Protocol Report 2015

April 2015
Emissions Estimation Protocol

for Petroleum Refineries
Version 3
Submitted to:
Office of Air Quality Planning and Standards

U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Submitted by:
RTI International
3040 Cornwallis Road
Research Triangle Park, NC 27709-2194
[This page intentionally left blank.]
Emissions Estimation Protocol for Petroleum Refineries
Version 3 Table of Contents
Table of Contents
1. Introduction ......................................................................................................................................... 1-1
1.1 Completeness ............................................................................................................................ 1-1
1.2 Data Quality............................................................................................................................ 1-10
1.3 Calculations and Significant Digits ........................................................................................ 1-10
2. Equipment Leaks ................................................................................................................................. 2-1
2.1 Methodology Rank 1 for Equipment Leaks.............................................................................. 2-2
2.2 Methodology Rank 2 for Equipment Leaks.............................................................................. 2-3
2.2.1 Speciating Equipment Leak Emissions ....................................................................... 2-5
2.2.2 Calculating Hourly and Annual Equipment Leak Emissions ..................................... 2-7
2.3 Methodology Rank 3 for Equipment Leaks............................................................................ 2-10
2.3.1 Speciating Equipment Leak Emissions ..................................................................... 2-12
2.3.2 Calculating Hourly and Annual Equipment Leak Emissions ................................... 2-12
2.4 Methodology Ranks 4 and 5 for Equipment Leaks ................................................................ 2-13
2.4.1 Speciating Equipment Leak Emissions ..................................................................... 2-14
2.4.2 Calculating Hourly and Annual Equipment Leak Emissions ................................... 2-17
3. Storage Tanks ...................................................................................................................................... 3-1
3.1 Methodology Rank 1 for Storage Tanks................................................................................... 3-1
4. Stationary Combustion Sources .......................................................................................................... 4-1
4.1 Methodology Rank 1 for Stationary Combustion Sources ....................................................... 4-2
4.2 Methodology Rank 2 for Stationary Combustion Sources ....................................................... 4-5
4.3 Methodology Rank 3A for Stationary Combustion Sources .................................................... 4-9
4.4 Methodology Rank 3B for Stationary Combustion Sources .................................................. 4-10
4.5 Methodology Rank 4 for Stationary Combustion Sources ..................................................... 4-11
4.5.1 Default Emission Factors for Process Heaters .......................................................... 4-12
4.5.2 Default Emission Factor for Internal Combustion Engines ...................................... 4-18
4.5.3 Default Emission Factors for Combustion Turbines................................................. 4-18
5. Process Vents ...................................................................................................................................... 5-1
5.1 Catalytic Cracking Units .......................................................................................................... 5-1
5.1.1 Methodology Ranks 1 and 2 for Catalytic Cracking Units ......................................... 5-2
5.1.2 Methodology Ranks 3 and 4 for Catalytic Cracking Units ......................................... 5-2
5.1.3 Methodology Rank 5A for CCU Metal HAP Emissions Estimates ............................ 5-7
5.1.4 Methodology Rank 5B for Catalytic Cracking Units................................................ 5-10
5.2 Fluid Coking Units ................................................................................................................. 5-11
5.2.1 Methodology Ranks 1 and 2 for Fluid Coking Units ................................................ 5-12
5.2.2 Methodology Ranks 3 and 4 for Fluid Coking Units ................................................ 5-12
5.2.3 Methodology Rank 5 for Fluid Coking Units ........................................................... 5-12
5.3 Delayed Coking Units ............................................................................................................ 5-12
5.3.1 Methodology Ranks 1 and 2 for Delayed Coking Units ........................................... 5-14
5.3.2 Methodology Rank 3 and 4 for Delayed Coking Units............................................. 5-14
5.3.3 Methodology Rank 5 for Delayed Coking Units ...................................................... 5-16
5.4 Catalytic Reforming Units ...................................................................................................... 5-17
5.4.1 Emissions Estimation Methodology for Catalytic Reforming Units ........................ 5-17
5.5 Sulfur Recovery Plants ........................................................................................................... 5-19
5.5.1 Methodology Ranks 1 and 2 for Sulfur Recovery Plants .......................................... 5-20
iii
5.5.2 Methodology Ranks 3 and 4 for Sulfur Recovery Plants .......................................... 5-20
5.5.3 Methodology Rank 5 for Sulfur Recovery Plants ..................................................... 5-21
5.6 Other Miscellaneous Process Vents ....................................................................................... 5-21
5.6.1 Hydrogen Plant Vents ............................................................................................... 5-21
5.6.2 Asphalt Plant Vents................................................................................................... 5-22
5.6.3 Coke Calcining.......................................................................................................... 5-24
5.6.4 Blowdown Systems ................................................................................................... 5-27
5.6.5 Vacuum Producing Systems ..................................................................................... 5-27
6. Flares ................................................................................................................................................... 6-1
6.1 Methodology Rank 1 for Flares ................................................................................................ 6-2
6.5 Methodology Ranks 5 and 6 for Flares .................................................................................... 6-7
7. Wastewater Collection and Treatment Systems .................................................................................. 7-1
7.1 Methodology Rank 1 for Wastewater Treatment Units ............................................................ 7-2
7.2 Methodology Rank 2 for Wastewater Treatment Units ............................................................ 7-2
7.2.1 Wastewater Collection Systems .................................................................................. 7-3
7.2.2 Primary Weirs ............................................................................................................. 7-5
7.2.3 Oil-Water Separators .................................................................................................. 7-5
7.2.4 Dissolved Air Flotation Units ..................................................................................... 7-6
7.2.5 Equalization Tanks...................................................................................................... 7-6
7.2.6 Biological Treatment Units ......................................................................................... 7-6
7.2.7 Polishing Ponds........................................................................................................... 7-8
7.2.8 Site-Specific Factors ................................................................................................... 7-8
7.2.9 Model Validation ........................................................................................................ 7-9
7.3 Methodology Rank 3 for Uncovered Units .............................................................................. 7-9
7.3.1 Engineering Estimates Based on Wastewater Treatment Plant Load ......................... 7-9
7.3.2 Engineering Estimates Based on Process Capacities .................................................. 7-9
8. Cooling Towers ................................................................................................................................... 8-1
8.1 Methodology Ranks 1 and 2 for Cooling Towers .................................................................... 8-2
8.2 Methodology Rank 3 for Cooling Towers ................................................................................ 8-8
8.3 Methodology Rank 4 for Cooling Towers .............................................................................. 8-10
8.4 Methodology Rank 5 for Cooling Towers .............................................................................. 8-11
8.4.1 VOC and Volatile Organic HAP ............................................................................... 8-12
8.4.2 Particulate Matter Emissions .................................................................................... 8-13
8.4.3 Non-volatile Organic HAP Emissions ...................................................................... 8-16
8.4.4 Chlorine Emissions ................................................................................................... 8-16
9. Loading Operations ............................................................................................................................. 9-1
9.1 Data Available on Product Composition and Properties .......................................................... 9-2
9.2 Methodology Rank 1 for Loading Operations .......................................................................... 9-2
9.3 Methodology Ranks 2 and 3 for Loading operations ............................................................... 9-3
9.4 Methodology Rank 4 for Loading operations........................................................................... 9-4
9.4.1 AP-42 Emission Factors for “Product” or “Total Hydrocarbon” Emissions .............. 9-5
9.4.2 Estimate Uncontrolled Emissions and Speciate .......................................................... 9-6
9.4.3 Capture Efficiency ...................................................................................................... 9-6
9.4.4 Overall Control Efficiency .......................................................................................... 9-7
iv
10. Fugitive Dust Sources ....................................................................................................................... 10-1

11. Startup and Shutdown ....................................................................................................................... 11-1
11.1 Gaseous Process Vessel Depressurization and Purging ......................................................... 11-1
11.2 Liquid Process Vessel Depressurization and Purging ............................................................ 11-2
12. Malfunctions/Upsets .......................................................................................................................... 12-1
12.1 Control Device Malfunctions ................................................................................................. 12-2
12.2 Vessel Overpressurization ...................................................................................................... 12-3
12.3 Spills ....................................................................................................................................... 12-6
13. References ......................................................................................................................................... 13-1
Appendix A Average Stream Compositions ......................................................................................... A-1
Appendix B Wastewater Treatment System Equations ........................................................................ B-2
Appendix C Primer for TANKS Model Users ...................................................................................... C-1
Appendix D Emission Factors from Air Toxic Emission Factors for Combustion Sources Using
Petroleum Based Fuels – Final Report, Volume 2 – Development of Emission
Factors Using CARB Approach ....................................................................................... D-1
v
List of Figures
Figure 2-1. Illustration of alternative methods to determine equipment leak emissions from
routine monitoring data. ................................................................................................................ 2-8
Figure 7-1. Typical refinery wastewater treatment system. ....................................................................... 7-1
Figure 7-2. Typical refinery wastewater collection system process drainage areas. ................................. 7-4
Figure 7-3. Simplified drawing of a constituent mass balance in a biological treatment unit. .................. 7-7
List of Tables
Table 1-1. Summary of Pollutants and Emission Sources Inclusion in a Petroleum Refinery’s
Emission Inventory ....................................................................................................................... 1-3
Table 2-1. Summary of Equipment Leak Emission Estimates .................................................................. 2-2
Table 2-2. Equipment Leak Rate for Petroleum and SOCMI Equipment Componentsa ........................... 2-4
Table 2-3. Screening Ranges Emission Factorsa ...................................................................................... 2-11
Table 2-4. Median Equipment Leak Component Counts for Small Model Processesa............................ 2-15
Table 2-5. Median Equipment Leak Component Counts for Large Model Processesa............................ 2-16
Table 2-6. Refinery and SOCMI Average Component Emission Factorsa .............................................. 2-17
Table 2-7. Concentration of HAP in Refinery Process Unit Streamsa ..................................................... 2-18
Table 3-1. Summary of Typical Hierarchy of Storage Tank Emission Estimates ..................................... 3-1
Table 3-2. Default Control Efficiencies for Different VOC Control Devices ........................................... 3-4
Table 3-3. Default Emissions Factors for Petroleum Refinery Storage Tanks .......................................... 3-5
Table 4-1. Summary of Typical Hierarchy of Stationary Combustion Source Emission Estimates.......... 4-2
Table 4-2. Molar Exhaust Volumes and Molar Heat Content of Refinery Fuel Gas Constituents ............ 4-5
Table 4-3. Summary of Emission Factors for Boilers and Process Heaters Firing Various Fuels .......... 4-13
Table 4-4. Summary of Emission Factors for Internal Combustion Engines Firing Various Fuels ........ 4-19
Table 4-5. Summary of Emission Factors for Combustion Turbines Firing Various Fuelsa ................... 4-23
Table 5-1. Summary of Typical Hierarchy of Process Vent Emissions Estimates .................................... 5-1
Table 5-2. Default Size Distribution for Filterable PM from CCU ........................................................... 5-6
Table 5-3. Default Ratio of Metal HAP Composition of CCU Finesa ....................................................... 5-8
Table 5-4. Organic HAP Emissions Factors for CCU Catalyst Regenerator Vent .................................. 5-10
Table 5-5. Emissions Factors for Delayed Coking Unit Decoking Operations ....................................... 5-16
Table 5-6. Emissions Factors for CRU Catalyst Regeneration Vent ....................................................... 5-18
Table 5-7. Emissions Factors for Sulfur Recovery Plants ....................................................................... 5-21
Table 5-8. Emission Factors for Asphalt Blowing (U.S. EPA, 1995a) .................................................... 5-23
Table 5-9. Asphalt Blowing – Nonmethane Volatile Organic Compounds Speciationa .......................... 5-23
Table 5-10. Summary of Emissions Factors for Controlled Asphalt Blowing ........................................ 5-24
Table 5-11. Summary of Emissions Factors for Controlled Coke Calcining........................................... 5-25
Table 5-12. Default Emissions Factors for Blowdown Systems.............................................................. 5-27
Table 5-13. Default Emissions Factor for Vacuum Producing Systems .................................................. 5-28
Table 6-1. Summary of Flare Emissions Estimate Methodologies ............................................................ 6-2
Table 6-2. Flare Energy Consumption-Based Emission Factorsa .............................................................. 6-5
Table 6-3. Emission Factors for Soot from Flares ..................................................................................... 6-5
Table 6-4. Flare General Emission Factorsa............................................................................................... 6-8
Table 7-1. Summary of Wastewater Treatment Emission Estimates ......................................................... 7-2
Table 7-2. Critical Inputs and Chemical Properties Specific to Wastewater Collection System
PDA Air Emission Calculations ................................................................................................... 7-5
Table 7-3. Critical Inputs, Variables, and Chemical Properties Specific to Primary Weir Air
Emission Calculations................................................................................................................... 7-5
vi
Table 7-4. Critical Inputs, Variables, and Chemical Properties Specific to Oil-Water Separators
Air Emission Calculations ............................................................................................................ 7-5
Table 7-5. Critical Inputs, Variables, and Chemical Properties Specific to DAF Air Emission
Calculations .................................................................................................................................. 7-6
Table 7-6. Critical Inputs, Variables, and Chemical Properties Specific to Equalization Tank Air
Emission Calculations................................................................................................................... 7-6
Table 7-7. Critical Inputs, Variables, and Chemical Properties Specific to Biological Treatment
Unit Air Emission Calculations .................................................................................................... 7-8
Table 7-8. Model Process Unit Characteristics for Petroleum Refinery Wastewatera ............................. 7-10
Table 7-9. Refinery Wastewater Contaminant Concentrations as a Ratio to Benzene ............................ 7-10
Table 7-10. Default Mass Emission Factors for Refinery Wastewater Collection and Treatment
Systems ....................................................................................................................................... 7-11
Table 8-1. Summary of Cooling Tower Emissions Estimation Methodologies ........................................ 8-2
Table 8-2. Data Requirements for VOC or Speciated VOC Emissions, Methodology Ranks 1
and 2.............................................................................................................................................. 8-3
Table 8-3. Data Requirements for Speciated Compound Emissions, Methodology Rank 3 ..................... 8-9
Table 8-4. Data Requirements for Speciated Compound Emissions, Methodology Rank 4 ................... 8-11
Table 8-5. Methodology Rank 5 Default Emission Factors ..................................................................... 8-11
Table 8-6. Data Requirements for VOC or Speciated VOC Emissions, Methodology Rank 5 ............... 8-12
Table 9-1. Summary of Loading Operations Emission Estimates ............................................................. 9-2
Table 9-2. Data Requirements for VOC or Speciated Emissions, Methodology Rank 1 or 2 ................... 9-3
Table 9-3. Saturation Factorsa .................................................................................................................... 9-5
Table 9-4. VOC Emission Factors for Marine Vessel Loading of Gasoline at Marine Terminalsa ........... 9-6
Table 9-5. Capture Efficiencies for Vapor Collection Systemsa ................................................................ 9-7
Table 9-6. Sample Calculation Methodology Rank 4—Summary of Emissions (When Property
Data Are Not Available) ............................................................................................................. 9-10
Table 10-1. Default Values for Fugitive Dust Emission Estimates ......................................................... 10-1
Table 12-1. Control Device Efficiency and Multiplier Factors for Control Device Malfunctions .......... 12-2
Table 12-2. Values of k for Various Gasesa ............................................................................................. 12-4
vii
viii
Draft Version 3 List of Acronyms and Abbreviations
List of Acronyms and Abbreviations
µg/L micrograms per liter

