),
specify which HAP are covered (the commenter understands the rule to mean organic HAP and
HC1), exclude streams that are routed to a fuel gas system, and delete the phrase "has the
potential to be."
In the proposed rule, a Group 1 process vent was defined as a "process vent from a
process at an existing or new affected source for which the uncontrolled emissions from the sum
of all process vents are greater than or equal to 150 kilograms per year (kg/yr) (330 lb/yr)."
Three commenters (IV-D-21, PV-D-28, and IV-D-29) requested that EPA change this definition
to match the HON. Commenters IV-D-21 and IV-D-29 suggested the following language:
". . . Group 1 process vent means any process vent from a continuous process at an existing or
new affected source for which the flow is greater than or equal to 0.005 cubic feet per minute
5-7
-------
(scfm), the total organic HAP concentration is great than or equal to 50 ppmv, and the total
resource effectiveness index value, calculated according to § 63.115 of subpart G, is less than or
equal to 1.0." The commenters suggested the following for batch process vents: "Group 1
process vent means any process vent at an existing or new affected source with uncontrolled
emissions greater than 330 Ib/yr."
Response: The definition of "process vent" in the final rule incorporates several of the
suggested changes. Specifically, the definition excludes streams with low HAP concentrations
(20 ppmv was selected instead of the suggested 50 ppmv), the phrase "has the potential to be"
has been deleted, and the type of HAP has been identified. Although concentration data are not
available from the surveyed plant, streams with concentrations below 20 ppmv are likely to be
uncontrolled because that level is considered to be the practical limit of control. Furthermore,
such streams are likely to have low annual emissions, and thus have little impact on the
applicability determination for a process. The exemption for fuel gas systems has not been
included for reasons discussed in the response to comments in section 6.3 of this document. In
the final rule, a process vent is defined as:
a point of emission from processing equipment to the atmosphere or a control device.
The vent may be the release point for an emission stream associated with an individual unit
operation, or it may be the release point for emission streams from multiple unit operations that
have been manifolded together into a common header. Examples of process vents include, but
are not limited to, vents on condensers used for product recovery, bottom receivers, surge control
vessels, reactors, filters, centrifuges, process tanks, and product dryers. A vent is not considered
to be a process vent if the undiluted and uncontrolled emission stream that is released through the
vent contains less than 20 ppmv HAP, as determined (1) through process knowledge that no HAP
are present in the emission stream; (2) using an engineering assessment as discussed in
§ 63.1365(b)(2)(ii); (3) from test data collected using Method 1818 of 40 CFR part 60,
appendix A; or (4) from test data collected using any other test method that has been validated
according to the procedures in Method 301 of appendix A of this part. Process vents do not
include vents on storage vessels regulated under § 63.1362(c), vents on wastewater emission
sources regulated under § 63.1362(d), or pieces of equipment regulated under § 63.1363.
The other suggested changes to the definition of Group 1 process vent have not been
included in the final rule because they would be inconsistent with the MACT floor for process
vents. The MACT floor was developed by evaluating all PAI processes in one group, and
considering the emissions from the sum of the process vents within each process. However, as
-------
noted above, EPA did change the definition of "process vent" to exclude emission streams with
HAP concentrations less than 20 ppmv; this change also affects Group 1 process vents.
The definition of a product dryer vent was changed to indicate that it is a process vent.
Thus, the 20 ppmv cutoff also applies to product dryer vents. However, even if the organic HAP
and HC1/C12 concentrations are below 20 ppmv, the product dryer vent is still subject to the
0.01 gr/dscf standard for paniculate matter emissions.
Comment 9: Commenter FV-D-28 supports the definitions of recapture device and
recovery device.
Response: These terms were mistakenly used in several places in the proposed rule;
except for the concept of recovery from water streams, they are not used in the final rule. These
terms are not used in the process vent provisions in the final rule because they conflict with the
terms "process condenser" and "air pollution control device." An EPA survey of the industry
specified that respondents distinguish between process condensers and air pollution control
devices, and the MACT floor was based on this information. This approach is simpler than the
three-term approach used m the HON, and it draws a clearer distinction between equipment that
is essential to the operation of the process and equipment that is used primarily to control
emissions. There may be incidental recovery of compounds that are collected in some absorbers
and carbon adsorbers, but the intended purpose of such equipment is likely to be as an air
pollution control device.
The HON specifies that water is not wastewater until after it exits the last recovery
device. Because the final rule, like the proposed rule, continues to cross-reference the
wastewater provisions in the HON, the concept of recovery also is retained in the wastewater
provisions in the final rule.
5.3 REVISED DEFINITIONS CONSISTENT WITH COMMENTERS SUGGESTIONS
Comment 1: A number of commenters (IV-D-16, IV-D-20, IV-D-21, IV-D-28, and
IV-D-29) requested that the definition of "cover" be made consistent with the definition in the
HON by deleting the requirement that covers be sealed.
Response: The requirement that covers be sealed has been deleted from the definition in
the final rule to be consistent with the definitions in previous rules.
5-9
-------
Comment 2: Commenters IV-D-21 and IV-D-29 stated that EPA has not explained or
justified the reason why hoses should be considered containers in the PAI industry, but not in the
SOCMI. The commenters recommend that EPA remove hoses from the definition of containers
in this rule.
Response: Hoses have been deleted from the definition of containers in the final rule
because they are not used as containers and were mistakenly included in the proposed rule.
Comment 3: Commenter IV-D-28 stated that EPA should not use the term "isolated
intermediate" because this term has a different meaning under TSCA, which will cause
confusion.
Response: The EPA agrees with the commenter's concern. The term "isolated
intermediate" has been deleted from the final rule to avoid confusion.
Comment 4: Commenter IV-D-28 requested that the definition of "hard-piping" be
revised to mean pipe as well as tubing, which would be consistent with the HON.
Response: The EPA agrees that the definition should be consistent with the HON. The
definition of "hard-piping" has been changed in the final rule to mean pipe as well as tubing.
Comment 5: Commenter IV-D-28 stated that the definition of Publicly-Ownted
Treatment Works (POTW) is too wordy and only needs to state "as defined at 40 CFR part
403.3(O)". The commenter stated that the rest of the existing definition serves no purpose and
can do harm if, for example, one regulation or the other is amended in a nonuniform way.
Response: The EPA agrees that the definitions should be identical and avoid potential
discrepancies that might result from amendments to the Clean Water Act regulations. The final
rule states that POTW is defined in 40 CFR part 403.3(O).
Comment 6: Commenter IV-D-28 stated that contrary to the definition, a "continuous
process" is not "typically steady state;" it typically "approaches" steady state, which is an ideal
condition that does not exist most of the time.
Response: The commenter is correct; the word "approaches" has been added to the
definition of "continuous process" in the final rule.
Comment 7: Commenters FV-D-21 and IV-D-29 stated that although the basic definition
of point of determination (POD) is a paraphrase of the HON, the NOTE attached to the POD
definition has a reference to TABLE 8 compounds in addition to TABLE 9 compounds. The
5-10
-------
commenters stated that TABLE 8 is not mentioned anywhere else in the proposed rule and the
commenters believe this was a transcription error and should be deleted; TABLE 9 is the only list
needed for both new and existing PAI manufacturing operations. Commenter IV-D-28 stated
that the proposed definition of POD is unclear, primarily because it uses the term "process,"
which is also unclear. The commenter stated that the definition of POD should be changed to
match the HON.
Response: For the final rule, the definition of POD has been changed to mean "each
point where a wastewater stream exits the PAI process unit." This definition is consistent with
the definition in the HON. All references to Table 8 compounds in the proposed rule were
inadvertent and have been deleted from the final rule.
Comment 8: Commenter FV-D-16 requested that nonwastewater "waste" tanks under the
definition of "storage tank" be exempt since this type of tank was not included in the storage tank
database. Commenter IV-D-27 stated that EPA should include the following exclusions to the
definition of storage tank: tanks regulated under 40 CFR 260-270 RCRA provisions; and other
waste (nonwastewater) tanks.
Response: For the final rule, the definition of storage vessels states that nonwastewater
waste tanks are not considered to be storage vessels. Tanks regulated under 40 CFR parts
260 through 270 are a subset of all nonwastewater waste tanks.
5.4 DEFINITIONS UNCHANGED SINCE PROPOSAL
Comment 1: Commenter IV-D-28 requested that the definition of "internal floating
roof should use the term "fixed roof," not "permanently affixed roof."
Response: The suggested change would not change the meaning of the definition. To be
consistent with other regulations, the definition has not been changed in the final rule.
Comment 2: Commenter IV-D-28 stated that the definition of liquid-mounted seal must
be incorrect because there should never be any liquid between the wall of the storage tank and the
floating roof.
Response: The definition states that the seal must be "mounted in contact with liquid
between the wall of the storage vessel or waste management unit and the floating roof." This
means the base of the seal, which is between the wall and the roof, is also in contact with liquid.
Therefore, the definition of liquid-mounted seal has not been changed for the final rule.
5-11
-------
Comment 3: Commenter IV-D-16 requested that the definition of "Group 1 wastewater
stream" be revised to mean either process wastewater meeting the criteria in § 63.132(c) of the
RON or the criteria in the proposed pharmaceuticals regulation, not just the HON criteria. The
commenter recommended this change as a way to exclude streams with low emission potential
from control requirements. In addition, the commenter believes that not having definitions for
"partially soluble HAP" and "soluble HAP" (i.e., terms used in the pharmaceuticals regulation)
could result in confusion and misapplications during implementation by State authorities.
Therefore, the commenter requested that EPA define these terms by listing the compounds in
each category.
Response: Applying only the criteria from the proposed pharmaceuticals regulation
would result in a different group of streams being subject to treatment requirements. Allowing
both sets of criteria would result in fewer streams being treated, and less emissions reduction
would be achieved. Because the standard as proposed was determined to be cost effective, the
EPA decided not to change the definition of Group 1 wastewater stream. The EPA also did not
add definitions for partially soluble HAP and soluble HAP because these terms are not used in
the rule.
Comment 4: Three commenters (IV-D-16, IV-D-22, and IV-D-27) stated that EPA needs
to supply a definition for "particulate HAP" or specifically identify which of the HAP listed
under CAA Section 112(b) are particulates covered under the bag dump and product dryer
standard. The commenters suggested that the definition of "particulate HAP" should be as
follows: "any air pollutant listed in or pursuant to section 112(b) of the Act which is a solid
material at ambient conditions and which exists as discrete particles over a wide range of sizes."
Response: As described in section 11.1 of this document, the standards for bag dumps
and product dryers in the final rule have been changed to be for particulate matter rather than
particulate HAP. Therefore, a definition of particulate HAP is not needed.
Comment 5: Commenter IV-D-28 requested that EPA provide a definition for the term
"planned, routine maintenance."
Response: The final rule states that in the Notification of Compliance Status report and
each Periodic report, the owner or operator must describe the maintenance that is anticipated for
5-12
-------
the storage vessel .control device during the next 6 months. The EPA believes that this provision
is clear, and that a definition of planned routine maintenance is not necessary.
Comment 6: Commenter FV-D-28 requested that the definitions of "batch process" and
"continuous process" be deleted because it is the unit operations that comprise a process that are
either batch or continuous, and both types may exist within a process.
Response: The final rule retains definitions for batch and continuous processes because
the terms are used in certain provisions. For example, the pollution prevention provisions
require the owner or operator to calculate HAP and VOC factors every 30 days for continuous
processes and every 10 batches for batch processes. However, to address the situation of a
process that consists of both batch and continuous unit operations, the final rule specifies that
such a process is considered to be a continuous process for the purposes of the pollution
prevention provisions.
Other instances where the term "batch process" was used in the proposed rule have been
changed to focus on unit operations within the process. For example, the block averaging period
for monitoring in the proposed rule was defined as equal to, at a maximum, the period from the
beginning to the end of a batch process. For the final rule, this provision has been changed to
specify that the block averaging period is equal to the time, at a maximum, from the beginning of
the first batch unit operation to the end of the last batch unit operation in a process.
Similarly, the recordkeeping provisions in the proposed rule would require the owner or
operator to keep a record of the number of batches produced during a year. For the final rule, this
provision was changed to specify that the owner or operator must keep records of the number of
batches for batch processes, the number of operating hours for continuous processes, and both
the number of batches for batch unit operations and the operating hours for the continuous unit
operations within a process that contains both types of unit operations.
Comment 7: Commenter IV-D-28 found the definition of "unit operation" in the
proposed rule to be confusing and requested that EPA change it to match the definition in the
HON. The commenter considers the definition in the HON to be clear because it refers to
equipment, and it specifies that a unit operation makes a single physical or chemical change to a
stream. The proposed rule defined a unit operation as a processing step, and did not refer to
physical or chemical changes in the stream. The commenter also explained that the term
5-13
-------
"distillation column" was replaced with the term "distillation unit" in the HON because two
columns may be designed to work together to make a single change to a stream, in which case
both columns, together, are within the same unit operation. In addition, each distillation column
has a reflux condenser, which is part of the "unit."
Response: A unit operation is an activity, not a piece of equipment. For example, unit
operations include filling, mixing, absorption, reaction, extraction, heating, distillation, washing,
decanting, filtering, and drying. The term "unit operation" is used in the definitions of "batch
emission episode," "process condenser," and "process vent" to characterize the type of activity
that causes emissions. It is the activity, not the equipment, that changes the characteristic (or
possibly multiple characteristics) of the stream and that may result in an emissions episode.
Furthermore, multiple activities with different emission levels may occur in a given piece of
equipment (e.g., filling followed by pressurization, reaction, and depressurization). Because
compliance demonstrations are based on determination of the emissions from each emissions
episode (which ait then summed as appropriate for a given process vent), EPA believes it is
reasonable to link the unit operation with the activity that causes the emission rather than the
equipment from which the emission occurs. The term "distillation column" has been replaced
with "distillation unit."
Comment 8: Commenter IV-D-28 requested that EPA add a definition of control device
consistent with the HON.
Response: The definition in the HON refers to combustion, recovery, and recapture
devices, and for process vents, recovery devices are not considered to be control devices. As
described above, the final rule considers add-on devices to be either control devices or process
condensers. Otherwise, the definition in the final rule is comparable to the definition in the
HON.
5-14
-------
6.0 STANDARDS-GENERAL
6.1 HEAT EXCHANGER SYSTEMS
Comment: Commenter IV-D-16 stated that the requirements for heat exchanger systems
should be deleted. The commenter stated that the EPA has not justified the high costs of
sampling that would be required by the proposed rule; however, if sampling is required, the
proposed reference to subpart F is the most reasonable available. Commenter FV-D-28 supports
the decision to use the HON requirements for heat exchangers, but believes the rule should
simply cross-reference the HON, not modify and spread out the requirements among the
standards, compliance, monitoring, recordkeeping, and reporting sections of this rule. If the final
rule is not revised to cross-reference subpart F, Commenter IV-D-28 requested that EPA correct
the last sentence in § 63.1362(g)(3) to be consistent with the HON. Specifically, the commenter
stated that the word "and" should be deleted from the phrase "ion specific electrode monitoring,
pH, and conductivity or other representative indicators" because only one surrogate indicator is
needed, not all of them.
Response: The EPA disagrees with the commenter's assertion that the heat exchanger
system provisions impose a high cost for sampling. The rule allows considerable flexibility in
the type of sampling or other monitoring that an owner or operator may perform, and the amount
of required sampling or monitoring is minimal. The owner or operator may elect to sample for
one or more HAP or other substances whose presence in the cooling water indicates a leak.
Alternatively, the owner or operator may elect to monitor for any surrogate indicator that reliably
identifies the presence of a leak. If the owner or operator elects to comply by monitoring for a
surrogate indicator, the owner or operator must develop a plan that specifies what parameter or
condition will be monitored, the level that constitutes a leak, and an explanation of how the
selected parameter or condition will reliably identify a leak. In the first year, sampling or
6-1
-------
monitoring is required eight times; in subsequent years, sampling or monitoring is required only
four times per year. If the heat exchangers are all part of a single system, only one set of inlet
and outlet samples are required. These requirements also are not considered burdensome
because many facilities in the chemical processing industry, and presumably the PAI production
industry as well, conduct such sampling or monitoring as a common maintenance practice.
Furthermore, sampling for the detection of heat exchanger system leaks is a general requirement
of some State permits (e.g., Texas Natural Resources Conservation Commission).
The Agency agrees with the comment that cross referencing the heat exchanger
provisions in subpart F of the HON would simplify the rule. Therefore, as noted in the response
to the first comment in section 20.3 of this document, the final rule cross-references all of the
heat exchanger system provisions in subpart F rather than incorporating some of the provisions in
the rule and cross-referencing others. However, the heat exchanger system provisions are
contained in more than one section in the PAI rule because this rule is structured differently from
the HON. In the HON, all of the requirements for a specific type of emission point were
presented in a single section or in consecutive sections. In the PAI rule, the standards for all
types of emission points are presented in one section, the initial compliance provisions for all
types of emission points are presented in the next section, and so on. Therefore, each section in
the final rule cross-references the appropriate heat exchanger system provisions from subpart F.
This change also corrects the editorial error that one commenter identified.
6.2 OUTLET CONCENTRATION STANDARDS
Comment 1: Commenters IV-D-16 and IV-D-27 requested that the 20 ppmv outlet
concentration standards be the same for all existing and new source process, storage tank, and
wastewater vents and that they be based on both TOC and total organic HAP. The commenters
requested this clarification because § 63.1364(d)(5) of the proposed rule only used organic HAP,
and § 63.1364(c)(l)(viii) only used TOC. The commenters also asked for clarification that TOC
excludes methane and ethane. Commenter W-D-27 requested that the 20 ppmv concentration
apply to both uncontrolled and controlled vents.
Response: The Agency agrees with the commenters that the 20 ppmv compliance option
should be the same for all emission streams at both new and existing sources. In the final rule,
the control device (or last control device in series) must achieve an outlet, undiluted TOC
6-2
-------
concentration of 20 ppmv or less, as calibrated on methane or the predominant HAP. The
reference to organic HAP in § 63.1364(d)(5) of the proposed rule was an error. The intent at
proposal was that the standards would be based on TOC for all emission points; in addition, for
wastewater emissions, the standard included total organic HAP as well as TOC because the
wastewater provisions were based on the HON. To make the standards consistent for all
emission points, the final rule states that when the phrase 'TOC concentration or total organic
HAP concentration" is used in §§ 63.139 and 63.145, the phrase 'TOC concentration" applies for
the purposes of the final rule.
The final rule also includes the option to subtract out methane and ethane from the TOC
concentration. Though sources have this option, EPA believes the use of a continuous TOC
analyzer, such as a FID, will be complicated by having to subtract out methane and/or ethane.
Because the inclusion of methane and ethane, which should only be significant for combustion
devices, will only yield a higher TOC value, the simplification of the method to use a continuous
FTD without such a correction is allowed for all devices.
Lastly, the final rule also specifies that uncontrolled emission streams containing less than
20 ppmv HAP are not considered to be process vents subject to the control requirements in the
rule. This provision is incorporated in a revised definition of "process vent," as described in the
response to comment 8 in section 5.2.
Comment 2: Commenter FV-G-03 contends that the standards are arbitrary and
capricious because smaller producers will be forced to comply with the 98 percent reduction
requirement, whereas larger producers are more likely to be able to comply with the "volume
limitation," which the commenter believes is less stringent. Commenter FV-D-17 stated that the
limits proposed in the rule are readily achievable using proven, commercially available
technologies.
Response: The EPA does not understand what the commenter means by the volume
limitation, or why small producers are at a disadvantage to large producers. The process vent
standards require either 98 percent reduction or control to an outlet concentration of 20 ppmv for
individual vents that meet certain flow rate and HAP load thresholds. New source standards also
specify either 98 percent reduction or control to an outlet concentration of 20 ppmv. Possibly the
commenter believes the 20 ppmv outlet concentration option is less stringent than the 98 percent
6-3
-------
reduction requirement. This is true if the uncontrolled concentration is less than 1,000 ppmv.
However, the uncontrolled concentration is a function of the individual production process,
which may be operated by either a small producer or a large producer. Furthermore, the rule
prohibits intentional dilution as a means of achieving this outlet concentration standard. Finally,
the 20 ppmv limit is not intended to be a less stringent option; it is simply a recognition that
20 ppmv is the practical limit of control for control devices.
6.3 COMPLIANCE OPTIONS
Comment 1: Commenter FV-D-28 requested that the final regulation allow owners and
operators to comply by routing emissions to a fuel gas system or to a process, as in the HON.
Commenters P/-D-21 and IV-D-29 requested that EPA exempt from regulation all gaseous
streams routed to fuel gas systems by changing the definition of process vent to match the HON.
Response: The EPA has not changed the final rule to allow compliance by routing
streams to fuel gas systems. Under the HON, a fuel gas system is defined as a combustion
device. One common type of combustion device is a boiler. The final rule already contains
requirements for boilers as control devices. Therefore, to avoid the confusion of whether the
owner or operator should comply with requirements for boilers or fuel gas systems, the final rule
contains provisions only for boilers. However, for large boilers (>44 megawatts [MW]), there
are no performance testing or monitoring requirements, which is the same outcome as allowing
compliance via fuel gas systems.
The HON allows emissions from storage vessels to be routed to a process. Specifically,
the owner or operator must conduct a design evaluation to demonstrate that such emissions
predominately meet one or more of the following ends: recycled and/or consumed in the same
manner as a material that fulfills the same function in that process, transformed by chemical
reaction into a material that is not a HAP, incorporated into a product, or recovered. The EPA
agrees that routing emissions from a storage vessel to a process is also acceptable for storage
vessels in the PAI source category, and the final rule has been revised accordingly. However, the
provisions in the PAI final rule differ from the HON in two ways. First, the recovery option is
not included because, as noted in the response to the comment in section 3.4, the concept of
recovery as used in the HON is not used in the PAI final rule. Second, the design evaluation
6-4
-------
must demonstrate that less than 5 percent of the HAP routed to the process are ultimately
emitted; otherwise, routing to a process is less stringent than routing to a control device.
Comment 2: Commenter IV-D-28 stated that if emission streams from different types of
vents (i.e., process vents, storage vessels, and wastewater) are all manifolded to a common
device, the rule should specify which requirements take precedence. For example, the
commenter stated that under the HON, the owner or operator may comply with the first
applicable set of requirements from the following hierarchy: process vents, transfer racks,
storage vessels, waste management units, and in-process equipment subject to § 63.149.
Response: For situations in which different streams are manifolded into the same control
device, the owner or operator would have to demonstrate that the device meets the most stringent
control efficiency applicable to the emission sources. In the HON, the control level required for
process vents and transfer racks was highest, at 98 percent, followed by 95 percent control for
storage and wastewater emission sources. Therefore, in a situation where all types of emission
streams would be manifolded to a device, the "hierarchy" would be according to process vents
and transfer racks, followed by storage and wastewater, because process vents and transfer racks
require higher control efficiencies.
6.4 ROUTINE MAINTENANCE
Comment: Several commenters (IV-D-16, IV-D-21, F/-D-27, IV-D-28, and IV-D-29)
requested either an extension in the 240 hr/yr allowance for routine maintenance or greater
flexibility in its application. Commenter IV-D-27 suggested that EPA allow up to a 30-day
extension for control devices (like RCRA incinerators) that require more than 10 days of
maintenance per year, or allow a facility to compensate for longer downtime by overcontrolling
at other times (this would also require a change in the compliance averaging period-see
comments in section 7.5 of this document). In addition, commenters IV-D-21 and IV-D-29
recommended that the 240 hr/yr be allowed for each PAI process unit because maintenance may
be required prior to each campaign. Alternatively, commenter IV-D-29 suggested that, based on
standard maintenance work practices, the startup, shutdown, and malfunction requirements in
subpart A of part 63 should be allowed in lieu of the proposed 240 hr/yr time period.
Commenter IV-D-29 explained that the standard work practice for many companies is to isolate
all equipment upstream of control devices where planned maintenance will occur to eliminate all
6-5
-------
safety hazards to personnel and minimize any impact to the environment. Commenter IV-D-28
supported the provision, but suggested it be expanded to cover controls for waste management
units, controls used on equipment leaks, and recovery devices (if applicable).
Response: The proposed 240 hr/yr for planned routine maintenance was mistakenly
applied to all control devices in the proposed rule; it should only have been applied to storage
vessels. The startup, shutdown, and malfunction provisions prohibit the shutdown of control
devices during operation; however, EPA recognizes that for storage vessels, it is impossible to
"not operate" (i.e., not have breathing losses) during a period of time in which an add-on control
device would be undergoing planned maintenance. Therefore, EPA has in the final rule allowed
an amount of time (240 hr/yr) in which the control devices for storage vessels only can be non-
operational due to planned routine maintenance. All other situations (i.e., those that require
unplanned, emergency maintenance) should be addressed through the startup, shutdown, and
malfunction provisions. This change makes the final rule consistent with other MACT standards.
The rationale for the 240 hr/yr allowanceis that EPA determined that routine maintenance for
certain control devices may take as much as 10 days to complete, and that this time frame is
consistent with State permitting activities (see 59 FR 19441 for a more detailed discussion of the
time allowance).
6.5 MACT AND MACT FLOOR
Comment: Commenter IV-D-28 stated that the 20 surveyed plants represent only
26 percent of the estimated number of affected sources; thus, it is possible that EPA may not
have collected data representing the best controlled source, the best controlled 12 percent of
existing sources, and the impacts of going beyond the floor for each of the initial source
categories and the combined source category. The commenter suggested that it is possible that
EPA has no data at all on many of the source categories that were combined in the proposed rule.
Commenter IV-D-28 also stated that the applicability thresholds for process vents and storage
vessels are tiny. The commenter had not determined whether the thresholds were for the floor or
options more stringent than the floor, but either was a concern. For example, if they are for the
floor, the commenter believes it is likely that they are from facilities that are not major sources
due to PAI operations alone and if so, the floor should not be based on them. Alternatively, if
they are for options more stringent than the floor, the commenter believes the resulting cost
6-6
-------
effectiveness values must be in a range that has been considered unacceptable for other
standards.
Response: Some of the above comments imply that the database that was used to
determine the MACT floor is not representative of the industry. The EPA acknowledges that
data are not available for every process included in the source category, nor for every affected
source. However, EPA believes the processes at the surveyed plants are representative of all
processes in the source category, that the surveyed plants include the best controlled plants in the
source category, and that using the data from the surveyed plants to develop the MACT floor is
appropriate.
The response to the comment in section 3.1 describes a few changes that have been made
since proposal to exclude from the source category some processes that are not representative of
those that were surveyed. Because the remaining processes are believed to have similar product
use, process unit operations, types of HAP emitted, emission rates, and controls, they were
combined in a single source category. As a result, separate evaluations of each process (or each
initial source category) are unwarranted.
Although only a subset of the estimated total number of affected facilities in the entire
source category were surveyed, the MACT floor was based on actual data obtained from a
number of facilities, nine, that is equal to 12 percent of the total number of affected facilities in
the source category. Even if data from the entire source category had been collected, the MACT
floor dataset would still consist of nine facilities. In addition, because EPA strived to identify
and survey the best controlled sources, as described in the Basis and Purpose document, EPA
believes that the surveyed facilities include both the best controlled source and the best
performing 12 percent of existing sources. The EPA also rejects the notion that facilities whose
PAI processes alone do not exceed the major thresholds should not be part of the MACT floor.
The EPA's implementation of the definition of major source, which requires the aggregation of
all HAP emissions within a plant site, was affirmed in a July 21, 1995 D.C. Circuit Court
decision (National Mining Association vs. U.S. Environmental Protection Agency), where NMA
had challenged EPA's definition of a major source. Thus, EPA believes the MACT floor
developed from the survey data is representative of the source category.
6-7
-------
Finally, the standard was based on regulatory alternatives more stringent than the MACT
floor only when the cost was determined to be reasonable. For storage vessels, the standard is
based on the MACT floor because no alternative was determined to be cost effective. For
process vents at existing sources, the standard is more stringent than the floor for certain large
vents for which control at 98 percent was determined to be cost effective. The standard for all
other process vents at existing sources and at new sources is based on the MACT floor.
6.6 CLARIFICATION
Comment: Commenter F/-D-28 noted that § 63.1362(a) of the proposed rule
requires control of HAP emissions to the level specified in Table 2 and paragraphs (b) through
(g). The commenter stated that because these are not different sets of requirements, it would be
better to require control of HAP emissions "to the levels specified in paragraphs (b) through (g),
as summarized in Table 2.
Response: The EPA agrees with the commenter, and has made the suggested change in
the final rule.
6-8
-------
7.0 STANDARDS-PROCESS VENTS
7.1 MACT AND MACT FLOOR
Comment 1: Three commenters (IV-D-26, IV-D-20, and W-D-27) requested that sources
be able to use vents meeting the criteria for 98 percent control in determining 90 percent overall
process control requirements. Commenters stated that EPA determined that the MACT floor was
90 percent on a process wide basis. Commenter IV-D-16 also indicated that vents required to be
controlled to 98 percent should be used in establishing the overall 90 percent reduction for the
process because the data analysis supporting the MACT floor for existing sources (90 percent
reduction requirement for process vents) is based on control achieved on all process vents by the
best controlled 12 percent of all processes. According to commenter FV-D-27, excluding these
vents increases the stringency of the floor.
Response: The MACT floor was determined to be 90 percent control for process vents at
existing sources. In addition to the MACT floor, the EPA is required to develop regulatory
alternatives beyond the floor and to select MACT based on the cost effectiveness of these
alternatives. For specific process vents, Regulatory Alternative 1 would require 98 percent
control efficiency, i.e., those vents that meet the flow and annual uncontrolled emissions criteria
described in § 63.1362(b)(2)(iii) of the final rule, and would require 90 percent control efficiency
for the sum of emissions from all other vents within the process. The cost of Regulatory
Alternative 1 was judged to be acceptable, and this alternative was selected as MACT. The EPA
agrees that this requirement is more stringent than the floor. If a vent that must be controlled to
98 percent is included in determining 90 percent control for all process vents within the process,
the owner or operator would only be complying with the MACT floor, not the more stringent
regulatory alternative. Thus, the final rule does not allow an owner or operator to use process
7-1
-------
vents that are subject to the 98 percent control requirement in determining compliance with the
90 percent overall control level for the sum of other process vents in the process.
Comment 2: Commenters IV-D-21 and IV-D-29 stated that the data used to set the
98 percent level of performance was obtained during off-line tests during which liquids (for the
most part hazardous waste surrogates) were burning during continuous injection. Furthermore,
the commenters stated that there were no data for the efficiency of combustion of dilute
concentrations of gaseous HAP streams emitted from batch processes in a discontinuous fashion.
The commenters believe that EPA has insufficient data to support the establishment of a PAI
batch process combustion control equipment MACT floor. Commenters suggested that EPA
should reexamine this control requirement, and determine a true MACT floor based on testing
that is conducted on actual batch processes, over the typical 1-hour batch cycle times, with
gaseous vent line concentrations ranging from 100 percent HAP down to 1 Ib/yr HAP.
Response: Some of the surveyed PAI manufacturing facilities reported control
efficiencies for RCRA incinerators that were based on performance tests conducted for liquid
burning, and these facilities all reported control efficiencies that exceeded 98 percent. Although
these data were not from tests conducted when only gaseous streams were incinerated, a
considerable amount of performance testing for thermal incinerators has been conducted over the
years. Based on these tests, EPA has concluded that incinerators that operate above certain
temperatures and with certain residence times consistently achieve 98 percent emission reduction
or, if inlet concentrations are low, to an outlet concentration of no more than 20 ppmv. Typical
operating temperatures for the RCRA incinerators in the PAI data base are at least 1800°F and
residence times are in the range of 1 to 3 seconds. Because these temperatures and residence
times are higher than would be expected for a typical thermal incinerator, it is expected that these
RCRA incinerators also achieve at least 98 percent for process vent streams. Therefore, EPA
believes the 98 percent control efficiency assigned to RCRA incinerators in the MACT floor
analysis are reasonable.
Comment 3: Commenters FV-D-16 and IV-D-27 noted an inconsistency that should be
resolved. The commenters pointed out that in the proposed standards, integral intermediate
processes are combined with PAI processes to define a single "process," but they were evaluated
separately in the MACT floor analysis. Commenter IV-D-27 further noted that this change
7-2
-------
would result in an increase in the applicability cutoff of the MACT floor because part of the
emissions from intermediate process No. 36 should be combined with the emissions from active
ingredient process No. 35 that had the lowest uncontrolled emissions and was used to establish
the applicability cutoff of 0.15 Mg/yr.
Response: The response to the comment in section 3.3 describes changes in the final rule
to clarify the definitions of "integral intermediate" and "process." That response explains that
the intent in the proposed rule was to consider production of integral intermediates to be separate
processes. As the commenters noted, this is also the approach used to develop the MACT floor.