acfm actual cubic feet per minute
API American Petroleum Institute
atm atmosphere
atm/in. of Hg atmospheres per inch of mercury
bbl/cd barrels per calendar day
bbl/sd barrels per stream day
bbl/hr barrels per hour
BID Background Information Document
BOX test Batch test with oxygen addition (a type of aerated reactor test specified in
Appendix C to 40 CFR Part 63)
Btu/mol British thermal units per mole
Btu/scf British thermal units per standard cubic feet
BWON Benzene Waste Operations NESHAP
CARB California Air Resources Board
CCU catalytic cracking unit
CEMS continuous emission monitoring system
CERMS continuous emission rate monitoring system
CFR Code of Federal Regulations
CH4 methane
Cl2 chlorine
CO carbon monoxide
COS carbonyl sulfide
Cr+6 hexavalent chromium
CRU catalytic reforming unit
CS2 carbon disulfide
DAF dissolved air flotation
DIAL Differential Absorption Light Detection and Ranging
DNF dissolved nitrogen flotation
dscf/MMBtu dry standard cubic feet per million British thermal unit
dscf/mol dry standard cubic feet per mole
dscfm dry standard cubic feet per minute
E-cat equilibrium catalyst
EPA U.S. Environmental Protection Agency
ESP electrostatic precipitator
FCCU fluid catalytic cracking unit
FCU fluid coking unit
FID flame ionization detection
g/hr grams per hour
ix
g/mol grams per mole

gal/bbl gallons of wastewater per barrel of capacity at a given process unit
GC gas chromatograph
GC/MS gas chromatography/mass spectrometry
GHG greenhouse gases
H2S hydrogen sulfide
HAP hazardous air pollutants
HCl hydrogen chloride
hepta-CDD heptachloro-dibenzo-p-dioxin
Hg Mercury
hr/yr hours per year
IAF induced air flotation
ICR information collection request
in. Inches
kg TOC/hr kilograms of TOC per hour
kg/hr kilograms per hour
kg/hr/source kilograms per hour per source
kg/kg-mol kilogram per kilogram mole
kg/yr kilograms per year
kPa kilopascals
lb/dscf pounds per dry standard cubic foot
lb/lb-mol pounds per pound mole
lb/lton pounds per long ton
lb/Mgal pounds per thousand gallons
lb/MMbbl pounds per million barrels
lb/MMBtu pounds per million British thermal unit
lb/MMcf pounds per million cubic feet
lb/ton pounds per ton
lb/ton coke pounds per ton of petroleum coke
LDAR leak detection and repair
LIDAR Light Detection and Ranging
m3/day cubic meters per day
m3/min cubic meters per minute
MACT Maximum Achievable Control Technology
mg/kg milligrams per kilogram
min/hr minutes per hour
mL/min milliliters per minute
mL-atm/mol-K milliliters-atmospheres per mole-Kelvin
MMBtu/scf million British thermal units per standard cubic foot
MW molecular weight
Ni Nickel
x
NMOC nonmethane organic compounds

NMVOC nonmethane volatile organic compounds
NO nitric oxide
NO2 nitrogen dioxide
NOx nitrogen oxides
NSPS new source performance standards
O2 Oxygen
OCDD octachloro-dibenzo-p-dioxin
OCDF octachloro-dibenzo-furan
OVA organic vapor analyzer
PAH polycyclic aromatic hydrocarbons
PCB polychlorinated biphenyls
PM particulate matter
PM10 PM emissions that are 10 µm in diameter or less
PM10-FIL filterable (or front-half catch) portion of the PM emissions that are 10 µm in
diameter or less
PM10-PRI “primary” PM emissions that are 10 µm in diameter or less
PM2.5 PM emissions that are 2.5 µm in diameter or less
PM2.5-FIL filterable (or front-half catch) portion of the PM emissions that are 2.5 µm in
diameter or less
PM2.5-PRI “primary” PM emissions that are 2.5 µm in diameter or less
PM-CON condensable PM (or back-half catch)
POM polycyclic organic matter
POTW publicly owned treatment works
ppmv parts per million by volume
ppmw parts per million by weight
psi pounds per square inch
psia pounds per square inch absolute
psig pounds per square inch gauge
QA quality assurance
QC quality control
scf/kg-mol standard cubic feet per kilogram mole
scf/lb-mol standard cubic feet per pound mole
SCR selective catalytic reduction
SNCR selective non-catalytic reduction
SO2 sulfur dioxide
SRU sulfur recovery unit
SSM startup, shutdown, or malfunction
SV screening value
TCCU thermal catalytic cracking unit
TCDD tetrachlorodibenzo-p-dioxin
xi
TDS total dissolved solids

TEQ toxic equivalents
THC total hydrocarbons
TOC total organic compounds
tons/hr tons per hour
tons/kg tons per kilogram
tons/yr tons per year
VOC volatile organic compounds
vol% volume percent
WebFIRE Internet version of the Factor Information Retrieval (FIRE) data system
WF weight fraction
WSPA Western States Petroleum Association
wt% weight percent
°C degrees Celsius
°F degrees Fahrenheit
°R degrees Rankine
xii
Version 3 Section 1—Introduction
1. Introduction
This Refinery Emissions Protocol document is intended to provide guidance and instructions to petroleum
refinery owners and operators and to federal, state, and local agencies for the purpose of improving
emission inventories for the petroleum refining industry. This document presents a hierarchy of emission
measurement or estimation methods for various petroleum refinery emission sources and provides a
listing of pollutants for which emissions are anticipated for each source type.
For each emission source, the various emission measurement or estimation methods specific to that
source are ranked in order of preference, with “Methodology Rank 1” being the preferred method,
followed by “Methodology Rank 2,” and so on. Refinery owners and operators and other inventory
developers are requested to use the highest ranked method (with Methodology Rank 1 being the highest)
for which data are available. Methodology Ranks 1 or 2 generally rely on continuous emission
measurements. When continuous measurement data are not available, engineering calculations or site-
specific emission factors (Methodology Ranks 3 and 4) are specified; these methods generally need
periodic, site-specific measurements. When site-specific measurement or test data are not available,
default emission factors (Methodology Rank 5) are provided.
In the development of this Refinery Emissions Protocol document, U.S. Environmental Protection
Agency (EPA) reviewed available source test data to verify or refine existing emission factors and
develop new emission factors for sources that currently do not have default emission factors. EPA also
provided guidance on characterizing and quantifying emissions associated with start-up, shut-down, and
malfunction events. The “peer review draft version” (Version 1) of the Refinery Emissions Protocol
document was posted on the EPA Web site (http://www.epa.gov/ttn/chief/efpac/esttools.html) on January
7, 2010, for initial public comment. Public comments were received from three different commenters. A
special version of this Refinery Emissions Protocol document was then developed specific to the EPA’s
2011 information collection request (ICR) for the petroleum refining industry (Version 2). This “version”
(Version 3) of the Refinery Emissions Protocol document provides updates to certain emissions factors
and methodologies developed using the additional test data collected as part of the 2011 ICR. While
efforts have been made to coordinate the revisions of this Refinery Emissions Protocol document with the
updates to the emissions factors in AP-42 (U.S. EPA, 1995a) and WebFIRE (the internet version of the
Factor Information Retrieval (FIRE) data system), this Refinery Emissions Protocol document may
include additional methodologies and emissions factors that are not included in AP-42, particularly for
hazardous air pollutants (HAP). The EPA website (http://www.epa.gov/ttn/chief/efpac/esttools.html)
should be checked prior to preparing an emissions inventory to ensure the most recent Refinery Emissions
Protocol document is used.
1.1 Completeness
Emission estimates should be provided for each emission source at the refinery, including ancillary
sources and non-refinery process units. While this Refinery Emissions Protocol document attempts to
identify and provide methodologies for each emission source at a typical petroleum refinery, there may be
certain sources located at the refinery that are not specifically addressed within the Refinery Emissions
Protocol document. Additionally, there are sources included in this Refinery Emissions Protocol
document for which no emission data are available to provide default (Methodology Rank 5) methods. It
is the EPA’s policy that emission estimates should be provided for every emissions source known to be
present at a facility, even when emissions factors are not specifically included in AP-42 or this Refinery
Emissions Protocol document (U.S. EPA, 1995a).
Similarly, emission estimates should be provided for each pollutant (except for greenhouse gases [GHGs],
which are required to be reported under the Mandatory Greenhouse Gas Emissions Reporting Rule [Final
Rule, 74 FR 56260]) emitted from a given emission source. Table 1-1 provides a listing of the pollutants
1-1
expected to be emitted by various sources described in this protocol document. Filled circles indicate
compound/emission source pairings for which emissions are expected to occur and for which emissions
estimates should be developed. Hollow circles indicate pairings for which data may be reported (if the
chemical is present or if data are available to speciate to that extent), but the inventory can be deemed
complete without these estimates. In general, overall VOC emission measurement methods via EPA
Methods 25, 25A-E, or 305 are preferred for reporting VOC emissions. Speciated VOC emissions should
also be reported when that data are available; however, speciation of VOC emissions is not required.
While emission estimates are desirable for every compound/emission source pairing where there is a
filled circle in Table 1-1, nothing in this Refinery Emissions Protocol document should be construed to
require additional sampling or analyses in order to provide these emissions estimates. Rather available
data, supplemented with engineering analyses (following the guidance provided in this Protocol, where
applicable), may be used to provide the desired speciation.
Some criteria pollutants, such as PM10 or PM2.5 have special reporting nomenclatures to indicate the
fraction of the particulate matter (PM) emissions that are filterable or condensable (see PM Emission
Inventory Nomenclature text box). Other criteria pollutants, such as nitrogen dioxide (NO2), are often
determined or regulated as a combination of chemicals. For example, nitrogen oxides (NOx), is the sum of
NO2 and nitric oxide (NO) emissions. The inclusion of these additional nomenclatures or groupings in
Table 1-1 is not intended to suggest that these compounds are criteria pollutants, but that these
“pollutants” should be included in the refinery emission inventory.
PM Emission Inventory Nomenclature

PM emissions inventories have their own nomenclature and structure. A complete PM emissions inventory
includes the following components:
 PM10-PRI: “Primary” PM emissions that are 10 µm in diameter or less. PM10-PRI = PM10-FIL + PM-CON.
 PM10-FIL: Filterable (or front-half catch) portion of the PM emissions that are 10 µm in diameter or less.
 PM-CON: Condensable PM (or back-half catch). All condensable PM is assumed to be less than 2.5 microns
(µm) in diameter (PM2.5).
 PM2.5-PRI: “Primary” PM emissions that are 2.5 µm in diameter or less. PM2.5-PRI = PM25-FIL + PM-CON.
 PM2.5-FIL: Filterable (or front-half catch) portion of the PM emissions that are 2.5 µm in diameter or less.
Although a complete PM emissions inventory includes PM emissions that are 10 µm in diameter or less, some
measurement methods also collect PM particles that are greater than 10 µm in diameter. The following
nomenclature is used to designate PM emissions that include PM greater than 10 µm in diameter:
 PM-PRI: “Primary” PM emissions of any particle size. PM-PRI = PM-FIL + PM-CON.
 PM-FIL: Filterable (or front-half catch) portion of the PM emissions of any particle size.
1-2
Table 1-1. Summary of Pollutants and Emission Sources Inclusion

in a Petroleum Refinery’s Emission Inventory
Process Vents
Stationary Combustion
Fugitive Dust Sources

Catalytic Cracking
Startup/Shutdown
Fluid Coking Unit
Equipment Leaks
Vacuum Systems
Product Loading
Sulfur Recovery
Cooling Towers
Delayed Coking
Hydrogen Plant
Coke Calcining
Reforming Unit
Storage Tanks
Asphalt Plant
Malfunctions
Wastewater
Blowdown
CAS
Catalytic
Systems
Number or
Plants
Flares
Pollutant
Unit
Unit
Code Substance
Criteria Pollutants
630-08-0 Carbon monoxide ● ● ● ● ● ● ○ ○
7439-92-1 Lead ● ● ● ○ ○ ○ ○
Particulate matter (PM)

PM10-PRI ● ● ● ● ● ● ● ● ● ○ ○
≤ 10 micrometers (µm)
PM10-FIL Filterable PM ≤ 10 µm ● ● ● ● ● ● ● ● ● ○ ○
PM25-PRI PM ≤ 2.5 µm ● ● ● ● ● ● ● ● ● ○ ○
PM25-FIL Filterable PM ≤ 2.5 µm ● ● ● ● ● ● ● ● ● ○ ○
PM-CON Condensable PM ● ● ● ● ● ● ● ● ○ ○ ○
10102-44-0 Nitrogen dioxide ● ● ● ● ● ● ○ ○
NOX Nitrogen oxides ● ● ● ● ● ● ○ ○
7446-09-5 Sulfur dioxide ● ● ● ● ● ● ● ●
Volatile organic
VOC ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
compounds
Specific VOC Constituents (Compounds listed below plus those listed under “Volatile Organic HAPs”)
74-85-1 Ethylene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
74-86-2 Acetylene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
74-98-6 Propane ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
115-07-1 Propylene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
463-49-0 Propadiene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
106-97-8 n-Butane ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
75-28-5 Isobutane ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
106-98-9 1-Butene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
107-01-7 2-Butene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
115-11-7 Isobutene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
(continued)
1-3

in a Petroleum Refinery’s Emission Inventory (continued)
Process Vents

Catalytic Cracking
Startup/Shutdown
Fluid Coking Unit
Equipment Leaks
Vacuum Systems
Product Loading
Sulfur Recovery
Cooling Towers
Delayed Coking
Hydrogen Plant
Coke Calcining
Reforming Unit
Storage Tanks
Asphalt Plant
Malfunctions
Wastewater
Blowdown
CAS
Catalytic
Systems
Number
Plants
Flares
or Pollutant
Unit
Unit
Code Substance
590-19-2 1,2-Butadiene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
109-66-0 n-pentane ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
78-78-4 2-Methylbutane ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
287-92-3 Cyclopentane ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
591-95-7 1,2-Pentadiene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
1574-41-0 1-cis-3-Pentadiene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
2004-70-8 1-trans-3-Pentadiene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
591-93-5 1,4-Pentadiene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
591-96-8 2,3-Pentadiene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
598-25-4 3-Methyl-1,2-butadiene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
78-79-5 2-Methyl-1,3-butadiene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
542-92-7 Cyclopentadiene ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
110-82-7 Cyclohexane ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
108-87-2 Methylcylcohexane ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
142-82-5 Heptane (and isomers) ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
111-65-9 Octane (and isomers) ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
78-93-3 Methyl ethyl ketone ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
25551-13-7 Trimethylbenzene(s) ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
Hazardous Air Pollutants (HAPs)
Volatile Organic HAPs
75-07-0 Acetaldehyde ○ ○ ● ● ● ● ● ○ ● ● ● ○ ○ ● ○ ○ ○ ○ ○
107-02-8 Acrolein ○ ○ ● ● ● ● ● ○ ○ ○ ● ○ ○ ● ○ ○ ○ ○ ○
62-53-3 Analine ○ ○ ● ● ● ● ● ○ ○ ○ ● ○ ○ ● ○ ○ ○ ○ ○
71-43-2 Benzene ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
74-83-9 Bromomethane ○ ○ ○ ○
106-99-0 1,3-Butadiene ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
75-00-3 Chloroethane ○ ○ ○ ○ ○
(continued)
1-4

Process Vents

Catalytic Cracking
Startup/Shutdown
Fluid Coking Unit
Equipment Leaks
Vacuum Systems
Product Loading
Sulfur Recovery
Cooling Towers
Delayed Coking
Hydrogen Plant
Coke Calcining
Reforming Unit
Storage Tanks
Asphalt Plant
Malfunctions
Wastewater
Blowdown
CAS
Catalytic
Systems
Number
Plants
Flares
or Pollutant
Unit
Unit
Code Substance
67-66-3 Chloroform ○ ○ ● ○ ○
74-87-3 Chloromethane ○ ○ ○ ○ ○
98-82-8 Cumene ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
106-93-4 1,2-Dibromoethane ○ ○ ○ ○ ○ ○
106-46-7 1,4-Dichlorobenzene ○ ○ ○ ○ ○ ○
75-34-3 1,1-Dichloroethane ○ ○ ○ ○ ○
107-06-2 1,2-Dichloroethane ○ ○ ○ ○ ○
75-35-4 1,1-Dichloroethylene ○ ○ ● ○ ○
78-87-5 1,2-Dichloropropane ○ ○ ○ ○ ○
542-75-6 1,3-Dichloropropene ○ ○ ○ ○ ○
111-42-2 Diethanolamine ● ● ● ○ ○ ○ ○
100-41-4 Ethylbenzene ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
50-00-0 Formaldehyde ○ ○ ● ● ● ● ● ● ● ● ○ ○ ● ○ ○ ○ ● ●
67-72-1 Hexachloroethane ○ ○ ○ ○
110-54-3 n-Hexane ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
67-56-1 Methanol ● ● ○ ○
108-10-1 Methyl isobutyl ketone ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
1634-04-4 Methyl tert-butyl ether ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
100-42-5 Styrene ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
1,1,2,2- ○ ○
79-34-5 ○ ○ ●
Tetrachloroethane
127-18-4 Tetrachloroethylene ○ ○ ● ○ ○
108-88-3 Toluene ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
79-00-5 1,1,2-Trichloroethane ○ ○ ● ○ ○
79-01-6 Trichloroethylene ○ ○ ● ○ ○
121-44-8 Triethylamine ● ● ● ○ ○ ○ ○
540-84-1 2,2,4-Trimethylpentane ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
593-60-2 Vinyl bromide ○ ○ ○ ○ ○

(continued)
1-5

Process Vents

Catalytic Cracking
Startup/Shutdown
Fluid Coking Unit
Equipment Leaks
Vacuum Systems
Product Loading
Sulfur Recovery
Cooling Towers
Delayed Coking
Hydrogen Plant
Coke Calcining
Reforming Unit
Storage Tanks
Asphalt Plant
Malfunctions
Wastewater
Blowdown
CAS
Catalytic
Systems
Number
Plants
Flares
or Pollutant
Unit
Unit
Code Substance
75-01-4 Vinyl chloride ○ ○ ● ○ ○
95-47-6 o-Xylene ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
108-38-3 m-Xylene ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
106-42-3 p-Xylene ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
1330-20-7 Xylenes (total) ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
Semi-volatile and Non-volatile Organic HAPs (except dioxins, furans, and polychlorinated biphenyls [PCBs])
83-32-9 Acenaphthene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
208-96-8 Acenaphthylene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
120-12-7 Anthracene POM

● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
56-55-3 Benzo(a)anthracene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
50-32-8 Benzo(a)pyrene POM

● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
205-99-2 Benzo(b)fluoranthenePOM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
192-97-2 Benzo(e)pyrene POM

● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
191-24-2 Benzo(g,h,i)perylenePOM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
207-08-9 Benzo(k)fluoranthene POM

● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
92-52-4 Biphenyl POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
Bis(2-ethyl hexyl)
117-81-7 ○ ● ● ○
phthalate
91-58-7 2-ChloronaphthalenePOM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
108-39-4 m-Cresol ● ● ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ● ● ●
95-48-7 o-Cresol ● ● ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ● ● ●
106-44-5 p-Cresol ● ● ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ● ● ●
1319-77-3 Cresols (total) ● ● ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ● ● ●
218-01-9 Chrysene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
Dibenz(a,h)
53-70-3 ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
anthracenePOM
84-74-2 di-n-Butyl phthalate ○ ● ● ○
84-66-2 Diethyl-phthalate ○ ● ● ○
(continued)
1-6

Process Vents

Catalytic Cracking
Startup/Shutdown
Fluid Coking Unit
Equipment Leaks
Vacuum Systems
Product Loading
Sulfur Recovery
Cooling Towers
Delayed Coking
Hydrogen Plant
Coke Calcining
Reforming Unit
Storage Tanks
Asphalt Plant
Malfunctions
Wastewater
Blowdown
CAS
Catalytic
Systems
Number
Plants
Flares
or Pollutant
Unit
Unit
Code Substance
57-97-6 7,12-Dimethylbenz(a) ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
anthracene POM
206-44-0 Fluoranthene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
86-73-7 Fluorene POM

● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
Indeno(1,2,3-cd)
193-39-5 ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
pyrenePOM
56-49-5 3-MethylchloranthrenePOM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
91-57-6 2-Methylnaphthalene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
218-01-9 Chrysene POM

● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
91-20-3 Naphthalene POM ● ● ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ● ○ ● ●
198-55-0 Perylene POM

● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
85-01-8 Phenanthrene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
108-95-2 Phenol ● ● ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ● ● ●
129-00-0 Pyrene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●
Dioxins/Furans/PCBs
1746-01-6 Dioxin: 4D 2378j ● ● ● ● ● ○ ○ ○
40321-76-4 Dioxin: 5D 12378 j

○ ○ ○ ○ ○ ○ ○ ○
39227-28-6 Dioxin: 6D 123478j ○ ○ ○ ○ ○ ○ ○ ○
57653-85-7 Dioxin: 6D 123678 j

○ ○ ○ ○ ○ ○ ○ ○
19408-74-3 Dioxin: 6D 123789j ○ ○ ○ ○ ○ ○ ○ ○
35822-46-9 Dioxin: 7D 1234678 j

○ ○ ○ ○ ○ ○ ○ ○
3268-87-9 Dioxin: 8Dj ○ ○ ○ ○ ○ ○ ○ ○
132-64-9 Dibenzofurans ● ● ● ● ● ○ ○ ○
51207-31-9 Furan: 4F 2378k ○ ○ ○ ○ ○ ○ ○ ○
57117-41-6 Furan: 5F 12378 k

○ ○ ○ ○ ○ ○ ○ ○
57117-31-4 Furan: 5F 23478k ○ ○ ○ ○ ○ ○ ○ ○
70648-26-9 Furan: 6F 123478 k

○ ○ ○ ○ ○ ○ ○ ○
57117-44-9 Furan: 6F 123678k ○ ○ ○ ○ ○ ○ ○ ○

(continued)
1-7

Process Vents

Catalytic Cracking
Startup/Shutdown
Fluid Coking Unit
Equipment Leaks
Vacuum Systems
Product Loading
Sulfur Recovery
Cooling Towers
Delayed Coking
Hydrogen Plant
Coke Calcining
Reforming Unit
Storage Tanks
Asphalt Plant
Malfunctions
Wastewater
Blowdown
CAS
Catalytic
Systems
Number
Plants
Flares
or Pollutant
Unit
Unit
Code Substance
72918-21-9 Furan: 6F 123789k ○ ○ ○ ○ ○ ○ ○ ○
60851-34-5 Furan: 6F 234678k ○ ○ ○ ○ ○ ○ ○ ○
67562-39-4 Furan: 7F 1234678 k

○ ○ ○ ○ ○ ○ ○ ○
55673-89-7 Furan: 7F 1234789k ○ ○ ○ ○ ○ ○ ○ ○
39001-02-0 Furan: 8F k
○ ○ ○ ○ ○ ○ ○ ○
Polychlorinated biphenyls
1336-36-3 ● ● ● ● ● ○ ○ ○
(total)
Metal HAPs
7440-36-0 Antimony ● ● ● ○ ○ ○ ○
7440-38-2 Arsenic ● ● ● ○ ○ ○ ○
7440-41-7 Beryllium ● ● ● ○ ○ ○ ○
7440-43-9 Cadmium ● ● ● ○ ○ ○ ○
18540-29-9 Chromium (hexavalent) ● ● ● ○ ○ ○ ○
7440-47-3 Chromium (total) ● ● ● ○ ○ ○ ○
7440-48-4 Cobalt ● ● ● ○ ○ ○ ○
7439-92-1 Lead ● ● ● ○ ○ ○ ○
7439-96-5 Manganese ● ● ● ○ ○ ○ ○
7439-97-6 Mercury ● ● ● ○ ○ ○ ○
7440-02-0 Nickel ● ● ● ○ ○ ○ ○
7782-49-2 Selenium ● ● ● ○ ○ ○ ○
Other Inorganic HAPs
75-15-0 Carbon disulfide ○ ○ ○ ○ ● ○ ○ ○ ○ ○ ○ ○ ○ ○
463-58-1 Carbonyl sulfide ○ ○ ○ ○ ● ○ ○ ○ ○ ○ ○ ○ ○ ○
7782-50-5 Chlorine ● ●
7647-01-0 Hydrogen chloride ○ ○ ○ ○ ● ○ ○ ○ ○ ○ ○ ○ ○ ○
Hydrogen cyanide (&

74-90-8 ● ● ● ● ○ ○ ● ● ● ○ ○ ● ○ ○ ○ ● ●
cyanide compounds)
7664-39-3 Hydrogen fluoride ○ ○ ○

(continued)
1-8

Process Vents

Catalytic Cracking
Startup/Shutdown
Fluid Coking Unit
Equipment Leaks
Vacuum Systems
Product Loading
Sulfur Recovery
Cooling Towers
Delayed Coking
Hydrogen Plant
Coke Calcining
Reforming Unit
Storage Tanks
Asphalt Plant
Malfunctions
Wastewater
Blowdown
CAS
Catalytic
Systems
Number
Plants
Flares
or Pollutant
Unit
Unit
Code Substance
7723-14-0 Phosphorus ○ ○ ○
Other Compounds of Interest
7664-41-7 Ammonia ○ ○ ○ ● ● ○ ○ ○
74-84-0 Ethane ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
7783-06-4 Hydrogen sulfide ● ○ ○ ○ ○ ● ● ○ ○ ○ ● ○ ○ ○
7440-39-3 Barium ○ ○ ○ ○
7440-50-8 Copper ○ ○ ○ ○
7439-98-7 Molybdenum ○ ○ ○ ○
7440-62-2 Vanadium ○ ● ● ○
7440-66-6 Zinc ○ ○ ○ ○
● Designates compound/source pairings for which emission estimates should be developed.

○ Designates compound/source pairings for which emission estimates may be developed depending on the available data.
POM
Designates compounds that meet the HAP definition of polycyclic organic matter (POM).
j
The listed HAP is 2,3,7,8-Tetrachlorodibenzo-p-dioxin (2378-TCDD); other dioxin isomers are listed because they can be
used to calculate a 2378 TCDD toxicity equivalence. Abbreviations used in the table are as follows: 4D 2378 = 2,3,7,8-
Tetrachlorodibenzo-p-dioxin; 5D 12378 = 1,2,3,7,8-Pentachlorodibenzo-p-dioxin; 6D 123478 = 1,2,3,4,7,8-
Hexachlorodibenzo-p-dioxin; 6D 123678 = 1,2,3,6,7,8-Hexachlorodibenzo-p-dioxin; 6D 123789 = 1,2,3,7,8,9-
Hexachlorodibenzo-p-dioxin; 7D 1234678 = 1,2,3,4,6,7,8-Heptachlorodibenzo-p-dioxin; 8D = Octachlorodibenzo-p-dioxin.
k
The listed HAP is dibenzofurans. Abbreviations used in the table are as follows: 4F 2378 = 2,3,7,8-Tetrachlorodibenzofuran;
5F 12378 = 1,2,3,7,8-Pentachlorodibenzofuran; 5F 23478 = 2,3,4,7,8-Pentachlorodibenzofuran; 6F 123478 = 1,2,3,4,7,8-
Hexachlorodibenzofuran; 6F 123678 = 1,2,3,6,7,8-Hexachlorodibenzofuran; 6F 123789 = 1,2,3,7,8,9-
Hexachlorodibenzofuran; 6F 234678 = 2,3,4,6,7,8-Hexachlorodibenzofuran; 7F 1234678 = 1,2,3,4,6,7,8-
Heptachlorodibenzofuran; 7F 1234789 = 1,2,3,4,7,8,9-Heptachlorodibenzofuran; 8F = Octachlorodibenzofuran
Carbon dioxide, methane (CH4), and nitrous oxide are GHGs expected to be emitted from petroleum
refineries, but are not listed in this table. The Mandatory Greenhouse Gas Emissions Reporting Rule (74
FR 56260) requires detailed GHG emission reporting from a variety of industry sectors and emission
sources, including petroleum refineries and stationary combustion sources (40 Code of Federal
Regulations [CFR] Part 98, Subparts Y and C, respectively). Consequently, this Refinery Emissions
Protocol document focuses primarily on criteria and toxic air pollutants (i.e., the pollutants listed in
Table 1-1). Emission estimates for GHG will be calculated and reported according to the methodologies
and requirements in the GHG reporting rule. Note that the “tiers” used in 40 CFR Part 98 Subpart C are
listed in opposite order from the “ranks” used in this Refinery Emissions Protocol document. That is, the
Tier 4 method for stationary combustion sources is equivalent to Methodology Rank 1 for combustion
sources in this protocol document; the Tier 3 method is equivalent to Methodology Rank 2, and so on.
While Table 1-1 is intended to provide a comprehensive list of pollutants for each emission source for
which emission estimates should be provided, there may be pollutants released from some sources for
1-9
which we have little or no information. If there is credible information that emissions of other pollutants
are released from a given emission source (e.g., from a source test that was conducted on a particular
process unit), then emission estimates for these additional pollutants should also be provided for that
process unit, even if Table 1-1 does not include a bullet for that pollutant/emission source combination.
1.2 Data Quality