As a result, EPA has not changed the basic approach used to develop the MACT floor. However,
in re-examining this approach since proposal, EPA realized that some of the active ingredient
processes at the surveyed facilities included production of intermediates; in addition, some of the
reported intermediate processes may satisfy one of the criteria for storage and thus not be integral
intermediates. If all of the intermediates are integral intermediates, the floor would increase to
92 percent. If none of the intermediates are integral intermediates, the floor would decrease to
88 percent. Thus, EPA considers the proposed floor of 90 percent control to still be appropriate.
The applicability cutoff also is unchanged because the active ingredient production and
intermediate production are not combined into a single PAI process unit.
Comment 4: Commenter IV-D-15 stated that the control technologies at their facility
were "taken out of context" and inappropriately used in the MACT floor calculation. The
commenter stated that their intermediate process is not representative of other processes within
the PAI industry; in addition, it may not be an "integral intermediate," depending on what
"storage" means (the intermediate could be stored from several hours to several days prior to the
active ingredient process). The commenter stated that the regulation would require burdensome
administrative requirements that do not result in emission reductions, and the equipment required
for control of volatile organic HAP is prohibitively expensive ($48,500/ton of HAP removed).
Response: While the commenter indicates that their intermediate and active ingredient
processes are not "representative," these processes are located at a MACT floor facility (i.e., the
top 12 percent of sources). The commenter's basic concern for the active ingredient process is
that both gaseous HAP and particulate HAP are emitted from the process, and paniculate HAP is
well controlled. Good control of particulate HAP at this plant provides a high overall control
7-3
-------
efficiency and is the factor that makes it a MACT floor plant. While the control efficiency for
paniculate HAP is high, it is not included in the MACT floor calculation for process vents for
organic HAP. It is also important to remember that the presence of multiple planks for a process
(i.e., process vents, storage tanks, equipment leaks, wastewater, and product dryers) varies
throughout the industry from process to process. For example, some processes emit HAP from
process vents and wastewater but not from the other three planks. In addition, use of product
dryers in the PAI industry is common. At the commenter's facility, the product dried is a HAP
compound, and HAP is emitted from the process and controlled; these uncontrolled and
controlled emissions would be included in the overall HAP emission calculation at the facility
just like any other plank. At other facilities, however, the product dried was not a HAP and there
were no HAP emissions from this process step. See chapter 6.5 for additional discussion.
The commenter's concerns regarding the intermediate process are related to the fact that
it is a single vent with low flow that is controlled with a flare. While this may not be
representative of all processes in the industry, it is important to remember that this intermediate
process in and of itself is not the basis for the MACT floor for existing source process vents, but
was included in the average control efficiency calculation along with other processes at the
MACT floor facilities. The commenter expressed concern regarding the control costs not being
representative either (i.e., a flare is a less expensive control device than some other devices and is
not technically feasible for all processes). However, flare costs were not estimated for control of
process vents, and the cost to control this particular process vent was not determined directly
anyway. Model streams were developed based on "average" streams in the industry that are
subject to the standard, and the cost to control these model or average streams was estimated for
thermal oxidizers and condensers. See chapter 19 for additional discussion of costs and impacts.
7.2 APPLICABILITY CUTOFFS
Comment I: Commenters IV-D-16, IV-D-20, and IV-D-27 requested that an
uncontrolled emission concentration cutoff of 50 ppmv be established for existing and new
sources (to be demonstrated on either a HAP or a TOC basis). According to
commenter IV-D-27, this could be done by revising the definition of Group 1 process vent or by
creating an exemption in § 63.1360(d)(4). According to commenter IV-D-16, this would be a
powerful incentive for some sources to fully implement pollution prevention practices to achieve
7-4
-------
compliance. Commenters FV-D-16 and IV-D-27 indicated the cutoff (or Group 1 process vent
definition) should also include a cutoff of 0.005 standard cubic meters per minute (scmm).
Commenters F/-D-21, IV-D-27, and IV-D-29 stated that to be consistent with the RON,
EPA should set an applicability cutoff based on "total resource effectiveness" that would provide
incentives for sources to implement pollution prevention practices (such as product recovery
devices) in order to avoid triggering process vent standards. Commenters FV-D-21 and IV-D-29
recommended that the flow rate and concentration cutoffs from the HON process vent definition
be used.
Response: Under the proposed rule, the applicability cutoff of 0.15 Mg/yr of HAP per
process was contained in the definition of Group 1 process vent. For the final rule, the definition
of "process vent" was revised to exclude streams containing less than 20 ppmv HAP, and the
definition of "Group 1 process vent" was revised to include a mass cutoff of 6.8 Mg/yr for total
hydrogen chloride and chlorine emissions as well as the 0.15 Mg/yr cutoff for organic HAP
emissions. The rationale for changes to these definitions is provided in the response to
comment 8 in section 5.2.
It is not clear that the higher cutoffs suggested by the commenters would lead to increased
implementation of pollution prevention practices (and it certainly would not result in increased
compliance with the pollution prevention alternative in the rule). However, higher cutoffs would
result in lower overall control efficiencies, which would be inconsistent with the MACT floor.
Finally, as noted in the response to the comment in section 3.4, the concept of recovery devices,
as used in the HON, was not intended to be in the proposed rule and is not in the final rule.
Comment 2: Commenter FV-D-15 suggested a higher threshold level (greater than or
equal to 0.15 Mg/yr) needs to be developed either for a process as a whole or for the individual
process entities that make up the commenter's Captan process. Commenters FV-D-16 and
FV-D-20 stated that the process mass emissions cutoff should be set at 2,000 Ib/yr and be
applicable to both controlled and uncontrolled emissions. The commenters stated that to account
for the limited nature of the PAI data base, EPA should use 2,000 Ib/yr process cutoff that was
established for the proposed Pharmaceutical MACT standards. In addition, the commenters
stated that a 330 Ib/yr cutoff is far lower than the 10,000 Ib/yr cutoff that was determined to be
the cost effective cutoff in the Batch Processes Alternative Control Techniques (ACT) document.
7-5
-------
Commenter IV-D-27 stated that the applicability cutoff for process vents should be set at
10,000 Ib/yr (or higher) as was done in the Batch Processes ACT document; this is the minimum
value for which 90 percent control was determined to be cost effective. The commenter noted
that in many cases, controls on processes with small HAP emissions were added for odor control
or VOC's. The commenter disagreed with EPA's assertion during Partnership group meetings
that the CAA does not allow the Agency to consider the reason controls were added.
Specifically, the commenter states that there is no statutory limitation on how EPA defines the
"affected source"; for example, EPA has already provided exclusions in § 63.1360(d), and a
higher applicability cutoff could be another. Commenters FV-D-16 and FV-D-28 claim that the
lack of meaningful cutoffs for existing and new source process vents will discourage pollution
prevention (because if pollution prevention doesn't reduce emissions below cutoffs, it will be
become more difficult to meet the required percent reduction from the lower starting point) and
result in disproportionate impact on process vents with extremely small emissions; other MACT
standards exclude low flow and low HAP to mitigate this concern.
Response: The EPA attempted to collect information on the best controlled facilities in
the PAI industry; EPA believes that the best controlled facilities are contained in its PAI data
base and that the processes contained in the data base are representative of the industry. Based
on the PAI data base, many processes with uncontrolled emissions that were significantly less
than the cutoffs mentioned by the commenters were controlled to levels of 90 percent or greater.
Because the emission cutoffs mentioned by the commenters were not supported by the process
vent data, these cutoffs would not have been defensible because they would have been less
stringent than the cutoff prescribed by the MACT floor.
The commenters indicate that the cutoff determined at proposal is not cost effective and
suggested other cutoffs that have been demonstrated as cost effective. However, there is no
provision in the amended CAA for consideration of cost effectiveness in setting the MACT floor.
Therefore, it is conceivable that the standards, which are set based on the practices of the
industry, will require a level of control that is greater than what was determined to be cost
effective for other CAA programs. For example, the 10.000 Ib/yr cutoff contained in the draft
Batch Processes ACT that was referenced by the commenters was intended to simplify
applicability of presumptive Reasonably Available Control Technology (RACT) control
7-6
-------
measures, which are applied to the reduction of criteria pollutants (in this case, VOC's), and can
include the consideration of cost effectiveness.
Finally, the amended CAA also contains no provisions for considering reasons why
certain processes are controlled and others are not in assembling the group of sources that will
make up the best 12 percent of the source category. Therefore, the issue of facilities controlling
HAP's for odor control or other purposes is not a consideration in setting the floors.
Comment 3: Commenter IV-D-28 also noted that, unlike other regulations, many of the
process vent provisions apply to a collection of vents rather than to individual vents. The
commenter is unsure whether the collective approach will work and stated that the rule should
also include an option that allows for the determination of Group status for each individual vent
and control of emissions on a vent-by-vent basis.
Response: The Group 1 process vent emissions cutoff (i.e., 0.15 Mg/yr) was determined
based on the uncontrolled HAP emission level from the collective or entire process, not on an
individual vent basis. The rule allows an applicability determination for an individual process
vent based on a concentration cutoff; emission streams containing less than 20 ppmv HAP are
not considered process vents.
Comment 4: Commenter IV-G-01 asserted that the applicability equation found in
§ 63.1362(b)(2)(iii)(A) used to determine which vents must be controlled to 98 percent is
inappropriately applied to batch operations. The commenter explained that the flow rate used in
the computer model to develop the 98 percent applicability regulatory alternative (Batch
Processes ACT) is a constant flow rate, which is inconsistent with batch processing.
Response: In the Batch Processes ACT, EPA developed costs for an incinerator to
estimate the cost effectiveness of controlling emissions from batch process vents. Although flow
rates from batch processes vary, the control device must be capable of handling the maximum
flow rate possible. Therefore, the incinerator was sized and costed for the maximum flow rate
that occurs from the process, even though venting from batch processes will include periods of
lower flow rates.
7.3 HC1 STANDARDS
Comment 1: Commenters FV-D-21 and FV-D-29 are concerned that EPA's approach to
determining the MACT floor for the HC1 emission limit criteria (e.g., the 6.8 Mg/yr cutoff) in
7-7
-------
§ 63.1362(b)(3) of the proposed rule considers only a limited number of process vents emitting
HC1 which may not be representative of the entire source category. The commenters recommend
that EPA consider setting the HC1 cutoff for existing sources at least as high as the average of the
two lowest HC1 emission rates from controlled processes at the MACT floor facilities (i.e.,
(6.8 Mg/yr + 11.0 Mg/yr)/2 = 9.0 Mg/yr), or that the control device for the process vent emitting
HC1 meet a minimum 90 percent efficiency if installed and in operation before November 7,
1997. (Note: EPA assumes the commenter means the proposal date of November 10, 1997.)
The commenters believe these changes will improve incentives for pollution prevention, and
allowing 90 percent control would reduce the cost burden on existing facilities because
retrofitting to achieve an incremental improvement in control is very expensive.
Response: The EPA disagrees with the commenter that the proposed cutoff for HC1
emissions is inappropriate. As described in the Basis and Purpose document and summarized
below, EPA believes the cutoff of 6.8 Mg/yr is a very clear and obvious breakpoint. Also, even
though the MACT floor plants have fewer processes with HC1 emissions than organic HAP
emissions, this is representative of the industry as a whole. Thus, one would expect that the HC1
floor would be based on less data than the floor for organic HAP emissions. The EPA also notes
that if the floor were determined by evaluating the best controlled processes throughout the
industry rather than the processes at the best performing 12 percent of existing facilities that the
applicability cutoff might be lower than 6.8 Mg/yr. It certainly would not be higher.
To develop the MACT floor for the proposed rule, all of the processes at the nine MACT
floor facilities were ranked by uncontrolled HC1 emissions. All processes with uncontrolled
emissions below 6.8 Mg/yr were uncontrolled, and processes with higher emissions were
controlled to various levels. Therefore, the MACT floor was determined to be no control for
processes below this threshold, and 94 percent for processes above it.
The EPA believes there is no basis for setting a cutoff at 9.0 Mg/yr or for setting a control
level of 90 percent for control devices installed before November 10, 1997. Because the MACT
floor consists of both a control efficiency and a cutoff, the cutoff cannot be changed
independently of the control efficiency. A cutoff of 9.0 Mg/yr would be inappropriate because it
is not associated with the determined MACT floor control efficiencies. Furthermore, it would
not make sense to include one controlled process (i.e.. the process with emissions of 6.8 Mg/yr)
7-8
-------
in the group with all of the uncontrolled processes; this is a very clear and natural cutoff. If the
standard were based on an alternative more stringent than the floor, the rule might allow
emission points that are already controlled to the level of the MACT floor to comply with that
level (as was done for organic emissions from process vents). However, there is no basis for a
90 percent control level, regardless of the installation date, because the 94 percent control level
for HC1 is the MACT floor. Finally, the EPA recognizes that the incremental cost effectiveness
will be high for a facility with control just below the required level. However, this would be true
no matter where the level was set.
Comment 2: Cornmenters IV-D-16 and IV-D-27 stated that the HC1 standards for new
sources should be set at 99 percent removal for consistency with the HON requirements.
Commenter FV-D-27 stated that since there is no actual test data from the pesticide
manufacturing industry demonstrating a 99.9 percent removal of HC1; a change to 99 percent
would provide consistency with HON rule requirements.
Response: The EPA agrees with the commenters. The proposed control level was based
on a value reported by a surveyed facility. This value was not supported by test data or other
documentation. However, a control level of at least 99 percent is likely for this scrubber because
HC1 control levels of 99 percent are widely accepted as achievable by scrubbers, and several
other facilities reported this level. Therefore, for the final rule, the required control level for new
sources has been changed to 99 percent. Although being consistent with the HON is not a
priority, this change, as one commenter observed, does make the two rules consistent.
7.4 COMPLIANCE AVERAGING PERIOD
Comment: Commenter FV-D-27 stated that the process vent standards should be based
on "annual average" removals to reflect the data used in developing the standards. The
commenter stated that these data were annual values that mask variability in uncontrolled
emissions (especially for batch processes) and control device operating parameters, thus, it is
unreasonable to expect that the annual efficiency can be met on a daily basis.
Response: The EPA believes that the data used to develop the MACT floor are consistent
with compliance determinations over 24-hour periods (or, as allowed in the final rule, the
duration of a batch process). One consideration is that emission stream conditions from batch to
batch should be essentially the same. Even if temperatures, reaction times, purge flows, or other
7-9
-------
operating characteristics varied slightly from one batch to the next, EPA assumes that the
reported annual uncontrolled emissions for the MACT floor facilities were calculated, quite
logically, assuming identical emissions for each batch of a given process. A second
consideration is the reported control efficiencies for most of the control devices were values that
likely are consistent from day-to-day. For example, many of the processes are controlled with
combustion devices, which operate with a consistent reduction efficiency. For other devices
(e.g., a condenser), it is likely that the efficiency varies for different episodes, but if emissions for
a given episode are consistent from one batch to the next (as described above), the control
efficiency for that episode should also be consistent, assuming the control device operating
parameters are consistent. There may be a few situations where control parameters vary for a
given emission episode over the course of a year (e.g., if a control device is used only part time
or if the vents manifolded to the control device change as the mix of processes change). At the
MACT floor plants, several processes are known to be subject to part-time control, but EPA has
no evidence of changes in control settings for different combinations of manifolded vents.
Processes subject to part-time control that have average control efficiencies below the level of
the standard are no different from processes with a consistent control level below the level of the
standard; both must implement additional control. The response to the comment in section 16.1
presents rationale for monitoring on a daily or block basis.
7.5 OVERLAP WITH OTHER STANDARDS
Comment: Commenter FV-D-27 requested that EPA specifically state that regulated
emissions from pesticide manufacturing process vents that are routed to Resource Conservation
and Recovery Act (RCRA) incinerators are deemed to comply with the standards of this rule with
no additional performance testing, monitoring, recordkeeping, or reporting requirements.
Response: The proposed rule exempted an owner or operator from performance test
requirements for RCRA control devices, and this provision has been retained in the final rule. In
addition, the final rule exempts the owner or operator from monitoring, recordkeeping, and
reporting provisions when emissions are routed to a RCRA control device.
7.6 CLARIFICATIONS
Comment 1: Commenter IV-D-28 asserted that § 63.1362(b)(2)(iii)(B) of the proposed
rule allows an owner or operator to demonstrate that a control device was "designed" to reduce
7-10
-------
organic HAP emissions by between 90 and 98 percent; however, the commenter suggests that the
focus should be on the reduction the device is capable of achieving instead of its design. The
commenter has a similar concern on § 63.1362(c)(l)(ii) of the proposed rule.
Response: The Agency agrees with the commenter regarding the term "design" of a
control device. This subparagraph in both § 63.1362(b) for process vents and (c) for storage
tanks has been revised to include "If the owner or operator can demonstrate that a control device
installed on a process vent subject to the requirements ... on or before November 10,1997
reduces inlet emissions of total organic HAP by greater than or equal to 90 percent but less than
98 percent..."
Comment 2: Commenter FV-D-16 stated that the flowrate equation in
subparagraph 63.1362(b)(2)(iii)(A) of the proposed rule should be revised to contain a "less than
or equal to" sign rather than an "equal to" sign to agree with the regulatory text.
Response: The equation in § 63.1362(b)(2)(iii)(A) of the proposed rule for the flow rate
is correct. For the 98 percent requirement to apply, the actual weighted average flow rate for the
vent must be less than 01 ~qual to the flow rate calculated using this equation. In the final rule,
an additional equation for determining the weighted average flow rate for the vent has been
added.
7.7 REQUIREMENTS FOR SURGE CONTROL VESSELS AND BOTTOMS RECEIVERS
Status at proposal. Under the proposed rule, emissions from surge control vessels and
bottoms receivers would be regulated as process vents. This approach differs from the HON,
which regulated these emissions under subpart H. Because the equipment leak standards in the
proposed rule also cross-referenced subpart H, the proposed rule specified that the provisions for
surge control vessels and bottoms receivers in subpart H would not apply.
Comment: Commenter FV-D-28 opposes the proposed requirement to regulate surge
control vessels and bottoms receivers as process vents because it introduces a third way to
regulate such emissions under the MACT standards. The commenter would prefer that these
emissions be regulated as equipment leaks, as under the HON. If that is not acceptable, the
commenter's second choice is to regulate the emissions as storage vessels, as under P&R IV.
The commenter believes that additional inconsistency is confusing and likely to lead to
inadvertent compliance mistakes.
7-11
-------
Response: The EPA notes that there is essentially no difference between regulating
emissions from these equipment as "equipment leaks" (as in subpart H) versus as "storage tanks"
(as in subpart G). Both the applicability and control requirements for these sources in the HON
are identical. The reason EPA departed from this rationale in the proposed rule, is that surge
control vessels and bottoms receivers typify the processing equipment, in capacity and function,
found in these industries. Especially in the case of batch processing (where the HON does not
regulate process vents), the characteristics of emission streams from these equipment are not
significantly different than any other equipment. Emission streams from bottoms receivers and
surge control vessels result from the displacement of saturated gases from incoming materials.
Displacement emissions are very common in both the Pharmaceuticals and this industry.
Therefore, EPA decided to regulate them in a manner consistent with the remainder of processing
equipment found in these industries.
In response to the commenter's concern about possible confusion from the inconsistent
application of requirements across different source categories, EPA believes that the consistent
treatment described above will actually eliminate a great deal of confusion in the implementation
of the rule, because all equipment associated with a process will be treated in the same manner,
and the control requirements, which are process based, can be evaluated over all equipment in the
process. Additionally, because of the similarities of these equipment with other process vessels,
the confusion related to defining a surge control vessel or bottoms receiver from another process
vessel will also be averted.
7-12
-------
8.0 STANDARDS-STORAGE VESSELS
8.1 STORAGE VESSEL MACT FLOOR
Status at proposal: Under the proposed rule, the MACT floor for storage vessels
consisted of applicability cutoffs and a control efficiency for vessels that exceeded the cutoffs.
To develop the floor, the storage vessels at the best performing 12 percent of facilities were
ranked by decreasing uncontrolled emissions. The vessels were divided into two groups based
on an uncontrolled emissions cutoff below which the median control efficiency was no control.
This cutoff was 108 kg/yr, and the median control above the cutoff was 41 percent. A vessel size
cutoff was established at 38 m based on the smallest vessel with uncontrolled emissions greater
than 108 kg/yr that was controlled at least to 41 percent. For new sources, the smallest vessel
with the best level of control was determined. The floor for new sources was determined to be
98 percent control efficiency for storage vessels 26 m3 or greater with uncontrolled emissions of
at least 0.45 kg/yr.
Comment: Commenter F/-D-27 stated that the control levels originally provided by the
commenter for Tank Nos. 10 and 15 are inaccurate due to incorrect coolant temperatures used by
the commenter. The commenter stated that the impact of this change is that the existing source
MACT floor based on the median control level for vessels with greater than 240 Ib/yr of
uncontrolled emissions becomes 21 percent (now based on Tank No. 10 instead of Tank No. 15),
instead of 41 percent. Commenter IV-D-14 stated that the new source MACT floor should be
revised to include consideration of vapor pressure of the stored HAP to be a primary parameter.
Response: The EPA has corrected the control efficiencies for each of the storage vessels
mentioned by the commenter. The EPA also reexammed the data base since proposal and
removed several vessels that should not have been included because they do not meet the
definition of storage vessel. Changes to the storage vessel data base, and changes to the MACT
8-1
-------
definition of storage vessel. Changes to the storage vessel data base, and changes to the MACT
floor and the final standard that are summarized in the remainder of this response and in the
responses to comments in section 8.2, are discussed in the memorandum "Explanation of Options
for Re-evaluating the Storage Tank MACT Floor for the Production of Pesticide Active
Ingredients NESHAP," (Docket A-95-20, Docket Item No. IV-B-2).
The proposed approach to developing the MACT floor for storage vessels was
significantly different than the approach used to develop the floor for other rules (e.g., the HON,
polymers & resins, and pharmaceuticals). Since proposal, EPA has reevaluated the revised data
base and determined that an approach consistent with that used for the other rules is feasible and
appropriate for this rule. One of the commenters also recommended that the floor include vapor
pressure cutoffs as in other rules. As a result, EPA decided to revise the MACT floor. The
revised approach established vapor pressure cutoffs at the same storage vessel capacity cutoffs
and control efficiency cutoffs as were used in the previous rules. Specifically, the approach
"3. 7 -3
examined storage vessel cutoffs at 38 cubic meters (m ), 75 m , and 151 m . (In English units,
these capacities correspond with 10,000 gallons [gal], 20,000 gal, and 40,000 gal, respectively,
and the data base includes at least one storage vessel at each of these sizes.) Within these size
ranges, the vapor pressure cutoff at which the majority of storage vessels were controlled to
95 percent or more was determined; the 95 percent level is consistent with the efficiency of
floating roofs, which are the most cost-effective controls.
Under the revised approach, at liquid vapor pressures of 3.45 kPa and higher, the median
o
control efficiency was found to be at least 95 percent in both the 75 m and larger range and the
•3
151 rrr and larger range; at all vapor pressures, the majority of storage vessels with capacities
smaller than 75 m were found to be uncontrolled. The vapor pressure of 3.45 kPa is the vapor
pressure of toluene, which is the predominant HAP in the industry and the most common organic
HAP stored in storage vessels. Therefore, the revised MACT floor for storage vessels at existing
sources was determined to be 95 percent control for storage vessels with a capacity greater than
or equal to 75 nr that store material with a vapor pressure greater than or equal to 3.45 kPa. In
addition, the MACT floor was determined to be no control for all storage vessels with a capacity
less than 75 m .
8-2
-------
be 95 percent. The capacity of the smallest vessel controlled to 95 percent was determined to be
40 m3, and the vapor pressure of the compound stored in this vessel was 16.5 kPa. The MACT
floor for new sources must be at least as stringent as the floor for existing sources. Therefore, the
MACT floor for new sources is 95 percent control for storage vessels with (1) a capacity of 40
m or greater that store material with a vapor pressure of 16.5 kPa or greater and (2) a capacity of
75 m or greater that store material with a vapor pressure of 3.45 kPa or greater.
8.2 STORAGE VESSEL STANDARD
Status at proposal: Under the proposed rule, one regulatory alternative more stringent
than the floor was developed. The regulatory alternative would require 95 percent control of
storage vessels with capacity of 75 m or greater that have uncontrolled emissions of 108 kg/yr or
greater. Storage vessels smaller than 75 m3 (and greater than 38 m3 ) that have uncontrolled
emissions of 108 kg/yr or greater would require control to the floor level (41 percent). This
regulatory alternative was determined to be cost effective. Therefore, the proposed standard for
storage vessels at existing sources was established at 95 percent control for vessels with a
capacity greater than or tqual to 75 m that have uncontrolled emissions greater than or equal to
108 kg/yr.
No regulatory alternatives more stringent than the MACT floor were developed for
storage vessels at new sources. Therefore, the proposed standard for storage vessels at new
sources was determined to be 98 percent control efficiency for storage vessels with a capacity of
26 m or greater with uncontrolled emissions of at least 0.45 kg/yr.
8.2.1 Applicability Cutoffs
Comment 1: Several commenters (IV-D-16, IV-D-21, IV-D-22, IV-D-27, and FV-D-29)
requested that EPA increase the lower emission cutoff for existing and new storage vessels.
Most of the commenters recommended increasing it to at least 500 Ib/yr; this level corresponds to
the level in the Batch Processes ACT document for which manifolding to an existing control
device was shown to be cost effective. Commenter IV-D-27 suggested adding an exemption in
§ 63.1360(d)(4) for such vessels. Several of the commenters also noted that combustion would
be the only feasible means of controlling HAP emissions of only 0.45 kg/yr, and that secondary
emissions would increase significantly as a result.
8-3
-------
Response: The Agency has determined that including the higher cutoff suggested by the
commenter would have been less stringent than the cutoff prescribed by the MACT floor. The
emission cutoffs mentioned by the commenters are not supported by the storage vessel data base.
As discussed in the response to the comment in section 8.1, EPA has revised the MACT
floor since proposal. This revised MACT floor uses storage vessel capacity and the vapor
pressure of stored material as the parameters for determining applicability for storage vessels,
and no uncontrolled emissions cutoff is included in the floor. The Agency expects that
implementing standards based on this format will be considerably easier than implementing the
proposed standards, because no ongoing emission tracking will be required to demonstrate
compliance with a cutoff for the standard. Use of these parameters is consistent with
requirements for storage vessels in other rules.
Comment 2: Commenters IV-D-16 and IV-D-27 stated that the minimum applicability
size cutoff for existing Group 1 storage vessels should be changed to correlate with the NSPS
subpart Kb size cutoff to simplify compliance. The commenters stated that the cutoff for storage
vessels at existing sources would change from 38 m3 to 40 m3. In addition, the commenters
pointed out that the 38 m3 cutoff is below the smallest controlled to the median control
efficiency in the study (i.e., 39 m ).
Response: For the final rule, EPA based the standards for new and existing sources on
the MACT floor because the cost to go beyond the floor was determined to be unreasonable. As
a result of the changes to the database discussed in the response to the comment in section 8.1,
the capacity cutoffs in the final rule are higher than the cutoffs suggested by the commenters. For
existing sources, the cutoff is 75 m3 instead of the 40 m3 suggested by the commenters. For new
sources, the cutoff is 40 m3 instead of the 39 m3 suggested by the commenters.
8.2.2 Control Level of the Standard
Comment: Commenter IV-D-27 requested that EPA keep the existing source standard for
storage vessels with capacities greater than 75 m3 the same as that for smaller storage vessels,
unless floating roof technology is already in-place. The commenter asserted that the EPA's
"beyond the floor" standard of 95 percent organic HAP control for existing "large" storage
vessels is not justified for storage vessels that were not already equipped with floating roof
8-4
-------
technology. The commenter stated that EPA's assumption that any existing vessel larger than
75 m3 can be cost-effectively retrofitted with a floating roof is unrealistic.
Response: For the proposed rule, the MACT floor was based on a control efficiency of
41 percent. As discussed in the response to the comment in section 8.1, the revised MACT floor
is based on 95 percent control. The final standards also are based on a control of 95 percent
because the cost to control to a higher level was determined to be unreasonable. Now that both
the MACT floor and the standard are based on the same control efficiency, the commenter's
concern about going beyond the floor is no longer relevant. However, EPA wishes to emphasize
that the costs for the proposed standard were determined to be reasonable. The costs estimated at
proposal reflect the costs for retrofitting an existing fixed roof vessel with an internal floating
roof (IFR). These costs were based on equations developed from vendor information on the cost
of installing a new aluminum, noncontact IFR with a vapor-mounted primary seal and a
secondary seal; this cost estimate included retrofit costs for modifying the vessel, including any
corrections to vessel shell deformations or obstructions, any special structural modifications, and
cutting vents or openings if necessary. The estimated costs also included the costs to empty,
clean, and degas the vessel prior to retrofit.
8.2.3 Format of the Standard
Comment: Commenters IV-D-21, IV-D-28, and IV-D-29 stated that EPA should allow
floating roofs as a control option for storage vessels at new sources. Commenters IV-D-21 and
IV-D-29 stated that it is possible to reduce emissions of some HAP by 98 percent using a floating
roof, with the efficiency calculated using TANKS3.
Response: As discussed above, the control level for storage vessels at new sources is 95
percent under the final rule. Floating roof technology is allowed to meet this limit, just as it is for
existing sources.
8.3 COMPLIANCE AVERAGING PERIOD
Comment: Commenter IV-D-27 requested that the storage vessel standards be on an
annual average basis to reflect the data used in developing the standards (the commenter included
data for the median controlled vessel, which is controlled with a condenser; the data show the
variation in emissions and control efficiency throughout the year for this vessel).
8-5
-------
Response: The EPA does not agree with the commenter's assertion that the data used to
develop the MACT floor reflect only an annual compliance period. All of the storage vessels in
the revised MACT floor analysis (see section 8.1) that are controlled to at least 95 percent are
controlled with technologies (i.e., combustion devices) that achieve their control efficiency at all
times (i.e., as averaged over a period as short as the 3-hour period of a performance test). The
storage vessel that is controlled with a condenser is not part of the majority controlled to at least
95 percent in the revised MACT floor analysis. Therefore, the final rule retains the requirement
to demonstrate compliance with the percent reduction or outlet concentration limit for an add-on
control device on a daily basis. Alternatively, the owner or operator may install a floating roof,
which must be inspected annually and as otherwise specified in § 63.120 of the HON.
8.4 STORAGE VESSEL DEFINITION
Comment 1: Commenters FV-D-21 and IV-D-29 recommended that EPA adopt the
complete SOCMI MACT storage vessel definition set in the PAINESHAP.
Response: The definition for storage vessels for the PAI industry is similar to the
definition used in the HON. Additional language contained in the HON definition for storage
vessels includes a vessel "... that has been assigned ... to a chemical manufacturing process
unit that is subject to this subpart." This language regarding the assignment of storage vessels to
the source category or a PAI process unit that is subject to subpart MMM has been added to the
definition in the final rule. The only remaining difference between the two definitions is that the
HON excludes only bottoms receivers and surge control vessels, and the PAI draft rule excludes
all process vessels. Under the PAI rule, bottoms receivers and surge control vessels are process
vessels. Thus, they are not regulated as storage vessels under either rule. The exclusion of
process vessels has been retained in the definition for storage vessels. The response to comments
in section 7.7 of this document provides rationale for regulating bottoms receivers and surge
control vessels as process vessels.
Comment 2: One commenter (IV-D-27) suggested that EPA add two exclusions to the
storage vessel definition: (1) vessels regulated under 40 CFR 260 through 270 RCRA
provisions, and (2) other waste vessels (nonwastewater). Because EPA's data base only contains
vessels used for storing raw materials and products, other types of vessels should be excluded.
8-6
-------
Response: Storage vessels containing waste materials were not intended to be included
under the storage vessel definition. The PAI data base includes only those storage vessels that
contain raw materials, solvent, or product. Therefore, an exclusion for nonwastewater waste
vessels has been added to the definition of storage vessels in the final rule. This exclusion
includes vessels subject to 40 CFR parts 260 through 270.
8.5 OVERLAP WITH OTHER STANDARDS
Comment: Commenter IV-D-27 requested that EPA specifically state that regulated
emissions from pesticide manufacturing storage vessels that are routed to RCRA incinerators are
deemed to comply with the standards of this rule with no additional performance testing,
monitoring, recordkeeping or reporting requirements.
Response: The proposed rule exempted an owner or operator from performance test
requirements for RCRA control devices, and this provision was retained in the final rule. In
addition, the final rule exempts the owner or operator from monitoring, recordkeeping, and
reporting provisions when emissions are routed to a RCRA control device because these RCRA
provisions also are at lea^, as stringent as the provisions in the final rule.
8.6 CLARIFICATIONS
Comment 1: Commenter IV-D-28 suggested that EPA change every use of "control
device" to "control device(s)." The commenter stated that the proposed rule appears to require
that only a single control device be used to control emissions from a storage vessel, however,
combinations of control devices should be allowed, as appears to be the case for other streams.