The consistent use of standardized methods and procedures is essential in the compilation of reliable
emission inventories. Quality assurance (QA) and quality control (QC) of an emission inventory are
accomplished through a set of procedures that ensure the quality and reliability of data collection and
analysis. These procedures include the use of appropriate emission estimation techniques, applicable and
reasonable assumptions, accuracy/logic checks of computer models, and checks of calculations and data
reliability. Depending upon the technical approach used to estimate emissions, a checklist with all of the
particular data needs should be prepared to verify that each piece of information is used accurately and
appropriately.
Appropriate metadata (data about the data) should be maintained to assist data users with assessing the
accuracy of the reported emissions. QA/QC and other metadata records should also be maintained to
allow verification of the reported emissions, although this information does not need to be reported unless
specifically requested. For measured emissions, these metadata include manufacturer’s design
specifications for accuracy, initial calibrations, periodic calibration checks, and other QA/QC procedures
used to ensure the accuracy of the measurement device(s). For source tests used to develop site-specific
emission factors, the metadata include results of field and laboratory blanks, duplicate analyses, method
detection limits, isokinetic and cyclonic flow checks (if applicable), and key process operating data (e.g.,
throughput, temperature, material processed). For some pollutants, there may be different methods by
which the emissions can be determined. For example, VOC emissions may be determined using a “total
organics” method (e.g., using EPA Method 25, 25A through 25E, or 305) and subtracting any non-VOCs
present or by speciating individual VOCs and summing the emissions of these compounds to determine
the overall VOC emissions. When reporting VOC emissions, therefore, it should be clearly indicated how
the emissions were determined. If the emissions are determined as TOC or from a TOC measurement, it
must be indicated how the emission are being reported, i.e., “as methane” (or “as” whatever compound
was used to calibrate the total organic analyzer). These metadata assist users of the inventory data and
help to ensure that the inventory data are correctly used when performing subsequent analyses.
1.3 Calculations and Significant Digits

The methodology ranking presented in this Refinery Emissions Protocol document is designed to
highlight and promote those methods that are expected to yield the most accurate emission data. We
recognize that the Methodology Rank 5 methodologies may only provide emission estimates that are
within a factor of 2 or 3 from the actual emission rate. Nonetheless, the emission factors presented in this
document are generally presented with two significant digits. The two significant digits should not be
construed as an expectation that these emission factors are more accurate. The emission factors are
provided with two significant digits because it is recommended that all calculations be performed carrying
at least one additional significant digit to minimize round-off errors. The emissions calculated using
default emission factors may be rounded to one significant digit when reporting the emissions, but at least
two significant digits should be carried in the calculations. For methodologies that may have uncertainties
in the range of ±10 percent, at least three significant digits should be carried when performing the
calculations, even though the final emission estimate may only warrant two significant digits.
1-10
Version 3 Section 2—Equipment Leaks
2. Equipment Leaks
Equipment leaks are small emission sources that occur throughout the process area of the refinery from
various equipment components and connections that develop leaks that allow process fluids to escape into
the atmosphere. Leaks are typically identified using EPA Method 21 (via an organic vapor analyzer
[OVA]) or using optical leak imaging techniques; other remote sensing techniques can also be used to
identify leaks. Although direct measurement methods provide the most accurate means of quantifying
equipment leak emissions, few, if any, refineries have or will implement direct measurement of
equipment leak emissions. Instead, mass emissions for several types of equipment can be estimated using
correlation equations that relate mass emissions to leak concentrations that are obtained using an OVA. In
the absence of concentration measurements, mass emissions can also be estimated using the number of
equipment components and emission factors. Typically, these procedures estimate either total organic
compound (TOC) or non-methane organic compound mass emissions. To estimate either total VOC or
constituent-specific emissions, the process streams being monitored must be characterized at least to the
point of identifying the typical VOC concentration.
The most common optical leak imaging technique at this time uses passive infrared spectral imaging at a
wavelength that is strongly absorbed by the gas of interest to produce a real-time video image of the
emission plume. Although this technology is very useful for quickly and easily identifying the presence of
leaks, particularly large leaks, it has not yet been developed to the point of being able to quantify
emissions. Therefore, it must be combined with other techniques as described below to quantify
emissions. Other remote sensing techniques include Differential Absorption Light Detection and Ranging
(LIDAR) and Solar Occultation Flux. These techniques measure either the volumetric or mass
concentrations of a compound or mixture of compounds in a vertical cut through a plume. Combining
these data with wind speed can be used to estimate mass flux. However, these remote sensing techniques
are not yet approved by EPA as a method of quantifying emissions from equipment leaks or any other
sources. Furthermore, because the measurement is conducted some distance downwind from a source, the
techniques alone are not practical for identifying specific leaking equipment components. When
appropriate, we will update this document to include methodologies for any optical leak imaging or other
remote sensing techniques that develop to the point of being able to quantify equipment leak emissions.
Table 2-1 summarizes the hierarchy of equipment leak emission estimation techniques. The methods are
ranked in terms of anticipated accuracy. Within a given measurement method (or rank), there may be
alternative methods for determining the constituent-specific emissions; these compositional analysis
methods are also provided in order of accuracy. It is anticipated that each refinery will use a mixture of
different methods. For example, Methodology Rank 2a for equipment leaks may be used for certain
components and Methodology Rank 2b for equipment leaks may be used for other components that are
monitored using Method 21, depending on the availability of equipment-specific or process-specific
concentration profiles for a given component or group of components. Additionally, Methodology Ranks
4 or 5 for equipment leaks may be used to estimate emissions from other components that are not
routinely monitored.
The remainder of this section provides additional details and guidance regarding the ways to implement
these methods. Most of the methods outlined in this section are based on the revised equipment leak
protocol developed specifically for the petroleum refinery industry. For more information regarding the
way in which the correlations were developed, please refer to EPA’s document, Protocol for Equipment
Leak Emission Estimates (U.S. EPA, 1995b).
For heavy liquid leaks (e.g., fuel oil, heavy gas oil, residual fuel oil, bitumen) that create a pool or puddle
of liquid, emissions from the accumulated liquid pool should be estimated using the methods for spills in
Section 12, Malfunctions, of this Refinery Emissions Protocol document in addition to using the methods
2-1
presented in this section to estimate the emissions that occur at the leaking component. The spill
methodology is needed in these cases for several reasons. First, for some refinery heavy liquids, the low
volatility of heavy liquids will result in a relatively low OVA reading using EPA Method 21, even with
large visible leaks, so that the equipment leak methodologies are expected to understate the potential
emissions from the pooled material. Second, emissions from the pooled material will be dominated by the
more volatile components of the heavy liquid, and the spill methodology will more accurately assess the
speciated emissions from the pooled material. Finally, the spill methodology will better account for the
dimensions and duration of the liquid pool, which can be affected by the clean-up measures used.
Table 2-1. Summary of Equipment Leak Emission Estimates

Correlation Equations
Rank Measurement Method or Emission Factor Compositional Analysis Dataa
1 Direct measurement (high-volume Not necessary Speciation of collected gas samples
sampler or bagging)
2 EPA Method 21 Correlation equation a) Process-specific, equipment-
specific concentrations
3 EPA Method 21 Default screening
ranges factors b) Process-specific average
concentrations
c) Refinery average stream
concentrations
4 No monitoring; facility-specific Default average a) Process-specific, service-specific
component counts emission factors concentrations
5 No monitoring; default model Default average b) Process-specific average
process component counts emission factors concentrations
c) Refinery average stream
concentrations
d) Default process compositions
a The letters represent ranking sublevels. For example, rank 2a consists of using the correlation equation to
estimate total VOC emissions and using process-specific and equipment-specific process fluid concentration
data to estimate speciated emissions.
2.1 Methodology Rank 1 for Equipment Leaks

There are two primary quantitative leak measurement methods: the bagging method and high-volume
sampling. Typically, EPA Method 21 would be used to initially screen and identify leaking components,
and then one of these methods would be used to quantify the mass emission rate of the leak. Direct leak
rate measurement, using either of these techniques, is accurate within ±15 percent (U.S. EPA, 2003).
In the bagging method, the leaking component or leak opening is enclosed in a “bag” or tent. An inert
carrier gas (e.g., nitrogen) is conveyed through the bag at a known flow rate. Once the carrier gas attains
equilibrium, a gas sample is collected from the bag, and the TOC concentration of the sample is
measured. That collected gas can also be analyzed for individual compound concentrations. The mass
emission rate is calculated from the measured concentrations of the bag sample and the flow rate of the
carrier gas. Although bagging techniques are useful for the direct measurement of larger leaks, bagging
may not be possible for equipment components that are inaccessible, unusually shaped, or very large, and
it is a relatively slow process (i.e., only two or three samples per hour).
High-volume samplers are essentially vacuums that capture all of the emissions from a leaking
component to accurately quantify leak emission rates. Leak emissions and a large-volume sample of the
air around the leaking component are pulled into the instrument through a vacuum sampling hose. High-
volume samplers are equipped with dual hydrocarbon detectors, which measure the concentration of
hydrocarbon gas in the captured sample and the concentration of ambient hydrocarbon gas. Sample
2-2
measurements are corrected for the ambient hydrocarbon concentration, and a mass leak rate is calculated
by multiplying the flow rate of the measured sample by the difference between the ambient gas
concentration and the gas concentration in the measured sample. Chemical speciation of the leak, using
vacuum canisters or similar methods, can be used to apportion the TOC emission rate to individual
constituents. High-volume samplers measure leak rates up to 0.23 cubic meters per minute (m3/min),
which is a rate equivalent to 330 cubic meters per day (m3/day), and can be used to quantify 10 to 20
sources per hour. Leak rates greater than 0.23 m3/min must be measured using bagging techniques (U.S.
EPA, 2003).
While bagging or high-volume sampling are more accurate than other equipment leak emission estimate
methods, they are time consuming and impractical for routine screening of the large number of equipment
components present at a refinery. However, some directed inspection and maintenance programs use, for
example, optical imaging techniques to identify leaking components and then use high-volume sampling
to quantitate the limited number of leaks identified. While these methods are more common at oil and gas
production operations, they could also be applied at some refineries for certain sources. As the emission
rate measured by the high-volume sampler (or bagging method) will be more accurate for that specific
leak than emission estimates developed using any of the lower-ranked methodologies, the high-volume
sampling (or bagging) results should be used for that specific leak if high-volume sampling (or bagging)
is conducted.

Most leak detection and repair (LDAR) programs require periodic monitoring using EPA Method 21 to
identify leaking components. The preferred methodology for estimating equipment leak emission rates
directly from EPA Method 21 data (i.e., Methodology Rank 2 for equipment leaks) is to use the screening
value correlations in the right column of Table 2-2 (U.S. EPA, 1995b) for each individual component as
screened via EPA Method 21. When a screening value of zero is registered, the default zero value in the
second column in Table 2-2 is used to estimate TOC emissions. If the monitoring instrument measures
concentrations only up to 10,000 parts per million by volume (ppmv) or 100,000 ppmv, then the
applicable values for pegged emission rates in Table 2-2 are used to estimate emissions. Table 2-2
includes screening value correlations for both the petroleum industry (applicable for petroleum refinery
operations, marketing terminals, and oil and gas production) and the synthetic organic chemical
manufacturing industry (SOCMI). Generally, most petroleum refineries will use the petroleum industry
correlations; however, the SOCMI correlations are provided here for convenience because some refineries
may also have chemical manufacturing processes that must be included in the facility’s emission
inventory.
Many refineries or monitoring specialists use software programs that directly record the TOC reading (or
screening value) for each component. Many of these software programs will directly calculate the TOC
emissions for each component using the screening value correlations and will even calculate component-
specific emission rates when composition data are entered for the components. The TOC rates calculated
by the correlation equation (and the default zero and pegged emission rate values) include non-VOC
organic compounds, primarily methane and ethane. The uncertainty of the correlations for any single
measurement may be as much as a factor of 3 higher or a factor of 10 lower than the actual emissions for
that component, but when summed over thousands of components, the uncertainty in the cumulative total
emissions is expected to be much less. For example, based on Monte Carlo simulations of 100 leaking
components and using an uncertainty of plus or minus a factor of 10, the uncertainty in the cumulative
emissions is approximately plus or minus a factor of 1.4. The advantage of Methodology Rank 2 for
equipment leaks is that it is not based on a presupposed distribution of equipment leaks, as are the lower
ranked methodologies. Consequently, Methodology Rank 2 for equipment leaks is much more accurate
than any of the lower ranked equipment leak emission estimation methods.
2-3
Table 2-2. Equipment Leak Rate for Petroleum and SOCMI Equipment Componentsa
Pegged Emission Rates

Default Zero (kg/hr/source)
Equipment Type Emission Rate Correlation Equationb
(All Services) (kg/hr/source) 10,000 ppmv 100,000 ppmv (kg/hr/source)
Leak Rates for Petroleum Industry (Refinery, Marketing Terminals, and Oil and Gas Production)
Valve 7.8E-06 0.064 0.14 2.29E-06×SV0.746
Pump 2.4E-05 0.074 0.16 5.03E-05×SV0.610
Otherc 4.0E-06 0.073 0.11 1.36E-05×SV0.589
Connector 7.5E-06 0.028 0.030 1.53E-06×SV0.735
Flange 3.1E-07 0.085 0.084 4.61E-06×SV0.703
Open-ended line 2.0E-06 0.030 0.079 2.20E-06×SV0.704
Leak Rates for Synthetic Organic Chemical Manufacturing Industry (SOCMI)
Gas valve 6.6E-07 0.024 0.11 1.87E-06×SV0.873
Light liquid valve 4.9E-07 0.036 0.15 6.41E-06×SV0.797
Light liquid pumpd 7.5E-06 0.14 0.62 1.90E-05×SV0.824
Connector 6.1E-07 0.044 0.22 3.05E-06×SV0.885
Note: kg/hr/source = kilograms TOC per hour per source
a Data reported in U.S. EPA, 1995b.
b SV is the screening value (SV, ppmv) measured by the monitoring device.
c The “other” equipment type was developed from instruments, loading arms, pressure relief devices, stuffing
boxes, vents, compressors, dump lever arms, diaphragms, drains, hatches, meters, and polished rods. This
“other” equipment type should be applied to any equipment other than connectors, flanges, open-ended lines,
pumps, or valves.
d The light liquid pump factors can also be applied to compressors, pressure relief valves, agitators, and heavy
liquid pumps.
When an optical gas imaging camera is used to identify leaks, the emissions can be quantified under
Methodology Rank 2 (or Methodology Rank 1 for certain components) only if both of the following
conditions are met: (1) the equipment is monitored in accordance with the procedures in §63.11(e), and
(2) all leaks identified by the camera are monitored before repair using Method 21 or are measured using
high-volume sampling or bagging methods. For equipment found to be leaking when monitoring with the
camera, either the subsequent Method 21 screening values in the applicable correlation equation should
be used to estimate the emissions (Methodology Rank 2) or the leak rake rates measured via the high-
volume sampling or bagging method (Methodology Rank 1) should be used, as applicable. For all
equipment not found to be leaking when monitoring with the camera, emissions should be estimated
using the screening values obtained when conducting the annual Method 21 monitoring required by
§63.11(d)(7) in the applicable correlation equations.
2-4
Example 2-1: Calculation for Methodology Rank 2 for Equipment Leaks

A refinery catalytic reforming unit (CRU) operating 8,000 hours per year (hr/yr) has 600
valves. To keep the example simple, assume Method 21 monitoring registers the screening
value readings in the following table, and assume the average weight percents of methane and
ethane in all streams are known or estimated to be equal to 3 percent and 1 percent of the
TOC, respectively. Also assume the TOC content of each stream is 100 percent. Using
Methodology Rank 2b for equipment leaks (correlation approach), what is the cumulative
hourly VOC emission rate from the valves in this process unit at the time the monitoring is
conducted?
To calculate the emissions, the default zero value for valves on Table 2-2 (7.8E-06) is used to
estimate the TOC emissions from the 580 valves with a screening value of 0 ppmv. The
pegged emission rate for valves in Table 2-2 (0.140) is used to estimate the TOC emission
rate for the two valves with pegged readings. The correlation equation for valves in Table 2-2
(2.29E-06 x SV^0.746) is used to estimate the emissions for each of the valves with a
measured screening value. In each case, the calculated TOC emissions are multiplied by (100-
4)/100 to calculate the VOC emissions.
Emissions, kg/hr
Method 21 Screening
Number of Valves Value, ppmv TOC VOC
580 0 0.00452 0.00434
5 200 0.00012 0.00011
5 400 0.00020 0.00019
2 1,500 0.00054 0.00051
2 7,000 0.00169 0.00162
2 20,000 0.00370 0.00355
2 50,000 0.00733 0.00704
2 Pegged at 100,000 0.28000 0.26880
Total 0.30 0.29
2.2.1 Speciating Equipment Leak Emissions

In developing constituent-specific emission estimates, the composition of the process stream in contact
with the equipment is used to estimate the speciated equipment leak emissions. To the extent that
compositional data are available for individual process streams, each equipment component associated
with that process stream should be tagged with the average composition of that process stream
(Methodology Rank 2a for equipment leaks). Although an ideal situation would be if chemical speciation
data were available for each equipment component (i.e., each process stream associated with each
equipment component), in many cases, this level of detail may not be available. Alternatively, average
compositional data may be determined for all streams in a specific process unit, for groups of streams in
different portions of a process unit, or for groups of streams in a particular service type in a process unit.
Under this approach, all equipment associated with the applicable group of streams would be tagged with
the same concentration profile (Methodology Rank 2b for equipment leaks). For example, an average
composition could be determined for all gas streams, all light liquid streams, and all heavy liquid streams
at a given refinery process unit (resulting in three composition profiles per unit). Alternatively, an overall
average composition could be determined for all streams (regardless of the type of service) for that
process unit (resulting in one composition profile per unit). As a last resort, a single, overall average
2-5
process stream composition can be estimated on a refinery-wide basis (e.g., one concentration profile for
the entire refinery, which is Methodology Rank 2c for equipment leaks). Methodology Rank 2c for
equipment leak emissions is very easy to implement in that the TOC emissions can be aggregated for all
components first, and then the chemical-specific emissions can be calculated from the cumulative TOC
emissions. However, this method greatly reduces the accuracy of the chemical-specific emission rates.
With the automated software programs that are now available for logging equipment leak readings and
calculating equipment leak emissions, most refineries should be able to implement Methodology Rank 2a
or 2b for equipment leaks.
The correlation equations for equipment leaks provided in Table 2-2 provide emissions in terms of TOC
(including methane and ethane). To calculate the emission rate for VOC (i.e., to exclude methane and
ethane), use Equation 2-1.
EVOC = ETOC × (WFVOC/WFTOC) (Eq. 2-1)
where:
EVOC = Emission rate of VOC for a specific type of equipment (kilograms per year [kg/yr])
ETOC = Emission rate of TOC for a specific type of equipment (kg/yr)
WFVOC = Average weight fraction of VOC in the stream (typically TOC minus methane and
ethane)
WFTOC = Average weight fraction of TOC in the stream.
Table 1-1 in Section 1, Introduction, lists the specific organic compounds that should be included in an
inventory of equipment leak emissions. Either of the following equations (Equation 2-2a or Equation
2-2b) is used to speciate emissions for specific organic compounds from a single equipment piece:
Ei = ETOC × (WFi/WFTOC) (Eq. 2-2a)

Ei = EVOC × (WFi/WFVOC) (Eq. 2-2b)
where:
Ei = Mass emissions of organic chemical “i” from the equipment (kg/yr)
WFi = Concentration of organic chemical “i” in the equipment (weight fraction).
2-6
Example 2-2: Calculation for Speciating Equipment Leak Emissions

For Example 2-1, the measured composition of the process stream associated with one of the
valves (Methodology Rank 2a) for which a screening value of 7,000 ppmv was obtained is:
Hexane 10 wt%
Toluene 8 wt%
Benzene 2 wt%
Other volatile organic compounds (VOCs) 60 wt%
Methane and ethane 4 wt%
Nitrogen 10 wt%
Water 5 wt%
Hydrogen 1 wt%
The TOC weight fraction is calculated as the sum of all of the organic compounds. The VOC
weight fraction is the sum of all of the organic compounds, minus methane and ethane, as
shown in the following equation:
WFVOC = (%Hexane + %Toluene + %Benzene +% Other VOC)/100%
= (10 + 8 + 2 + 60)/100 = 0.80
From Example 2-1, the valve’s VOC emission rate was 0.81 grams per hour (g/hr).
Equation 2-1 is used to attribute these emissions to individual components as follows:
EHexane = 0.81 × (0.1/0.8) = 0.10 g/hr
EToluene = 0.81 × (0.08/0.8) = 0.08 g/hr
EBenzene = 0.81 × (0.02/0.8) = 0.02 g/hr
EOtherVOC = 0.81 × (0.60/0.8) = 0.61 g/hr
2.2.2 Calculating Hourly and Annual Equipment Leak Emissions

The emissions estimated based on EPA Method 21 measurement data (i.e., Methodology Rank 2 for
equipment leaks) represent the emission rate at the time when the measurements were made (i.e., hourly
emission estimates). Thus, the direct emission rate calculated based on the monitored screening values
should be used for the hourly emission estimate. If the components are monitored multiple times per year,
the hourly emissions for each process unit should be calculated for each monitoring period (as the
summation of the emissions of all components for that process unit), and the monitoring period resulting
in the highest overall emission rate is to be reported as the hourly emission rate for that unit. There may
be components that are monitored at different frequencies; for example, pumps may be monitored
monthly, while valves are monitored either quarterly or semi-annually. In general, the hourly emissions
for the process unit should only be calculated for the periods where a significant number of the
components are monitored (in the example, quarterly or semi-annually). One should not take the highest
hourly emission rate for each individual component, regardless of when it was monitored during the year,
and then sum the maximum value for the individual components because this will tend to overstate the
actual hourly emission rate from the process unit.
The first time an LDAR program is implemented, the emission estimates from the component screening
measurements should be used as the emission rate for the facility (or an individual component or set of
components) for all periods prior to the screening measurements (i.e., the portions of the inventory year
prior to the screening measurements). However, most equipment components at a refinery are expected to
be monitored for leaks on a routine basis (i.e., monthly, quarterly, or semiannually) as part of an ongoing
LDAR program. Leaks greater than a certain threshold are required to be repaired within certain time
frames (although many LDAR programs allow some repairs to be delayed). After repairs are made, the
2-7
repaired components should be re-screened to verify that the leak has been repaired and to provide a new
starting emission rate for that component. During the next routine monitoring period, new screening
values will be determined for each component, providing new instantaneous emission rate estimates for
each component. The variations in component emissions during the year need to be accounted for when
developing annual emission estimates for equipment components that are routinely monitored. Figure 2-1
illustrates the three acceptable methods for estimating annual emissions from routinely (monthly,
quarterly, semiannually) monitored equipment components.
Figure 2-1. Illustration of alternative methods to determine equipment leak emissions

from routine monitoring data.
The mid-period method assumes that the initial reading represents the emission rate for the first half of
the period between monitoring events and that the subsequent reading represents the emission rate for the
second half of the period between monitoring events. When a leak is detected for subsequent repair, the
“leak” monitor reading is used from the time the leak is detected to the time it is repaired (i.e., re-
screened). The modified trapezoid method assumes that the mass leak rate changes linearly between any
two monitoring points, except for periods between leak detection and repair; the leak monitor reading is
used from the time the leak is detected to the time it is repaired (i.e., re-screened). The average period
method uses the arithmetic average emission rate of two adjacent instantaneous emission rate estimates
2-8
(except for periods between leak detection and repair) for the emission rates between the measurement
periods. As in the mid-period and modified trapezoid method, the average period method uses the leak
monitor reading from the time the leak is detected to the time it is repaired (i.e., re-screened). In all of
these alternatives, it is important that the emission rate be determined for each component and each
monitoring event before extrapolating the emissions to the intervening periods. It is invalid to average the
monitored screening values first and then to calculate the emission rate based on the average screening
value.
Mathematically, the calculated emissions for periods wholly within the inventory year will be identical
regardless of the alternative selected. Provided that one can elect the “inventory year period” to start and
end with a monitoring event, then the method selected is immaterial. The only difference in the annual
emissions determined using the three alternative methods for routinely monitored components will occur
if emissions must be determined for a specific time period (e.g., a calendar year), and the emissions
determined for an intervening interval (between monitoring events at the end of one year and start of the
next year) must be parsed between the two years. While there can be differences in the annual emissions
calculated using these different methods for a particular component, these differences will tend to cancel
out when emissions are summed over a large number of components. Therefore, it is only important that
there is consistency in the application of the selected method (i.e., all components use the mid-period
method or all components use the modified trapezoid method or all components use the average period
method).
If the equipment is taken out of service (e.g., no process fluid is in the piping at the location of the
specific component), the emission rate for that component can be assumed to be zero for the time period
the equipment in out of service. If the process unit is not operating, but fluid remains in the components,
then no correction for operating hours should be made. There can be some differences in the emissions
calculated for the operating periods between monitoring intervals when the components are out of service,
but these differences are expected to be small. Again, consistency in the application of the selected
method is of key importance.
When emission inventories are required for a set calendar year, there are also some practical matters to
consider when selecting an annual estimation method for components, especially when components are
monitored semiannually or less often. The midpoint method has the advantage of not relying on the
subsequent year’s first semiannual monitoring if the last semiannual monitoring event for the inventory
year occurred in October or later. If the last semiannual monitoring event for the inventory year occurred
prior to October, then the subsequent year’s first semiannual monitoring would likely occur in March or
earlier, so that the inventory could be developed in a timely fashion, even if the subsequent year’s
monitoring data are needed to complete the inventory for the current year. Similarly, for annual
component monitoring, the midpoint method would not require the subsequent year’s monitoring results
if monitoring occurred in July or later. On the other hand, the application of either the modified trapezoid
method or the average period method requires the subsequent year’s monitoring data, regardless of when
that monitoring event occurs; the modified trapezoid method is further complicated in that interpolation is
required to correctly account for emissions between two inventory years. If allowed, an “equipment leak
year” (like a fiscal year) could be established based on the typical timing of the monitoring events, so that
the emissions for that equipment leak year can be determined based entirely on whole monitoring periods.
2-9
Example 2-3: Calculation for Annual Emissions Using the Midpoint Method
The following monitoring data for a pump were recorded (Columns 1 and 2). The correlation
equation from Table 2-2 for pumps is used (i.e., 5.03E-05×SV0.610) to calculate the emissions
rate in Column 3 when the OVA reading is greater than zero, and the default zero rate is used
when the OVA reading is zero. The hours between intervals were calculated; half of the hours
were assigned to the first reading, and half were assigned to the second reading, except for
periods before repair, which were assigned the “leak” emission rate.
3
2 TOC Emission 5
1 OVA Reading Rate 4 TOC Emissions
Date and Time (ppm) (kg/hr) Hoursa (kg/period)
January 2, 8:00 a.m. 200 1.27E-3 32+372 0.51
February 2, 8:00 a.m. 300 1.63E-3 372+336 1.15
March 2, 8:00 a.m. 280 1.56E-3 336+372 1.10
April 2, 8:00 a.m. 22,000 2.24E-2 372+74 9.99
April 5, 10:00 a.m.b 150 1.07E-3 323 0.35
May 2, 8:00 a.m. 140 1.02E-3 323+372 0.71
June 2, 8:00 a.m. 200 1.27E-3 372+360 0.93
July 2, 8:00 a.m. 180 1.19E-3 360+372 0.87
August 2, 8:00 a.m. 500 2.23E-3 372+372 1.66
September 2, 8:00 a.m. 45,000 3.47E-2 372+241 21.27
September 12, 9:00 a.m.b 0 2.4E-5 239.5 0.006
October 2, 8:00 a.m. 0 2.4E-5 239.5+372 0.015
November 2, 8:00 a.m. 0 2.4E-5 372+360 0.018
December 2, 8:00 a.m. 200 1.27E-3 360+372 0.93
January 2, 8:00 a.m. 250 1.46E-3 372-32 0.50
Annual Totals 8,760 40.0 kg/yr
a Each emission rate applies over half of the hours since the previous screening and over half of the hours
until the next screening. For example, the results from the July 2 measurement apply to half of the hours
since the June 2 measurement and to half of the hours before the next measurement on August 2. The
July 2 measurements occur 30 days (720 hours) after the June 2 measurement and 31 days (744 hours)
before the August 2 measurements. Therefore, the July 2 results apply to 732 hours during the year
(720/2 + 744/2 = 732).
b Indicates special Method 21 measurement reading to verify repair.