Response: The Agency agrees with the comment regarding the use of multiple control
devices on a single process; it was not EPA's intention to limit the control scenario implemented
by a facility. Either a single control device or a series of control devices is acceptable as long as
the overall control efficiency specified by the regulation is achieved. The term "control
device(s)" has been added to the pesticide regulation in place of "control device" to make it clear
that use of multiple control devices is acceptable.
Comment 2: Commenters IV-D-21 and IV-D-29 point out that § 63.1362(c)(l)(ii) refers
to control devices "designed to reduce emissions of organic HAP by greater than 41 percent but
less than 95 percent." The commenters assert that including the words "95 percent" is not
8-7
-------
necessary and makes this statement confusing. The commenters recommend that the words "but
less than 95 percent" be removed to help clarify this section.
Response: This provision in the proposed rule referred to storage vessels that are already
controlled at least to the level of the floor, but do not exceed the level of the standard. Because
the standard in the final rule is equivalent to the floor, this provision is no longer needed and has
been deleted.
Comment 3: One commenter (FV-D-16) pointed out that in both the definitions of
Group 1 Storage Vessel (§ 63.1361) and the standard (§ 63.1362), the conversion from metric
units to English units are rounded off. The commenter requests that EPA provide a more precise
conversion to English units.
Response: In an effort to reduce confusion over the conversion from English to metric
units (or vice versa), only metric units have been included in the final rule. This is consistent
with the approach used in the HON.
8-8
-------
9.0 STANDARDS FOR EQUIPMENT LEAKS
9.1 LEVEL OF THE STANDARD
Comment: The MACT floor for equipment leaks was determined to be no control, and
the proposed standard was based on a more stringent regulatory alternative that consisted of
implementation of the leak detection and repair (LDAR) program from the HON. Six
commenters (F/-D-14, IV-D-16, IV-D-20, IV-D-21, IV-D-27, and IV-D-29) stated that the
standard should not be based on the LDAR program from the HON because their analyses show
it is not cost effective. Several commenters provided data contradicting both the emission rates
and costs used in EPA's cost analysis.
All six commenters asserted that initial equipment leak frequencies and thus the actual
emissions rates are lower than EPA estimated using the SOCMI average emission factors.
According to commenter IV-D-16, recent screening data obtained to demonstrate compliance
with the HON, P&R IV, and State programs (from 7 companies and 21 processes, not necessarily
PAI) show current industry maintenance, design, and materials selection result in far lower leak
rates and emissions than EPA assumed; commenter FV-D-21 made a similar summary statement.
Commenter IV-D-16 provided a tabular summary of leak rates for more than 100,000 connectors,
64,000 valves, and 1,400 pumps. Estimated emissions based on these data are approximately
6 percent of the level used in the EPA analysis. Commenter IV-D-27 also stated that leak rates
are typically less than 5 percent of the value obtained using the SOCMI average emission factors.
Commenter IV-D-16 noted that this level is comparable to the controlled level estimated by EPA,
and concluded that any additional reductions achieved by implementing the standard would be
extremely small. Commenters IV-D-16 and IV-D-27 believe EPA's assumption that the fluid in
contact with each component is 100 percent HAP is overly conservative; commenter IV-D-27
suggested 50 percent would be a better value.
9-1
-------
Commenters IV-D-14, IV-D-16, and IV-D-27 asserted that the costs used in the EPA
analysis were underestimated. Commenter IV-D-27 provided the following specific examples
based on a quote the commenter obtained from Team Industrial Services to conduct an LDAR
program: (1) initial and annual monitoring costs should be at least $4.50/component and
$2.95/component, respectively, instead of $2.50/component and $2.00/component; and (2) labor
costs to repair components should be at least $30.00/hour, not $22.50/hour. Commenter IV-D-27
also noted that the lower emissions estimated by the commenters results in a smaller product
recovery credit than EPA estimated.
Based on the available information, commenters IV-D-14, IV-D-16, and IV-D-27
estimated cost-effectiveness values that are much higher than EPA's estimate. Based on
unspecified data and assumptions, commenter IV-D-14 estimated the cost effectiveness to be
more than $35,000/Mg for a facility with 215 pumps, 4,460 valves, and more than
16,000 connectors. Commenter IV-D-27 estimated the cost effectiveness for the nationwide
population of components in EPA's analysis to be more than $64,000/Mg; this estimate was
based on the revised monitoring and labor costs described above, assuming fluid in contact with
the components contains 50 percent HAP, and assuming that the LDAR program would achieve
the same percentage emission reductions that were used in the EPA analysis.
Commenter IV-D-16 calculated a cost effectiveness of $78,000/Mg for the nationwide
population of components in EPA's analysis; this estimate was based on the costs from EPA's
analysis, assuming the LDAR program would achieve a 10 percent reduction in emissions from
the lower baseline described above.
Commenters IV-D-16 and IV-D-20 recommended that EPA consider setting a standard
based on a sensory (visual, olfactory, audible) LDAR program or other alternatives that are under
consideration for other rules (i.e., for formulation processes covered by the MON or the draft
Generic MACT standards, part 63 subparts TT and UU).
Response: In recent regulatory development efforts involving similar industries, the EPA
has generally found equipment leaks to be a significant source of emissions. The EPA's
approach has been to require industries to identify leaks and fix them as soon as possible. The
EPA is sensitive to the recordkeeping burden associated with an LDAR program for this industry
and has strived to minimize the number of activities that have to be conducted and documented
9-2
-------
while still requiring sources to identify and eliminate equipment leaks. Relative to earlier rules,
the Agency developed the HON to focus most of the recordkeeping and reporting burden on
those processes and types of equipment that have the most significant leaks, in terms of HAP
emissions. Since the development of the HON, the Agency has proposed the consolidated air
rule (CAR) that is designed to minimize the reporting and recordkeeping burden even further
(see 63 FR 57748). The EPA believes that, in addition to consolidating many LDAR programs,
the CAR addresses many concerns regarding the burden placed on industry to implement LDAR
programs with little environmental benefit. The proposed CAR is specifically focused on
identifying and fixing leaking components, and leaves out many of the recordkeeping
requirements that are focused on nonleakers. For example, the proposed CAR includes options
for identifying groups of equipment, such as valves, that are located within an area or length of
pipe without individually listing each component. The proposed CAR also allows much less
frequent monitoring of components, depending on leak rates identified during an initial survey.
If less than 0.25 percent of connectors are found to be leaking, monitoring is only required every
8 years (however, at leas; 50 percent of the connectors must be monitored during the first 4 years,
and the remainder must be monitored within the next 4 years). For valves, the required
monitoring frequency can be as low as every 2 years, if the percentage of leakers for a given
group is less than 0.25 percent.
In general, commenters contend that EPA has overstated the reductions achieved by the
LDAR program and understated the costs in estimating cost effectiveness. In response to
commenter concerns, EPA reviewed the original analysis and also calculated the cost
effectiveness of a revised LDAR program based on the requirements of the proposed CAR. The
revised analysis used the data supplied by the commenters; the EPA also combined recently
obtained initial leak rate data for components in pharmaceutical processes with the data provided
by commenters. The EPA does not consider the emission estimates in the original analysis to be
invalid. However, for the revised analysis, EPA used the leak rate data provided by the
commenters and other recently obtained data to determine a lower bound on the baseline
emissions (and a corresponding upper bound on cost effectiveness for a given set of assumptions
regarding subsequent leak frequencies and the number of monitoring instruments that are
needed). The remainder of this response summarizes the procedures and results of the revised
9-3
-------
analysis. This summary includes a description of models used in the analysis, procedures to
calculate emissions, specific aspects of the costing methodology, and the estimated cost
effectiveness. Details of the analysis are documented in reference number 1 (Docket
No. A-95-20, Docket Item No. IV-B-3). As a result of this analysis, the standards for equipment
leaks have been changed since proposal; the requirements in the final rule are consistent with the
proposed CAR requirements rather than the subpart H requirements.
Models. Batch and continuous model processes were developed for the original analysis
and these models were also used in the revised analysis. The batch process consisted of 65 gas
valves, 340 liquid valves, 14 pumps, and 1,100 connectors. The continuous process consisted of
240 gas valves, 1,100 liquid valves, 33 pumps, and 1,500 connectors. For this analysis, all
valves were considered to be liquid valves because the commenters provided leak rate data only
for "valves." Components in the batch process were assumed to be in service for 2,800 hr/yr, and
components in the continuous process were assumed to be in service for 5,000 hr/yr.
Emission': Estimates. Uncontrolled emissions for the model processes were estimated
based on initial leak rate data. The EPA started with the initial leak rate data provided by the
commenters. Most of these data were from facilities in the SOCMI or polymers and resins
industry. The EPA also combined recently obtained initial leak rate data for components in
Pharmaceuticals processes with the data provided by the commenters. These data were combined
because EPA believes pharmaceuticals processes are at least as representative of PAI processes
as are SOCMI or polymers and resins processes due to the prevalence of batch processing,
similar process equipment, and similar HAP in the pharmaceutical and PAI industries.
Combining the two data sets approximately doubled the amount of available data to 200,000
connectors, 87,000 valves, and 2,600 pumps. All of the leak rates for a particular type of
component in the data base were summed. To estimate average leak rates, each of these overall
leak rates was divided by the total number of that type of component in the data base.
Uncontrolled annual emissions for the two model processes were estimated by multiplying these
average leak rates by both the number of components and the operating hours for the models.
Controlled emissions for the two model processes were based on several assumptions
about subsequent leak rates because leak rate data are not available. For valves and connectors,
the leak frequency occurrence rates after implementation of LDAR were assumed to be equal to
9-4
-------
the performance levels required in the proposed CAR (i.e., 0.25 percent). The EPA also assumed
that repairs are 100 percent effective, and that there are no recurrences of leaks. These
assumptions are similar to those in the HON, except that the occurrence levels are lower than
those assumed in the HON analysis. These appear to be reasonable assumptions given the fact
that the initial leak frequencies for valves and connectors were in some cases lower than the
specified performance level. For pumps, the CAR does not specify a performance level.
Therefore, it was assumed that the leak occurrence rate after implementation of the LDAR
program is equal to 50 percent of the initial leak frequency. This assumption appears reasonable
based on the leak frequency reductions that have been achieved by other LDAR programs. The
initial leak frequency for pumps was estimated to be 7.44 percent by using the estimated
uncontrolled average leak rate for pumps (as estimated above) in the average leak rate (ALR)
equation for a leak definition of 1,000 ppmv. (The ALR equations are presented on page 5-46 of
the EPA document entitled "Protocol for Equipment Leak Emission Estimates,"
EPA-453/R-95-017, November 1995.) The occurrence rate for pumps, therefore, was assumed to
be 3.72 percent. Controlled average leak rates were estimated using one-half of the occurrence
rates in the appropriate ALR equations (i.e., equations for leak definitions of 500 ppmv for
valves and connectors, and 1,000 ppmv for pumps). One-half of the occurrence rate is assumed
to be the average leak frequency over the monitoring cycle. Controlled annual emissions for the
two model processes were estimated by multiplying these average leak rates by both the number
of components and the operating hours for the models.
Unlike the original analysis, the revised analysis does not depend on the assumption that
components are in 100 percent HAP service. The leak-rate methodology used in the revised
analysis estimates the amount of organic compounds (VOC or HAP) emitted. Thus, the analysis
will overstate the HAP emissions if the leak includes non-HAP VOC. In these situations,
however, the program will still control the remaining non-HAP VOC, and the cost effectiveness
will reflect the control of both HAP and VOC.
Costing Methodology. In general, the LDAR costs are based on procedures and unit costs
that were used in the analysis for the HON. The approach identifies both initial costs to establish
the monitoring program, recurrent annual costs to monitor and repair leaks, and a credit for
product recovered by reducing leaks. Initial LDAR monitoring and repair costs and the cost of a
9-5
-------
monitoring instrument are annualized using a capital recovery factor. Monitoring and repair
costs are assumed to be the same every year. The total annual cost of the LDAR program is the
sum of the annualized initial costs plus the annual monitoring and repair costs minus product
recovery credits.
A review of the costing methodology indicated that costs for the monitoring instrument
needed to be updated, but no changes are warranted in other elements of the methodology. The
original analysis was based on costs for a monitor that is no longer available. Capital costs for a
currently available monitor that is widely used are higher than the capital costs for the original
monitor, but maintenance costs are lower. See reference number 2 (Docket No. A-95-20, Docket
Item Numbers IV-B-6 and IV-B-7). As a result, total annual costs for the new monitoring
instrument are lower. A discussion of other costs in the analysis is provided below.
One commenter stated that initial and annual monitoring costs should be at least
$4.50/component and $2.95/component, respectively, instead of $2.50/component and
$2.00/component. The commenter indicated that the higher values were based on costs typically
charged by contractors who monitor equipment for leaks as a service. Unit monitoring costs in
the EPA analysis are lower because they include only labor costs. When a monitoring contractor
is hired, the cost per component must consider annualized costs of implementing the overall
program, including overhead costs and the costs of purchasing and maintaining a monitoring
instrument. Adding 40 percent to account for overhead costs increases the unit monitoring costs
used by EPA to $3.50/component and $2.80/component. The annual maintenance charge plus
the annualized purchase cost of the monitoring instrument adds another $2.00/component to both
values; this value is likely higher than the cost to a monitoring contractor because it assumes the
monitoring instrument is only fully utilized in the first year (this was not an issue in the original
analysis because all components were assumed to be monitored the same number of times each
year). The resulting costs of $5.50/component and $4.SO/component are higher than the values
quoted by the commenter. Therefore, EPA did not change the monitoring costs for this analysis.
One commenter stated that labor costs to repair components should be at least $30.00/hr,
not $22.50/hr. The actual cost in the EPA analysis is $22 50/hr, multiplied by a factor of 1.4 to
add another 40 percent for administrative and support costs, which yields a value of $31.50/hr.
9-6
-------
Again, this value is comparable to the value quoted by the commenter. Therefore, EPA did not
change the labor cost for this analysis.
The total annual costs (TAG), after accounting for product recovery, were estimated to be
$2,066/yr for the batch model process and $5,226/yr for the continuous model process. The total
capital investment (TCI) is $7,914 for the batch model process and $15,277 for the continuous
model process. The TCI includes both the cost of the monitoring instrument and the cost for the
initial round of monitoring and repair.
Impacts. Nationwide emissions were estimated by scaling the emissions for the model
processes based on the number of batch and continuous processes in the PAI industry.
Nationwide, there are an estimated 146 batch processes and 43 continuous processes. The
nationwide uncontrolled emissions for equipment components in the PAI industry are estimated
to be 503 Mg/yr, and the nationwide controlled emissions following implementation of the
LDAR program in the CAR are estimated to be 117 Mg/yr; the nationwide emission reduction is
estimated to be 386 Mg/yr.
Nationwide costs to implement an LDAR program for the PAI industry consistent with
the program in the CAR were estimated using two approaches. In the first approach, the
monitoring instrument capital cost and annual maintenance cost were prorated based on the ratio
of the nationwide number of components as characterized by the two model processes to the
number of components that a fully utilized instrument could be used to monitor (i.e., about
9,000 components). The second approach, a more conservative one, assumes one instrument is
needed for every 9,000 components (or fraction thereof) at each facility. Under second approach,
a total of 80 monitoring instruments would be needed (58 for the 58 modeled facilities, and 22
for the 20 surveyed facilities (some of the surveyed facilities needed two instruments because
they had more than 9,000 components in PAI processes).
The nationwide costs under both approaches were estimated by scaling the unit costs for
the model processes based on the number of batch and continuous processes in the PAI industry.
As a result, the estimated nationwide TCI and TAG using the first approach are $1.81 million and
$526,000/yr, respectively. Using the second approach, the TCI and TAG increase to
$2.26 million and $692,000/yr, respectively.
9-7
-------
Cost Effectiveness. The cost effectiveness associated with implementation of an LDAR
program consistent with the requirements of the proposed CAR using the first costing approach
is estimated to be $l,400/Mg at existing sources ($526,000/386 Mg). The cost effectiveness at
existing sources using the second approach is estimated to be $l,800/Mg ($692,000/386 Mg).
Both of these values are considered to be reasonable.
9.2 CROSS-REFERENCING THE CONSOLIDATED AIR RULE
Comment: The preamble to the proposed rule states that EPA will consider
cross-referencing the CAR if the CAR is complete before this rule is promulgated.
Commenter IV-D-28 strongly opposes cross-referencing requirements in the CAR (unless the
CAR requirements are only one option) because the CAR has become more and more
complicated, still contains flaws, and has not been promulgated.
Response: The EPA has decided to incorporate requirements consistent with those of the
proposed CAR into this final rule, as opposed to cross-referencing them. The EPA intends, with
this action, to minimize confusion and reduce the potential for errors and inconsistencies that
could be created by cross referencing.
9.3 REFERENCES
1. Memorandum from B. Shine and K. Schmidtke, MRI, to Project File. October 21,1998.
LDAR program cost effectiveness.
2. Memorandum from K. Barnett, EPA:ESD, to Equipment Leaks Costing Project File.
October 18, 1998. Equipment leaks costing.
9-8
-------
10.0 STANDARDS FOR WASTEWATER
10.1 GENERAL
Comment: Commenter IV-D-20 believes the wastewater provisions should be based on
the proposed pharmaceuticals rule, not the HON, because: (1) both the pharmaceuticals and PAI
manufacturing industries produce biologically active materials and are relatively small in scale
compared to SOCMI facilities (the commenter noted that the surveyed facilities are larger than
the typical manufacturing processes within the PAI industry), (2) the evaluation for the proposed
pharmaceuticals rule was based on more recent emission models and a more thorough assessment
of emissions and control costs, and (3) the HON wastewater requirements are extremely complex
and difficult to understand. Other commenters requested that only specific provisions from the
proposed pharmaceuticals rule be included in this rule, as noted in the comments below.
Response: The EPA disagrees with the commenter's claim that the proposed
pharmaceuticals wastewater provisions are more appropriate than the provisions in the HON for
the PAI source category. First, although the scale of PAI operations typically is closer to
pharmaceuticals than to SOCMI, the applicability cutoffs in both the HON and pharmaceuticals
regulations are comparable. Second, the analysis for the pharmaceuticals proposed standard was
not better, it was just different. For example, the pharmaceuticals analysis was based on the use
of methanol as a model compound because of the prevalence of that compound in the
pharmaceuticals industry, but the HON analysis modeled each HAP. Comparisons of specific
aspects of the provisions in the two rules are discussed in many of the responses in the remainder
of this chapter.
10.2 AFFECTED WASTEWATER
Comment 1: Several commenters (IV-D-14, IV-D-21, IV-D-28, and IV-D-29) stated that
maintenance wastewater streams should either be excluded from the regulation or subject to the
same requirements as in § 63.105(b)(2) of the HON. All of the commenters cited the variability
10-1
-------
and unpredictable nature of maintenance wastewater streams (which makes it difficult to
determine whether a stream is Group 1 or Group 2) and the low potential for substantial
emissions (because such streams are typically due to rinsing or flushing equipment) as reasons to
regulate maintenance wastewater streams differently. Commenter IV-D-28 added that
maintenance wastewater streams cannot be controlled like process wastewater streams. For
example, the commenter explained that trying to pump the small amount of water generated
when bleed lines or pumps are drained would cause equipment problems if there was not enough
flow to keep material running through the pump itself. Commenter IV-D-28 also stated that the
cost to comply with conveyance requirements would be enormous, especially if an enclosed
system has to be installed for every piece of equipment because someday a maintenance
wastewater stream might be generated there.
Response: The EPA considered the above comments and is persuaded by the
commenters' arguments that the variability of maintenance activities makes characterization of
these wastewaters difficult, and that there is fairly low potential for substantial emissions for
most of these wastewater streams. However, EPA has no data on typical quantities of
maintenance wastewater streams generated, or the characteristics of these wastewater streams.
Therefore, EPA's approach, in resolving this issue, was to specify characteristics of maintenance
wastewater streams that have significant emission potential. The EPA also sought to minimize
the burden of characterization of all maintenance wastewater streams. Based on this approach,
EPA evaluated three possible options for regulating maintenance wastewater streams. The first
option was to adopt the same requirements as in § 63.105 of the HON, which is the option
suggested by the commenters. The EPA believes that maintenance wastewater streams may
warrant a different treatment in this industry than what was done under the HON because the PAI
industry is expected to generate process wastewater streams in discrete batches, due to the batch
nature of the industry. These process wastewater streams are expected to have properties similar
to those for maintenance wastewater streams in terms of the quantities generated, the frequency
of generation, and the options for management, suppression and treatment. Therefore, for
streams with significant emissions potential, whether generated because of maintenance activities
or by the process operations, EPA believes that proper management and treatment is warranted.
10-2
-------
The second option evaluated was to require the same management and treatment for both
maintenance and process wastewater streams, as in the proposed rule. Under this option, the
applicability thresholds are the same as in the HON for both types of streams. However, because
information on maintenance wastewater streams is unavailable, it is not clear how many such
streams would be subject to management and treatment requirements. For example, one extreme
is that industry would be required to characterize numerous maintenance wastewater streams
with no environmental benefit. Conversely, it is possible that many streams with low flow rates
and low emission potential would meet the Group 1 applicability thresholds of 10,000 ppmw at
any flow rate. Another concern with this option is the extent of dedicated maintenance
wastewater conveyance systems that will need to meet emission suppression requirements on the
chance that a Group 1 maintenance wastewater stream might be discharged in the processing area
served by that part of the conveyance system.
Finally, the third option considered and incorporated into the final rule is a modification
of option 2 that does not require characterization, suppression, and treatment of small
maintenance wastewater sireams with low emission potential. The HON includes two thresholds
for triggering Group 1 applicability: the first, which has already been discussed, captures any
streams with greater than 10,000 ppmw HAP load and does not consider emissions potential; the
second applicability threshold, however, considers emission potential by adding a quantity
(greater than 10 liters per minute [L/min]) in addition to the HAP concentration (1,000 ppmw
HAP). When converted to HAP load, the second applicability threshold is equivalent to
approximately 5.3 Mg of HAP. This load was used as the applicability threshold in the definition
of maintenance wastewater in the final rule. The final rule also specifies that the maintenance
wastewater definition applies to individual discharge events, not the sum of all events occurring
from a single POD over the course of a year. By defining maintenance wastewater streams in
this manner, only the largest, most significant maintenance wastewater streams would be subject
to suppression and treatment. These large streams should be easier to identify and may occur
only at certain POD's. The definition of Group 1 wastewater streams also includes maintenance
wastewater streams with this same load; thus, there are no Group 2 maintenance wastewater
streams, and there is no burden to characterize and track any maintenance wastewater streams
other than Group 1 streams.
10-3
-------
It is conceivable that there are no maintenance wastewater streams with characteristics
approaching this definition in the industry. However, because EPA has no data on the quantities
or characteristics of these maintenance wastewater streams, EPA believes the best approach is to
define a threshold of concern rather than to exempt from suppression and treatment all
maintenance wastewater streams.
The EPA considered the example brought up by the commenters of needing to drain
maintenance wastewater streams onto diked floor areas and sumps prior to sending to enclosed
drains. The commenters argue that it would be unreasonable to build enclosed conveyance
systems for all maintenance wastewater streams that require control. However, EPA believes
that those few maintenance wastewater streams that could trigger Group 1 applicability would be
routed to a process or chemical sewer to begin with, regardless of whether they were considered
affected streams for purposes of this MACT standard, and believes that most facilities are
capable of routing wastewaters through temporary flexible hoses to a conveyance system that
would serve a chemical or process sewer especially for these situations.
Comment 2: Because the MACT floor for wastewater at existing sources is no control,
commenter FV-D-16 recommended that EPA exclude from control requirements those streams
that have low emission potential. Therefore, the commenter recommended that EPA establish an
alternative Group 1 threshold for soluble HAP with low vapor pressures, similar to the threshold
in the proposed pharmaceuticals rule. Another commenter (IV-D-27) also requested that EPA
add alternative Group 1 thresholds based on differences in the volatility of HAP compounds, like
those for partially soluble and soluble HAP compounds in the proposed pharmaceuticals rule. A
third commenter (FV-D-28) supports the proposed approach to classify process wastewater (but,
as noted above, not maintenance wastewater streams) as Group 1 or Group 2, and require control
only for wastewater classified as Group 1.
Response: Under the proposed rule, streams with low emission potential were excluded
by the 1,000 ppmw and 10 Urnin cutoffs. In addition, a facility would not be required to control
any wastewater streams if the total load of Table 9 compounds in Group 1 wastewater (untreated
or treated to levels less than required by the standards) from the "total source" (for the final rule,
this term is defined to mean the affected source) is less than 1 Mg/yr. The EPA determined that,
on average, treatment of streams with concentrations and flows that exceed the cutoffs was cost
10-4
-------
effective. If a stream contains compounds with low Henry's law constants, the "low potential"
for emissions is accounted for by requiring less stringent control based on low Fr values. This
approach also is consistent with the HON, which should facilitate compliance for many facilities,
and is consistent with one commenter's request that provisions be like the HON wherever
possible. Thus, the final rule does not include alternative Group 1 thresholds.
10.3 TREATMENT OPTIONS FOR DISCHARGES TO POTW's
Comment 1: Three commenters (IV-D-16, IV-D-20, and IV-D-27) requested that the
enhanced biological treatment option in the proposed pharmaceuticals MACT standard be
included in this rule (i.e., for wastewater that contains soluble HAP and less than 50 ppmw of
partially soluble HAP) for discharges to a POTW. According to commenter IV-D-20, the HON
provisions essentially preclude discharge to POTW's because POTW's could not reasonably be
expected to understand, implement, and certify compliance with this regulation. Furthermore,
the commenter stated that the detailed analysis performed for the proposed pharmaceuticals rule
indicated that air emissions for certain wastewater streams would be negligible; thus, there is no
need to essentially ban discharge to POTW's.
Response: Except for minor differences in applicability cutoffs, one of the treatment
options in the HON (and thus in the proposed rule) is similar to the enhanced biotreatment option
under the proposed pharmaceuticals rule. Both the HON and the proposed pharmaceuticals rule
regulate two groups of HAP compounds in wastewater. For the HON, the groups are called "list
1" and "list 2" compounds. For the proposed pharmaceuticals standard, they are called "partially
soluble HAP" and "soluble HAP." All 52 of the compounds on list 2 are also classified as
partially soluble HAP. List 1 contains all 14 soluble HAP as well as the 10 remaining partially
soluble HAP. (Note that for the final pharmaceuticals rule, epichlorohydrin has been moved
from the solubles list to the partially solubles list.) Under the HON, an owner or operator is
exempt from the performance test requirement if (1) wastewater is treated in an enhanced
biological treatment process and (2) compounds on list 1 comprise at least 99 percent by weight
of the HAP compounds (list 1 plus list 2) in the wastewater. Under the proposed
pharmaceuticals standard, an owner or operator would be exempt from the performance test
requirement if wastewater containing soluble HAP and less than 50 ppmw of partially soluble
HAP is treated in an enhanced biological treatment unit, and the owner or operator demonstrates
10-5
-------
that less than 5 percent of the soluble HAP is emitted from the municipal sewer system. The
definition of an enhanced biotreatment unit is also the same under both rules, and waste
treatment units that qualify as enhanced biotreatment units are subject to the same compliance
requirements under both rules. Therefore, EPA disagrees with the commenter's assertion that the
treatment provisions in the proposed rule reduce the burden on POTW's, and EPA has not
revised the treatment provisions for today's final rule.
Comment 2: Commenter IV-D-16 cited the results of a study conducted by
Pharmaceutical Research and Manufacturers of America (PhRMA) (and discussed in detail in
PhRMA's comments on the proposed Pharmaceuticals rule) showing that streams discharged to
POTW's have the potential for significant emissions only from "totally open" collection and
municipal sewer systems. Therefore, if the collection and municipal sewer system is totally
open, the commenter recommended adding a provision that would allow an owner or operator to
use the enhanced biotreatment option only if the owner or operator demonstrates that less than 5
percent of the soluble HAP is emitted from the system (and would not impose this requirement if
the system is not "totally open").
Response: Under the proposed rule, an offsite facility that treats wastewater would be
required to comply with the same requirements as an affected source, including the emission
suppression requirements from the collection system. The EPA has reexamined municipal sewer
systems and determined that the primary potential for emissions from the collection system is
from the headworks at the POTW. Thus, the final rule specifies that either the waste
management units up to the activated sludge unit must be covered, or the owner or operator must
demonstrate that less than 5 percent of the total list 1 HAP is emitted from these units.
10.4 STANDARDS FOR NEW SOURCES
Comment: Several commenters (IV-D-16, FV-D-21, FV-D-27, IV-D-28, and IV-D-29)
consider the proposed wastewater standards for new sources with HAP loading greater than
2,100 Mg/yr to be too restrictive. Commenter IV-D-27 believes only Group 1 wastewater, not all
wastewater, should be subject to the standards. The commenter claims that requiring control of
all wastewater will result in negligible additional environmental benefits, but would likely cause
greater secondary air and resource impacts (e.g., from fuel usage and emissions of combustion
products).
10-6
-------
All five commenters requested that additional treatment options be allowed.
Commenter IV-D-28 requested that EPA add a treatment option that allows an owner or operator
to reduce the mass flow rate by the Fr values; the commenter stated that a 99 percent reduction
might be achievable for an individual facility with a certain combination of HAP, but it would
not be achievable by all facilities. Commenters IV-D-16, IV-D-21, IV-D-27, and IV-D-29
recommended adding at least an enhanced biotreatment option. Commenter IV-D-27 believes all
of the treatment options for existing sources should be allowed for new sources.
Commenters IV-D-16, IV-D-21, IV-D-27, and IV-D-29 requested the additional options because
they believe that limiting treatment options significantly impacts compliance flexibility with
little, or no, environmental benefit. For example, commenter IV-D-27 realizes that a steam
stripper would not meet the standard for compounds that have Fr values less than 0.99, but
believes that because the remaining HAP in the treated streams are less volatile, they would have
negligible air impacts. Several commenters (IV-D-16, IV-D-21, and IV-D-29) stated that EPA
had agreed during the development of revised wastewater provisions for the HON that the
various treatment options under the HON are equivalent from an air emissions standpoint (e.g.,
95 percent reduction in a biological treatment unit is equivalent to 99 percent reduction in a
non-biological treatment unit).
Response: According to the Act, the MACT floor for new sources is to be based on the
emission control that is achieved by the best controlled similar source. In the PAI production
industry, the best controlled source is achieving 99 percent control. This source also is treating
all of its wastewater from PAI processes, the HAP load in this wastewater is 2,100 Mg/yr, and
this wastewater contains a mixture of compounds with a range of Henry's law constants. Thus,
the proposed MACT floor for new sources with a HAP load exceeding 2,100 Mg/yr consisted of
the requirements to treat all wastewater and to achieve a 99 percent reduction in the HAP content
in the wastewater; for new sources with lower HAP loadings, the MACT floor is no control, as
for existing sources. The EPA continues to stress that the proposed MACT floor is consistent
with the Act, and it is retained in the final rule.
If a facility has a HAP load that exceeds the cutoff, the enhanced biotreatment option
(i.e., the option that exempts an owner or operator from initial compliance demonstrations) is not
allowed because the EPA does not have information showing that enhanced biotreatment units
10-7
-------
achieve 99 percent removal for mixtures of compounds with low Fr values. Otherwise, the final
rule allows any treatment option (including enhanced biotreatment) for such affected sources,
provided the owner or operator demonstrates that it achieves 99 percent removal of all HAP in
the wastewater. The EPA also points out that the requirement to achieve 99 percent removals
applies only to facilities that have extremely high HAP loads and thus, high potential for
emissions. Few new sources are likely to exceed the applicability cutoffs for the MACT floor
because 2,100 Mg/yr was more than three times higher than the load at any other surveyed
facility.
Finally, the commenter's statement about the equivalence of treatment options needs
clarification. Under the HON, the 95 percent option for biological treatment units requires that
the reduction be achieved from all wastewater sent to the treatment unit, not just the Group 1
wastewater. The 95 percent reduction also applies to all Table 9 compounds (in subpart G) in the
wastewater, not just compounds with high Fr values. Thus, on average, this option is considered
equivalent to other treatment options in the HON. This option is not considered equivalent to the
99 percent option for new sources described above because the 99 percent reduction is required
for all wastewater and all compounds.
10.5 OVERLAP WITH OTHER REGULATIONS
Comment 1: Commenter IV-D-27 believes the rule should state that routing regulated
wastewater (and associated emissions and residuals) to a RCRA incinerator constitutes
compliance with the standards, and no additional performance testing, monitoring,
recordkeeping, or reporting is required.
Response: The proposed rule cross-referenced § 63.138(h) of the HON, which specifies
that RCRA units used to treat wastewater streams and residuals are exempt from design
evaluation or performance test requirements, monitoring requirements, and associated
recordkeeping and reporting requirements. This cross-reference is retained in the final rule.
Comment 2: Commenter IV-D-28 suggested adding a paragraph to § 63.1362(d) to avoid
an overlap between this rule and subpart DD (Off-Site Waste and Recovery Operations). The
commenter noted that all organic HAP-containing wastes that are sent to an offsite facility for
treatment become subject to subpart DD, unless stated otherwise in subpart DD. Subpart DD
contains an exemption for wastewater from a HON facility. However, it does not contain an
10-8
-------
exemption for wastewater from other facilities subject to MACT standards, even those, like the
proposed rule, that require compliance with the wastewater provisions in the HON. Therefore,
the additional provision recommended by the commenter would stipulate that an offsite facility is
not subject to the requirements of subpart DD for wastewater streams and residuals received
from an affected source.
Response: The PAI rule cannot include a provision that modifies the requirements of
another rule. However, EPA will consider amending subpart DD at a later date.
10.6 CLARIFICATION OF CROSS-REFERENCES
Comment: Commenter IV-D-28 believes § 63.1362(d) of the proposed rule should cross-
reference only §§ 63.131 through 63.147 because §§ 63.148 and 63.149 are not wastewater
provisions. Commenter FV-D-28 also requested clarification of how to deal with provisions for
Table 8 compounds in the cross-referenced sections. Similarly, other commenters (FV-D-16 and
IV-D-27) believe EPA should clarify that references to Table 8 compounds are not applicable to
this rule.
Response: The EPA agrees with the commenters, and these changes have been
incorporated in the final rule. However, cross-references to § 63.148 from §§ 63.131 through
63.147 are still applicable, and the provisions in § 63.149 have been incorporated directly into
the final rule.
10-9
-------
11.0 STANDARDS FOR BAG DUMPS AND PRODUCT DRYERS
11.1 JUSTTFICATION FOR THE STANDARDS
Comment: Four commenters (IV-D-16, F/-D-21, IV-D-22, and IV-D-29) stated that EPA
should not issue regulations for bag dumps and product dryers because: (1) data are available for
only one source, which is not (commenter FV-D-21 said "may not be") representative of the
source category; (2) the terms paniculate HAP and bag dump are not defined; and (3) no test data
are available to verify estimates. Commenters IV-D-16, IV-D-21, and IV-D-29 also quoted the
following statement from the ruling in Portland Cement Association v. Ruckleshaus. 486 F. 2d
375, 396 (D.C. Cir. 1973): "[A] significant difference between techniques used by the Agency in
arriving at standards, and requirements presently prescribed for determining compliance with
standards, raises serious questions about the validity of the standard." As a result, the three
commenters noted that the Court remanded the rule, and the commenters concluded that the test
method used for compliance must be closely linked to the test method used as the basis for the
standard. Commenters FV-D-21 and IV-D-29 also recommended that EPA repeal the proposed
provisions for bag dumps and product dryers because they believe the MACT floor was not
established properly per EPA protocol.
Other commenters (IV-D-14, IV-D-20, and IV-D-27) stated that bag dumps should not be
regulated. Commenter FV-D-14 stated that it was inappropriate to include bag dumps in this
standard when no bag dump data were used to determine the MACT floor. According to
commenter IV-D-27, basing a standard for bag dumps on the data from a single product dryer is
inappropriate because the two types of emission points are not comparable. The commenter
explained that a dryer vent is typically a point source that can be easily vented to a control
device, but a bag dump typically is a fugitive emission source that is often difficult to capture and
route to a control device. According to commenter IV-D-20, EPA should not regulate bag dumps
11-1
-------
unless they are a found to be a significant source of emissions; the commenter noted that there
was no data on bag dumps in the PAI data base, and therefore no basis on which to regulate such
emissions. This commenter also cited the following reasons why regulating bag dumps would be
difficult and would not result in any meaningful emission reduction: (1) use of bag dumps is
avoided or minimized to the greatest extent practicable for ergonomic and worker exposure
reasons; (2) any particulate emissions are small and would be controlled to reduce workplace
exposure; (3) if dusting and workplace exposure are not a concern for a particular material, there
may be no process vent associated with the bag dumping operation; and (4) PM10 regulations are
generally applied to larger sources with actual vents. Another commenter (IV-D-16) also stated
that bag dumps should not be regulated as part of this rule because they are a source of fugitive
emissions.
Response: Standards for product dryers and bag dumps were included in the proposed
rule because these emission points can be a source of HAP emissions, specifically particulate
matter HAP emissions. The MACT floor for these emission points was developed for equipment
that emits particulate matter HAP; this equipment was limited to product dryers and bag dumps
because these are the only known sources of particulate matter HAP emissions at PAI facilities.
The MACT floor also was based on the level of control for these emission points at the MACT
floor facilities (i.e., the nine facilities with the best overall control of PAI process units). One of
the MACT floor facilities dried a PAI that is also a HAP. Emissions from this product dryer
were controlled with a fabric filter, and emissions tests showed the outlet PM concentration was
less than 0.01 grains per dry standard cubic foot (gr/dscf). The floor for particulate matter HAP
emission sources was based on this value because both .product dryers and bag dumps are
controlled with fabric filters, and 0.01 gr/dscf is a typical level for fabric filters.
The EPA is not persuaded by the commenter's argument that bag dumps should not be
regulated because they are (or may be) sources of fugitive emissions and are thus not comparable
to product dryers. The EPA knows of two bag dumps where a HAP raw material is added to a
PAI process, and both are controlled with fabric filters. At a minimum, a hood or partial
enclosure can be placed above or around a bag dump to capture the emissions and route them to
the control device. Furthermore, one of the commenters also stated that particulate emissions
from bag dumps would be controlled to reduce workplace exposure.
11-2
-------
Uncontrolled emissions (i.e., the pre-control emissions) from one of the two known bag
dumps exceed 1.6 Mg/yr. The EPA considers this to be a significant source, and the required
emission reduction to be meaningful. The fact that some facilities may have found more
desirable alternatives to the use of bag dumps does not justify exempting facilities that still use
them from regulation.
As a result, EPA maintains that standards are appropriate for bag dumps and product
dryers that emit HAP, that the MACT floor is valid, and that the standard should be based on the
MACT floor. However, EPA has decided to make one change for the final rule. At proposal, the
standard was for "particulate matter HAP." For the final rule, the standard is for "paniculate
matter" because the material captured in the fabric filters is essentially all HAP, and test methods
are for "particulate matter," not "particulate matter HAP." (The EPA assumes this is why the
commenters mentioned linking the test method used as the basis of the standard with the method
used to demonstrate compliance.) The final rule also specifies that the particulate matter
standards are for product dryers that dry a PAI or integral intermediate that is a HAP and for bag
dumps that introduce a HAP to a PAI process unit. See section 5.1 for definitions of "product
dryer" and "bag dump," and see section 16.3 for a discussion of changes in the monitoring
provisions for these types of emission points.
11.2 LEVEL OF THE STANDARD
Comment: Commenters FV-D-22 and IV-D-27 believe the rule should include an
applicability cutoff. Commenter IV-D-22 stated the cutoff level should be derived from data.
Commenter IV-D-27 stated that it should be set at the lowest uncontrolled emissions level for a
source that has MACT controls (i.e., the uncontrolled emissions from the single product dryer
used to establish the MACT floor control level) because EPA has no data to support the
0.01 gr/dscf standard on sources with lower emissions. Another commenter (FV-D-14) stated
that the 0.01 gr/dscf standard is not readily achievable using commercially available equipment
and that this level is not typical of fabric filter control. The commenter stated that a typical
vendor guarantee is an outlet particulate concentration of 0.02 gr/dscf for particles above
2 microns (^m) in diameter, with no guarantee for smaller particles. Consequently, the
commenter recommended that EPA consult with manufacturers of particulate control devices to
11-3
-------
ensure that the emission standard is set at a level consistent with vendor guarantees for readily
available equipment.
Response: No mass emission rate cutoff was established because all known bag dumps
that introduce a HAP raw material to a PAI process unit, and all product dryers that dry a product
that is a HAP, are controlled with fabric filters, and EPA believes 0.01 gr/dscf is a reasonable
level for all fabric filters in such applications. An emissions test for the fabric filter used to
control the product dryer at the MACT floor facility provides evidence that this concentration is
achievable. The outlet concentration was less than 0.01 gr/dscf for each of the 12 runs in the test.
The EPA expects that the existing fabric filters were designed to meet this outlet concentration,
but the standard and associated monitoring requirements are included in the rule to provide
assurance that they will continue to perform at this level. As a result, EPA did not change the
level of the standard, or add an applicability cutoff, for the final rule.
11-4
-------
12.0 POLLUTION PREVENTION STANDARD
Status at proposal. The proposed rule contained a pollution prevention alternative
standard that would allow owners and operators the opportunity to comply with the standards by
demonstrating reductions in the amount of HAP used per unit of product for a given process.
The pollution prevention alternative standard consisted of two options. Under the first option,
the owner or operator would be required to demonstrate an 85 percent reduction in the
consumption of HAP. The second option would allow a 50 percent reduction in consumption
combined with traditional emission controls to achieve an overall reduction equivalent to the
85 percent reduction specified under option 1. By demonstrating compliance with either of the
pollution prevention options, an owner or operator would satisfy the requirements for all process
vents, storage vessels, wastewater, equipment leaks, and heat exchanger systems associated with
a PAI process unit.
12.1 FORMAT OF THE STANDARD
Comment 1: Two commenters, FV-D-16 and 20, asserted that the 85 percent reduction in
HAP consumption factor should be changed to 75 percent for both pollution prevention options
to be consistent with the Pharmaceutical MACT proposal.
Response: The 85 percent reduction was not changed in the final rule to be consistent
with the value specified in the Pharmaceutical MACT standard because both values were
developed using industry-specific data. The basis for the 85 percent reduction target is the
overall nationwide reduction from uncontrolled emissions that is estimated as a result of the
implementation of the standards in this industry. Although the required reduction "target" was
calculated using the same methodology as that in the Pharmaceuticals MACT standard, the
difference in numerical value is simply due to differences in the impact of the two rules on each
respective industry. For the PAI production industry, the standards achieve slightly greater
12-1
-------
reductions relative to the uncontrolled baseline, which is carried forward to the reduction target
for the P2 alternative.
Comment 2: Commenter IV-G-03 stated that the mechanism to realize pollution
prevention reductions must be maintained in a system that can be managed and provide data that
regulated entities and EPA can use. The commenter asserted that States may not be prepared to
support this regulation with the training requirements of their already overworked staffs.
Response: The Agency agrees with the commenter that the information necessary to
demonstrate compliance with the pollution prevention alternative should be identified, collected,
and managed in a way that minimizes burdens on both the industry and the regulatory agencies
charged with enforcement. Therefore, the final rule requires sources seeking to comply with the
P2 alternative to submit, as part of the Precompliance pian, a P2 demonstration summary that
describes how the P2 alternative will be applied at the facilities, and what tracking mechanisms
will be used to demonstrate compliance with the alternatives. This summary should include
descriptions of how the facility will measure and record HAP consumption and production on a
daily, monthly, and annual basis. The summary should also include appropriate documentation
of how consumption will be tracked such as, but not limited to, operator log sheets, daily,
monthly, and annual inventories of materials and products, and shipment and purchasing records.
The P2 demonstration summary report allows the owner or operator some flexibility in deciding
the most reasonable and efficient way to demonstrate compliance, while incorporating the
regulatory agency's review and approval prerogative. Regarding the agency burden, the Agency
believes that compliance with the P2 alternative may actually reduce much of the burden on the
enforcement agency, in that the monitoring, reporting and recordkeeping burden will be reduced
to a material tracking effort, potentially minimizing the amount of data needed to demonstrate
continuous compliance (e.g., monitoring data) for an entire process.
Comment 3: Commenter IV-G-03 requested that EPA strengthen and encourage
pollution prevention alternatives in lieu of command and control regulations. Several
commenters, IV-D-01, IV-D-27, and IV-D-28, recommended allowing any combination of P2
and end-of-pipe controls to meet the required 85 percent reduction (or lower value, as noted in
comment 1 above) for P2 because:
12-2
-------
1. The requirements of option 2 would be impossible to achieve if less than 70 percent of
the remaining HAP after P2 is in the air stream (even if it were all controlled). Also, if more than
90 percent control is needed, there is no incentive to use P2.
2. Implementing P2 practices that achieve less than 50 percent of required reductions in
conjunction with add-on controls would bring overall removals to levels equal to or greater than
that required by the standard.
3. Option 2 places unnecessary constraints on the type of control device that can be used
to obtain required reductions, beyond that achieved by P2.
4. Proposed P2 is only viable if a fundamental change has occurred; sources should be
allowed to take credit towards the conventional standards for more limited P2 activities.
5. If P2 achieves reduction of 65 to 85 percent, it would be impossible to achieve the
additional requirement for 35 percent reduction from add-on controls.
6. Option 2 provides incentives for unsound environmental practices (destruction of
recovered material instead of reuse).
Commenter IV-D-27 suggested allowing sources to comply with 90 percent of any
applicable standard if at least 50 percent of the reductions are the result of P2, as an alternative to
the above comment.
Response:. In response to the comments, EPA stresses that the P2 alternative is an
alternative to the standards in the rule. As such, the Agency has flexibility in developing
requirements that may provide alternative approaches for compliance, but is charged with
preserving the reductions that would have been achieved through compliance with the standards
themselves. Under Option 2, EPA required that a significant portion (50 percent) of the
reductions be achieved using P2 techniques, not add-on controls. Without such a restriction,
owners and operators could attempt to use add-on controls entirely in meeting the P2 target
reductions, which might result in reductions that are less than those required by the standards.
For example, the process vent standard requires a 90 percent reduction in the HAP emissions
from affected processes, not an 85 percent reduction.
In an effort to ensure the emission reductions from the P2 alternative are at least
equivalent to the emission reductions achieved by the standards, the reduction target for the P2
consumption factor was linked to the predicted reductions from the nationwide uncontrolled
12-3
-------
emissions through implementation of the standards. It was always the Agency's intent that these
reductions would be achieved primarily through P2 techniques. In recognition of the difficulties
associated with achieving such high consumption reduction targets (85 percent), however, the
Agency developed Option 2 to allow some of the reductions to be achieved using add-on
controls. For these reasons, the Agency disagrees, in general, with the comments suggesting
lesser reductions in both the overall target of 85 percent and the requirement that at least
50 percent of the reduction be attributed to the P2 alternative. However, the Agency agrees with
the comments that option 2 as proposed is unworkable if the reduction achieved by P2 exceeds
50 percent of the required amount. For the final rule, option 2 was revised to require that at least
50 percent of the reduction be achieved using P2 and that the remainder of the 85 percent,
however much is needed, to be achieved using conventional controls.
Secondly, the Agency stresses that the restrictions on the types of add-on controls allowed
to be considered in addition to the P2 reductions in meeting the overall target are in place to
guard against double-counting of emission reductions; for example, control via a technique that
recycles HAP material back to the process is an environmentally beneficial technique and is
encouraged. However, the recycling effect will also reduce the consumption : HAP; therefore,
the recycling is already considered. To further reduce the consumption factor by the control
achieved by the condenser would result in double counting of emissions.
12.2 POLLUTION PREVENTION FOR REACT ANTS AND GENERATED HAP
Comment: Commenter IV-D-27 stated that P2 should be allowed for HAP generated in a
process. Commenter IV-D-28 indicated that the rule was not clear on how to comply when the
HAP generated in the process is the same as that introduced. Commenter IV-D-15 noted that
these exclusions would prevent them from using P2 and suggested that the rule include TRE
calculations like in the HON as a way to provide more cost-effective alternatives for processes
that are prohibitively expensive to control (i.e., that would exclude such processes from the
requirements of the conventional standards).
Response: The Agency reviewed the language contained in the proposed standard and
has revised it to capture the Agency's intent in restricting the use of the alternative in situations
where HAP's are generated, without prohibiting its use altogether. The Agency's concern, in
adding the restriction to the proposed standard, was that HAP generated in a process would not
12-4
-------
be addressed through the pollution prevention alternative because it requires only a reduction in
the consumption of HAP that are actually brought into the process. Therefore, a situation could
exist in which a process could be exempted from control because the production-indexed
consumption factors were reduced by adequate amounts (85 percent), while a potentially
significant amount of HAP, which happened to be generated in the process, could still be
emitted. The EPA agrees with the commenter that sources that generate HAP should be eligible
for the P2 standard, provided the HAP generated by the source are included in the analysis.
Therefore, the final rule allows owners and operators to use the P2 alternative for processes that
generate HAP that are not part of the production-indexed consumption factor (e.g., the HAP
generated are different from the HAP brought into the process), provided the following
conditions are met: (1) emissions of generated HAP are controlled to the levels required by the
applicable provisions for storage vessels, process vents, wastewater, and equipment leaks; and
(2) the P2 requirements are applied to the HAP that are added to the process. For the final rule,
the definition of consumption has been revised to consider quantities of HAP that are generated
by the process as well as those that are brought into the process, provided the HAP generated in
the process is the same as one (or more) of the HAP added to the process.
A related issue is the tracking of the VOC consumption-indexed production factor and the
proposed rule's requirement that this factor should not be increased as a result of pollution
prevention. Although this issue was not specifically commented on, EPA also revised the
language of the final rule regarding the production-indexed VOC consumption factor. In
developing the pollution prevention alternative, EPA's intention was to recognize those
processes that have reduced or will reduce the amount of HAP solvents used in the manufacture
of PAI's as viable alternatives to add-on controls. By preventing affected sources from
increasing the production-indexed VOC consumption factor, EPA intended to prevent solvent
substitutions that merely replaced HAP with VOC. After reviewing the proposed pollution
prevention standard, EPA realized that the proposed standard gave an unfair advantage to
affected sources that use VOC-HAP solvents as opposed to non-VOC-HAP solvents. As
proposed, the rule did not allow affected sources using non-VOC-HAP solvents to switch to low-
VOC solvents and still qualify under the pollution prevention alternative because such a switch
would increase the production-indexed VOC consumption factor. However, affected sources that
12-5
-------
use VOC-HAP solvents could switch to low-VOC solvents as long as the production-indexed
VOC consumption factor did not increase. The EPA's intention in the final rule is that pollution
prevention be accomplished through reductions in solvent usage as opposed to solvent
substitution. After consideration, EPA changed the final rule to require an equivalent reduction
in the production-indexed VOC consumption factor, if the reduction in the production-indexed
HAP consumption factor is achieved by reducing a HAP that is also a VOC. If the reduction in
the production-indexed HAP consumption factor is achieved by reducing HAP that is not VOC,
the consumption-indexed VOC factor may not be increased. In making these changes to the final
rule, EPA essentially eliminated the possibility of receiving credit, through the pollution
prevention alternative, for substituting VOC for HAP.
12.3 COMPLIANCE PROCEDURES
Comment: Commenter FV-D-28 stated that the P2 requirements have insufficient
safeguards against cheating. As examples, the commenter suggested that a facility could:
(1) claim a reduction for shutting down one process unit, (2) calculate current usage using
detailed information while inflating baseline emissions by using conservative assumptions in the
calculation, or (3) falsely assigning an inflated proportion of a site's historic emissions to a PAI
process without detailed supporting data. Commenter IV-D-28 also asserted that toxic release
inventory (TRI) data are not accurate enough to use for compliance determinations.
Response: As stated previously, the final rule requires owners and operators electing to
use the P2 alternative standard to submit a plan for demonstrating compliance with the
alternative. This plan must be reviewed and approved by the reviewing agency prior to the
implementation of the standard. Claiming a reduction for shutting down a process would be
rejected because a facility must show either consumption reductions or emission reductions for a
given process; consumption reductions on one process cannot be used to demonstrate compliance
for a different process. Also, the Agency believes that TRI data can be used to set the baseline
consumption factor if it can be linked to the process. The data contained in the TRI reports
essentially provide a mass balance of TRI chemicals brought onsite or produced, and their
environmental fates (air emissions, discharges to water, onsite or offsite destruction, etc.).
However, because TRI data are reported from the facility as a whole, it may not be feasible to
link data to individual processes, negating this approach for many facilities. The EPA also notes
12-6
-------
that TRI data have been allowed to demonstrate applicability with respect to synthetic minor
status for other rules such as the HON. In any case, the regulatory agency will be charged with
assessing the adequacy of the compliance method prior to approval.
12.4 RECORDKEEPING
Comment: Commenter IV-D-28 suggested that the rule allow at least a couple of weeks
from the end of each 30-day period before the owner or operator has to complete the rolling
average calculation.
Response: The EPA believes that it is important to determine if a violation has occurred
as soon as possible. For the P2 alternative, compliance is determined by calculating the
production-indexed consumption factors for operation over every 30 days for continuous
processes and every 10 batches for batch processes. These calculated factors must be reported in
the semiannual periodic reports, and each periodic report is due 60 days after the end of the
applicable reporting period. The consumption and production data must be collected for each
applicable 30-day or 10-batch period, but the calculations may be performed at any time prior to
submission of the periodic report. However, a violation will be charged to all days in which the
P2 factor was exceeded. Therefore, it is in the operator's best interest to discover an exceedance
and correct the problem as soon as possible to minimize the number of days that are included in
the period of violation.
12.5 BASELINE YEAR
Comment: Commenter IV-D-28 stated that the compliance options appear to reward
"laggards" while penalizing companies that implemented pollution prevention early. A company
that has not paid as much attention to pollution prevention may be able to comply by
implementing only pollution prevention, whereas a company that implemented pollution
prevention changes early may not have room for further improvement by pollution prevention
and thus be forced to install expensive emission control equipment to comply with the standard.
Response: During the development of the standard, EPA considered 1987 to be the
earliest year to set the baseline factor. The final rule has been revised to comprise a 3-year
average beginning no earlier than 1987, or the first year the process was operating, if 3 years of
data are not available, to account for sensitivity in data quality. The Agency notes that the P2
standard actually rewards people for making a noticeable improvement over time. For
12-7
-------
companies that have been proactive in minimizing wastes, including air emissions, from periods
of time prior to 1987, and, as a result, may not be able to take advantage of the P2 alternative, the
Agency believes that the rule also provides relief in the form of the cutoff based on controlled
emissions that is contained in the final rule for process vents. The Agency also notes that the
compliance date for the standard, in 2002, is approximately 15 years later than the 1987 baseline
and should provide an adequate window for demonstrating reductions.
12-8
-------
13.0 EMISSION AVERAGING
13.1 COMPLEXITY OF METHODOLOGY
Comment: Commenter FV-D-28 supported the concept of emissions averaging, but noted
that the provisions are so complex and burdensome that many owners and operators may be
deterred from using this option.
Response: The emissions averaging provisions provided in the proposed rule are
identical to those included in the HON. Further, the requirements are necessarily complex
because of the increased flexibility of the compliance approach provided by the provisions. As
stated in the HON promulgation preamble discussion, the EPA's goal in crafting the emissions
averaging provisions was to make emissions averaging available to sources faced with
controlling emission points that are particularly difficult or costly to control, while maintaining
the ability to demonstrate compliance with the standard.
13.2 RESTRICTIONS ON EMISSIONS AVERAGING
Comment 1: Commenters FV-D-16 and IV-D-27 suggested that EPA set a nominal
control efficiency for combustion devices used for air emission control for storage tanks and/or
wastewater at 98 percent. Commenter IV-D-27 asserted that EPA's wording in § 63.1362(k)(2)
of the proposed rule inappropriately restricts sources equipped with controls listed in that section
from generating emissions averaging credits.
Response: The Agency believes that the commenters would like to equate 98 percent
control to the performance specifications provided in the proposed rule for combustion devices
used for air emission control for storage tanks and/or wastewater sources. The Agency does not
agree that a nominal 98 percent should be assigned to these devices. Although EPA did establish
these performance specifications, EPA maintains that testing is important to ensure that a control
device can achieve the reported efficiency. For these reasons, EPA has required performance
13-1
-------
testing on combustion devices that control greater than 9.1 Mg/yr of HAP. Therefore, EPA will
not allow credits based on a control efficiency that has not been demonstrated. Secondly, the
provisions of § 63.1362(k)(2) incorrectly referred to the 98 percent and 95 percent control levels
as "nominal" control efficiencies. These efficiencies must be demonstrated via performance
testing and therefore should not be restricted from obtaining credits in emissions averaging. The
final rule corrects this oversight.
Comment 2: Commenters IV-D-16 and IV-D-27 believe EPA should delete the
restrictions that prohibit a source from calculating emission averaging credits for emission
reductions achieved prior to November 15, 1990 or with equipment installed to comply with
other State/Federal rules. The commenters believe these restrictions (1) are arbitrary, (2) are not
dictated by the Act, (3) unfairly limit economic incentives and thus impose unreasonable costs,
(4) penalize progressive companies, and (5) are inconsistent with procedures to develop the floor
(i.e., emission points that would be excluded from emissions averaging are used in setting the
standard). In addition, commenter IV-D-27 believes EPA's response to comments in the April
22, 1994 Federal Register notice on the HON are inadequate to justify the restriction.
Response: The Agency's policy on not allowing averaging of emission reductions for
controls in place prior to the passage of the 1990 CAA amendments was explained in the
April 22, 1994 Federal Register notice for the promulgated HON (59 FR 19426), and this
rationale is still applicable. In general, the emissions averaging provisions are designed to
provide an owner or operator with flexibility in designing a compliance strategy that optimizes
the use of existing controls, rather than replacing them. However, the final rule does not allow
credit for emissions reductions achieved by control devices installed before November 15, 1990
because EPA policy is that regulations must achieve additional reductions beyond what would
have occurred in the absence of the amended Act. Emission reductions achieved by controls that
were in place prior to November 15, 1990 would have occurred regardless of whether or not the
CAA was amended. If the rule allowed a source to take credit for these preexisting emission
reductions, the source could increase its emissions above the 1990 baseline levels. Regarding the
commenter's view that the restrictions penalize progressive companies, the Agency notes that, at
least for process vents that meet the applicability criteria for 98 percent control, owners and
operators who can demonstrate that controls achieving the MACT floor level of control
13-2
-------
(90 percent) were in place prior to the proposal date of these standards, are not required to
achieve the higher efficiency requirement of 98 percent. In this manner, companies who have
taken proactive measures to control emissions are actually rewarded. Additionally, the pollution
prevention alternative standard also rewards facilities who have demonstrated significant
reductions in their production-indexed consumption factors. Finally, these provisions have been
included in numerous regulations beginning with the HON, and they have been reviewed and
approved by OMB.
Comment 3: Commenters FV-D-16 and FV-D-27 objected to restrictions on emissions
averaging for "new sources." The commenters disagreed with EPA's rationale in the preamble
that this approach holds new sources to a stricter standard and that flexibility is unnecessary for
new sources. The commenters argued that using emissions averaging is the more stringent
approach because of the 10 percent discount factor that is applied to credits. Furthermore, the
commenters stated that new sources also need flexibility to comply with the standard in the most
economical and efficient manner; for example, if a new source is added to an existing facility
there may be opportunities to route emissions from the new source to existing controls, or to
overcontrol certain existing or new emission points to provide equal or greater environmental
benefit at lower cost. Also, commenters believe this restriction unfairly limits economic
incentives and imposes unreasonable costs.
Response: The EPA's policy on not allowing averaging of emission reductions for new
sources was explained in the April 22, 1994 Federal Register notice for the promulgated HON
(59 FR 19427), and this rationale is still applicable. As noted above, EPA designed emissions
averaging provisions to provide existing sources with flexibility in achieving compliance.
Instead of requiring the replacement of all existing controls that do not meet the level of the
standard, the emissions averaging provisions allow an existing source to optimize the use of
existing controls in the most economical and technically feasible fashion. The EPA maintains
that this concern does not apply to new sources because the owner or operator of a new source
would be able to integrate state-of-the-art controls into the design of the new source. However,
nothing in the rule prevents an owner or operator from routing emissions from a new PAI process
unit to an existing control that meets the required control levels. Finally, as noted in the previous
13-3
-------
response, these provisions have been included in numerous regulations beginning with the HON,
and they have been reviewed and approved by OMB.
Even if emissions averaging were allowed for new sources, certain other factors may limit
its feasibility for new sources. For example, new sources are subject to PSD/NSR review that
may require levels of control similar to those in the rule for new sources. In addition, because the
level of stringency in the new source standards is high (98 percent), achieving credit above and
beyond the 98 percent levels is probably unrealistic in most situations.
13-4
-------
14.0 TESTING PROVISIONS
14.1 TESTING CONDITIONS
Status at Proposal: Under the proposed rule, inlet and/or outlet emissions testing on a
control device would be required if the HAP load to the control device from all sources exceeded
10 tons/yr. This testing would be required under absolute, representative, or hypothetical peak-
case conditions. Any device could be tested under absolute or hypothetical peak-case conditions;
testing under representative conditions would be limited to thermal incinerators, wet scrubbers,
and carbon adsorbers.
Absolute peak-case conditions were defined by any of three conditions: (1) the period in
which the inlet to the control device contains at least 50 percent of the maximum HAP load
capable of being vented to the control device over any 8-hour period, (2) a 1-hour period in
which the inlet to the control device contains the highest HAP mass loading capable of being
vented to the control device, or (3) for a condenser as the control device, the 1-hour period of
time in which the gas stream capable of being vented to the condenser would require the
maximum heat removal capacity to cool the stream sufficiently, based on calculations, to achieve
the required removal efficiency. Hypothetical peak-case conditions would be simulated test
conditions that contain the highest hourly HAP load predicted to be vented to the control device.
For both absolute and hypothetical peak-case, the owner or operator would be required to
develop an emissions profile to identify the appropriate time period, and the profile would be
based on all streams vented to the control device, not just streams from PAI processes. The
owner or operator would develop the emissions profile based on process knowledge, engineering
analyses, or test data. Existing test data could be used, provided the data characterize current
process vent stream conditions.
14-1
-------
Representative peak-case conditions would be based on the 1-hour period that contains
the highest HAP mass loading rate from a single process, or for a condenser, the 1-hour period in
which the vent from a single process requires the maximum heat removal to cool the stream
sufficiently to achieve the required removal efficiency. In both cases, testing would be
conducted while that single process is venting, with or without simultaneous vents from other
processes.
Compliance would be demonstrated based on the results of a single run for runs that
exceed 1 hour, and three runs would be required if the duration of a run was less than 1 hour.
The owner or operator would not be required to conduct test runs exceeding 8 hours, even if the
duration of emissions from a process vent exceeded 8 hours.
Comment 1: Three commenters (IV-D-16, IV-D-20, and IV-D-27) requested that EPA
simplify the performance testing requirements. Commenter IV-D-20 stated that the performance
testing requirements for process vents must be simplified and aligned with the annual average
survey data on which the standards are based. Commenters IV-D-16 and IV-D-27 found the
"absolute," "representative," and "hypothetical" peak-case testing conditions in
§ 63.1364(b)(l l)(ii) of the proposed rule to be extremely confusing. They suggested simplifying
the provisions; commenter IV-D-27 provided the following language, and commenter IV-D-16
provided a slight variation:
Testing shall be performed under actual or simulated conditions (based on an emissions
profile) that includes any of the following at a minimum:
(1) at least the 1-hour time period with the highest mass loading rate (or maximum heat
removal requirement for condensers) of regulated pollutant to the control device, or
(2) the time period that contains at least 50 percent or the maximum mass loading rate (or
maximum heat removal for condensers) of regulated pollutant vented to the control device over
any 8-hour period.
Response: The Agency agrees that the wording in this portion of the proposed rule is
confusing. Therefore, in the final rule, EPA has made several changes to the testing language
that generally cover the commenter's suggested revisions, but also allow more flexibility in
defining the peak-case. These changes include the elimination of the option to test under
representative peak-case conditions and the elimination of testing requirements for condensers.
Additionally, EPA has expanded the testing language to cover factors other than the highest HAP
load that also impair control efficiencies (i.e., the most challenging conditions for the control
14-2
-------
device). These other factors that limit control efficiencies relate to characteristics of components
and the operating principles of the control devices. For example, the solubility of an emission
stream component in scrubbing media, or the affinity of an emission stream component for
carbon can also define the most challenging conditions for a particular control device. These
changes are described in more detail in the responses to comments 2 and 3 in this section and to
the response to comment 1 in section 14.5. The EPA is not sure what the commenter means by
aligning the performance testing with annual average survey data; however, as discussed in
responses to comments in sections 7.5 and 16.1, EPA does not believe the survey data are
applicable only on an annual basis.