In some older LDAR programs, the only information that is recorded is whether a leak was found. For
these LDAR programs, leaks were defined as screening value readings greater than 10,000 ppmv. In these
cases, generally the available data are the number of components of each type at the plant and the number
of components of each type found with TOC readings less than 10,000 ppmv and the number of
components of each type with TOC readings greater than or equal to 10,000 ppmv. Methodology Rank 3
for equipment leaks estimates the emission rate using these data and the screening ranges emission factors
in AP-42 (U.S. EPA, 1995b). Table 2-3 provides the screening ranges emission factors for refinery and
SOCMI sources; see AP-42 for additional screening ranges emission factors for marketing terminals or oil
and gas production operations if these sources are also part of the facility. If the number of components
and the number of leakers are recorded by process, then this provides a more accurate method for using
2-10
process-specific compositional data to estimate constituent-specific emission rates (Methodology Ranks

3a and 3b). Alternatively, the average percentage of leakers facility-wide can be used, and these data can
be coupled with process-specific component counts to estimate emissions by process (Methodology
Rank 3c).
Table 2-3. Screening Ranges Emission Factorsa
Refinery Factorsb SOCMI Factorsc

≥10,000 ppmv <10,000 ppmv ≥10,000 ppmv <10,000 ppmv
emission emission emission emission
factor (kg/hr/ factor (kg/hr/ factor (kg/hr/ factor (kg/hr/
Equipment type Service source)b source)b source)c source)c
Valves Gas 0.2626 0.0006 0.0782 0.000131
Light liquid 0.0852 0.0017 0.0892 0.000165
Heavy liquid 0.00023 0.00023 0.00023 0.00023
Pump and agitator Light liquid 0.437 0.0120 0.243 0.00187
seals Heavy liquid 0.3885 0.0135 0.216 0.0210
Compressor seals All 1.608 0.0894 1.608 0.0894
Pressure relief All 1.691 0.0447 1.691 0.0447
valves
Connectors All 0.0375 0.00006 0.113 0.000081
Open-ended lines All 0.01195 0.00150 0.01195 0.00150
a Data reported in U.S. EPA, 1995b.
b These factors are for non-methane organic compound emissions.
c These factors are for total organic compound emissions.
The uncertainty of the leak rates calculated using the screening ranges emission factors for any single
measurement may be a factor of 10 or more. Although this methodology is intended to account for
reduced emissions gained by an LDAR program, the underlying default leak rates are based on the
distribution of leaks prior to the implementation of an LDAR program. This method is expected to be
biased high for facilities that have implemented an LDAR program, especially an LDAR program with
leak action levels less than 10,000 ppmv. First, the implementation of any LDAR program is expected to
alter not only the prevalence of leaks, but the relative magnitude of leaks above the action level (i.e., the
average component leak rate for components with screening values of 10,000 ppmv or more).
Furthermore, if an LDAR program uses a 1,000 ppmv leak action level, the relative magnitude and
number of leaks with screening values between 1,000 ppmv and 10,000 ppmv is expected to be much less
than for facilities that use a 10,000 ppmv screening level or do not have an LDAR program. The
distribution of leaks is also affected by the monitoring frequency, so components that are monitored
quarterly are expected to have a different leak frequency distribution than components that are monitored
annually. As such, the emissions estimated using Methodology Rank 3 for equipment leaks is expected to
be an upper-range emission estimate of equipment leak emissions. The accuracy of this method is
dependent on the level of LDAR program used; the more frequent the monitoring and the lower the leak
action level is, the more likely this method will be to significantly overestimate emissions (by as much as
a factor of 10 to 50, based on data from Lev-On et al. [2007]).
The leak/no-leak factors presented by Lev-On et al. (2007) were evaluated for facilities that use an optical
gas imaging camera to identify leaks as a potential Methodology Rank 3 for equipment leaks. However,
for the same reasons the screening ranges emission factors presented in Table 2-3 are expected to
overestimate actual emissions for facilities that have implemented a stringent LDAR program, the
leak/no-leak factors presented by Lev-On et al. (2007) are expected to underestimate the emissions from
2-11
facilities that only use an optical gas imaging camera. The leak/no-leak factors presented by Lev-On et al.
(2007) are based on a distribution of leaks for refineries that used quarterly monitoring with a leak
definition of 1,000 ppmv. Most optical gas imaging cameras have leak detection sensitivities much higher
than this level. As such, the distribution of leaks by screening range is expected to be much different than
the baseline distribution used to develop the leak/no-leak factors for optical imaging cameras. Until
leak/no-leak factors are available for a more representative distribution of leaks, there is no acceptable
Methodology Rank 3 for quantifying emissions using an optical gas imaging camera.

The refinery screening ranges emission factors are for non-methane organic compound emissions. To
apply the refinery screening ranges emission factor approach, use Equation 2-3 to calculate the total
annual TOC emissions per type of equipment for which Methodology Rank 3 is used. Note: Equation 2-3
should be applied separately to groups of equipment in streams with significantly different methane
weight fractions. The SOCMI screening ranges emission factors are presented in terms of TOC, so
Equation 2-3 does not apply to the SOCMI factors. For both refinery and SOCMI screening ranges
emission factors, use Equations 2-1 and 2-2a or 2-2b (presented previously in Section 2.2.1) to calculate
the VOC and specific compound emissions, respectively.
  WFTOC     WFTOC  
ETOC =   FG ×   × N G  +  FL ×   × N L   (Eq. 2-3)
  WFTOC − WFm  WFTOC − WFm 
      
where:
ETOC = Emission rate of TOC for a specific type of equipment (kg/hr)
FG = Applicable emission factor for specific type of equipment type with screening values
greater than or equal to 10,000 ppmv (kg/hr/source)
FL = Applicable emission factor for specific type of equipment with screening values less
than 10,000 ppmv (kg/hr/source)
WFTOC = Average weight fraction of TOC in the stream
WFm = Average weight fraction of methane in the stream
NG = Equipment count (specific equipment type) for sources with screening values greater
than or equal to 10,000 ppmv
NL = Equipment count (specific equipment type) for sources with screening values less than
10,000 ppmv.

Typically, facilities that use Methodology Rank 3 have limited data. The instantaneous emission rate
measured during a given monitoring event is used directly as the hourly emission rate for the process unit
at that time. If the facility monitors annually, the instantaneous emission rate calculated for a given
process or facility is used directly as the hourly emission rate for the process unit. The annual emissions
for the process unit are calculated using the hourly emission rate and annual operating hours of the
process equipment as shown in Equation 2-4. If more than one monitoring event occurs during the year,
the instantaneous (hourly) emission rates calculated for each monitoring period are calculated and the
highest value is reported for the hourly emissions. The annual emissions are calculated as the average of
the calculated instantaneous emission rates multiplied by the operating hours of the equipment.
ETOC , annual = ETOC × H (Eq. 2-4)
2-12
where:
ETOC,annual = Annual emission rate of TOC for a specific type of equipment (kilograms per year
[kg/yr])
ETOC = Emission rate of TOC for a specific type of equipment (kg/hr)
H = Operating hours per year (hr/yr).
Example 2-4: Calculation for Methodology Rank 3 for Equipment Leaks

For the valves in the same reforming unit described in Example 2-1, what is the cumulative
hourly VOC emission rate when Methodology Rank 3 is used? What are the annual VOC
emissions? Assume that the distribution of valves in gas, light liquid, and heavy liquid service
is as shown in the table below, and assume that the valves were monitored once during the
year.
To calculate hourly TOC emissions for valves in gas service with screening values less than
10,000 ppmv, use the second half of Equation 2-3.
ETOC = [FL × (WFTOC/(WFTOC – WFm)) × NL]
= [0.0006 × (100/(100-3)) × 236]
= 0.146 kg TOC/hr
Use the same procedure to calculate emissions for valves in light liquid service and heavy
liquid service of 0.514 kilograms of TOC per hour (kg TOC/hr) and 0.015 kg TOC/hr,
respectively. Thus, the total TOC emissions for all valves with screening values less than
10,000 ppmv are 0.675 kg/hr. Use Equation 2-1 to calculate the VOC emissions from these
valves.
EVOC = ETOC × (WFVOC/WFTOC)
= 0.675 × (100-4)/100
= 0.648 kg VOC/hr
Use the same procedure with the first half of Equation 2-3 to estimate TOC emissions of 1.076
kg/hr from the valves with screening values equal to or greater than 10,000 ppmv, and apply
Equation 2-1 to calculate VOC emissions of 1.033 kg/hr. Total annual emissions are
calculated to be 13,400 kg VOC/yr (1.68 kg/hr x 8,000 hr/yr).
Number of Valves by Type Emission Factors,
of Service kg/hr/source
Screening TOC VOC
Value, Light Heavy Light Heavy Emissions, Emissions,
ppmv Gas Liquid Liquid Gas Liquid Liquid kg/hr kg/hr
<10,000 236 293 65 0.0006 0.0017 0.00023 0.675 0.648
≥10,000 3 3 0 0.2626 0.0852 0.00023 1.076 1.033
1.68
2.4 Methodology Ranks 4 and 5 for Equipment Leaks

Methodology Ranks 4 and 5 for equipment leaks should be used only for components that are not being
routinely monitored for leaks. For refineries, this might be processes that have low hazardous air pollutant
(HAP) and low VOC content, or it might be specific types of components that are either classified as
unsafe to monitor (e.g., certain pumps and valves) or are not subject to the monitoring requirements (e.g.,
2-13
connectors) in the new source performance standards (NSPS) or Petroleum Refinery Maximum
Achievable Control Technology (MACT) I rules. Facility- or process-specific component counts
(Methodology Rank 4 for equipment leaks) should be used when these data are available. Methodology
Rank 4 is preferred over default component counts (Methodology Rank 5 for equipment leaks) because
these defaults may not account for facility-specific factors (e.g., control of pressure relief valves).
Default component counts (Methodology Rank 5 for equipment leaks) are dependent on the size of the
refinery and the process units present. The median (default) equipment component counts for small
refineries (those less than 50,000 barrels per calendar day [bbl/cd]) 1 and large refineries (greater than
50,000 bbl/cd) as presented in EPA’s document, Locating and Estimating Air Emissions from Sources of
Benzene (U.S. EPA, 1998a), are presented in Tables 2-4 and 2-5, respectively.
For both Methodology Ranks 4 and 5 for equipment leaks, default emission factors per component for
uncontrolled or unmonitored components must be used. It is assumed that these components are not
monitored because emissions from monitored components should be estimated using a higher ranked
methodology (Methodology Rank 1, 2, or 3 for equipment leaks). The default emission factor per
component for uncontrolled or unmonitored components is provided in EPA’s document, Protocol for
Equipment Leak Emission Estimates (U.S. EPA, 1995b). These emission factors for refinery and SOCMI
components are summarized in Table 2-6.

The emission factors for refinery components in Table 2-6 were developed to estimate emissions of non-
methane organic compounds (NMOCs). Therefore, to estimate VOC emissions, the emission factor must
be multiplied by the ratio of the VOC-to-NMOC weight fractions. Additionally, the emission factors
apply specifically to streams that are 100 percent TOC. To calculate VOC emissions from equipment in
contact with a stream that contains both organic compounds and other compounds, such as nitrogen or
water vapor, the emission factor must be multiplied by the TOC weight fraction in the stream. As a result,
TOC, methane, and ethane weight fractions must be known or estimated in order to calculate the VOC
emissions as shown in Equation 2-5.
 WFTOC − WFethane − WFmethane 

EVOC = FA ×   × WFTOC × N (Eq. 2-5)
 WFTOC − WFmethane 
where:
EVOC = Emission rate of VOC from all equipment in the stream of a given equipment type
(kg/hr)
FA = Applicable average non-methane organic compounds emission factor for the
equipment type (kg/hr/source)
WFTOC = Average weight fraction of TOC in the stream
WFethane = Average weight fraction of ethane in the stream
WFmethane = Average weight fraction of methane in the stream
N = Number of pieces of equipment of the applicable equipment type in the stream.
1
Barrels per calendar day is the amount of crude that a refinery can process under usual operating conditions and is
expressed in terms of actual capacity during a 24-hour period (i.e., actual annual capacity divided by 365 days). The
other typical capacity measure, barrels per stream day (bbls/sd), is the maximum number of barrels of crude that a
refinery can process within a 24-hour period when running at full capacity.
2-14
Table 2-4. Median Equipment Leak Component Counts for Small Model Processesa
Valves Pumps Pressure Relief Valves Flanges

Open-
Light Heavy Light Heavy Light Heavy Light Heavy Ended Sampling
Process Unit Gas Liquid Liquid Liquid Liquid Compressors Gas Liquid Liquid Gas Liquid Liquid Lines Connections
Crude distillation 75 251 216 8 8 2 6 6 5 164 555 454 39 10
Alkylation (sulfuric acid) 278 582 34 18 10 1 12 15 4 705 1,296 785 20 16
Alkylation (HF) 102 402 62 13 3 2 12 13 0 300 1,200 468 26 8
Catalytic reforming 138 234 293 8 5 3 5 3 3 345 566 732 27 6
Hydrocracking 300 375 306 12 9 2 9 4 4 1,038 892 623 25 10
Hydrotreating/hydrorefining 100 208 218 5 5 2 5 3 5 290 456 538 20 6
Catalytic cracking 186 375 450 13 14 2 8 8 7 490 943 938 8 8
Thermal cracking (visbreaking) 206 197 0 7 0 0 4 0 0 515 405 0 0 4
Thermal cracking (coking) 148 174 277 9 8 2 7 16 13 260 322 459 13 8
Hydrogen plant 168 41 0 3 0 2 4 2 0 304 78 0 8 4
Asphalt plant 120 334 250 5 8 2 5 10 9 187 476 900 16 6
Product blending 67 205 202 6 11 1 10 6 22 230 398 341 33 14
Sulfur plant 58 96 127 6 6 3 3 88 15 165 240 345 50 3
Vacuum distillation 54 26 84 6 6 2 2 5 2 105 121 230 16 4
Full-range distillation 157 313 118 7 4 2 5 4 6 171 481 210 20 6
Isomerization 270 352 64 9 2 2 7 10 1 432 971 243 7 8
Polymerization 224 563 15 12 0 1 10 5 3 150 450 27 5 7
MEK dewaxing 145 1,208 200 35 39 3 10 14 4 452 1,486 2,645 19 17
Other lube oil processes 153 242 201 7 5 2 5 5 5 167 307 249 60 6
a Process component counts as presented in EPA’s document, Locating and Estimating Air Emissions from Sources of Benzene (U.S. EPA, 1998a), for refineries with crude capacities less than
50,000 bbl/cd.
2-15
Table 2-5. Median Equipment Leak Component Counts for Large Model Processesa
Valves Pumps Pressure Relief Valves Flanges

Open-
Light Heavy Light Heavy Light Heavy Light Heavy Ended Sampling
Process Unit Gas Liquid Liquid Liquid Liquid Compressors Gas Liquid Liquid Gas Liquid Liquid Lines Connections
Crude distillation 204 440 498 15 14 2 7 5 12 549 982 1,046 75 9
Alkylation (sulfuric acid) 192 597 0 21 0 2 13 4 0 491 1,328 600 35 6
Alkylation (HF) 104 624 128 13 8 1 9 11 1 330 1,300 180 40 14
Catalytic reforming 310 383 84 12 2 3 8 11 0 653 842 132 48 9
Hydrocracking 290 651 308 22 12 2 10 12 0 418 1,361 507 329 28
Hydrotreating/hydrorefining 224 253 200 7 6 2 9 4 8 439 581 481 49 8
Catalytic cracking 277 282 445 12 12 2 11 9 13 593 747 890 59 15
Thermal cracking (visbreaking) 110 246 130 7 6 1 6 3 15 277 563 468 30 7
Thermal cracking (coking) 190 309 250 12 11 1 8 5 10 627 748 791 100 10
Hydrogen plant 301 58 0 7 360 3 4 139 0 162 148 0 59 21
Asphalt plant 76 43 0 4 0 0 3 7 0 90 90 0 24 24
Product blending 75 419 186 10 10 2 9 16 6 227 664 473 24 8
Sulfur plant 100 125 110 8 3 1 4 4 4 280 460 179 22 7
Vacuum distillation 229 108 447 2 12 1 5 1 4 473 136 1,072 0 7
Full-range distillation 160 561 73 14 2 2 7 8 2 562 1,386 288 54 6
Isomerization 164 300 78 9 5 2 15 5 2 300 540 265 36 7
Polymerization 129 351 82 6 2 0 7 12 28 404 575 170 17 9
MEK dewaxing 419 1,075 130 29 10 4 33 6 18 1,676 3,870 468 0 7
Other lube oil processes 109 188 375 5 16 3 8 6 20 180 187 1,260 18 9
a Process component counts as presented in EPA’s document, Locating and Estimating Air Emissions from Sources of Benzene (U.S. EPA, 1998a), for refineries with crude capacities greater than
50,000 bbl/cd.
2-16
Table 2-6. Refinery and SOCMI Average Component Emission Factorsa
SOCMI Emission
Refinery Emission Factor Factor
Equipment Type Service (kg/hr/source)b (kg/hr/source)c
Valves Gas 0.0268 0.00597
Light liquid 0.0109 0.00403
Heavy liquid 0.00023 0.00023
Pump sealsd Light liquid 0.114 0.0199
Heavy liquid 0.021 0.00862
Compressor seals Gas 0.636 0.228
Pressure relief valves Gas 0.16 0.104
Connectors All 0.00025 0.00183
Open-ended lines All 0.0023 0.0017
Sampling connections All 0.0150 0.0150
Note: kg/hr/source = kilograms per hour per source
a Source: U.S. EPA, 1995b.
b The refinery emission factors are for non-methane organic compound emission rates.
c The SOCMI emission factors are for TOC (including methane).
d The light liquid pump seal factor can be used to estimate the leak rate from agitator seals.
If a process stream contains no ethane, the non-methane organic fraction is equal to the VOC fraction,
which means the emission factor can be used without the correction ratio noted above, and the equation
simplifies to Equation 2-6:
EVOC = FA × WFTOC × N (Eq. 2-6)
Two guidelines when correcting the “FA” term when applied to refineries are as follows:
 The correction should only be applied to equipment containing a mixture of organics and
methane.
 The maximum correction for the methane weight fraction should not be greater than 0.10, even if
the equipment contains greater than 10 weight percent of methane. (This reflects the fact that the
equipment for which the average refinery emission factors in Table 2-6 were developed typically
contained 10 weight percent or less of methane).
When using Methodology Ranks 4 and 5, speciated emissions are calculated using the same procedures
described in Section 2.2.1 of this document. Preferably, facility- or process-specific compositional
analyses would be available to develop constituent-specific emission estimates. When facility-specific
data are not available, average stream compositions presented in Table 2-7 should be used. The average
stream compositions presented in Table 2-7 were calculated based on the Petroleum Environmental
Research Forum (PERF) refinery process stream speciation study (API, 2002).