Comment 2: Commenter FV-D-28 believes the requirement to test under peak-case
conditions should be deleted from the rule and replaced with a provision that allows testing under
representative conditions. The commenter cited the following reasons: (1) the peak-case
conditions might mistakenly be interpreted to mean they apply during startup, shutdown, and
malfunction; (2) the validity of a performance test depends more on the operating conditions of
the control device than on the operating conditions of the process; (3) other regulations have
allowed testing under representative conditions; (4) planned production schedules may not
produce the peak-case conditions before the deadline to demonstrate compliance, which would
mean operating for a test at higher or lower production rates than demand would dictate; and
(5) typically, the organic HAP reduction efficiency of a control device is fairly stable across a
wide range of HAP concentrations. Alternatively, if the peak-case concept is retained in the final
rule, the commenter recommends adding three provisions limiting the timing of testing.
Specifically, the commenter seeks assurances that a facility would not be required to operate
(1) at a rate, or in a manner, that could reasonably be expected to damage the process, control
device, or monitoring system; (2) in a manner that would produce unmarketable product (i.e.,
does not meet customer specifications); or (3) at a rate that would produce product in excess of
demand.
Response: The intent of compliance testing under peak-case conditions is to document
the reduction efficiency of the control device under the most challenging conditions. This
documentation is necessary to assure compliance in cases where the process operations yield
emission stream characteristics that may vary significantly over time, and where conditions
14-3
-------
approaching absolute peak-case may occur. Subsequent to the initial compliance test, continuous
monitoring of operating parameters established during the test is a reasonable measure of
continuous compliance with the efficiency requirement under all conditions. Presumably, the
control device should function as well or better under conditions that are not as challenging.
Although EPA is sensitive to unnecessarily increasing the burden associated with testing
of control devices for little benefit, the Agency still has concern that testing under
"representative" conditions (where "representative" is defined either as in the proposed rule for
representative peak-case or as a more general concept suggested by the commenter) may not be
sufficient to demonstrate that the control device will achieve required efficiencies under all
conditions. This is especially important as it relates to the continuous compliance demonstration
provision. Therefore, the option to test under representative peak-case conditions has been
eliminated for the final rule, and testing under representative conditions has not been added.
The final rule, however, does allow more flexibility in defining absolute and hypothetical
peak-case conditions. The definition of absolute peak-case load in the final rule incorporates the
possibility that conditions other than the highest HAP loading constitute the most challenging
conditions for the device. These conditions include, but are not limited to, periods when the
emissions to the device may contain the highest combined VOC and HAP load, periods when the
streams contain HAP constituents that approach limits of solubility for scrubbers, or periods
when the streams contain HAP that approach limits of adsorptivity for carbon systems.
The hypothetical peak-case conditions have also been expanded. In addition to
establishing hypothetical peak-case testing conditions based on a calculation of maximum actual
emissions, the final rule allows hypothetical peak-case conditions to be defined based on
equipment design features that limit the maximum hourly emissions that can be routed to the
control device. For example, a fan may limit the flowrate, and the concentration may be limited
to a certain percentage of the lower explosive limit before a bypass valve opens.
The Agency does not believe that the testing provisions in the final rule require operation
in a manner that could damage equipment, because the testing is only required for conditions that
have some reasonable likelihood of occurring. Thus, the design of the system should have
considered the possibility of operating under these conditions.
14-4
-------
Regarding the comment that the testing provisions should not require operation in a
manner that produces excess or unmarketable product, or in a manner that will not occur within
the time frame allotted prior to the compliance date, the Agency concedes that some
inconvenience to the source may occur, but believes that in most situations, facilities will be able
to work within the confines of the definitions to arrive at a set of testing conditions that minimize
production disruptions. The EPA notes that the requirement for submittal of the site-specific test
plan is also an opportunity for the facility to present site-specific information that may influence
the selection of testing conditions. The EPA encourages owners and operators to work with the
permitting agencies to arrive at solutions that meet the intent of this regulation.
Comment 3: Two commenters (FV-D-16 and IV-D-27) believe testing under
representative peak-case should not be limited to only thermal incinerators, carbon adsorbers, and
wet scrubbers.
Response: As discussed above, EPA has not allowed representative peak-case testing
under any circumstances because the Agency believes that the initial compliance demonstration,
which is used, in part, to set conditions for the continuous compliance demonstration, must be
conducted at the most challenging conditions. Therefore, testing is required under absolute peak-
case conditions or under simulated hypothetical peak-case conditions.
Comment 4: According to commenter FV-G-03, peak-case conditions that are not
relevant to an 8-hour or 1-hour condition need to be addressed.
Response: As described above, the final rule requires that control devices used to control
emissions from batch operations must be conducted under either absolute or hypothetical peak-
case conditions. These peak-case conditions are identified based on the results of an emissions
profile. All sources of emissions to the control device, including scheduling restrictions, are
considered in the emissions profile; alternatively, limitations of the equipment are identified.
From the emissions profile, the 1-hour or 8-hour period that defines the peak-case conditions is
determined. Thus, by definition, peak-case conditions are relevant to 1-hour or 8-hour periods.
If the most challenging conditions are not based on the highest load, the emissions profile and the
peak-case conditions are still determined over hourly periods.
14-5
-------
14.2 TEST PLAN
Comment: Commenter IV-D-28 believes that preparation of a site-specific test plan is
unduly burdensome, produces no discernible benefit, and should not be required. Therefore,
Table 1 also should indicate that § 63.7(c) of the General Provisions is not applicable.
Response: All section 112 MACT standards require the owner or operator to develop a
site-specific test plan in accordance with § 63.7(c) of the General Provisions. In the PAI rule
EPA requires that this site-specific plan be submitted. The test plan is needed because of the
complexity of identifying the most challenging conditions for testing.
14.3 EMISSION PROFILE
Comment 1: Section 63.1364(b)(l l)(iii) of the proposed rule specified how to develop
an emission profile. One sentence in that section was the following: "Previous test results may
be used provided the results are still relevant to the current process vent stream conditions."
Commenter IV-D-28 requested that the phrase "with or without adjustment" be added to the end
of that sentence.
Response: Generally, the Agency's position regarding the requirement of testing under
the most challenging conditions will not allow the use of data that have been collected under test
conditions that do not meet those specified in the final rule. However, if the conditions under
which the testing was originally conducted represent more challenging conditions than the
current operation, the rule provides a mechanism for owners and operators to extrapolate the
results of the initial test down to current operations, provided the methodology is presented in the
Precompliance plan and approved.
Comment 2: According to two commenters (IV-D-16 and IV-D-27), the rule should state
that development of an emission profile is not required if testing will be performed over the
duration of a complete batch cycle.
Response: An emission profile is needed to identify the period of time that constitutes
the most challenging conditions for the control device. The content of the emission profile will
vary depending on the circumstances. The most straightforward control scenario is one in which
a control device is dedicated to vent streams from a single process, and a test run can encompass
all of the emission events that are vented to the control device. For this situation, an acceptable
emission profile would describe the situation and explain how it meets at least one of the criteria
14-6
-------
that define the absolute peak-case conditions. This simplified approach would not be acceptable
if streams from multiple processes are routed to the control device, even if the test were to be
conducted over the duration of the batch.
14.4 TEST DURATION
Comment: Several commenters (IV-D-16, IV-D-27, and IV-D-28) requested clarification
of the test duration requirements in §§ 63.1364(b)(l l)(iv) and (v) of the proposed rule.
Commenters IV-D-16 and IV-D-27 stated that for continuous process vents, sources should be
allowed to test for longer than 1 hour per run, and commenter IV-D-28 asked for clarification of
this provision because it sounds as if three runs must be conducted in the same period of 1 hour
or less. These two commenters also stated that for batch process vents, sources should be
allowed to test for the entire length of a batch, even if it is longer than 8 hours. Similarly,
commenter IV-D-28 requested that EPA revise the language of the provision for batch process
vents to indicate that sources are not required to test for more than 8 hours per run, but that they
are not prohibited from conducting longer tests. Commenter IV-D-28 requested clarification that
the 8 hours of testing applies only to vents from batch processes, not continuous processes.
Commenter IV-D-28 also requested that EPA edit the phrase "... process vents of duration ..."
because it is the emission episode, not the vent, that has a duration.
Response: The final rule requires that three test runs with a minimum duration of 1-hour
each be conducted during the initial compliance demonstration. For consistency, each run must
be conducted over the same set of conditions.
14.5 CLARIFICATIONS
Comment 1: Commenters IV-D-16 and IV-D-27 requested clarification of the ways to
characterize the various peak conditions for condensers. They noted that in
§§ 63.1364(b)(l l)(ii)(A) and (B) of the proposed rule absolute and representative peak-case
conditions for condensers are based on the maximum heat load, rather than the HAP load, that is
used to define these conditions for other types of control devices. However, in
§ 63.1364(b)(l l)(ii)(C) of the proposed rule, hypothetical peak-case for all control devices is
only based on the HAP load. The commenters believe maximum heat load should be used to
characterize hypothetical peak-case for condensers.
14-7
-------
Response: The commenters identified an oversight in the proposed rule. However, rather
than implement the change the commenters requested, EPA has decided, for other reasons, to
delete the requirement to identify the period of maximum heat removal in an emissions profile
for absolute peak-case conditions (and as noted above, the final rule does not allow testing under
representative peak-case conditions). See the response to the comment in section 15.2.5 for a
discussion of the initial compliance procedures specified in the final rule for condensers.
Comment 2: Commenters FV-D-16 and FV-D-27 requested that EPA modify the language
in the testing conditions section of the rule to clarify that use of the terms "HAP load" and "HAP
concentration" means the specific class of HAP (either organic HAP or HC1/C12) for which the
performance test is conducted.
Response: Section 63.1365(b)(l 1) in the final rule includes a statement specifying that
the term "HAP mass loading," as used in that section, refers to the class of HAP that the control
device is intended to control. As a result of the revisions to the provisions for condensers (see
section 15.2.5), the term "HAP concentration" is not used in § 63.1365(b)(ll) of the final rule.
Comment 3: Commenter IV-D-28 requested clarification of the phrase "capable of being
vented" in the description of maximum HAP loads for absolute peak-case in
§§ 63.1364(b)(l l)(ii)(A)(2) and (3) of the proposed rule. The commenter speculated that it could
mean legal limits, maximum production, abnormal venting during a malfunction, hypothetical
scenarios, or control device capacity.
Response: The intent of the phrase "capable of being vented" is that the absolute peak-
case conditions are to be based on the most challenging conditions for the control device. The
emissions profile is to be used to determine and document these conditions, and in developing
the emissions profile the owner or operator may consider factors such as legal limits, maximum
production rates, hypothetical scenarios, and/or control device capacity. All of these factors are
related to production scheduling. Therefore, the final rule clarifies the description of absolute
peak-case by specifying that the emission profile for absolute peak-case is to consider production
scheduling. The most challenging conditions for absolute peak-case conditions are not based on
"abnormal venting during a malfunction" because the emission limitations do not apply during
periods of malfunction. In § 63.1365(b)(l l)(i) of the final rule, the phrase "capable of being
vented" has been changed to "that may be vented."
14-8
-------
Comment 4: Commenter IV-D-28 believes the testing requirements are covered so
completely in § 63.1364 of the proposed rule that there seems to be no need to say § 63.7(e)(3) of
the General Provisions is also applicable. If anything from § 63.7(e)(3) not yet in § 63.1364 is
needed, the commenter suggests that the rule more precisely identify it.
Response: Section 63.7(e)(3) of the General Provisions contains additional requirements
for test data reduction that are not completely covered in § 63.1364 of the regulation. The
Agency commonly references various portions of the General Provisions without re-writing the
contents of each reference verbatim in the regulation. Therefore, the final rule retains a reference
to the applicable portions of the General Provisions.
Comment 5: Section 63.134(b) of the proposed rule would require that Methods 2, 2A,
2C, 2D, 3, and 4 be performed twice during each test period. Commenter IV-D-28 asked why
this provision is necessary.
Response: For the final rule this provision was deleted to be consistent with the sampling
times in Appendix A of 40 CFR Part 60.
Comment 6: Commenter IV-D-28 requested that the rule not require the use of specific
units except when necessary. For example, in § 63.1364(b)(l 1), the commenter indicated that
there is no reason to require that maximum heat removal, mass loading rates, and duration be
expressed only in kilowatts, kg/hr, and hours, respectively.
Response: The Agency agrees with the commenter that many of the quantities that an
owner or operator is required to calculate (such as mass loading rate) do not need to be expressed
in only one set of units. The Agency has modified the final rule where appropriate to allow
owners or operators the option to choose the specific units that they want to use, and that are
acceptable to the implementing agency.
14-9
-------
15.0 INITIAL COMPLIANCE DEMONSTRATIONS
15.1 GENERAL
Comment 1: Commenter IV-D-28 believes that an owner or operator should be allowed
to deem a stream to be a Group 1 stream, and that for such streams there should be no
requirement to calculate, estimate, or otherwise determine Group status.
Response: Since proposal, EPA re-examined the procedures for determining Group 1
process vents, storage vessels, and wastewater streams. The proposed rule would allow owners
or operators to designate a wastewater stream to be a Group 1 stream, and this provision is
retained in the final rule. As a result of the review, EPA changed the final rule to allow owners
and operators the option to designate any storage vessel as a Group 1 storage vessel, but
determined there is no reason to designate Group 1 process vents. Rationale for the decisions
regarding storage vessels and process vents are described in the remainder of this response.
For storage vessels, Group status is required for determining applicability of the standard,
and it is used in emissions averaging. As noted in chapter 8, the applicability cutoffs for the
storage vessel standard have been changed from both the annual mass emissions and tank size
under the proposed rule to both the maximum true vapor pressure and tank size under the final
rule. These characteristics are also used to determine Group status. Designating that a storage
vessel stores material with a maximum true vapor pressure above the applicability cutoff (or
below the size cutoff) is acceptable because it results in control that is equal to or more stringent
than the standard and has no impact on emissions averaging calculations. Thus, the final rule
allows an owner or operator to designate a storage vessel as a Group 1 storage vessel, but the
owner or operator may still need to determine the maximum true vapor pressure for a compliance
demonstration. For example, if the owner or operator conducts a design evaluation or controls
15-1
-------
emissions from the storage vessel with a condenser, the owner or operator may elect to use the
maximum true vapor pressure in emissions calculations.
Determining Group status for process vents is needed only to identify Group 1 process
vents for which debits must be calculated; it is not needed to comply with the regular standard.
To determine Group status, the owner or operator must calculate the uncontrolled emissions from
all process vents within a process. Because this same information is needed to calculate debits
for emissions averaging, designating process vents does not reduce the burden to comply with the
emissions averaging provisions. However, there are situations for which the owner or operator
would not be required to calculate uncontrolled emissions as part of the compliance procedures.
For example, uncontrolled emissions are not needed if the owner or operator complies with the
alternative standard; thus, this compliance procedure has the same effect as designating the vents
as Group 1 process vents.
Comment 2: Commenter IV-D-18 believes the regulation should include compliance
procedures for catalytic incinerators. The commenter indicated that the design evaluation would
include a minimum temperature requirement for the destruction efficiency, the stream's flow
rate, and a minimum residence time.
Response: Catalytic incineration is a viable control option for the PAI production
industry, just as it is for the synthetic organic chemicals, pharmaceuticals, and other industries.
The rules that apply to these other industries include design evaluation criteria (and monitoring
criteria) for catalytic incinerators. To be consistent, these criteria have also been added to the
final PAI rule,
Comment 3: Commenters IV-D-18 and IV-D-28 addressed the provision in
§ 63.1364(c)(3)(ii)(B) of the proposed rule that requires an owner or operator to correct the outlet
TOC emissions to 3 percent oxygen. Commenter IV-D-18 disagrees with this provision because
many thermal and catalytic incinerators properly operate v, ith higher oxygen levels in the exhaust
stream. Therefore, this commenter believes that the measured outlet concentration of TOC
should be corrected to the design oxygen content in the outlet stream. Commenter FV-D-28
stated that § 63.1364(c)(3)(ii)(B) should require concentration corrections to 3 percent oxygen
only for combustion devices.
15-2
-------
Response: The general provisions prohibit the use of dilution as a means of achieving
compliance with a standard (see 40 CFR 63.4(b), Circumvention). However, EPA also
recognizes that there are valid reasons for introducing air or inert gases into manifolds for safety
or design considerations. For example, supplemental combustion air may be required for proper
operation of an incinerator. The intent of the proposed requirement for correction to 3 percent
oxygen was to allow an owner or operator to add supplemental combustion air, but only take
credit for the amount that is needed for proper operation. As one commenter noted, this
correction was not intended to apply to other types of control devices.
The correction to 3 percent oxygen concentrations was drawn from the HON and the
earlier SOCMINSPS. Under these rules, this correction is required for purposes of
demonstrating compliance with a 20 ppmv outlet concentration standard. The value of 3 percent
originates from good engineering practices. For oxygen deficient streams, if the proper amount
of supplemental combustion air is added, the outlet stream would contain approximately 3
percent oxygen. Typically, SOCMI facilities have low oxygen, high VOC/HAP concentration
streams that generally require supplemental combustion air when they are combusted. Therefore,
a correction to prevent dilution was needed in rules for the SOCMI industry.
A similar requirement to correct the outlet concentration was included in the Polymer
Manufacturing NSPS. Commenters on the proposed NSPS asserted that an oxygen correction
may be appropriate for oxygen deficient streams to which supplemental combustion air is added
to ensure combustion of the emissions, but it is not appropriate for high oxygen, low VOC
concentration streams. The commenters on the proposed NSPS further stated that requiring an
oxygen correction for processes with inherently high oxygen concentrations would prevent
facilities from being able to use the 20 ppmv outlet concentration compliance option. Because at
some point the combination of low VOC/HAP concentration and technology limitations of
control devices makes it impossible to achieve a high percentage reduction (98 percent in the
case of the Polymers NSPS), the 20 ppmv outlet concentration may be the only compliance
option for some streams. As a result of considering these comments, the final rule for the
Polymer NSPS was changed to require a correction to 3 percent oxygen only if supplemental air
was used to combust emissions.
15-3
-------
Other available information indicates that for some Pharmaceuticals processes, dilution is
needed for safety or design considerations other than for use as supplemental combustion air.
Typically, this dilution occurs in manifolds conveying emission streams from unit operations that
already have high oxygen concentrations, and it occurs for control devices other than
incinerators. Although EPA does not have similar information for the PAI production industry,
the information from the surveyed plants supports the commenters contention that there are
process vent streams with high oxygen concentrations. It is also possible that some of these
streams are diluted for reasons other than to supply supplemental combustion air.
It is not EPA's intent to prohibit the introduction of dilution air or other gases, only to
ensure that outlet concentrations are corrected for such dilution. As a result, EPA made a
number of changes in the requirement to correct outlet concentrations to prevent dilution. First, a
definition of "supplemental gases" has been added to the final rule; this term includes any
nonaffected streams that contain less than 20 ppmv TOC and less than 20 ppmv HC1/C12 that are
combined with affected streams. Second, the final rule clarifies that the correction to 3 percent
oxygen applies only for incinerators, and only if supplemental gases are added. Third, the final
rule explicitly describes procedures to correct for dilution in noncombustion devices.
Comment 4: Three commenters addressed the issue of compliance procedures when the
control device is in compliance with RCRA provisions. Commenter FV-D-27 believes the rule
should specifically state that test methods and compliance provisions under RCRA can be
followed instead of the corresponding provisions in this rule for air emissions and/or wastewater
routed to RCRA incinerators covered under 40 CFR part 264, subpart O or 40 CFR part 265,
subpart O (40 CFR 264/265 subpart O). Commenter IV-D-16 echoed this position for
wastewater and associated residuals and air emissions. According to commenter IV-D-27, this
change would avoid overlapping or contradictory requirements. Commenter IV-D-16 requested
the change because the cross-referenced sections of the HON provide exemptions from
performance testing for other types of units (e.g., BEFs), but do not mention RCRA incinerators.
Commenter IV-G-05 asked if an affected source that conducts a trial burn to demonstrate
compliance with RCRA regulations can use this same demonstration to comply with the
performance testing requirements of the PAI NESHAP.
15-4
-------
Response: Under the proposed rule, the owner or operator would be exempt from the
requirement to conduct performance tests to demonstrate compliance with the process vent and
storage tank standards if the emission streams are vented to a RCRA control device.
Specifically, the exemption would apply for hazardous waste incinerators subject to subpart O
and boilers and process heaters subject to subpart H. The cross-referenced wastewater provisions
in § 63.138(h) of the HON would exempt an owner or operator from design evaluation and
performance test requirements (as well as monitoring requirements, and associated recordkeeping
and reporting requirements) for wastewater and residuals treated in RCRA incinerators, boilers,
process heaters, and underground injection wells. In addition, cross-referenced provisions for
wastewater vent streams in § 63.139(d)(4) of the HON are identical to the provisions described
above for process vents and storage tanks, except that they exempt the owner or operator from
design evaluations as well as the performance test requirements.
All of the provisions in the proposed rule have been retained in the final rule. In addition,
to be consistent with other standards, the final rule exempts an owner or operator from all initial
compliance requirements (i.e., emission calculations and design evaluations as well as
performance tests) for process vents and storage tank vents controlled by a RCRA unit. After
making this change, EPA believes that the initial compliance provisions in the final rule are
consistent with the concepts requested by the commenters. Similar comments for monitoring,
recordkeeping, and reporting provisions are discussed in sections 16.10, 17.8, and 18.8,
respectively.
Comment 5: If a facility is using a combustion control device that is capable of burning
liquids, commenters IV-D-21 and FV-D-29 recommended that EPA allow initial compliance
demonstrations based on a metered liquid performance test. According to the commenters, this
approach would be more consistent with the data provided by the industry in response to the
section 114 information request and used to develop the MACT floor. They also note that it
would be a way to streamline the compliance process at lower cost to the facilities.
Response: The EPA rejects the commenters suggestion. The final rule requires that
testing be conducted under the most challenging conditions for the control device, which are
intended to be either the highest hourly HAP load (in vapor phase), or, if the most challenging
condition is not defined by load, other challenging conditions (such as component solubility for
15-5
-------
liquid scrubbers or affinity for carbon). Unless the owner or operator can demonstrate that a test
conducted under conditions in which liquid waste is metered represents the most challenging
conditions for a control device, it may not be used for a compliance determination.
A metered liquid test would not be consistent with the MACT floor determination. As
for many standards, the MACT floor was determined based on reported control efficiencies from
surveyed facilities in the industry. The surveyed facilities based reported control efficiencies on
a variety of information. At most, the reported process vent control efficiencies for only two
processes at one facility were based on the results of a metered liquid test. These reported values
were comparable with those for hazardous waste destruction efficiencies in a RCRA incinerator.
However, for the MACT floor analysis, EPA considered the efficiencies to be only 98 percent.
Based on historical data, this is a typical value for a well designed and operated
combustion-based control device. Even if this value is still high, it would be only slightly high,
and the difference would not affect the MACT floor determination.
15.2 INITIAL COMPLIANCE WITH PROCESS VENT STANDARDS
15.2.1 Emission Estimation Calculations
Status at Proposal: The proposed rule included equations for calculating emissions from
five types of emissions episodes: vapor displacement, purging, heating, depressurization, and
vacuum systems. Alternatively, if the owner or operator could demonstrate that the proposed
methods were inappropriate, the proposed rule specified that an owner or operator conduct an
engineering assessment to estimate emissions. The proposed rule stated that the specified
equations would be considered inappropriate if either of two criteria are met: (1) if available test
data and the estimated value differ by more than 20 percent, or (2) the owner or operator
developed any other means to show the equations are not appropriate for a given batch emissions
episode.
Comment 1: One commenter (IV-D-14) believes facilities should be allowed to calculate
emissions based on all available information, including, but not limited to, the equations in the
proposed rule, and they should not have to demonstrate that the equations in the rule are
inappropriate. According to the commenter, it is not logical to require facilities that produce a
variety of products, only a small portion of which are PAI's, to modify their calculation
methodology; nor is it logical to require recalculation on a large scale when the existing
15-6
-------
emissions estimates are based on fundamentally sound principles. The commenter also noted
that facilities may already have invested significant resources to develop methodologies for
calculating emissions.
Another commenter (TV-D-28) had several comments on the provisions for estimating
emissions if the emission estimation equations are inappropriate. First, the commenter requests
clarification of the statement that specifies the emission estimation equations shall be considered
inappropriate if certain criteria are met because it could have several interpretations. For
example, the commenter suggested it could mean "not applicable for any one equation," "not met
for any equation at all," or "not met for every equation at the same time." Second, the
commenter supports the provision that allows any method of demonstrating the emission
estimation equations are inappropriate. Finally, the commenter noted that the concept that the
emissions estimation equations are inappropriate if there is a greater than 20 percent discrepancy
between a test value and the calculated value was also used in the P&R I and P&R IV rules,
which are in litigation. If the litigation results in changes to this concept, the commenter believes
EPA should take comment on incorporating the same changes into this rule.
Response: For the final rule EPA did not change the requirement to use equations to
estimate emissions when the emission episodes fit the descriptions provided in the rule. The
EPA believes that the equations in the rule are the most appropriate methods to estimate
emissions from seven specific types of emission episodes. The requirement to use the equations,
when appropriate, also is important in standardizing compliance procedures for the industry and
in providing replicable procedures which the regulated community and EPA can follow to assure
compliance. However, the rule also allows owners or operators to request approval to use
alternatives for estimating emissions. The EPA believes it is important that the owner or
operator be able to make a case for any alternative approach. Finally, as noted in the response to
comment 1 in section 15.1, there are situations where uncontrolled emissions do not have to be
estimated.
To clarify when an engineering assessment must be conducted and when it may be
conducted, § 63.1365(c)(2)(ii) of the final rule has been changed to read as follows:
(ii) Engineering assessments. The owner or operator shall conduct an engineering
assessment to determine uncontrolled HAP emissions for each emission episode that is not due to
vapor displacement, purging, heating, depressurization, vacuum operations, gas evolution, or air
15-7
-------
drying. For a given emission episode caused by any one of these seven types of activities, the
owner or operator also may request approval to determine uncontrolled HAP emissions based on
an engineering assessment. All data, assumptions, and procedures used in an engineering
assessment shall be documented in the Precompliance plan in accordance with § 63.1367(b). An
engineering assessment includes, but is not limited to, the information and procedures described
in paragraphs (c)(2)(ii)(A) through (D) of this section.
(A) Test results, provided the tests are representative of current operating practices at the
process unit. If test data show a greater than 20 percent discrepancy between the test value and
the estimated value, the owner or operator may estimate emissions based on the test data; the
results of the engineering assessment shall be included in the Notification of Compliance Status
report, not the Precompliance plan.
The information and procedures in subparagraphs (B) through (D) are the same as in the
proposed rule. (Note that equations for gas evolution and air drying have been added since
proposal. Also, the second criterion from the proposed rule has been deleted from the final rule
because it is redundant. Note that nothing in the rule prohibits an owner or operator from
requesting approval for any alternative method, but only the 20 percent discrepancy criterion is
"preapproved.") Following a decision concerning the litigation on other rules that contain the
concept that test data may be used if there is more than a 20 percent discrepancy between a test
value and a calculated value, EPA will determine if any changes to the final rule are necessary
and determine the procedure for implementing any such changes.
Comment 2: Commenter IV-D-28 requested clarification of the requirements for
calculating emissions from vacuum systems. The proposed rule states that emissions from
vacuum systems "may be calculated if the air leakage rate is known or can be approximated,
using Equation 11." According to the commenter, this could be interpreted to mean that
emissions from vacuum systems do not have to be calculated, but if that is true, it is not clear
what would be required. The commenter also asked what an owner or operator is supposed to do
if the air leakage rate is not known or cannot be approximated.
Response: For the final rule this provision has been simplified to state that "emissions
from vacuum systems shall be calculated using Equation 24." The revised statement is more
consistent with the statements associated with equations for the other types of emission episodes.
The owner or operator is responsible for determining how to estimate the value for this variable,
just as the owner or operator is responsible for determining the temperature or volume to use in
other equations.
15-8
-------
Comment 3: For purge streams with flow rates greater than 100 scfm,
commenter IV-D-28 believes an owner or operator should not be required to assume the mole
fraction is 25 percent of the saturated value if they have data or have conducted an engineering
assessment to support another value. The commenter also recommended that EPA consider
adding an assumed value (and allow the use of test data or engineering assessments) for
situations where the purge is beneath the liquid level; below the contact area of a scrubber,
absorber, or distillation column; or another location with sufficient liquid/vapor contact.
Response: The proposed and final rules acknowledge that emission episodes may arise
that are not covered by the equations provided in the rule or where the assumptions used to
develop the equations do not apply. Under these circumstances the rule allows the owner or
operator to calculate uncontrolled emissions by conducting an engineering assessment. For
example, if an owner or operator has information to support a different mole fraction than the
default of 25 percent, the owner or operator may submit an engineering assessment in the
Precompliance plan requesting approval to use the alternative value. The Agency decided not to
include an equation for the sparging situations described by the commenter because these are
likely to be site-specific events that do not easily lend themselves to a generalized equation.
Comment 4: Commenter IV-D-28 believes EPA needs to clarify the procedures for
demonstrating compliance when PAI and non-PAI emissions are combined prior to control. If
the PAI contribution to the combined emission streams entering a control device is small, the
commenter wants to know how to demonstrate that the control device achieves the required
reduction for the HAP from the PAI process, not the other process(es). To address this issue, the
commenter suggested adding provisions similar to those in the distillation NSPS (subpart NNN).
As a related point, the commenter stated that the rule should specify how to determine the
uncontrolled emissions if PAI and non-PAI emissions are combined prior to the final recovery
device.
Response: The distillation NSPS specifies procedures for establishing sampling sites and
determining outlet emissions when both nondistillation and distillation vent streams are
combined before a recovery device. There are two inlet sampling sites-one in the distillation
vent stream prior to the combination and the second in the combined stream. The concentration
at the second site is used with the outlet concentration to determine the efficiency of the recovery
15-9
-------
device. This value is applied to the concentration from the sampling site in the distillation vent
stream to determine the outlet emissions associated with that specific stream.
Under the proposed rule, the owner or operator would demonstrate compliance with the
percent reduction format of the standard by first using the specified equations or an engineering
assessment to determine the uncontrolled emissions from each emission episode. These values
are equivalent to the emissions based on testing at the sampling site in the distillation vent stream
under subpart NNN. For large control devices, the owner or operator would then determine the
efficiency based on testing at the inlet and outlet of the control device. The "inlet" in this case
would be after all streams to be controlled are combined. Thus, if streams are mixed before a
control device, the performance of the device is evaluated over the mixture. The resulting
efficiency, which is based on the most challenging conditions, regardless of whether they are due
to the PAI vent stream or other vent streams, would then be applied to the uncontrolled emissions
from individual emission episodes to determine the outlet emissions associated with those
episodes. Although the language in the proposed rule differed from subpart NNN, EPA believes
the effect is the same. Therefore, the procedures in the proposed rule have been retained in the
final rule.
The concept of recovery devices for process vent streams has been deleted from the final
rule; the reasons for this change are described in responses to the comment in section 3.4 and the
response to comment 4 in section 5.2. As a result, estimating uncontrolled emissions prior to a
recovery device is not an issue under the final rule. The final rule does retain the concept of
process condensers. However, by definition, a process condenser is an integral part of a process,
which means it would not receive combined streams from multiple processes.
15.2.2 Compliance with Outlet TOC Limit
Comment 1: Three commenters (IV-D-16, IV-D-27, and IV-D-28) believe EPA should
justify why a performance test to demonstrate compliance with the outlet TOC concentration
under § 63.1364(c)(l)(viii) of the proposed rule must be conducted only under absolute
peak-case conditions. Commenters IV-D-27 and IV-D-28 also believe this section unnecessarily
restricts the choice of test methods that may be used to demonstrate compliance with the outlet
TOC concentration. Both commenters requested that this section be modified to allow
combinations of test methods to measure TOC, consistent with § 63.1364(b)(6).
15-10
-------
Commenter IV-D-27 also requested that this section be modified to allow measurement of total
organic HAP using Method 18.
Response: The final rule has been modified to include two options for demonstrating
compliance with the outlet TOC concentration: (1) continuously monitor outlet concentration
using an FED or other device; or(2) perform an initial performance test at absolute or hypothetical
peak-case conditions and continuously monitor operating parameter levels to demonstrate
continuous compliance with outlet TOC concentration. Initial testing at the absolute or
hypothetical peak-case conditions is not necessary for Option 1 because continuous compliance
is determined through the use of an FID or other device that continuously monitors outlet
concentration (however, if the monitor is to be calibrated on the predominant HAP, it may be
necessary to perform an initial test to identify the HAP). Conversely, EPA believes testing under
the absolute or hypothetical peak-case conditions is necessary for the second option to ensure that
operating parameter levels are established that will ensure compliance under all operating
conditions. The monitoring requirements for Option 2 are the same as the monitoring
requirements for complying with the percentage reduction format of the standard. Therefore,
EPA believes the initial testing that is used to establish the monitoring parameters should also be
the same in both cases.
Finally, EPA has modified the final rule so as not to restrict the choice of methods that
the owner or operator may use to determine TOC (i.e., Method 18 is allowed for speciation).
However, EPA emphasizes that the concentration limit is based only on TOC, not total organic
HAP.
15.2.3 Exemption from Performance Testing
Comment 1: The provisions in §§ 63.1364(c)(5)(i) and (d)(5)(iii) of the proposed rule
would exempt a facility from the requirement to conduct a performance test if a previous test was
conducted using the same procedures. Three commenters (IV-D-21, IV-D-28, and IV-D-29)
believe these provisions are useless because it is unlikely that a previous performance test would
have been conducted using the same procedures and under the same peak-case conditions.
Commenter IV-D-28 believes that any test conducted on the subject control device to
demonstrate compliance under any EPA-supervised program (e.g., NSPS, NESHAP, RCRA, new
source review) should be sufficient to demonstrate compliance with this regulation.
15-11
-------
Furthermore, this commenter stated that a shared control device should not be subject to multiple
compliance demonstrations; one should be sufficient to show compliance with all standards.
Commenter F/-D-29 suggested adding another option in § 63.1364(b)(ll)(ii) which provides for
an alternative site-specific test and/or design evaluation plan developed by the company for the
following situations: (1) when batch process units share common abatement with continuous
process units (because batch emissions cannot be separated from continuous emissions), (2) for
batch processes which campaign different products with different HAP that use the same
equipment (perhaps even with years between production runs), or (3) for batch processes which
have unique circumstances where the prescribed methods in § 63.1364(b)(l l)(ii)(A) through (C)
are not feasible.
Response: The Agency maintains that performance testing under peak-case conditions is
necessary to ensure compliance over a range of conditions, especially when variability in
emission stream characteristics cannot be predetermined. If a test was not conducted under the
specified peak-case conditions, it does not provide assurance that the required level of control
can be met under such situations.
The EPA also is not convinced that there are circumstances under which it would be
necessary to allow an alternative to testing under peak-case conditions. As noted in section 14.1
of this document, the final rule allows testing under absolute or hypothetical peak-case
conditions, and these conditions are established based on an emissions profile. The final rule
also provides several options for developing the emissions profile. The goal is to identify the
most challenging conditions that the control device will encounter, regardless of the emissions
source. Thus, manifolding batch and continuous process vents is acceptable; the batch emissions
do not have to be evaluated separately. Also, the approach can accommodate different batch
processes operating at different times with different HAP. Even if the owner or operator
determines that the most challenging conditions only occur for a process that is run infrequently
and at a time that is not convenient for testing, the owner or operator may simulate conditions for
the test that are similar to (or more challenging than) the actual most challenging conditions (i.e.,
test under hypothetical peak-case conditions). Thus, the final rule was not changed to
incorporate the commenters suggestions.
15-12
-------
Comment 2: Commenter FV-D-28 supports the provision that allows initial compliance
demonstrations for small control devices to be based on design evaluations, but three
commenters (IV-D-14, IV-D-16, and IV-D-27) requested changes in the cutoff that defines the
minimum size of a control device for which a performance test must be conducted to
demonstrate compliance. Commenters IV-D-16 and IV-D-27 recommended that the cutoff be
changed from "greater than 10 ton/yr of HAP" to "greater than 10 ton/yr of any individual HAP
or greater than 25 ton/yr of aggregated HAP" because this would be consistent with rationale in
the Basis and Purpose document, which is that this criterion is based on the application of the
major source cutoffs. These two commenters also believe the rule should be clarified to indicate
that the cutoffs are for the class of pollutant for which the performance test is being conducted
(i.e., organic HAP or HC1/C12); this clarification is needed to avoid the situation where a
performance test would be required to demonstrate control of a negligible amount of organic
HAP simply because the quantity of HC1 present is above the cutoff. Commenter IV-D-14
believes the proposed cutoff should apply only to HAP emissions from PAI processes if the
control device is used to control emissions from both PAI processes and other types of processes.
Response: The EPA continues to believe that the testing cutoff for control devices is
proper. In developing the regulation, EPA could have required testing for all control devices.
However, EPA proposed the testing cutoff to decrease the testing burden on the industry. For
devices handling lesser loads, EPA believes that the design evaluation will be adequate to
demonstrate compliance.
The applicability cutoff for performance testing applies to the inlet stream to the control
device, regardless of the amount from PAI manufacturing contributing to the total emissions or
the type of HAP (i.e., organic HAP or HC1/C12). The intent of the cutoff is to identify larger
emissions sources where PAI and other emission occur, not only what is emitted from the source
category. However, if both organic HAP and HC1/C12 emissions are vented to a control device,
the owner or operator is only required to conduct a performance test for the class of pollutant that
the control device is being used to control. To clarify this point, the final rule states that the
owner or operator may assume the control efficiency of HC1/C12 to be zero percent if the control
device is intended to reduce only organic emissions and vice versa.
15-13
-------
15.2.4 Provisions for Flares
Comment: Section 63.1364(c)(4)(v) of the proposed rule would exempt an owner or
operator from an initial compliance test for a flare that complies with the criteria in § 63.1 l(b) of
the General Provisions. Commenter IV-D-28 questioned whether this provision requires an
initial compliance demonstration when the control device is a flare. According to the
commenter, this provision could be interpreted to mean that if the flare ever fails to meet the
criteria in § 63.11(b) that a performance test would have been required for initial compliance,
which would not be appropriate for flares anytime, and also would be an unfair retroactive
violation. The commenter notes that this provision also refers to the performance criteria for a
flare in the General Provisions. However, the commenter believes the General Provisions do not
require a compliance demonstration; therefore, that requirement needs to be stated in this
regulation. If it is required, the commenter recommends adding a new paragraph to the
section on testing provisions in the rule instead of revising paragraph (c)(4)(v) because a flare
could be used to control more than process vents. It should require an owner or operator to
demonstrate compliance by passing a visible emissions test and determining the exit velocity and
net heating value using the appropriate procedures from § 63.1 l(b).
Response: The Agency agrees with the commenter that § 63.1 l(b) is not clear on whether
an initial compliance demonstration is required. In response, the Agency has revised the final
rule to clearly state that Test Method 22 in appendix A of 40 CFR part 60 shall be used to
determine opacity, Test Method 18 in appendix A of 40 CFR part 60 shall be used to ensure the
proper heating value of the gas being combusted, and Method 2, 2A, 2C, or 2D of appendix A of
40 CFR part 60 shall be used to determine exit velocity. The EPA also agrees with the
commenter that the provisions should not be limited to flares used for process vent emissions.
As a result, paragraph (c)(4)(v) in the proposed rule has been replaced with revised provisions in
an introductory paragraph in § 63.1365 so that they are applicable to all flares.
15.2.5 Provisions for Condensers
Comment: Commenter IV-D-28 believes the rule should allow an owner or operator the
option to demonstrate compliance for a condenser using either the temperature measurement
option or a performance test. According to the commenter, the temperature measurement option
is beneficial because it may reduce the compliance burden in some cases, but the traditional
15-14
-------
testing method is also needed because the temperature measurement method may result in false
determination of noncompliance and is not necessarily always less burdensome.
Response: Under the proposed rule, EPA included three options for sources to determine
emissions and control efficiencies for condensers: (1) performance testing including
measurement of HAP concentration and flowrate under peak-case conditions, (2) direct
measurement of temperature of the outlet gas under peak-case conditions, or (3) emission
estimation. Since proposal, EPA identified the following problems with the proposed options:
(1) direct measurement of temperature is a procedure to demonstrate ongoing compliance, not
initial compliance; (2) for condensers, determining the control efficiency during the peak-case
conditions does not ensure that the same or higher control efficiencies will be achieved under
other conditions; (3) options 2 and 3 are not independent because the outlet temperature is
needed to estimate emissions from a condenser; and (4) performance testing is not a replicable
procedure for batch processing operations and is unnecessary for establishing the control
efficiency. To address these concerns, the final rule was revised to include only one procedure
for demonstrating initial compliance when using a condenser. This procedure requires
calculation of the outlet temperature that is needed to achieve the required control efficiency for
an emission episode (or group of episodes).
Determining the control efficiency for condensers under the peak-case conditions does
not ensure that the control efficiency under other conditions will be the same or higher. Under
the proposed rule, the peak-case conditions were defined based on the stream from which the
maximum amount of heat must be removed over a specified time period to achieve the required
emissions reduction. However, to achieve the required control efficiency for another emission
stream with a different pollutant and/or temperature may require a significantly lower outlet
temperature, even though less heat is removed. Basing the monitoring on the temperature for the
stream with the maximum heat removal requirement would not ensure that the lower outlet
temperature could be achieved for the other stream.
The revised procedure for the final rule is a replicable protocol in that for identical inlet
conditions, every source will estimate the same controlled emissions and control efficiency when
using the same outlet temperature. Performance testing for batch processing operations, on the
other hand, can be difficult and can lead to considerable v an ability in results. In addition to
15-15
-------
concerns about replicable results, the performance testing provisions in the proposed rule were
not structured to properly account for control efficiency of condensers under all conditions.
Under the performance testing option in the proposed rule, the control efficiency would be
determined for the peak-case conditions. Then, using the heat removal rate that occurred during
the test, the outlet temperatures, and thus control efficiencies, could be calculated for other inlet
conditions. However, a performance test is not needed because these temperatures can be
calculated based on the properties of the emission streams. For these reasons, the final rule does
not specifically require testing of condensers (e.g., measurement of flowrate and concentration to
generate a mass rate) as a means of compliance with the standards. However, as with other
practices, owners and operators can propose alternative means of demonstrating compliance with
the standards for approval on a case-by-case basis. 15.2.6 Provisions for Thermal Incinerators
Comment: Commenters IV-D-16 and IV-D-27 requested that EPA justify the provision
that restricts the maximum allowable flowrate into a thermal incinerator to design values if
testing is conducted under representative peak conditions. Commenter FV-D-28 believes this
provision is meaningless for control devices that are designed with excess capacity, which is
what the commenter typically does.
Response: The final rule specifies that performance testing shall be conducted under
peak-case conditions for all control devices. As discussed in the responses to comments in
section 14.1, the language allowing for testing of incinerators under representative peak-case
conditions has been deleted because of Agency concerns regarding operation at levels higher than
the levels under which the compliance demonstration occurred. Therefore, the requirement to
restrict maximum allowable flowrate into the incinerator is no longer necessary and has been
deleted from the final rule.
15.2.7 Clarifications
Comment: Commenter F/-D-28 requested each of the following editorial clarifications in
§ 63.1364(c) of the proposed rule. The first sentence in paragraph (c)(l) should cross-reference
the exemptions in paragraph (c)(5) as well as those in paragraph (c)(4). The second sentence in
paragraph (c)(l)(ii) should mention the 94 and 99.9 percent reduction requirement for HC1 as
well as the corresponding emission limit cutoffs of 6.8 and 191 Mg/yr.
Paragraphs (c)(2)(i)(A)(3) through (5) should be revised to state that the owner or operator is to
15-16
-------
determine HAP partial pressures in accordance with one of the following subparagraphs; as
written it sounds like all three must be used simultaneously. In the variables list for Equation 3,
the definitions of Tl and T2 should refer to the temperature of the material in the vessel, not to
the temperature of the vessel itself. Paragraph (c)(3)(ii) describes control devices that "meet" an
outlet concentration of 20 ppmv, but it should say "are intended to meet." Paragraph (c)(3)(ii)(B)
requires sampling for oxygen and TOC at the same time; however, it would be better to require
the sampling at "substantially" the same time because there will always be slight differences
between the starting (or ending) times of the two samples, which should not be considered
violations. Paragraphs (c)(3)(iii)(B), (C), and (D) indicate that an owner or operator may "also"
choose to test over representative peak-case conditions, but this suggests that two tests would be
conducted; these paragraphs should say testing under representative peak-case conditions is an
alternative to testing under absolute or hypothetical peak-case conditions.
Response: The percent reduction requirements for HC1 were specified in
paragraph (c)(l)(iv) of the proposed rule. However, the Agency agrees that this provision and
most of the other provisions identified by the commenter need to be clarified. Some of the
suggested changes have been incorporated verbatim into the final rule. In other cases, more
substantial changes to provisions make it impossible to incorporate the suggested changes
verbatim. For example, as noted elsewhere in this document, the final rule does not allow testing
under representative peak-case conditions. However, for the provisions that have been retained
in the final rule, the intent of each suggested change has been incorporated. The final rule has
not been changed to specify that sampling for oxygen and TOC is to be conducted at
"substantially" the same time because other rules do not use this qualifier. Also, slight
differences in sampling time are to be expected, and are not considered violations.
15.3 INITIAL COMPLIANCE WITH STORAGE TANK STANDARDS
Comment 1: If an owner or operator uses a floating roof to comply with the storage tank
standards, paragraph (d)(4) in § 63.1364 of the proposed rule would require the owner or
operator to demonstrate initial compliance by complying with various provisions in §§ 63.119
and 63.120 of the HON. Commenter IV-D-28 believes a sentence should be added to the end of
§ 63.1364(d)(4) to indicate that compliance with the provisions of this section (i.e., the
15-17
-------
cross-reference to floating roof provisions in the HON) shall be deemed to constitute compliance
with the percentage HAP emission requirements of this subpart.
Response: The EPA disagrees with the commenter's suggestion. The standard is to
install either a floating roof or a closed vent system with a control device. Thus, installing a
floating roof in accordance with provisions in the regulation demonstrates compliance with the
standard; it is not a means of demonstrating compliance with the percent reduction standard.
Comment 2: Paragraphs (d)(2)(iii) and (iv) in § 63.1364 of the proposed rule would
exempt sources from performance test requirements for storage vessel control devices if either:
(1) the control device is also used to control process vent emissions and the owner or operator
demonstrates initial compliance with the process vent standard, or (2) the control device is a
RCRA unit or a boiler or process heater that meets certain criteria. Commenter FV-D-28 believes
the following sentence should be added to the end of both paragraphs: 'This constitutes
compliance for the purposes of this subpart."
Response: The Agency believes the sentence the commenter wants to add to the rule is
either unnecessary or incorrect, depending on the meaning the commenter intended. For
example, the commenter may want to clarify that compliance with either of the subject
provisions constitutes initial compliance. However, EPA believes the additional sentence is
unnecessary for this purpose because an introductory sentence in the section that contains these
provisions clearly states that the owner or operator demonstrates initial compliance by fulfilling
the requirements of either of the subject paragraphs. Alternatively, the commenter may want
EPA to specify that compliance with the subject paragraphs constitutes compliance with all
provisions of the subpart. The EPA believes this would be confusing because § 63.1365 deals
only with initial compliance provisions, and it would be incorrect because other regulatory
requirements still apply to the control devices (although not specifically to the control of
emissions from the storage vessels). The commenter may believe the statement is needed to
ensure that an owner or operator is not subject to duplicative monitoring, recordkeeping, and
reporting requirements for multi-use control devices. However, the monitoring, recordkeeping,
and reporting provisions in §§ 63.1366 through 63.1368 focus on individual control devices, not
the type(s) of emission points venting to the device, and § 63.1362(1) of the final rule exempts
RCRA incinerators and boilers from the initial compliance, monitoring, and recordkeeping
15-18
-------
requirements of the final rule. In addition, monitoring is not required for the subject boilers and
process heaters. Finally, the commenter may be concerned that the subject paragraphs exempt
the owner or operator from only performance tests, but not design evaluations. For the final rule,
a number of changes have been made in § 63.1365, one of which was to clarify that the
exemptions apply to all initial compliance demonstrations for affected storage vessels (i.e., to
design evaluations as well as performance tests). Therefore, EPA did not add the sentence that
the commenter suggested.
Comment 3: Commenters FV-D-21 and FV-D-29 found the compliance procedures for
new Group 1 storage tanks to be unnecessarily convoluted. The commenters requested
confirmation of their interpretation that the provisions mean a design evaluation may be used to
demonstrate compliance.
Response: The commenters are correct; initial compliance with the standards for storage
tanks at new sources may be demonstrated using either a design evaluation or a performance test.
For the final rule, § 63.1365 has been reorganized and edited to clarify numerous provisions,
including the one cited by the commenter.
15.4 INITIAL COMPLIANCE WITH WASTEWATER STANDARDS
Comment: Commenter IV-D-28 stated that section 63.1364(e) of the proposed rule is
unclear because it says to "demonstrate compliance" with the wastewater requirements by
"complying" with the wastewater requirements. The commenter suggested that this
section should say compliance should be demonstrated by complying with the applicable
provisions of § 63.145.
Response: The Agency agrees with the commenter that the language in § 63.1364(e) of
the proposed rule is unclear; therefore, the suggested change has been incorporated into the final
rule.
15.5 PLANNED ROUTINE MAINTENANCE
Comment: The proposed rule specified that an owner or operator would demonstrate
compliance with the planned routine maintenance provisions by including the periods of planned
routine maintenance in each Periodic report. Commenter IV-D-28 requested clarification of
whether facilities are to report anticipated (planned) numbers, actual (as happened) numbers, or
both.
15-19
-------
Response: As noted in the response to the comment in section 6.4, the planned routine
maintenance provisions in the final rule only apply to storage tanks. Section 63.1365(d)(7) of the
final rule specifies that an owner or operator demonstrates initial compliance with these
provisions by reporting in the NOCS report the anticipated periods of planned routine
maintenance for the first reporting period (i.e., the 6 months after the NOCS report is due).
Ongoing compliance is demonstrated by including both the actual hours for the previous
reporting period and anticipated hours for the next period in each Periodic report. The sum of the
actual hours in the previous 12 months must be less than 240 hours for the facility to be in
compliance. The reports also must describe the type of maintenance to be performed and that
was performed. These provisions are similar to the provisions in §§ 63.120(d)(4) and
63.122(g)(l)oftheHON.
15-20
-------
16.0 MONITORING
16.1 IMPACT OF MONITORING REQUIREMENTS ON CONTROL LEVELS
Comment: Several commenters (IV-D-16, IV-D-20, and IV-D-27) believe that the
requirement to set compliance parameter levels based on performance testing at peak conditions
for batch processes will result in significant over control during most of the operation.
Commenters IV-D-16 and IV-D-20 indicated that the following requirements also increase the
stringency: (1) using the average of three test runs for setting parameter limits, and
(2) determining compliance on a daily basis. Commenters IV-D-16 and IV-D-27 suggested that
the impact could be minimized by not restricting the length of the test runs and by calculating
compliance on an annual basis rather than daily. The commenters suggested an annual basis
because they believe the data provided in response to the section 114 information request were
based on annual averages. Commenter IV-D-27 suggested daily calculations of rolling annual
averages, and commenter IV-D-16 suggested monthly calculations of rolling annual averages.
Another commenter (FV-D-28) supports the proposed provisions to base compliance on the daily
average of parameter values, not each individual monitored data point.
Response: As noted in the responses in sections 7.4 and 8.3, EPA believes the data
provided in response to the survey are as valid on a daily basis as on an annual basis. Testing for
initial compliance determinations may be as short as three 1-hour runs. In such cases, the
regulation could require that continued compliance be demonstrated using hourly average
monitoring parameter levels to be consistent with the data from the initial compliance
determination. The proposed rule would require a less stringent approach of averaging over an
operating day because EPA believes this minimizes the impact of momentary disruptions and
variability, yet also gives a reasonable assurance of continued compliance with the standard.
16-1
-------
Therefore, the final rule retains the requirement to demonstrate continued compliance with the
standard on a daily basis.
In the final rule, EPA requires that testing be conducted under absolute or hypothetical
peak-case conditions for all control devices because EPA firmly believes that if the conditions
under which the device will be operatedcannot be predetermined, the device must be
demonstrated to achieve the required efficiency over peak-case conditions. If inlet stream
conditions can be predetermined, the owner or operator has the option of setting multiple
monitoring levels for different operating conditions. This option was provided in the proposed
rule and has been retained in the final rule. Therefore, EPA does not believe the requirement
results in over control. Also, as noted in section 14.4, the final rule has been changed to remove
the provision restricting test runs to no more than 8 hours.
Regarding averaging periods, EPA has modified the compliance period of the standard to
allow averaging on either a 24-hour basis or a "block" basis, where the block may be any length
of time less than the time from the beginning to the end of a batch process. For batch operations,
an annual compliance period was determined by EPA to be too difficult to implement and
therefore not practical. The annual compliance period implies that owners and operators could
control a process to varying degrees during the course of a year, as long as the yearly percent
reduction target would be met. Although this format would offer flexibility to owners and
operators who want to change control strategies to accommodate production scheduling and
operational changes, EPA believes that the demonstration of compliance over such an extended
time period would result in delayed determination of exceedances and the possibility for
extended periods of violations. The EPA notes that the final rule offers numerous compliance
options to provide flexibility for owners and operators to address variability within their
processes.
Regarding the setting of parameter levels, the purpose of monitoring operating parameters
is to provide evidence of continued compliance with the rule. Monitoring parameters are set
based on test data, calculations, or information from the evaluation of the control device design.
The final rule requires sources to establish maximum or minimum operating parameter levels
based on the average of the average parameter values for each of the three test runs (i.e., average
values are to be determined for each of the three test runs, and the monitoring parameter level is
16-2
-------
to be based on the average of those three values). The Agency believes that setting monitoring
levels based on the average of three test runs is necessary because the control efficiency is also
based on the average from the three test runs. Basing the monitoring parameter on the results of
only one of the test runs would be inconsistent with the average control level.
16.2 ESTABLISHING PARAMETER LEVELS
Comment 1: Commenter FV-D-28 found it difficult to determine when parameter limits
are to be determined based on performance tests and when other methods may be used. For
example, the commenter believes § 63.1365(a) of the proposed rule could be interpreted to mean
every parameter limit has to be based on a performance test, even though a test may not have
been conducted to demonstrate initial compliance. The commenter urged EPA to clarify these
provisions using language similar to that in previous MACT standards (e.g., the Group I and
Group IV P&R rules).
Response: One sentence in § 63.1365(a) of the proposed rule stated that "test data,
calculations, or information from evaluation of the control device design shall be used to
establish the operating parameter." Another sentence specified how to establish the parameter
level if performance testing has been required. The intent of these statements was to indicate that
monitoring parameter levels are to be established during the initial compliance demonstration. If
the initial compliance demonstration is a performance test, the parameter levels are based on the
operation during the test (although, as noted in another response in this chapter, some levels may
also be based on extrapolation of data obtained during a test). If the initial compliance
demonstration is a design evaluation, the parameter levels are to be based on the design
evaluation. To clarify this requirement, the final rule specifies that monitoring parameter levels
are to be established based on test data, calculations, or other information from the initial
compliance demonstration.
Comment 2: Three commenters (IV-D-16, IV-D-20, and IV-D-28) questioned and
expressed concern about setting monitoring parameter levels based on the average of the values
from three test runs rather than the least stringent value. Commenter IV-D-16 stated that using
the average value to establish the limit is problematic because affected sources would have no
opportunity to compensate for potential process deviations that occur with the use of multi-
purpose equipment and other factors that would influence compliance with the measured
16-3
-------
parametric levels. Thus, commenter IV-D-16 believes that a source should be able to establish
parameter ranges other than those measured during a performance test. Commenter FV-D-20
believes there is no basis for using the average value because performance testing will always be
biased to indicate a parameter value more than sufficient to achieve compliance.
Response: The EPA disagrees with the commenters. The least stringent monitoring
parameter value is not necessarily indicative of conditions needed to demonstrate continued
compliance with the standard, just as demonstrating initial compliance using results from a
performance test with only a single run would not be considered acceptable. This is especially
true under certain situations. For example, to demonstrate compliance with the 90 percent
reduction standard for process vents, a test might yield control efficiencies of 88, 89, and
93 percent. Perhaps the least stringent monitoring parameter level would be associated with the
run that produced the 88 percent control efficiency. Using this least stringent level would not be
consistent with the conditions that were needed to achieve a 90 percent reduction. Furthermore,
if an owner or operator believes the control device will achieve an efficiency well above that
needed to demonstrate compliance, nothing in the rule prohibits the owner or operator from
conducting a test under less than optimum conditions in an effort to minimize the stringency of
the operating parameter level.
The proposed rule also included an error that has been corrected in the final rule. The
proposed rule states that a minimum is to be based on the average of the minimum values from
each of the three test runs, and that a maximum is to be based on the average of the maximum
values from each of the three test runs. The statements should have specified that the minimum
(or maximum) is to be based on the average of the values from each of the three runs (i.e., the
average of the averages for the three runs). This change has been made in the final rule.
Comment 3: Commenter IV-D-16 requested that additional control device parameter
monitoring options be added to the rule (e.g., pH for scrubbers and coolant temperature for
condensers) because the few parameters mandated by the proposed rule would be very expensive
and inappropriate for many sources.
Response: As suggested by the commenter, pH as a parameter for acid gas scrubbers was
added to the rule. In addition, parameters for catalytic incinerators were added. If an owner or
operator believes it would be appropriate to monitor parameters other than those specified in the
16-4
-------
rule, the owner or operator may request approval to monitor other parameters in accordance with
the General Provisions or by using the Precompliance plan. Monitoring the coolant temperature
for condensers was not included in the rule because coolant temperature is not a direct indicator
of condenser efficiency. The EPA believes that, while monitoring coolant flow to ensure
compliance with the standard may be a reasonable alternative for some systems, sources must
demonstrate that the measurement of coolant temperature for each condenser system will achieve
the required outlet gas temperature. This demonstration most likely would involve appropriate
heat transfer calculations and verified condenser (heat exchange) system specifications (i.e., heat
transfer coefficients and heat transfer area) in addition to providing actual temperature
measurements verifying the relationship between coolant temperatures and gas temperatures.
Comment 4: Commenters FV-D-16 and IV-D-27 support the provisions that allow
owners and operators the opportunity to request approval to monitor parameters other than those
specified in the rule (i.e., by following the procedures in § 63.8(f) of the General Provisions or by
using the Precompliance plan). However, commenter FV-D-16 believes that all facilities will use
the Precompliance plan because the procedures in § 63.8(0 are too narrow and restrictive. To
accommodate future changes, commenter FV-D-16 also requested that the rule state that
alternative monitoring parameters can be implemented via amendments to the Precompliance
plan.
Response: The final rule specifies how to request approval to change parameters or
procedures that were submitted for approval in the Precompliance plan, or that would have been
included in the Precompliance plan if the change had been implemented prior to the compliance
date. The new provision requires an owner or operator to submit a notification 60 days before
the scheduled implementation date. The notification must contain the same information that
would have been included in the Precompliance plan.
Comment 5: Commenter FV-D-29 urged EPA to exclude batch processes that vent to a
combustion device or share a common combustion device with continuous processes from
§ 63.1365(b)(2) of the proposed rule. The commenter explained that this section of the proposed
rule, which specifies the procedures to establish a parameter level, is meaningless for a control
device such as a flare because the monitoring parameter \vould be the presence of a pilot flame,
which is not related to the characteristics of the vent stream from a single emission episode.
16-5
-------
Response: The EPA agrees with the commenter that procedures to establish a parameter
level for flares is not needed because only the presence of a pilot flame is required. The final rule
clarifies this point. However, it is not clear why the commenter requested the elimination of
procedures for monitoring combustion devices that are used to control emissions from batch
processes. Monitoring is as feasible for these devices as for those that control vents from
continuous vents.
Comment 6: Commenter IV-D-28 requested clarification of the provision in
§ 63.1365(b)(2)(iv) of the proposed rule that monitoring levels be established at the conditions of
the test when the performance test is conducted at "routine" conditions.
Response: Section 63.1365(b)(2)(iv) is an artifact of a draft approach and should not
have been included in the proposed rule. It is not included in the final rule.
Comment 7: The EPA solicited comment on the use of alternative parameters without
the requirement of prior notification in the Precompliance plan. The preamble to the proposed
rule also requested comment on the adequacy of the following alternative parameters: (1) for
condensers, coolant temperature and flow (only with emissions testing); (2) for scrubbers,
pressure drop, scrubber fluid composition or pH; and (3) for carbon adsorbers, adsorption cycle
and regeneration frequency, bed temperature, regeneration stream flow, periodic test for bed
poisoning, and periodic vent testing and/or predetermined scheduled replacement.
Commenter IV-G-05 believes the suggested alternative parameters are adequate to demonstrate
continuous compliance with the rule. However, the commenter wondered why EPA solicited
comments on the use of alternative parameters without prior notification in the Precompliance
plan when the proposed rule also clearly would require the owner or operator to determine the
most appropriate method of verification and propose this method for approval in the
Precompliance plan.
Response: The alternatives for scrubbers and carbon adsorbers have been included in the
final rule. However, the alternative for condensers has not been included because, as noted in
section 15.2.5 of this document, EPA has decided that the most appropriate method for
estimating outlet emissions and the percent reduction is to use the results of temperature
measurements, not emissions tests.
16-6
-------
Under the proposed rule, the requirement to conduct a periodic verification to
demonstrate that the control device is operating as designed applied only to small control devices
(i.e., those with inlet HAP emissions less than 0.91 Mg/yr). For larger control devices, the
proposed rule specified default parameters to be monitored. If an owner or operator elected to
monitor these parameters, no approval would be necessary. However, an owner or operator
would have to request approval to monitor different parameters, either by including the request in
the Precompliance plan or by following the procedures in § 63.8(f) of the General Provisions.
This concept is also included in the final rule.
16.3 MONITORING FOR SPECIFIC CONTROL DEVICES
16.3.1 Bag Dumps and Product Dryers
Comment 1: Commenters FV-D-16 and FV-D-28 believe the requirement to initiate
corrective action within 1 hour of a bag dump alarm is unnecessarily rigid and will not always be
necessary. Commenter FV-D-28 believes it is too stringent because: (1) an alarm is
conservatively set so it does not always mean emissions are occurring; (2) there may be times
(especially in batch operations) when an alarm trips even though the system is not currently in
use; (3) there may be other things happening in the process unit that justifiably require priority
attention (e.g., a serious overpressure in the reactor area); and (4) depending on what EPA means
by "initiate," some corrective actions cannot be initiated within 1 hour (e.g., if a problem occurs
after normal business hours that requires a replacement part that is not on hand at the site).
Commenter IV-D-16 suggested changing the 1-hour time period to 3 hours.
Response: When an alarm on a bag leak detection system is triggered, the proposed PAI
rule would require owners or operators to inspect the control device to determine the cause of the
deviation and initiate corrective actions specified in the corrective action plan that is submitted
with the Notification of Compliance Status report. The definition of initiate is to be specified by
the owner or operator in the plan. To clarify this point, the final rule specifies that the owner or
operator is to initiate procedures to identify the cause of the alarm and take correction action as
specified in the corrective action plan. The final rule does not specify the time when these
actions are to be initiated or implemented; the owner or operator should specify that information
as part of the procedures for responses to different types of events that trigger the alarm. The
16-7
-------
final rule also specifies that the corrective action plan is to be submitted with the Precompliance
plan, which includes other procedures that must be approved by the implementing agency.
Comment 2: Commenter IV-D-28 believes the rule should allow either visual or audible
bag dump alarms, not just audible alarms.
Response: The Agency believes that both a visual and audible alarm are necessary
because staff may not be in a control room or other areas where visual alarms can be seen. In
addition, this requirement is consistent with other NESHAP requiring leak detection systems; no
changes have been made in the final rule.
Comment 3: Commenter IV-D-28 believes two of the monitoring provisions for bag
dumps and product dryers are unlawful. First, the commenter believes the requirement in
§ 63.1365(a)(7)(iv) of the proposed rule to install, operate, calibrate, and maintain the bag leak
detection system according to either available guidance from EPA or the manufacturer's written
specifications and instructions is unlawful because EPA cannot impose a legally binding
requirement to follow unspecified "guidance," and EPA cannot empower manufacturers to make
law. Therefore, the commenter suggested adding a third option to allow the owner or operator to
"follow other written procedures that provide reasonable assurance that the bag leak detection
system will function appropriately." Second, the commenter believes the requirement in
§ 63.1365(a)(7)(viii) of the proposed rule to develop a QIP consistent with the draft approach to
compliance assurance monitoring is unlawful because EPA cannot require compliance with a
draft document. The commenter believes EPA should either publish a supplemental proposal
specifying the features of an acceptable QIP and make it part of the rule or delete the
requirement.
Response: The EPA has completed guidance entitled "Fabric Filter Bag Leak Detection
Guidance." This document provides guidance on the use of triboelectric monitors as fabric filter
leak detection systems. Where this document is not applicable, EPA has allowed for the use of
manufacturer's written specifications and instructions. Proper setup and operation of a bag leak
detector can vary with site-specific conditions, and those conditions may dictate variances from
EPA guidance or manufacturer's specifications and instructions. For such cases, EPA has added
a provision that would allow for the development of site specific procedures. These procedures
must be included in the Precompliance plan and approved by the Administrator.
16-8
-------
The final rule does not require the development of a QEP program.