Equations 2-5 and 2-6 provide the hourly emission rate. Because the components for which Methodology
Ranks 4 or 5 for equipment leaks is applicable are not monitored, there are no additional considerations
other than operating hours of the equipment (as seen in Equation 2-4) to convert the hourly emission
estimate into an annual value.
2-17
Example 2-5: Calculation of Equipment Leak Emissions Using Refinery Average

Emission Factors
At a refinery, assume there are 100 gas valves in a stream that, on average, contains 80 weight
percent non-methane organic compounds, 10 weight percent water vapor, 10 weight percent
methane, and no ethane (thus, the TOC weight percent would be 90). If the process operates
8,000 hours per year, what are the hourly and annual VOC emissions from the 100 gas valves?
Because the process stream contains no ethane, the average hourly VOC emissions from the
gas valves in the stream can be calculated using the applicable emission factor from Table 2-5
and Equation 2-6.
EVOC = FA × WFTOC × N
= 0.0268 kg NMOC/hr/gas valve × 0.9 × 100 gas valves
= 2.412 kg VOC/hr
The annual emissions from the valves in the stream are calculated using Equation 2-4 (with ETOC
replaced by EVOC).
EVOC, annual = EVOC × H
= 2.412 kg VOC/hr × 8,000 hr/yr
Table 2-7. Concentration of HAP in Refinery Process Unit Streamsa
Average Weight Percent of Compound in Process Unit Stream

2,2,4-Trimethyl
Xylenes (total)
1,2,4-trimethyl
1,3-Butadiene
Ethylbenzene
Naphthalene
n-Hexane
Biphenyl
Benzene
benzene
Cumene
pentane
Toluene
Process Unit
Crude/Atmospheric 0.01 4.3 0.05 0.9 1.7 2.0 0.63 0.12 0.63 0.25 0.06
distillation
Vacuum distillation - 0.01 0.003 0.003 0.04 0.04 0.02 - 0.02 0.12 0.09
Coking 0.04 2.5 0.75 0.42 1.3 1.3 1.3 0.12 0.7 0.28 -
Hydrocracking 0.025 1.9 1.0 1.3 2.7 2.7 0.7 0.09 1.3 0.20 -
Catalytic cracking/FCCU 0.012 1.0 0.3 1.0 3.3 4.9 1.1 0.10 1.9 0.72 0.43
Catalytic reforming/CRU 0.009 2.8 0.25 6.3 17.4 17.6 3.9 0.42 5.9 0.87 -
Hydrotreating/ - 1.9 - 0.37 1.7 1.9 0.37 0.07 0.4 0.25 0.22
Hydrode sulfurization
Alkylation 0.22 1.6 25 0.03 2.0 0.08 - - - - -
Isomerization - 3.2 0.01 0.5 0.6 0.15 - - - - -
Polymerization - 0.5 0.7 1.2 - 1.8 1.2 0.09 - - -
a Source: API, 2002.
2-18
Version 3 Section 3—Storage Tanks
3. Storage Tanks
Storage tanks (which are also sometimes referred to as storage vessels) are used in refineries for storing
and dispensing various liquids used in the refining process. These various liquids are typically organic
liquids, also known as petroleum liquids, which are a mixture of hydrocarbons, such as gasoline and
crude oil. Depending on the specific design and construction of the tank and the characteristics of the
petroleum liquids, storage tanks may emit significant levels of VOC and HAP during typical operation,
venting, and tank filling or dispensing.
The basic designs for storage tanks include horizontal and vertical fixed-roof tanks, external floating roof
tanks, internal floating roof tanks, and domed external floating roof tanks. Chapter 7.1 (Organic Liquid
Storage Tanks) of EPA’s Compilation of Air Pollutant Emission Factors (AP-42) provides detailed
descriptions of the characteristics specific to each tank type (U.S. EPA, 1995a [some sections were
updated in 2006]).
The emission estimation methods for storage tanks are presented in Table 3-1. These methods are ranked
according to anticipated accuracy. There are three primary estimation methods: (1) direct measurement,
(2) tank-specific modeling, and (3) default tank emission factors. Direct measurements can only be taken
from storage tanks that are covered and vented to a control device. Tank-specific modeling can be
accomplished with emission equations, such as those presented in AP-42 or by computer-based models
designed to implement these equations (e.g., TANKS v4.09D; U.S. EPA, 2006). Most refinery owners
and operators will have sufficient data to use the tank-specific modeling approach and this method should
be used for most refinery storage vessels. In rare cases when tank-specific data are not available,
engineering estimates can be made using actual or estimated production quantities and default emission
factors. For the purposes of the refinery ICR, all storage tank emission estimates should be developed
using Methodology Ranks 1 or 2.
Table 3-1. Summary of Typical Hierarchy of Storage Tank Emission Estimates
Methodology
Rank Description Application Data Requirements
1 Direct measurement Covered and vented Constituent concentration and flow rate
storage tanks
2 Tank-specific modeling All petroleum liquid storage Tank type, tank dimensions, stored liquid
tanks properties and constituent concentrations,
tank condition/fitting information, throughput
3a Default tank modeling Common petroleum liquid Site-specific crude throughput, specific
storage tanks product production rates
3b Default tank modeling Common petroleum liquid Crude throughput and production capacities
storage tanks (from the U.S. Energy Information
Administration or other similar source)
3.1 Methodology Rank 1 for Storage Tanks

Emissions from fixed-roof storage tanks may be purged (e.g., tanks under a nitrogen blanket) and vented
to a control device; these emissions can be directly measured at the outlet of the control device using the
direct measurement methods for process vents or combustion sources as described in Sections 4,
Stationary Combustion Sources, and 5, Process Vents, of this Refinery Emissions Protocol document. It is
likely that the control device will serve a group of tanks; it is acceptable to report the cumulative
emissions from the group of storage tanks if this is the case.
3-1
For some control devices, such as flares, thermal oxidizers, or carbon adsorption systems, it may be
possible to measure the concentration of the gas upstream of the control device and estimate the emissions
using an assumed control device efficiency (for more details on these types of methods, see the discussion
at the end of Section 3.2 below; Section 4, Stationary Combustion Sources; and/or Section 6, Flares).
This method is also recommended as a means to estimate storage tank emissions when the emissions from
the storage tank(s) are routed to a control device that also serves other refinery emission sources. In this
case, the concentration and flow rate of the storage tank off-gas can be measured prior to combining the
stream with other non–storage tank emissions. The resulting constituent mass flow rates can be corrected
for the control device efficiency using the methods described in Chapter 6 (Flares) to estimate the
emissions associated with the storage tank(s) that are released to the atmosphere.
There are other direct measurement methods that have been used to measure emissions from storage tanks
even when the emissions from the tank are not vented (i.e., DIAL [Differential Absorption LIDAR]
techniques); however, these methods do not provide continuous monitoring and have additional
limitations (requiring consistent wind direction, etc.). Therefore, at the present time they are not
recommended as primary techniques for annual emission estimation.
If direct measurement methods are used, emissions from the tank during degassing, cleaning, or drained
idle periods may or may not be included, depending on how these gases are vented. If these gases are
routed to the same control device as used during normal operations, the measured emissions will account
for these periods. If these emissions are not monitored using direct measurement methods, the emissions
from the tank during degassing and cleaning should be estimated separately using the methods provided
in Section 11, Startup and Shutdown, of this Emissions Protocol Document.

Except for the limited number of storage tanks whose emissions are collected and controlled external to
the storage tank, EPA recommends that the emission estimation procedures detailed in Chapter 7.1 of AP-
42 (U.S. EPA, 1995a) be used to calculate air pollutant emissions from organic liquid storage tanks. There
are many tools available, such as TANKS v4.09D emission estimation software that can be used to
perform the necessary calculations. TANKS v4.09D software can be downloaded for free at:
http://www.epa.gov/ttn/chief/software/tanks/index.html under the How to Get TANKS 4.09D link.
Because TANKS v4.09D is widely used, Appendix C of this Refinery Emissions Protocol document
provides tips and insights on using the TANKS program.
Key things to consider when estimating emissions from storage tanks (whether using the AP-42 equations
directly or when using TANKS v4.09D or other software packages) include:
1. Each tank should be modeled individually using site-specific conditions. There may be instances in
which a set of tanks have identical properties and materials and very similar throughputs so that the
emissions from a single storage tank can be modeled and used as the emissions for each tank in the
set. However, this approach should be limited to situations where the tanks, contents, and
throughputs are effectively identical. In all other cases, it is recommended that individual tank data
be entered into the TANKS program (or other similar software), and the emission results for
individual tanks be reported.
2. Each tank should be modeled based on the vapor pressure and composition of the material stored in
the tank. The TANKS model has default vapor pressure and composition for certain streams;
however, these defaults should not be used unless the default parameters have been specifically
evaluated and determined to adequately represent the stored liquid. For example, it is possible that a
specific gasoline default stream (where different defaults are available for different Reid vapor
pressures) will match well the gasoline stored in certain tanks, but it is less likely that the single
crude oil default in TANKS will match the crude oil stored in the crude oil tanks.
3-2
3. For the internal floating roof tank, the external floating roof tank, and the domed external floating
roof tank, the tank fittings should be selected to represent the specific characteristics of each
individual storage tank. You should not use the “default” fitting settings when using TANKS
v4.09D, particularly for the type and control for guide poles. The default fitting settings should be
reviewed and revised to properly account for the number and control status of the specific tank’s
fittings.
4. The storage tanks should be modeled using monthly parameters, including average monthly
measured tank liquid temperatures, when available. In TANKS, selecting the city will set the
meteorological data that will be used in the calculations. Even if you select the monthly calculation
option, the liquid temperature remains at the annual average. As such, the TANKS program will not
adequately account for the monthly variations in emissions. Also, for intermediate process tanks
with floating roofs that store liquids that are generally warmer than ambient temperature, TANKS
will underestimate the emissions. 1 Therefore, for storage tanks with throughputs that vary
significantly with the seasons and floating roof tanks that store warm process fluids, the equations in
Chapter 7.1 of AP-42 should be used directly to more accurately estimate the annual emissions.
5. Special calculations should be performed to account for tank roof landings, tank degassing, and tank
cleaning, and these estimates should be included in the final annual emissions reported for each tank.
The TANKS model does not currently contain algorithms for estimating emissions from tank roof
landings. These emissions should be estimated separately for each landing event using the methods
provided in Section 7.1.3.2.2 of the 2006 update of Section 7.1 of AP-42. Tank degassing emissions
can be calculated using the liquid heal method described in Section 11, Startup and Shutdown, of
this Refinery Emissions Protocol document.
6. Emission estimates should be calculated and reported for individual pollutants. The TANKS
program typically provides total hydrocarbon emission estimates, but can also provide estimates for
individual pollutants.
7. Maximum hourly average emission rates for each tank should be calculated based on the reasonable
worst-case (high emission rate) situation for a given storage tank, which will generally correspond to
the emissions while the tank is actively filling. Factors that should be considered are the volatility of
the material stored, the filling rate, the bulk liquid and ambient air temperature, and the wind speed.
Higher emissions will occur when these parameters are at their highest values. Note that the TANKS
program is primarily designed to estimate long-term, annual average emissions and cannot be easily
manipulated to estimate the maximum hourly emission rate. Also note that the maximum hourly
average emission rate should not be based on the tank’s degassing emission estimate.
For fixed-roof tanks that are vented to a control device, but for which flow and composition data are not
measured (i.e., data are not available to use Methodology Rank 1for storage tanks), the pre-control
emissions from the fixed-roof storage tanks can be estimated using the appropriate equations for fixed-
roof storage tanks presented in Chapter 7.1 of AP-42 (U.S. EPA, 1995a). The post-control device
emissions are then estimated from the pre-control emission estimates and the efficiency of the control
device using Equation 3-1.
1
The TANKS program includes an algorithm for calculating emissions from heated tanks, but it is only applicable to
tanks with horizontal or vertical fixed roofs.
3-3
 CD eff 
E i = E unc ,i × 1 -  (Eq. 3-1)
 100% 
where:
Ei = Emission rate of pollutant “i” (tons/yr)
Eunc,i = Projected emission rate of pollutant “i” assuming storage tank or unit does not have an
add-on control device (tons/yr)
CDeff,i = Control device efficiency Feff for pollutant “i” (weight percent). See Table 3-2 for
default control efficiencies for various control devices.
Table 3-2. Default Control Efficiencies for Different VOC Control Devices
Control Device
Control Device Pollutants Efficiency
Refrigerated Condenser All VOC constituents Variable based on
constituents and
operating
temperaturea
Thermal oxidizer All VOC constituents 98%
Catalytic oxidizer All VOC constituents 98%
Carbon adsorption VOC constituents other than those listed in table note b 95%
Constituents listed in table note b 0%
a The control efficiency of a condenser should be determined based on the operating conditions of the condenser
and composition of the vent stream following the methods Methods for Estimating Air Emissions from Chemical
Manufacturing Facilities (EIIP, 2007, Section 4.2.3)
b The following compounds have extremely low adsorptive capacities on activated carbon: acetaldehyde,
acetonitrile, acetylene, bromomethane, chloroethane, chloromethane, ethylene, formaldehyde, methanol, and vinyl
chloride.