Comment 4: Commenter IV-D-28 believes it is both unnecessary and inconsistent with
other aspects of the rule to require written approval before adjusting the range, averaging period,
alarm set points or alarm delay time contained in the Notification of Compliance Status report.
According to the commenter, the rule should only require that changes be reported in the next
Periodic report, and, if prior approval is needed, it will be handled under the Operating Permit
program.
Response: The intended use of the bag leak detector is to identify upset conditions in the
baghouse operation. The EPA is concerned that unrestricted adjustment of the bag leak detector
could result in improper use, possibly resulting in the alarm and sensitivity settings being set
such that leaks or malfunctions could occur undetected. Based on further review, EPA has
determined that periodic adjustment may be necessary. Therefore, EPA has revised the bag leak
system adjustment requirements to: (1) allow for routine minor adjustments to the detector
system, (2) require owners and operators to identify all routine adjustments in an operating and
maintenance plan that is to be submitted with the Precompliance plan, and (3) require that
owners and operators perform complete baghouse inspections to ensure proper operation of the
baghouse prior to any significant adjustments to the sensitivity or range.
Comment 5: Commenter IV-D-28 asked for clarification of the requirement that a bag
leak detection system sensor must provide output of relative or absolute PM emissions.
Response: The final rule has clarified this statement by requiring that "The bag leak
detection system sensor must provide output relative to PM emissions."
16.3.2 Flares
Comment 1: Commenter IV-D-28 believes §§ 63.1365(a)(5) and (b)(7) of the proposed
rule should refer to the loss of "all" pilot flames, because losing one flame should not be a
violation if there are other flames at the pilot of the flare.
Response: The EPA agrees with the commenter. The final rule indicates that an
exceedance occurs upon the loss of all pilot flames and that records must be kept of periods when
all pilot flames are absent.
16-9
-------
Comment 2: Commenters IV-D-21 and IV-D-29 believe the rule should specify that a
pilot flame is not required when a process is shutdown. These commenters also pointed out that
this change would be consistent with the HON.
Response: As noted in the response to the comment in section 16.4 of this chapter, the
final rule specifies that monitoring is only required when the control device is functioning in
achieving the HAP removal required by the rule. Thus, the final rule does not require an owner
or operator to demonstrate the presence of a pilot flame during periods when the process is not
operating. However, while implementing shutdown, the owner or operator must follow the
procedures described in the startup, shutdown, and malfunction plan. These procedures may
specify that the pilot flame be maintained during the shutdown process.
16.4 MONITORING FREQUENCY
Comment: Commenter IV-D-16 suggested two changes to the frequency of monitoring.
First, the commenter believes that the requirement in § 63.1365(b)(3) of the proposed rule to
monitor batch episodes less than 15 minutes in duration should not be required. Second, the
commenter believes the monitoring frequency is excessive for control devices controlling less
than 10 ton/yr of an individual HAP or 25 ton/yr of aggregate HAP; "periodic" monitoring would
be sufficient because many parameters do not vary frequently, and it would allow for the use of
simpler monitoring systems that are less prone to design and maintenance problems.
Response: When only one monitoring level is established for a parameter, the EPA
agrees with the commenter that monitoring of batch episodes less than 15 minutes in duration
should not be required because the practical limit of monitoring frequency is one reading every
15 minutes of operation. Instead of requiring that each batch episode be monitored at least once,
the final rule requires an owner or operator to measure and record the parameter level at least
once every 15 minutes during the period in which the control device "is functioning in achieving
the HAP removal required" by the rule. This means that one reading must be taken for every
15-minute period of continuous venting from any combination of emission episodes manifolded
to the control device. Thus, even when individual emission episodes are shorter than 15 minutes,
one reading is required if venting occurs for at least 15 minutes due to overlapping or
"contiguous" episodes. On the other hand, no monitoring would be required if each of the
emission episodes that an owner or operator is controlling to comply with the rule last less than
16-10
-------
15 minutes, and they are separated by periods of no flow or venting from vents that do not need
to be controlled. For storage tanks, a control device is considered to be functioning in achieving
the HAP removal required at all times material is stored in the tank; although working losses
occur only during relatively short periods when the tank is being filled, breathing losses may
occur at any time. To identify periods of no flow, a flow indicator (not necessarily a flow
monitor) would be required.
An exception to the procedures described above exists if the owner or operator
establishes separate monitoring levels for different emission episodes. In this case, at least one
reading must be taken each time the level changes, even if episode lasts less than 15 minutes.
This exception is included to counteract the possibility of setting multiple levels in order to avoid
monitoring.
As a result of the change in monitoring frequency, the definition of a valid hour of data as
used in the definition of an excursion also has been modified in the final rule. At proposal,
monitoring data would not constitute a valid hour of data if measured values are unavailable for
any of the 15-minute periods within the hour. For the final rule, the word required has been
added before the phrase "15-minute period" to address the fact that less than four data points per
hour may be allowed in some situations.
The EPA believes that the requirement to take 15-minute readings for devices controlling
more than 1 ton/yr of HAP is reasonable. The cutoff for continuous monitoring was set because
EPA wanted to reduce the compliance burden on facilities with smaller control devices. The
EPA also notes that "periodic" monitoring could increase the potential for being out of
compliance with the standard, because a reduction in the number of data points places a
significantly higher emphasis on each reading for compliance determination. Additionally,
because emission stream characteristics in this industry are variable, the use of "periodic"
readings may not represent true conditions over the monitoring period.
16.5 QUALITY CONTROL PROGRAM AND PERFORMANCE EVALUATIONS OF
CONTINUOUS MONITORING SYSTEMS
Comment: According to commenter FV-D-28, §§ 63.8(d) and (e) and 63.10(e)(2)(i) of the
General Provisions should not apply because the quality assurance program and performance
standards for continuous monitoring systems (CMS) are directly regulated through the startup,
16-11
-------
shutdown, and malfunction plan and the requirement to report any exceedances in the Periodic
reports.
Response: The quality control and performance evaluation procedures for CMS in
§§ 63.8(d) and (e) are not covered in the startup, shutdown, and malfunction plan, nor are they
addressed by the requirement to report exceedances in the Periodic reports. Therefore, these
provisions are applicable in the final rule. The final rule also specifies the required accuracy and
calibration procedures for selected parameter monitoring devices that must be taken into
consideration in the quality control plan.
16.6 MONITORING DURING STARTUP, SHUTDOWN, AND MALFUNCTION
Comment 1: Commenter FV-D-16 stated that § 63.1365(d) of the proposed rule should be
revised to clarify that startup, shutdown, and malfunction provisions apply to monitoring
equipment as well as process and control equipment.
Response: The EPA agrees with the commenter. The final rule specifies that startup,
shutdown, and malfunction provisions apply to monitoring equipment as well as process and
control equipment. This change makes the final rule consistent with other regulations such as the
HON, pharmaceuticals and polymers and resins.
Comment 2: Commenter IV-D-28 raised several issues about how the startup, shutdown,
and malfunction provisions relate to data availability requirements and violations. First, the
commenter stated that monitoring data collected during any startup, shutdown, or malfunction
and during periods of nonoperation of the process should be excluded from the daily averages.
Second, the commenter believes the rule should clearly specify that there is no violation if an
event such as a malfunction results in insufficient data or an exceedance of a parameter. Third,
the commenter expressed concern about the requirement in § 63.1365(d) of the proposed rule,
which stated that an excursion is not a violation if it happens during a startup, shutdown, or
malfunction and the facility follows its startup, shutdown, and malfunction plan. The commenter
interprets this provision to mean that EPA could assess t\vo penalties if the plan is not followed-
one for not following the plan, and, because the plan was not followed, a second for the
excursion.
Response: If monitoring data obtained during a startup, shutdown, or malfunction result
in an exceedance, the exceedance is not a violation as long as the facility follows the startup,
16-12
-------
shutdown, and malfunction plan. However, if the facility does not follow the plan, an
exceedance would be a violation, but it would not be two violations. Thus, the final rule retains
the requirement to use data obtained during any startup, shutdown, and malfunction in daily
averages.
Similarly, if a startup, shutdown, or malfunction results in the inability to collect
monitoring data, it may cause an excursion. This excursion would not be a violation if the
facility followed its startup, shutdown, and malfunction plan, but it would be a violation if they
did not follow the plan.
As noted in the response to the comment in section 16.4, the final rule requires
monitoring when the control device is functioning in achieving the HAP removal required by the
rule. Thus, data obtained during time when the process is not operating are not to be used in
determining the daily average of the parameter level.
16.7 MONITORING FOR STORAGE VESSEL CONTROLS
Comment: Commenter FV-D-28 believes the proposed rule lacks appropriate monitoring
provisions for control devices that are used to control emissions from storage vessels. According
to the commenter, the proposed provisions address only continuous monitoring, which often will
not be appropriate for storage vessels because the emissions occur primarily during filling.
Furthermore, if emissions are controlled using a disposable carbon canister, the monitoring may
consist only of replacing the canister before the end of its rated life, not periodically checking a
parameter. Therefore, the commenter recommended that EPA include some of the concepts from
the storage tank monitoring provisions in § 63.120(d) of the HON. For example, these
provisions specify that the owner or operator must prepare a monitoring plan that describes how
the monitoring will be done. In addition, the commenter indicated that the rule needs to define
"excursion" for situations where monitoring is not continuous (e.g., the rule should specify that
the monitoring plan "shall define an excursion in terms of the relevant operating parameter").
Response: The monitoring provisions in § 63.1365(b) of the proposed rule were intended
to apply to all types of control devices except those used for continuous processes. In the final
rule, the provisions from §§ 63.1365(a) and (b) have been consolidated into one section that
specifies monitoring provisions for all control devices. The final rule also includes monitoring
provisions for nonregenerative carbon canisters; the owner or operator is required to determine
16-13
-------
the maximum time interval between replacement based on operation under absolute or
hypothetical peak-case conditions and to replace the canister before this time elapses.
Unlike the HON, the final PAI rule requires the same type of monitoring regardless of the
purpose for which the control device is used. The EPA does not believe it is necessary to have
different procedures for storage tank control devices because the types of emission episodes from
storage tanks are comparable to those from batch process vents. Furthermore, most storage tanks
at the surveyed PAI plants emit less than 0.91 Mg/yr. Under the final rule, if the total
uncontrolled HAP emissions entering a control device are less than 0.91 Mg/yr, the owner or
operator may elect to conduct a periodic (at least daily) verification that the control device is
operating properly. The verification procedures are to be described in the Precompliance plan.
This provision is comparable to the monitoring plan concept described in § 63.120(d)(2) of the
HON. As noted in the response to the comment in section 16.4, if the total uncontrolled HAP
emissions entering the control device exceed 1 ton/yr, the owner or operator must monitor the
appropriate parameter(s) every 15 minutes during which the control device is functioning in
achieving the HAP removal required by the rule. Based on information from the surveyed PAI
facilities, this situation would apply to very few storage tanks in the PAI industry. Most of the
few tanks with emissions greater than 0.91 Mg/yr are vented to the same control device that is
used to control process vent emissions. Thus, a separate set of monitoring requirements for
storage tank control devices is not needed.
For devices that control more than 0.91 Mg/yr of HAP, the definition of excursion in the
final rule is the same as that in the proposed rule, and it is applicable to all control devices.
Specifically, a valid hour of monitoring data must be obtained for 75 percent of the hours that a
control device operates during a day (or, if the control device operates less than 4 hours, at least
3 hours of valid data must be obtained). As noted in the response to the comment in section 16.4
of this chapter, the control device operation is based on the time when the control device is
functioning in achieving the HAP reduction required by the rule. For storage tanks, this means
all of the time that the storage tank contains material. When compliance for small control
devices is demonstrated by conducting a periodic verification, the final rule has been revised to
clarify that not conducting the verification is an excursion.
16-14
-------
The final rule also clarifies that exceedances of operating parameters are those times
when (1) the parameter level, averaged over the operating day, is above a maximum or below a
minimum established during the initial compliance demonstration, or (2) the required operating
characteristic is not met (e.g., loss of all pilot flames for a flare). If compliance is demonstrated
by conducting a periodic verification, an exceedance occurs any time the daily, or more frequent,
demonstration does not confirm that the control device is operating properly.
16.8 DATA AVAILABILITY REQUIREMENTS
Comment: Commenter IV-D-16 indicated that the applicability of § 63.8(c)(7) in Table 1
should be changed to "yes" to indicate that "recorded data shall not be used in data averages or
calculations, or to meet any data availability requirement when the data are recorded during out-
of-control periods, repairs, maintenance periods, calibration checks, and zero (low-level) and
high-level calibration drift adjustments."
Response: Section 63.8(c)(7) of the General Provisions describes out-of-control
situations for continuous monitoring systems. It also indicates that during out-of-control periods,
recorded data are not to be used in data averages or other data availability requirements
established under part 63. This section of the General Provisions was not applicable in the
proposed rule, and is not applicable in the final rule, because the out-of-control situations
described in this section (i.e., zero and high-level calibrations and audits) are not applicable to
parameter monitors. In addition, maintenance and repair periods are covered by the startup,
shutdown, and malfunction plan. As noted in the response to comment 2 in section 16.5, the rule
itself includes language requiring the use of data collected during any startup, shutdown, or
malfunction.
16.9 VIOLATIONS
Comment 1: Several commenters (FV-D-14, IV-D-16, and FV-D-28) addressed situations
that would result in violations under the proposed rule. All of the commenters asserted that
excursions or exceedances of an operating parameter should not be violations of the emission
standard; commenter FV-D-28 also stated that failure to take corrective action after a bag dump
alarm should be a violation of a work practice requirement, not the particulate HAP emission
standard. Commenter FV-D-28 stated that such excursions should not be a violation of an
emission standard because the parameters are only indicators of proper operation, they do not
16-15
-------
prove compliance with an emission standard. Commenter FV-D-14 stated that the proposed
provision conflicts with the basis of the compliance assurance monitoring (CAM) regulation.
Commenters IV-D-16 and IV-D-28 stated that the requirement in § 63.1365(a) to "operate
processes and control devices within the parameters" must be revised because they interpreted
this statement to mean that each datapoint must be within the established limit. Commenter IV-
D-16 indicated that the source must be allowed to demonstrate continued compliance with the
emission standard despite exceedance of a monitoring parameter. Commenter FV-D-28 indicated
there are startup, shutdown, and malfunction exceptions (as described in comment 2 in
section 16.6), and only the daily average value (not each data point) is used to demonstrate
compliance.
Response: The EPA's policy is that new part 63 rules, in particular those that require the
use of a control device to reduce pollutant emissions, will include compliance determinations on
two levels. The first level is the "traditional" performance test requirement that is based on the
use of a specific test method over a set period of time and operating conditions. A performance
test is generally conducted when the rule is effective (e.g., at facility startup or after an effective
date for an existing facility) and may be repeated periodically thereafter. The results of the
performance test are compared with an emission limitation (e.g., concentration, control
efficiency, or mass rate). The second level of the compliance determination in part 63 rules is the
continuous compliance obligation, which is implemented through monitoring.
In general, the EPA recognizes two basic approaches to monitoring. One method is to
establish monitoring as a direct measure of continuous compliance. Under this continuous
compliance monitoring approach, an enforceable value of the monitored parameters is defined
and measured. The Agency has adopted this approach in part 63 standards, and is committed to
following this approach whenever appropriate in future rulemakings. Another approach is to
establish monitoring to provide a reasonable assurance of compliance by documenting continued
proper operation of the control devices, indicating excursions from proper operating conditions,
and correcting the problems that cause excursions. This second approach is the basis of the
CAM rule, which applies to sources that are not currently subject to part 63 standards.
When a part 63 rule specifies a surrogate pollutant continuous emissions monitoring
system(s) (CEMS) or parameter monitoring for demonstrating continuous compliance, the rule
16-16
-------
includes specific limitations and averaging times for these alternative situations. The surrogate
pollutant or operating parameter limit becomes an enforceable limit for the rule. There is no
requirement that an alternative limit, whether a surrogate pollutant or an operational parameter,
be statistically correlated with emissions or the compliance level of the regulated pollutants(s).
The alternative limit is a separately enforceable requirement of the rule. The alternative is not
secondary to the emission limit; rather, it is applied in lieu of continuous emission limit
obligation.
The enforceable level for the surrogate pollutant or operating parameter may be based on
measurements made during a performance test or other conditions specified by the part 63 rule.
In any case, the alternative limit becomes the continuous compliance obligation and fulfills the
second level of compliance for the rule.
The EPA has considered the commenters' argument that an exceedance of a monitoring
parameter is not necessarily an exceedance of an emission limit. The Agency acknowledges that
a parameter exceedance does not necessarily mean that the source has exceeded the emission
limit. However, as discussed above, under EPA's approach to continuous compliance in part 63
rules, the continuous parameter monitoring limit is a separate requirement that is not rebuttable
through contrast with actual or estimated HAP emission values. In addition, EPA believes that
given the flexibility the owner or operator has to select operating parameters, including the
option that allows the owner or operator to set different parameter levels for different operating
conditions, the burden is on the source to remain within the operating limit defined for the
parameter or parameters.
To address the potential disparity between parameter limit exceedances and emission
limit exceedances, the final rule contains two different types of continuous compliance
violations. When a source is using a CEMS to monitor compliance with the 20 ppmv alternative
standard, an exceedance is defined as a violation of the emission limit. Similarly, because the
exit gas temperature of a condenser is so closely correlated with emissions, a condenser
temperature exceedance is considered a violation of the emission limit. Exceedances of other
types of parameter limits are defined as violations of an operating limit. Failure to initiate the
corrective action plan after a bag leak detector alarm also is a violation of an operating limit.
16-17
-------
Finally, for the final rule, §§ 63.1365(a) and (b) of the proposed rule have been combined
and revised to clarify the monitoring requirements. It was never EPA's intent to base compliance
on each individual datapoint. Therefore, the first paragraph in the monitoring section of the final
rule has been changed to read as follows:
To provide evidence of continued compliance with the standard, the owner or operator of
any existing or new affected source shall install, operate, and maintain monitoring
devices as specified in this section. During the initial compliance demonstration,
maximum or minimum operating parameter levels or other design and operating
characteristics, as appropriate, shall be established for emission sources that will indicate
the source is in compliance. Test data, calculations, or information from the evaluation of
the control device design, as applicable, shall be used to establish the operating parameter
level or characteristic.
The section then goes on to describe the types of parameters to monitor; it explains how to
establish the parameter levels and averaging periods; and it defines exceedances, excursions, and
violations.
Comment 2: Commenter FV-D-28 believes the rule should allow for a specified number
of excused excursions per reporting period, as in other MACT standards.
Response: The final rule does not included excused excursions. Excused excursions
were allowed in the HON to allow facilities time to become familiar with the new monitoring
provisions in the HON. The excursions were not meant to be precedent setting for all future
rules. The EPA believes that industry in general has had sufficient time to develop strategies for
complying with monitoring requirements, and that excused excursions are no longer necessary.
Other recent rules also have been issued without excused excursions.
Comment 3: Section 63.1366(d) of the proposed rule specifies that for unit operations
occurring more than once per day, exceedances of established parameter limits shall result in no
more than one violation per operating day for each monitored item of equipment utilized in the
unit operation. Commenter FV-D-28 supports the concept that there should be no more than one
violation per day for exceedances of operating parameters, but suggested incorporating it into a
different part of the rule. The commenter pointed out that § 63.1367(b)(2) addresses this concept
for excursions (i.e., periods of "insufficient data"). The commenter suggested that exceedances
of operating parameters be called a "parameter level" excursion, specify that § 63.1367(b)(2)
16-18
-------
applies to parameter level excursions, and delete § 63.1366(d). The commenter also believes
§ 63.1366(d) is unclear because "unit operations" are equipment, they do not "occur."
Response: Several changes have been made in the final rule to clarify the definitions of
exceedances, excursions, and violations. For the final rule, the definition of excursions has been
moved from the reporting section to the monitoring section, and a definition of exceedances has
been added to the monitoring section. The provision in § 63.1366(d) of the proposed rule has
been revised and moved to the monitoring section as well. For the final rule, this provision reads
as follows: "... for emission episodes occurring more than once per day, exceedances of
operating parameters or excursions will result in no more than one violation per operating day for
each monitored item of equipment utilized in the process."
Comment 4: Commenter IV-D-16 urged EPA to add a provision to protect an affected
source that discharges to a POTW from compliance violations under this rule if the POTW has a
violation of their permit.
Response: The final rule cross-references the offsite treatment provisions in § 63.132(g)
of the HON, which is identical to the proposed requirements. These provisions require a
"transferee" (e.g., a POTW) to submit a written certification to EPA stating that the transferee
will manage and treat any Group 1 wastewater stream or residual removed from a Group 1
wastewater stream in accordance with the requirements in subpart G. Furthermore, if a POTW
provides such a written certification, § 63.132(g)(3) states that the POTW accepts responsibility
for compliance with the wastewater provisions in the HON for any wastewater covered by the
certification, and is subject to enforcement action for violations.
16.10 MISCELLANEOUS
Comment 1: According to commenter FV-D-28, the provisions for heat exchangers and
equipment leaks in §§ 63.1365(d) and (e) are out of place because the monitoring for this
equipment is the emission limitation (or means of reducing emissions); the monitoring is not
used to demonstrate compliance with an emission limitation. For this equipment, there is no
"monitoring" to demonstrate compliance, except recordkeeping and reporting, which are stated
elsewhere.
Response: The commenter is correct that the standards for equipment leaks and heat
exchangers consist of monitoring requirements. Therefore, §§ 63.1366(d) and (e) in the final rule
16-19
-------
state this fact and indicate that no additional monitoring is required to demonstrate continued
compliance with the standards.
Comment 2: Commenters IV-D-14 and IV-D-27 requested changes in the monitoring
requirements to more closely match the requirements in the CAM rule. Commenter IV-D-27
believes that cutoffs in § 63.1365 for requiring continuous monitoring should be the same as
those in the CAM rule (i.e., 10 or 25 ton/yr, not 1 ton/yr). The commenter based this comment
on a statement in section n.B of the preamble to the final CAM rule. Commenter IV-D-14 stated
that the monitoring requirements should be revised to coincide with the principles of CAM.
Response: As discussed in the response to comment 1 in section 16.9, CAM is one of
two basic approaches to monitoring, but it was developed for sources that are not currently
subject to part 63 standards. Furthermore, as noted in the response to the comment in
section 16.4, the provision allowing a facility to demonstrate compliance by conducting periodic
verifications for devices that control less than 1 ton/yr of HAP was included in the proposed rule
to minimize the burden on small facilities. The EPA does not believe raising this cutoff is
warranted. Therefore, the cutoffs for the monitoring provisions in the proposed rule are retained
in the final rule.
Comment 3: Commenter IV-D-27 believes the rule should state that monitoring and
inspection provisions under RCRA can be followed instead of corresponding MACT provisions
for air and/or wastewater routed to RCRA incinerators covered under subpart O of either 40 CFR
part 264 or part 265.
Response: The RCRA provisions are as stringent or more stringent than the requirements
in the final rule. Therefore, § 63.1362(1) of the final rule specifies that a RCRA incinerator used
to control emissions is exempt from monitoring requirements of the rule. The final rule also
cross-references § 63.138(h) of the HON, which exempts RCRA units that are used to treat
wastewater from the monitoring provisions.
Comment 4: Commenter IV-D-28 supports the concept of allowing periodic verifications
rather than monitoring for small control devices, but suggested that this section also specifically
state that continuous monitoring is allowed.
Response: The Agency appreciates the support of the commenter in allowing periodic
verifications rather than monitoring for small control devices. In addition, the final rule states
16-20
-------
that, for small control devices, the owner or operator may elect to comply with the continuous
monitoring provisions rather than the periodic verification requirements.
Comment 5: Section 63.1365(b)(2)(iii) of the proposed rule specifies that an owner or
operator may establish monitoring parameters based on "the performance test supplemented by
engineering assessments and manufacturer's recommendations." Commenter FV-D-28 suggested
adding the words "if desired" after the word "supplemented" and change the word "and" to
"and/or." This commenter also recommended deleting the words "The procedures in this
section have not been approved by the Administrator" because this suggests that the provisions
are unacceptable or have been disapproved.
Response: The words "if desired" are not necessary because the provision already states
the owner or operator may use the procedure as an alternative to establishing parameter levels
based solely on performance test data. In the final rule, the word "and" has been replaced with
"and/or" because the test data may be supplemented using an engineering assessment,
manufacturer's recommendations, or both. The words 'The procedures in this section have not
been approved by the Administrator" were deleted because they are considered to be
unnecessary. The final rule retains the statement that any procedures an owner or operator uses
to develop additional monitoring parameter levels are subject to approval by the Administrator.
16.11 EDITORIAL CLARIFICATIONS
Comment: Commenter FV-D-28 requested all five of the following editorial clarifications
in § 63.1365; commenters IV-D-21 and IV-D-29 also requested the fifth clarification:
1. The term "parameter level" (not just "parameter" or "operating parameters") should be
used consistently to describe the limit within which the daily average parameter value must be
maintained.
2. The first sentence of § 63.1365(b) needs clarification because it is not clear how an
owner or operator may "choose" to comply with the emission limit or emission reduction.
3. The type of emissions needs to be specified every time a quantity of emissions is
specified (e.g., 9.1 tons of organic HAP, not just 9.1 tons of HAP).
4. Add "as specified in this section" to the end of the first sentence in § 63.1365(a).
5. Section 63.1365(a)(6) should be corrected to indicate that a deviation occurs when
combustion chamber temperatures are lower, not greater, than the parameter value.
16-21
-------
Response: The EPA agrees with the commenters' suggestions. As noted elsewhere in
this chapter, the monitoring provisions have been significantly rearranged and clarified in the
final rule. All of the editorial changes and corrections suggested by the commenters have been
incorporated in the final rule.
16-22
-------
17.0 RECORDKEEPING
17.1 RECORDKEEPING BURDEN
Comment: Two commenters (FV-D-28 and FV-G-03) commented on the burden to
comply with the recordkeeping requirements. Commenter IV-G-03 stated that over 100 different
records must be maintained daily, monthly, quarterly, or annually; therefore, the commenter
urged EPA to review the recordkeeping requirements to ensure that such extensive recordkeeping
is necessary. Commenter IV-D-28 supports provisions in §§ 63.1366(a) and (a)(3) that require
an owner or operator to maintain records of only the daily average of the parameter values, not
each datapoint, because this reduces the recordkeeping burden. However, this commenter also
stated that a reduced recordkeeping option, like that in § 63.152(g) of the HON, should be added
to the rule if EPA decides to change these provisions and require records of shorter term values.
Response: Detailed records are needed to demonstrate compliance with the regulation.
However, prior to proposal, EPA made a concerted effort to eliminate duplicative and
unnecessary recordkeeping requirements because EPA recognizes that these requirements would
burden both the affected sources and EPA enforcement agencies. Since proposal, EPA has
reviewed the recordkeeping provisions and made a number of changes. Many of the changes are
editorial revisions designed to clarify the requirements. Some of these clarifications are
discussed in more detail in other responses in this chapter. Other clarifications explicitly state
recordkeeping requirements that were merely implied in the proposed rule (e.g., records of
planned routine maintenance and records of the absolute or hypothetical peak-case conditions for
process vent testing).
The final rule also includes additional recordkeeping requirements to document
compliance with new or revised provisions in the rule. For example, the final rule includes
recordkeeping to document the primary use for material produced by PAI process units if the
17-1
-------
primary use is not as a PAI (see section 3.2 for a discussion of the new primary use provisions).
Another example in the final rule includes procedures to demonstrate ongoing compliance with
the annual emission limit for process vents by calculating an annual rolling summation every day,
and records of these calculations must be maintained. Finally, § 63.1362(j) was added to the
final rule to specify that bypass lines that could divert a vent stream away from a control device
must be monitored either with a flow indicator or by visual inspection of the seal or closure
mechanism that secures the valve in the closed position; records of any flow or the results of
inspections must also be maintained.
One additional change involves the parameter monitoring records in §§ 63.1366(a) and
(a)(3) that were cited by the commenter. After reviewing these requirements, EPA now believes
that, even when the daily average is in compliance, it is necessary to maintain all parameter
readings, not just the daily averages. This rule requires that owners and operators select only
parameter readings that are taken when the control device is controlling HAP emissions from
affected emission streams. Emission episodes from batch processes, which predominate in the
PAI production industry, are discontinuous. As a result, some monitoring readings may occur
during periods of no flow for affected streams (although there may be flow of nonaffected
streams). Readings taken during these periods must be excluded from the daily averages. In
order to verify that the daily average values were calculated correctly, the rule requires owners
and operators to keep all data. The EPA also does not believe that the approach in § 63.152(g) of
the HON would be appropriate for this rule because, unlike this rule, the HON regulates emission
streams with continuous flow.
17.2 RECORDS OF GROUP 1 DETERMINATIONS
Comment: Commenter IV-D-28 believes that records of Group 1 stream determinations
should not be required.
Response: The final rule requires an owner or operator to keep records of the results of
all Group determinations for process vents, storage vessels, and wastewater streams. The owner
or operator is not required to keep records of how streams were determined to be Group 1
streams; in fact, for storage vessels, the final rule allows the owner or operator to designate
Group 1 storage vessels. However, some of the same information may be required for other
reasons. For example, as described in section 17.4, records of uncontrolled process vent
17-2
-------
emission estimates are required if the owner or operator complies with the requirements to
reduce emissions from the sum of all process vents within a process by 90 percent. Records of
emission estimates are also required for all emission streams that are used in emissions
averaging, and records of wastewater stream concentrations and flowrates are needed for some of
the treatment options. Alternatively, an owner or operator complying with an outlet
concentration standard is not required to keep these records (although records of flow rates may
be needed to show compliance with corrections for supplemental gases).
By definition, if a stream is not Group 1, then it is Group 2. Therefore, the owner or
operator also must keep records of the Group 2 determinations because these data are needed to
demonstrate compliance with the applicability cutoffs.
17.3 RECORDS OF TURNOVERS
Comment: Section 63.1365(b)(5) of the proposed rule would require owners and
operators to keep records of tank turnovers. Three commenters (IV-D-21, FV-D-28, and
FV-D-29) opposed this provision. Commenter FV-D-28 stated that there is no substantive
requirement in the regulation that these records would support, and even if turnovers must be
recorded, this commenter questioned whether calculations rather than monitoring results would
be acceptable. Commenters IV-D-21 and FV-D-29 stated that turnovers should be required only
for tanks complying with the 110 kg/yr cutoff; such records are not needed when emissions are
controlled.
Response: The EPA agrees with the above comment for most cases, especially since the
storage tank mass emissions applicability cutoff was eliminated in the final rule. Actual
emissions from a tank are only needed for emissions averaging and for determining whether
emissions from a new PAI process unit exceed 10/25 tons/yr and thus would be subject to new
source standards. Turnovers are used to calculate working losses and thus would be one of the
parameters that must be recorded only as part of either the emissions averaging calculations or
new source determination.
17.4 RECORDS OF PROCESS VENT EMISSIONS
Comment: Section 63.1366(b)(l) of the proposed rule would require an owner or
operator to maintain records of "the emissions of gaseous organic HAP and HC1 per batch for
each process." Commenter FV-D-28 believes this provision implies that CEMS would be
17-3
-------
installed on every emission point, which conflicts with the compliance provisions. Thus, the
commenter believes this provision should be deleted.
Response: The provision is reworded in the final rule to clarify that records must be kept
of the initial calculations (and supporting data) used to estimate uncontrolled and controlled
emissions of each emission event from Group 1 process vents if the owner or operator is
complying with the 90 percent reduction requirement. These are the same data that would be
used to determine that the vents within a process are Group 1, but records of Group 1
determinations are not required (as noted in section 17.2) because the data are not needed if the
owner or operator is complying with the outlet concentration requirement of 20 ppmv, or is using
a flare. The provision is retained in the final rule because these data are needed to demonstrate
initial compliance with the percent reduction requirements. A copy of the NOCS report (and
updates to these data and calculations contained in Periodic reports) would satisfy this
requirement because the same information must be submitted in that report.
17.5 RECORDS OF WASTEWATER CHARACTERISTICS
Comment: Section 63.1366(b)(2) of the proposed rule would require an owner or
operator to maintain records of "wastewater concentrations and flowrates per POD and process."
Commenter FV-D-28 believes this statement could be interpreted either of two ways, but
regardless of which interpretation is correct, the provision should be deleted. For example, it
could mean annual average values, but if so, the commenter believes this record would be
unnecessary because there must be a requirement elsewhere in the regulation to keep records
supporting Group determinations (except for Group 1 determinations, as noted in the comment in
section 17.1). Alternatively, it could mean continuous "real time" records, which the commenter
points out conflicts with the compliance requirements and should not be required; thus there will
be no data from which to create the records.
Response: The provision is referring to the annual average values. Contrary to the
commenter's belief, recordkeeping requirements are specified only in § 63.1366. Other sections
of the rule specify that the concentrations and flowrates must be estimated for use in determining
compliance. Therefore, the provision is retained in the final rule.
17-4
-------
17.6 RECORDS OF BAG LEAK DETECTION ALARMS
Comment: Section 63.1366(a)(5) of the proposed rule would require an owner or
operator to maintain records of any bag leak detection alarm and the corrective action taken.
Commenter FV-D-28 believes corrective action would not always be necessary and, therefore, the
following phrase should be added to the end of the proposed provision: "or the reason why no
corrective action was taken."
Response: The provision has been clarified in the final rule but was not changed as
suggested by the commenter. In the final rule, the provision requiring the owner or operator to
keep records that document the date and time of the alarm, the cause of the alarm, and the
corrective action taken is in § 63.1367(b)(5). Note that under § 63.1368(e)(6), a corrective action
plan must be submitted with the Precompliance plan. The corrective action plan must describe
procedures for the proper operation and maintenance of fabric filters, procedures used to
determine and record the time and cause of an alarm, and corrective actions to be taken when the
alarm is triggered. If there are situations that the owner or operator believes do not require
corrective action (other than resetting the alarm), they must be identified in the corrective action
plan.
17.7 LOCATION OF RECORDS
Comment: Commenter FV-D-28 believes that the rule, like the HON, should include
provisions specifying where to keep records. For example, the HON defines "onsite" and
requires that records from the most recent 6 months be kept onsite (or accessible within 2 hours);
records from the preceding 4.5 years may be kept offsite.
Response: Section 63.1366(a) specified that "records shall be kept in accordance with the
requirements of applicable paragraphs of § 63.10 of subpart A of this part, as specified in the
General Provisions applicability table of this subpart (Table 1)." Section 63.10(b)(l) of the
General Provisions specifies that records must be retained for at least 5 years, and, at a minimum,
the most recent 2 years of data shall be retained onsite, and the remaining 3 years of data may be
retained offsite. Onsite means the same thing as the "plant site that is a major source," as stated
in the definition of the affected source in § 63.1360(a).
17-5
-------
17.8 OVERLAP WITH RCRA RECORDKEEPING REQUIREMENTS
Comment: Commenter IV-D-27 believes the rule should state that an owner or operator
may follow the recordkeeping provisions under RCRA instead of the corresponding
recordkeeping requirements in § 63.1366 for air emissions and/or wastewater routed to RCRA
incinerators covered under 40 CFR part 264/265 subpart O.
Response: The EPA agrees with this comment. Therefore, § 63.1362(1) of the final rule
exempts streams that are discharged to RCRA incinerators or boilers and industrial furnaces
meeting Subpart O from all requirements of the rule, except for identification in the NOCS
report.
17.9 CLARIFICATIONS
Comment 1: Commenter IV-D-28 believes the phrase "up-to-date" should be deleted
throughout § 63.1366 for two reasons. First, it means nothing for one-time records, and second,
it is unnecessary for continuous records because other portions of the rule already require records
of successive data.
Response: The EPA disagrees with the commenter. "One-time" records are subject to
change based on changes in the process. The requirement to maintain records of continuous
monitoring data or results of periodic calculations exists only in § 63.1366 of the proposed rule;
other sections of the rule specify only such things as the types of parameters that must be
monitored, the types of calculations that must be performed, and the frequency of these activities.
Therefore, the phrase "up-to-date" is retained in the final rule.
Comment 2: Section 63.1366(e) of the proposed rule specified that no more than one
violation per operating day would be assessed "for certain items of monitored equipment used for
more than one type of unit operation in the course of an operating day." Commenter IV-D-28
requested clarification of this section. For example, the "certain items" of monitored equipment
are not defined.
Response: The definitions of exceedances and excursions, and how these occurrences
constitute violations are clarified in § 63.1366(b)(6) of the final rule (see chapter 16 for
additional information about these changes). The term "certain items" is not used in the final
rule.
17-6
-------
18.0 REPORTING
18.1 PRECOMPLIANCE PLAN
Comment: Commenters IV-D-21 and FV-D-29 stated that the requirement for a
Precompliance report in § 63.1367(a)(2) of the proposed rule should be deleted, as it was for the
HON.
Response: The EPA believes the Precompliance report (or precompliance plan in the
final rule) is a valuable tool for the regulatory agency responsible for making compliance
determinations for the affected source. It provides an enforcement official or inspector with
some initial background information about the process being controlled, the types of emissions
associated with the process, corresponding control equipment, and the monitoring parameters
that have been or will be correlated to the process conditions.
The Precompliance plan is also the mechanism by which the affected source requests
approval to use alternative monitoring parameters and to use calculations or other compliance
procedures that differ from those prescribed in the rule. Because many of the compliance
procedures for this rule are more complicated than those for the HON, EPA believes the
Precompliance plan requirement is warranted for this industry and has retained the provision in
the final rule.
Comment: Section 63.1367(a)(2)(iii) of the proposed rule would require the owner or
operator to include in the Precompliance plan "a description of test conditions and limits of
operation for control devices tested under normal conditions . . ." Commenter IV-D-28 requested
clarification of what is meant by "normal" conditions.
Response: The language in this section of the proposed rule was an artifact of an earlier
approach and should have been deleted from the proposed rule. This provision has been
corrected in the final rule to require documentation of how monitoring parameter levels are
18-1
-------
established under § 63.1366(b)(3)(ii)(B), including the test data, any calculations, and rationale
for why the level indicates proper operation of the control device. This section allows an owner
or operator to establish monitoring levels based on performance test results supplemented with
engineering assessments and manufacturer's recommendations. Because the use of these
procedures is subject to approval by the Administrator, it must be included in the Precompliance
plan.
18.2 PERIODIC REPORTS
Comment 1: Several commenters (IV-D-21, IV-D-28, IV-D-29, and F/-G-03) addressed
the issue of the frequency of periodic reporting. Three commenters (IV-D-21, IV-D-28, and
FV-D-29) stated that periodic reporting should be changed from quarterly to semiannually.
Commenter IV-D-28 provided three reasons for the change: (1) quarterly reporting would be an
unwarranted increase in burden compared to other MACT standards; (2) consistency among
standards makes it easier for a facility subject to many standards to comply; and (3) EPA has
proposed to harmonize the paperwork burdens of a wide variety of NSPS, part 61 NESHAP, and
other part 63 MACT standards. Commenter IV-G-03 noted that the title of § 63.1367(b) is
"Quarterly reports" but suggested changing it because § 63.1367(b)(3) would require submittal of
a report within 180 days after a process change.
Response: The EPA re-evaluated the overall reporting requirements in the proposed rule
and compared the proposed reporting requirements with requirements in rules for similar
industries. Based on this evaluation, the Agency decided to change the periodic reporting from
quarterly to semiannual. In those cases where the continuous emission monitoring data are used
to demonstrate compliance with the 20-ppmv alternative standards, and the source experiences
excess emissions, quarterly reporting is required until a request to reduce reporting frequency is
approved. Section 63.1368(g) in the final rule is now titled "Periodic reports" and details the
submittal schedule and content of the required Periodic reports.
Comment 2: Commenter IV-D-28 believes the rule should indicate that Periodic reports
are due 2 months after the end of each reporting period.
Response: As mentioned in the response to comment 1 in this section, the Agency has
opted for semiannual Periodic reports. In concurrence with the comment, § 63.1368(g) of the
final rule details the submittal schedule and exceptions associated with the Periodic reports. In
18-2
-------
short, Periodic reports are to be submitted within 60 operating days after the end of the applicable
reporting period.
18.3 NOTIFICATION OF COMPLIANCE STATUS REPORT
Comment: Commenter FV-D-28 stated that the NOCS report submittal date in
§ 63.1367(a)(l) conflicts with the requirement in § 63.7(a)(2) to complete performance testing
within 180 days and the requirement in § 63.10(d)(2) to submit performance test reports 60 days
after the tests.
Response: The submittal date for the NOCS report in § 63.1368(f) of the final rule does
not conflict with the General Provisions requirements in §§ 63.7(a)(2) and 63.10(d)(2), it
supersedes it. As noted in Table 1 to Subpart MMM — General Provisions Applicability to
Subpart MMM, "[T]est results must be submitted in the NOCS due 150 days after the
compliance date." This means that the performance testing and the compilation of the test results
must be completed and submitted as part of the NOCS report which is due within 150 days after
the compliance date. Additional language was added to the final rule under § 63.1368(a) to
clarify which of the reporting requirements of subpart A (General Provisions) remain in effect for
this rule and which requirements have been superseded.
18.4 STARTUP, SHUTDOWN, AND MALFUNCTION REPORTS
Comment: Commenter FV-D-28 stated that § 63.8(c)(l)(ii) should not apply because if it
does apply there will be two sets of deadlines for reports of actions inconsistent with the startup,
shutdown, and malfunction plan—one affecting CMS and the other affecting all other periods of
startup, shutdown, and malfunction. According to the commenter, the deadlines should all be the
same and, ideally, the rule would require that all of this information be reported in the
semiannual periodic reports.
Response: The Agency agrees that two sets of inconsistent reporting deadlines are
cumbersome. As a result, the final rule overrides the 24-hour notification provisions of
§ 63.8(c)(l)(ii) and requires that all such notifications be reported within 2 days, consistent with
the provisions of § 63.6(e)(3)(iv). The Agency was not persuaded by the commenter's
suggestion that all such notifications can be reported in the Periodic reports and has maintained
the reporting requirements (schedules) from the General Provisions § 63.6(e)(3)(iv) (related to
events covered and those not covered in the facility's startup, shutdown, and malfunction plan).
18-3
-------
Furthermore, monitoring events or activities not covered by a source's startup, shutdown, and
malfunction plan should be rare and are likely to be of interest to the enforcement agency.
Reporting information to the enforcement agency within 2 working days will allow them to make
a timely evaluation (if needed or warranted) of the event, any associated excess emissions, and
the source's response to the event. Repeated occurrences of events not covered by a source's
startup, shutdown, and malfunction plan could prompt an enforcement agency to request a copy
of the startup, shutdown, and malfunction plan to review.
18.5 NOTIFICATION OF PROCESS CHANGE
Comment: Section 63.1367(b)(3) of the proposed rule would require notification of
process changes that cause an emission point to become a process vent with an emission rate of
1 Ib/yr or more. Two commenters opposed this requirement. Commenter IV-D-16 stated that
this provision is not applicable to this standard and should be deleted. Commenter IV-D-15
stated that EPA needs to review and revise the definition of de minimis because 1 Ib/yr is well
within the margin of error of the calculation methods and may not be significant. Another
commenter (FV-D-28) stated that the rule needs to specify how and when to report information
about changes that occur after the submittal deadline for a specific report. The commenter
suggested that it might be submitted no later than 60 days after it is obtained.
Response: The Agency decided to revise this section of the final rule based on these
comments. The final rule states that whenever a process change is made, or a change is made in
any of the information submitted in the NOCS report, a report must be submitted within 90
calendar days after the process change, unless the change requires approval prior to
implementation (such as a change that would consist of information submitted in the
Precompliance plan). The report may be submitted as part of the next Periodic report, if one is to
be submitted within the 90-day period. This more general or "generic" approach avoids the
problems associated with trying to define de minimis emission levels as raised by the
commenters. The information to be reported is to include: a brief description of the process
change, a description of any modifications to standard procedures or quality assurance
procedures, revisions to any of the information reported in the original NOCS report, and
information required by the NOCS report for changes involving the addition of processes or
equipment.
18-4
-------
18.6 EQUIPMENT LEAK REPORTS
Comment: According to commenter IV-D-28, § 63.1367(c) of the proposed rule should
specify the appropriate reporting deadlines for affected sources subject to the equipment leak
standards because the deadlines in the cross-referenced sections in subpart H are not appropriate.
Response: The commenter is correct in that the cross-referenced section in subpart H
does not match the rest of the reporting requirements in the proposed rule. With the reporting
frequency changed to semiannual (see previous comments and responses regarding Periodic
reports in 18.2), EPA decided to make the equipment leak reports consistent with the other
reporting requirements in the final rule. The equipment leak report(s) are to be included with the
NOCS report (due within 150 days of the compliance date) and with the Periodic reports (due
within 60 days after the end of each subsequent semiannual reporting period).
18.7 RECORDS OF REPORTS
Comment: Commenter IV-D-28 believes the requirement in § 63.1367(c) of the
proposed rule to maintain copies of properly-submitted reports as records is not appropriate for
two reasons. First, it imposes a paperwork burden with no environmental benefit and no purpose
except to be able to make a backup copy in the event EPA loses the original. Second, it creates
the potential for unfair "paperwork penalties" in the event the owner or operator cannot find it
when an inspector asks for it. The commenter recommended adding the following language,
based on an amendment to the HON: "If an owner or operator submits copies of reports to the
applicable EPA Regional Office, the owner or operator is not required to maintain copies of the
reports. If the EPA Regional Office has waived the requirement of § 63.10(a)(4)(ii) for submittal
of copies of reports, the owner or operator is not required to maintain copies of the reports."
Response: The General Provisions (§ 63.10(b)(l)) require the owner or operator of an
affected source to maintain files of all information (including all reports and notifications) in a
form suitable and readily available for expeditious inspection and review. At a minimum, the
most recent 2 years of data shall be retained onsite. The regulatory agency responsible for
compliance assurance will be the likely end-user of such information if and when an inspection
or site visit is conducted. The paperwork burden associated with keeping copies of the data and
reports for the last 2 years onsite is not considered a significant burden. After the 2-year period,
electronic copies of the reports may be maintained at an offsite location for the additional 3-year
18-5
-------
requirement. The EPA believes that the benefits of having the recent compliance reports
available onsite to an inspector (if needed) far outweigh the impacts or burden associated with
maintaining copies of the reports.
18.8 OVERLAP WITH RCRA REPORTING REQUIREMENTS
Comment: Commenter IV-D-27 believes the rule should state that an owner or operator
may comply with the reporting provisions under RCRA instead of the corresponding reporting
provisions in § 63.1367 of the proposed rule for air emissions and/or wastewater routed to a
RCRA incinerator covered under 40 CFR part 264/265 subpart O.
Response: The EPA agrees with the commenter. Therefore, § 63.1362(1) of the final rule
exempts streams that are discharged to RCRA incinerators or boilers and industrial furnaces
meeting subpart O from all requirements of the rule, except for identification in the NOCS
report.
18-6
-------
19.0 IMPACTS
19.1 ENVIRONMENTAL IMPACTS
Comment: Commenter F/-D-27 believes EPA did not adequately consider the secondary
air impacts of nitrogen oxides (NOX) formation caused by combusting nitrogen-bearing HAP
(and non-HAP VOC that may also be present) in process vent streams and wastewater.
Response: The impacts analysis was based on a small number of model streams with
characteristics that represent typical or average characteristics of streams at the surveyed
facilities. Very little nitrogen-bearing HAP is emitted from the surveyed facilities (less than
5 percent of both the total uncontrolled organic HAP emissions from process vents and the HAP
load in wastewater streams), and most of these HAP are controlled to the level of the standard.
Therefore, the model emission streams that were used to estimate secondary air impacts did not
include nitrogen-bearing HAP. In addition, any small underestimate in the NOX emissions from
nitrogen-bearing HAP is likely more than offset by the use of conservative estimates in the
original analysis. For example, the estimated increase in NOX emissions were based solely on the
emissions associated with operation of the more efficient controls needed to achieve the level of
the standards; emissions from existing controls that would be replaced were assumed to be
negligible. The EPA has no evidence that non-HAP VOC compounds in emissions streams
contain significant amounts of nitrogen.
19.2 COST IMPACTS
Comment: Commenters IV-D-15 and IV-D-27 believe EPA underestimated the costs to
comply with the proposed rule. Based on recent experience installing some of the control
devices that are used in the cost analysis, commenter IV-D-27 believes the costs are
"significantly" underestimated, especially when the standard is more stringent than the floor.
This commenter said their company would continue to review EPA's cost analysis and might
19-1
-------
provide their own detailed analysis at a later date; EPA has not received any additional
information from this commenter. Commenter FV-D-27 also indicated that, based on the
additional secondary air impact described in the comment above, the cost analysis should
consider the need to install best available control technology (BACT) or RACT technology to
control NOX emissions.
Commenter FV-D-15 believes none of the models used in the cost analysis adequately
address the situation at the cornmenter's facility. This facility emits carbon disulfide, which,
when burned, generates a significant amount of SOX. The SOX is not an issue under the proposed
regulation, but it is a criteria pollutant that would have to be controlled under state regulations.
As a result, the commenter believes EPA's cost analysis underestimates the cost the commenter
would face for two reasons. First, the model is based on a thermal incinerator with 70 percent
recuperative heat recovery, but the commenter could not use this control device because carbon
disulfide has a low auto-ignition temperature; they would have to use either a thermal incinerator
with no heat recovery or a regenerative thermal oxidizer with 85 percent heat recovery. Second,
the scrubber that follows the incinerator would need to be able to control the SOX emissions as
well as HC1 emissions. This commenter also stated that if their intermediate process meets the
definition of an intermediate (see comments in Section 3.3), either considerable changes to the
existing flare or a new flare would likely be needed to meet the provisions of the regulation.
Response: The cost impacts are based on models that represent a range of characteristics
at actual facilities. The models are expected to overestimate costs at some facilities and to
underestimate costs at others. It is possible that installing a control device could trigger the
requirement for a BACT or RACT analysis. Typically, to trigger BACT analysis the control
device would have to cause a net increase in NOX emissions of 40 tons/yr (or any amount that has
an impact of 1 microgram per cubic meter within 10 kilometers of a class I area). To increase
emissions by 40 tons/yr would require a very large incinerator; for example, the incinerator to
control the large model process with very low HAP concentrations was estimated to increase
NOX emissions by only about 9 tons/yr. Typically, a facility has only two PAI processes. Thus,
even if all emission streams are routed to the incinerator, and the emission streams contain
nitrogen-bearing HAP, it will be a very unusual situation for NOX emissions to increase by
40 tons/yr. In the unlikely event that BACT is required, BACT is likely to be the use of low-NOx
19-2
-------
burners, which are likely already part of the design of any new combustion devices. Although
RACT typically is applied to existing emission units, if it were to be applied to a new combustion
control device, RACT would also likely be low-NOx burners. As a result, EPA did not include
BACT or RACT technology in the models for the impacts analysis.
The SOX control also was not included in the cost analysis because it is not a typical
requirement, the amount of SO2 control that would be needed is unknown, and the cost is not
expected to be significantly different from that for an HC1 scrubber. The total annual cost of a
thermal incinerator with no heat recovery is approximately equal to that for a thermal incinerator
with 70 percent recuperative heat recovery. The annual auxiliary fuel costs would be higher for
the incinerator without heat recovery, but these costs are nearly offset by lower capital recovery
costs as a result of lower capital costs. Although the performance of a given scrubber will be
better for HC1 than for SC^, a scrubber can easily be designed to obtain excellent SO2 removal
efficiencies.
Regarding the use of a flare that may not meet the specifications in the rule, if the owner
or operator believes it is achieving a 90 percent reduction, the owner or operator may develop
and submit for approval a procedure to demonstrate that it is complying with the required
reduction efficiency.
19.3 ECONOMIC IMPACTS
Comment: Commenter FV-D-15 believes EPA has not adequately evaluated the impact
of the proposed rule on small businesses. The commenter notes that the regulatory flexibility
analysis finds minimal impact on small businesses, but the docket states that the two known
small firms for which data were available were not surveyed to find the impact of the regulation
on them. The commenter believes a survey of small businesses is needed; otherwise the impact
on them is unknown. This issue is important to the commenter because at the time facilities
responded to the section 114 information request, the commenter's plant was part of a large
business, but it has since been sold and is now classified as a small business.
Response: Prior to proposal, EPA estimated that the proposed regulation would affect
78 existing plants. The EPA confirmed company-level revenue data on 45 of the 78 existing
plants that were estimated to be affected; the 45 plants are owned by 29 firms. Additionally,
EPA selected 9 large PAI manufacturing companies with multiple plants to collect detailed
19-3
-------
information, which resulted in obtaining information for 20 plants altogether. This information
was used to estimated annualized costs of the proposed regulation. The EPA determined that 2
of the 29 firms are classified as small businesses, and each firm owns one plant. Because these
two small firms were not included in the more detailed survey of the 9 large PAI manufacturing
companies, direct costs associated with the proposed regulation were not available and average
control costs for modeled plants were used as an estimate of the control costs for the small firm
plants. Using the average control costs for model plants as an estimate for small firm plants'
control costs is a conservative approach (i.e., actual control costs for small firm plants are likely
to be smaller) because small firm plants are likely to be smaller in scale and have fewer distinct
processes per plant. The commenter's facility was one of the 20 plants for which the more
detailed information was collected.
Since proposal, EPA reevaluated the impacts for the commenter's facility. Revenue data
for the parent company of the new owner of this facility were obtained from Dun & Bradstreet.
Cost impacts were unchanged from the original analysis. The resulting cost-to-revenue ratio for
this small business was estimated to be approximately 2.3 percent. As noted at proposal, the
control costs for model small businesses were also estimated to be less than 3 percent of revenue.
This percentage suggests that the final rule does not significantly impact small firms in the PAI
manufacturing industry.
19-4
-------
20.0 MISCELLANEOUS
20.1 STANDARDS FOR ENDOCRINE DISRUPTORS
Comment: In the preamble to the proposed rule, EPA solicited comment on whether the
risk posed by possible endocrine disrupters warrants more stringent requirements than those
proposed. Six commenters (IV-D-16, IV-D-21, IV-D-26, IV-D-28, IV-D-29, and IV-G-05)
opposed the development of more stringent requirements; none supported the idea. The
commenters cited a variety of reasons for not developing more stringent requirements: (1) the
science for determining disrupting properties of chemicals and their risks is still under
development; (2) technology-based standards are not appropriate to address endocrine disruption;
(3) endocrine disruption is not an adverse endpoint, but a mechanism of action; (4) the
compounds are emitted in small quantities; and (5) this has not been an issue under other MACT
standards that address essentially the same materials.
Response: In the proposal preamble EPA indicated that available information shows
emissions of possible endocrine disrupters is very low relative to other HAP emissions. Based
on these data and the comments, EPA has decided not to include more stringent requirements for
possible endocrine disrupters in the final rule. However, this decision does not preclude the
possibility that EPA may take action on endocrine disrupters in the future as new information
becomes available.
20.2 RISK-BASED STANDARD FOR HC1
Comment: The preamble to the proposed rule explained that section 112(d)(4) of the Act
provides EPA with authority, at its discretion, to develop risk-based standards for HAP "for
which a health threshold has been established," provided that the standard achieves an "ample
margin of safety." Because HC1 is a threshold pollutant that is emitted from PAI manufacturing
facilities, EPA solicited comment on the adequacy, desirability, and feasibility of developing a
20-1
-------
risk-based standard instead of a MACT standard for HC1 emissions from PAI manufacturing
facilities. Three commenters addressed this issue. Commenter FV-D-17 opposed the
development of a risk-based standard for HC1 emissions because it would delay promulgation of
the rule. Commenter IV-G-05 opposed development of a risk-based standard because the
proposed requirements are very similar to those proposed in the NESHAP for Steel Pickling
Facilities-HCl Process. Furthermore, this commenter believes a risk-based standard is not
needed because existing permit limitations based on ambient concentrations are protective of the
environment and human health, and the NESHAP limitations will only increase the permit's
already protective nature. Commenter FV-D-28 supported EPA's determination of HC1 as a
threshold pollutant.
Response: The EPA agrees with the commenter that a risk-based approach would delay
promulgation of the rule. Given the relatively small potential difference between a MACT-based
standard and a risk-based standard, EPA believes that the small benefits are substantially
outweighed by the burden to EPA and the industry of collecting and analyzing the data needed
for a risk-based standard.
20.3 RELATIONSHIP BETWEEN THIS RULE AND OTHER RULES
Comment 1: Commenter IV-D-28 believes that wherever possible it would be better to
cross-reference other rules rather than to repeat or rephrase them in this rule because it will
(1) avoid unintentional changes in the language, (2) assure consistency if there are ever any
amendments, and (3) avoid tinkering with the wording in an effort to "improve" it. The
commenter wants to encourage consistency with other rules, especially the HON, because it
would facilitate compliance at the commenter's large manufacturing facilities, each of which is
subject to several standards. At a minimum, the commenter believes that additional
opportunities for cross-referencing exist in the sections that specify provisions for heat
exchangers (i.e., §§ 63.1362(g), 63.1365(f), 63.1366(g), and 63.1367(e) of the proposed rule).
Response: As with all new regulations that overlap with other regulations or have similar
or identical requirements for specific emission sources, EPA has to make a decision as to when
and where regulatory text is written out and when it is cross-referenced. The points raised by the
commenter are valid concerns and were considered in the rule development and in drafting the
proposed regulatory text. For the most part, EPA utilized cross-referencing with the General
20-2
-------
Provisions, the HON, and a few other rules. This provides uniformity and consistency for the
affected sources, as well as reducing the regulatory text to be included in the rule. However, in
some instances, specific language from other rules was incorporated directly into the proposed
rule in an effort to make it easier for the regulated community to understand the requirements.
Based on this comment, the Agency re-evaluated the heat exchanger provisions.
Although much of the language from the HON was incorporated directly into the proposed rule,
large sections of the provisions were also cross referenced. The EPA determined that this partial
incorporation of language did not have the intended effect of enhancing understanding because
not only would the regulated community have to refer to another rule for some of the provisions,
but they would also have to read both rules closely to check for differences. Therefore, the final
rule cross-references all of the heat exchanger provisions, and notes a few exceptions to those
provisions. This approach also is consistent with the approach used on several other recent
standards.
Comment 2: Commenter IV-D-28 stated that if any of the provisions that are borrowed
from other rules (i.e., restated rather than cross-referenced) are amended in the other rule, EPA
should reopen this rule and request public comment on making the same changes. Even if EPA
amends other rules that currently have no relationship to this rule (e.g., the P&R rules or the
Off-Site Waste and Recovery Operations rule), the commenter believes EPA should consider
whether the amendment might be beneficial in this rule as well, especially for the sake of
maintaining consistency.
Response: The EPA agrees with the commenter that subsequent amendments of other
rules may trigger a need to amend language from those other rules that has been incorporated in
this final rule. The decision, however, to amend any rule must be made on a case-by-case basis.
Comment 3: Because the rule references requirements from the HON that are periodic
(e.g., those for equipment leaks), commenter IV-D-28 believed there should be a statement as in
§ 63.100(k)(9) that specifies the meaning of periods of time.
Response: The language from § 63.100(k)(9) has been added to the final rule.
20-3
-------
20.4 RELATIONSHIP BETWEEN THE NOTIFICATION OF COMPLIANCE STATUS
REPORT AND TITLE V PERMITS
Comment: In the preamble to the proposed rule, EPA solicited comment on how to
incorporate the NOCS report into a facility's title V permit and on the types of changes that
should trigger review actions under title V. Commenter IV-D-28 believes that questions
involving title V programs are best addressed under title V. The commenter noted that States
have developed their own operating permits programs, which differ from one another. These
differences may include whether and how to incorporate various requirements into a permit. As
a result, personnel administering the permit program will need to decide whether, and how, to
incorporate the Notification of Compliance Status report into permits. Commenter F/-G-05 does
not recommend incorporation by reference of the NOCS report into the title V permits without
also requiring the permitting authority to specify the date of the incorporation. The commenter
believes the types of changes that should trigger review actions under title V are any process
changes, or operating and compliance procedures, that increase the emissions from the facility.
Response: The EPA agrees with commenter IV-D-28 that questions involving title V are
best addressed under the title V program.
20.5 OMB REVIEW
Comment: Commenter IV-D-28 noted that it is unusual for Office of Management and
Budget (OMB) to classify a rule as a "significant regulatory action" for "novel legal or policy
issues." Thus, the commenter is interested in understanding OMB's concerns and requests that
EPA make the OMB document publicly available and solicit comment on issues raised by OMB.
Response: No OMB document explaining the rationale behind their classification
decision is available. All interagency exchanges, including review material exchanged with
OMB, are included in the rulemaking docket.
20.6 EDITORIAL CLARIFICATIONS
Comment: Several commenters (P/-D-16, IV-D-21, IV-D-28, and IV-D-29) identified
typographical errors and suggested minor editorial changes.
Response: The commenters' editorial remarks/issues and EPA's responses are
summarized in Table 20-1.
20-4
-------
20.7 COMMENT PERIOD
Comment: Commenter FV-D-28 was concerned that the public comment period, even
with the 30-day extension, which was helpful, was not long enough to ensure that all issues were
identified. The commenter would prefer to deal with issues before the rule is promulgated
because making the provisions workable after promulgation can be time-consuming.
Response: The EPA believes sufficient time was provided for the public comment
period, especially in light of the 30-day extension. As part of the regulatory development
process, the Agency Jjas a schedule to meet in promulgating the rule as well. The Agency will
continue to work with the industry and the public commenters in finalizing the rule and resolving
any issues prior to the promulgation date.
20.8 SUPPORTIVE COMMENTS
Comment: Commenter IV-D-18 supported the rule as proposed and agrees with
identifying a broad range of control devices for compliance purposes. Commenter IV-G-01
supported EPA's proposed exclusion of research and development (R&D) facilities from the
requirements of the rule. Commenter FV-G-05 endorses EPA's approach for identifying PAI
processes subject to the standards because the EPA Form 3540-16 already identifies the affected
sources. However, the commenter cautions that a potential drawback of this approach is
uncertainty about the reliability of the data because it is self-reported. To lessen the effects of
this drawback, the commenter believes regulators can use permits and other information that
must be reported to comply with water or RCRA regulations to assist in identifying affected
sources.
Response: The EPA acknowledges and appreciates the comments.
20-5
-------
TABLE 20-1. EDITORIAL COMMENTS AND RESPONSES
Comment
In section 63.1360(d)(4)(i), change "from
segregated sewers" to "managed in segregated
sewers"
Use the term "storage vessel" rather than
"storage tank" for consistency with other
regulations
Paragraph (f)(4) in section 1360 is missing
Add the word "or" between "foam" and
"liquid" in the definition of liquid-mounted
seal
Reference to HON Table 8 compounds in the
definition of POD should be deleted
Change "bottom receiver" to "bottoms
receiver" in section 63.1362(e)
Delete the words "the use of from the first
sentence of section 63.1364(c)(l)(v)
In section 63.1364(c)(2)(i)(A)(6), change
"...can be obtained from standard reference
texts." to "...may be obtained from standard
references."
In section 63.1364(c)(3)(i)(G), add units for
the gas-to-cloth ratio
In the last sentence of section
63.1364(c)(3)(ii), add "at or" before "below 20
ppmv"
In section 63.1365(f)(2)(iv), the reference to
section 63.1362(f) should be changed to
63.1362(g)
In section 63.1367(a)(l), change "within
150 calendar days of the compliance date" to
"no later than 150 after the compliance date"
In section 63.1367(a)(2), add "at least" before
"12 months"
In section 63.1367(b)(l), replace the second
comma with "and"
In Table 1, delete the word "replace" wherever
it occurs and state what does or does not apply
Commenter
IV-D-28
IV-D-28
FV-D-16 and
IV-D-28
IV-D-28
IV-D-16,
IV-D-21,
IV-D-28, and
IV-D-29
IV-D-28
IV-D-28
IV-D-28
IV-D-28
IV-D-28
IV-D-16
IV-D-28
IV-D-28
IV-D-28
IV-D-28
Response
The proposed language is consistent with
the HON and has not been changed in the
final rule.
Change made as suggested in comment.
Paragraph (0(5) in the proposed rule
should have b«en numbered (f)(4).
Change made as suggested in comment.
This was an oversight in the proposed
rule. Change made as suggested in
comment.
Change made as suggested in comment.
Edited section to delete this phrase.
Change made as suggested in comment.
Example units have been included.
Edited this section and deleted this
sentence.
Commenter is correct, but this provision
has been deleted from the final rule and
replaced with cross reference to the
HON.
Change made as suggested in comment.
Change made as suggested in comment,
but submittal date changed from
12 months to 6 months before the
compliance date.
This provision has been deleted from the
final rule and replaced with cross
reference to the HON.
No changes made; several other rules use
this language to describe changes.
20-6
-------
TABLE 20.1 (continued)
Comment
The terms "HAP" and "organic HAP" are not
used consistently; check entire rule
In the definition of "Group 1 storage tank,"
change cutoffs from 37 m to 38 m and
replace <76 m3 with >38 nr to <76 m3 in the
applicability column of Table 2
Commenter
IV-D-28
IV-D-16
Response
Checked entire rule and made changes.
In some cases "total HAP" was added.
Commenter correctly identified
typographical errors, but the cutoffs have
changed in the final rule.
20-7
-------
Protection Agency
i
Chicago, IL 60604-3590
------- |