Refinery owners and operators are expected to have tank-specific information and should use
Methodology Rank 2 for storage tanks; however, state and local agency staff may not have access to this
information. When tank-specific information is not available, default emissions factors can be used. The
emissions factors presented in this section are primarily based on model external floating roof tanks for
various petroleum refinery liquids (e.g., crude oil, gasoline, jet fuel). These default emissions factors
required assumptions regarding the typical fitting controls and the average composition of crude oil,
gasoline, and other products at the refinery. The heavy distillate and aromatics emissions factors were
developed based on the permit applications, which were generally based on TANKS modeling. The
reported emissions, which were normalized by production capacity, were then averaged across a limited
number of storage tanks to develop the default emissions factors that are provided in Table 3-3.
Additional details regarding the assumptions used to develop these emissions factors are found in two
publications (i.e., RTI, 2002; Lucas, 2007). These emission factors typically provide emission estimates
within a factor of 2 of tank-specific modeling approach, but will be less accurate if most of the tanks are
internal floating roof tanks and/or external floating roof tanks with domed covers.
3-4
Table 3-3. Default Emissions Factors for Petroleum Refinery Storage Tanks
Emissions Factors for Petroleum Liquid Storage Tanks

(lb/MMbbl)
Gasoline Diesel Fuel Asphalt,
and Other and Other Lube Oils,
CAS Light Middle and Heavy
Number Chemical Name Crude Oil Distillates Distillates Distillates Aromatics
a
71432 Benzene 10. 70. 54 40
a
108883 Toluene 7.5 180 100 29
a
1330207 Xylene 6.2 140 70 26
100414 Ethylbenzene 1.6 31 18 5.3
100425 Styrene 0 66 0
98828 Cumene 0.5 15 10 0.4
95636 1,2,4-Trimethyl benzene 0.7 0 0 5.9
1634044 Methyl tertiary-butyl ether 0 310 0 0
106990 1,3 Butadiene 0 1.8 0 0
110543 Hexane 84 420 480 13
540841 2,2,4-Trimethylpentane 3.4 140 22 0
78933 Methyl ethyl ketone 0.3 0.33 0
108101 Methyl isobutyl ketone 320 320 0
67561 Methanol 3.8 3.8 0
108952 Phenol 0.9 0.9 0.67 0
1319773 Cresol 0.6 13 0.19 0
91203 Naphthalene 0.6 7.6 4 5.77
91576 2-Methyl naphthalene 3.5 3.5 0
92524 Biphenyl 0.2 0.17 0 0.7
POM Polycyclic organic matter/ 0 0 17
polynuclear aromatics/PAH
120127 Anthracene 0.24 0.24 0
218019 Chrysene 0.21 0.21 0
86737 Fluorene 0.36 0.36 0
85018 Phenanthrene 1.5 1.5 0
129000 Pyrene 0.39 0.39 0
Sum VOC 1,350 8,800 5,300 120 15,000

Note: lb/MMbbl = pounds per million barrels
a Use “Sum VOC” emissions factor and apply it to the production of each aromatic produced.
When facility-specific production data are available, these data should be used with the default emissions
factors to estimate emissions. When these data are not available, crude distillation capacity and
production capacities, as presented in the Refinery Capacity Report 2009 (EIA, 2009), can be used to
estimate these production rates as follows:
 The crude oil processing rate can be assumed to be the atmospheric crude oil distillation capacity.
If a refinery does not have atmospheric crude distillation columns, then the sum of the vacuum
distillation and coking capacities can be used as the crude oil processing rate.
3-5
 The aromatics production capacity is provided in the Refinery Capacity Report 2009 (EIA, 2009);
however, the specific aromatics produced cannot be determined. As a default assumption, equal
production of benzene, toluene, and xylenes can be assumed.
 Heavy distillate production can be estimated as the sum of lube oil production, asphalt
production, and coke production.
 Light distillate production can be estimated as 60% of the difference of the crude oil processing
rate minus the aromatics and heavy distillate production (0.6 × [Crude − Aromatics − Heavy]).
 Middle distillate production can be estimated as 40% of the difference of the crude oil processing
rate minus the aromatics and heavy distillate production (0.4 × [Crude − Aromatics − Heavy]).
3-6
Version 3 Section 4—Stationary Combustion Sources
4. Stationary Combustion Sources

Petroleum refineries include numerous stationary combustion sources, the most common of which are
process heaters, boilers, internal combustion engines, and combustion turbines. These combustion sources
are vent (point) sources that occur throughout the process area of the refinery. The size of the vent stack
varies with the size of the source (typically measured in terms of the rate that fuel is burned). Process
heaters are used to indirectly preheat feedstock or process fluids for a given process, to reheat
intermediates of a process, or to heat distillation columns (the latter are often termed “reboilers”). These
emission sources are typically localized at or near the process requiring the heater (or reboiler). Boilers
are used to generate steam for various refinery operations and to raise the temperature of feedstocks and
process streams. The primary fuel for nearly all petroleum refinery process heaters and boilers is refinery
fuel gas, which is a mixture of uncondensed overhead gases from distillation columns (also referred to as
“still gas”) and natural gas. Still gas is produced from a variety of refinery process units, including
atmospheric crude oil distillation units, fluid catalytic cracking units, catalytic reforming units, fluid and
delayed coking units, and hydrocracking units. Other fuels, such as natural gas only, fuel oil, and residual
oil, are also used. Internal combustion engines have a variety of uses, including powering fire suppression
systems, and supplying emergency back-up power. Internal combustion engines generally combust
natural gas, gasoline, or diesel fuel oil. Combustion turbines are used for cogeneration, and typically
combust natural gas or refinery fuel gas.
In addition to refinery fuel gas, some petroleum refineries also produce petroleum coke. The petroleum
coke produced by a coking unit is generally referred to as green petroleum coke or fuel grade coke. Some
refineries may use green petroleum coke as fuel in boilers to produce steam or electricity. Green
petroleum coke is expected to contain relatively high concentrations of sulfur and metals when compared
with other fuels used at a refinery. Green petroleum coke may be purified in a calcining unit for use in
industrial applications, most often as anode material in the manufacture of aluminum. Consequently,
calcined petroleum coke is often referred to as anode grade coke. Emissions from coke calcining are
described in more detail in Section 5.6.3, Coke Calcining, of this Refinery Emissions Protocol document.
Alternatively, green petroleum coke may be gasified with steam and either air or oxygen (O2) to form a
low (using air) to medium (using O2) heating value synthesis gas (or syngas) for subsequent use as a fuel
gas. There are no direct emissions from the coke gasification process; emissions from the combustion of
syngas can be estimated using the methods provided in this section.
Table 4-1 summarizes the hierarchy of emission estimation methods for stationary combustion sources.
Each refinery may use a combination of different methods, such as Methodology Rank 1 for criteria
pollutants such as sulfur dioxide (SO2) and NOx, and Methodology Rank 2 or 3 to estimate emissions of
organic or metal compounds. Different methodologies may be used for each combustion source for each
pollutant based on the type of monitoring data available for each source (e.g., Methodology Rank 1 may
be used for estimating NOx emissions for one combustion source for which continuous emission
monitoring systems (CEMS) data are available, while Methodology Rank 4 may be used to estimate NOx
emissions for another combustion source at the same facility where no CEMS data or emission test data
are available). It is also important to note that the selection of Methodology Rank 3A or Methodology
Rank 3B will depend on factors such as whether the pollutant being estimated has a reduced form (e.g.,
reduced sulfur compounds form SO2 when oxidized) that can be continuously monitored, the variability
of the combustion source, and the date of the most recent stack test. The remainder of this section
provides additional detail and guidance regarding the implementation of these methods.
A complete inventory for combustion sources will include emission estimates for PM, SO2, NOx, VOCs, 1
and carbon monoxide (CO). Depending on the fuel combusted, a complete inventory will also include
1
Total hydrocarbon [THC] may be estimated as a surrogate for VOC.
4-1
emission estimates for hydrogen sulfide; metals, including but not limited to arsenic, chromium, mercury,
lead, and selenium; organic compounds, including but not limited to benzene, formaldehyde, 1,3-
butadiene, and naphthalene and other polycyclic organic matter; and dioxins and furans. It is important to
note that this list is not exhaustive; if there is evidence that a combustion source at a refinery emits a
pollutant (e.g., site-specific emission test data for an additional compound or a relevant emission factor
for an additional compound), an emission estimate should be included in a complete inventory, even if
that pollutant is not specifically identified in this document.
Table 4-1. Summary of Typical Hierarchy of Stationary Combustion Source Emission Estimates
Rank Measurement Method Additional Data Needed

1 Direct measurement (continuous emission  Pressure, temperature, and moisture content
monitoring systems [CEMS]) for both flow rate and (depending on the monitoring system)
gas composition
2 Direct measurement (CEMS) for gas composition  Fuel usage
Use of F factors  Heat content of fuel (depending on units of
source-specific emission factors)
3A Fuel analysis/mass balance  Fuel usage
 Assumed destruction efficiency
3B Source-specific stack testing to calculate source-  Fuel usage
specific emission correlations or factors
 Heat content of fuel (depending on units of
source-specific emission factors)
4 Default emission factors  Fuel usage
There are many sources that describe emission methodologies for combustion sources, particularly boilers
and internal combustion sources, including Chapter 2 in the Emission Inventory Improvement Program
Technical Report Series (ERG, 2001), 2008 Emissions Inventory Guidelines (TCEQ, 2009), and A
Customized Approach for Quantifying Emissions for the Electric Power Generation Sector in Mexico
(The United States–Mexico Foundation for Science, 2008). These documents include many examples and
sample calculations for liquid and solid fuels and should be consulted for more information if needed. In
addition, the American Petroleum Institute (API) has posted a Compendium of Greenhouse Gas
Emissions Methodologies for the Oil and Natural Gas Industry on their Web site. While this document is
focused on estimating GHG emissions, it contains information on how to estimate combustion gas flows
and how to estimate emissions from various types of combustion sources, and it includes useful
conversion factor discussions and examples.
4.1 Methodology Rank 1 for Stationary Combustion Sources

Many stationary combustion sources will have a CEMS for NOx and/or SO2. A CEMS is a comprehensive
unit that continually determines gaseous or PM concentrations or emission rates using pollutant analyzer
measurements and a conversion equation, graph, or computer program to produce results in the desired
units. A CEMS that includes a flow rate monitor, which is needed to determine mass emission rates, such
as those needed for an emission inventory, is also referred to as a continuous emission rate monitoring
system (CERMS). The CEMS or CERMS continually determines the flow rate of the exhaust gas,
analyzes the composition or specific pollutant concentration in the exhaust gas, and records the results.
There are two main types of CEMS: in-situ and extractive. An in-situ CEMS measures and analyzes the
emissions directly in a stack. There is less sample loss associated with the in-situ CEMS compared to the
extractive CEMS because the sample lines of an extractive system can leak, freeze, or clog, or pollutants
can be lost because of adsorption, scrubbing effects, or condensation. In an extractive CEMS, the sample
gas is extracted from the gas stream and transported to a gas analyzer for the measurement of the
contaminant concentration. Because an extractive CEMS is located outside the stack, the sampling
4-2
instruments are not affected by stack conditions, maintenance and replacement are generally simpler, and
the cost is lower than with an in-situ CEMS, although extra costs are incurred for the sampling and
conditioning system for an extractive CEMS (Levelton Consultants, 2005).
The pollutant concentration recorded by a CEMS is generally on a concentration basis, such as parts per
million. The CEMS may also include a diluent monitor (e.g., O2) for correcting the concentrations to a
fixed excess air concentration. For in-situ CEMS, these measurements are made at stack conditions so
that the concentrations would be determined on a “wet basis.” That is, the concentrations are based on the
total amount of gas, including water vapor. For extractive CEMS, the gas is often conditioned to remove
water vapor before analysis, so the concentrations are commonly determined on a “dry basis.” Gas flow
measurements are made at stack conditions, so the flow rate will be in terms of actual gas volume on a
wet basis. If the gas composition is determined on a dry basis, then a moisture content measurement is
needed to convert the flow rate to a dry basis (or convert the composition to a wet basis) so that both
measurements are on the same basis, and many gas flow monitors contain temperature and pressure
monitors to allow conversion of the flow to standard conditions for this purpose. It is important to note
that care must be taken to ensure that the gas and flow measurements are made on the same basis and in
the same terms as the permitted limits, if applicable, or that appropriate ancillary measurements are made
to perform the necessary unit conversions.
The following general equation (Equation 4-1) is used for determining a mass emission rate from CEMS:
 (C ) MWi  To   Pn  
∑  (Q ) × [1 − ( f ) ]× 100%
N
Ei = i n
× ×  ×   × M N × K  (Eq. 4-1)
MVC  Tn 
n H 20 n
n =1    Po  
where:
Ei = Emission rate of pollutant “i” (tons per year [tons/yr]).
N = Number of measurement periods per year (e.g., for hourly measurements, N = 8,760).
n = Index for measurement period.
(Q)n = Volumetric flow rate for measurement period “n” (actual cubic feet per minute [acfm]).
If the flow rate meter automatically corrects for temperature and pressure, then replace
“To ÷ Tn × Pn ÷ Po” with “1.” If the pollutant concentration is determined on a dry basis
and the flow rate meter automatically corrects for moisture content, replace the term
[1-(fH20)n] with 1.
(fH2O)n = Moisture content of exhaust gas during measurement period “n,” volumetric basis
(cubic feet water per cubic feet exhaust gas).
(Ci)n = Concentration of pollutant “i” in the exhaust gas for measurement period “n” (volume
%, dry basis). If the pollutant concentration is determined on a wet basis, then replace
the term [1−(fH20)n] with 1.
MWi = Molecular weight of pollutant “i” (kilogram per kilogram mole [kg/kg-mol]).
MVC = Molar volume conversion factor = 849.5 standard cubic feet per kilogram mole
(scf/kg-mol) at 68°F (528°R) and 1 atmosphere (atm).
To = Temperature at “standard conditions” (528 °R).
Tn = Temperature at which flow is measured during measurement period “n” (°R).
Pn = Average pressure at which flow is measured during measurement period “n” (atm).
Po = Average pressure at “standard conditions” (1 atm).
MN = Minutes per measurement period (minutes per measurement period “n”). This term can
be calculated by dividing 525,600 minutes per year (min/yr) by the number of
measurement periods per year (“N”).
K = Conversion factor = 2.2046/2,000 (tons per kilogram [tons/kg] = 0.0011023 tons/kg.
4-3
A CEMS records multiple measurements per hour; the frequency depends on the pollutant being
measured and the type of CEMS. For example, a CEMS monitoring benzene concentration using gas
chromatography may only sample and record a measurement every 15 minutes, while a CEMS
monitoring SO2 concentration may record a measurement every minute. These individual measurements
can be used to calculate annual emissions in two ways. The most common method is for the CEMS to
average the measurements within each hour and develop 8,760 hourly average concentrations and flow
rates that can be summed. Example 4-1 demonstrates the calculation of NOx emissions for 1 hour for a
combustion unit firing refinery fuel gas based on an hourly average concentration and flow rate. This
method is best suited for measurements that are fairly consistent and stable over the course of an hour.
The other method is to determine the emission rate for each recorded measurement based on the
concentration and flow rate for that measurement. In other words, if the CEMS records measurements
every minute, then the emission rate is determined per minute and hourly emissions are determined by
summing the 60 applicable emission rates; if the CEMS records measurements every 5 minutes, then the
emission rate is determined for each 5-minute interval and hourly emissions are determined by summing
the 12 applicable emission rates. This method may be more accurate than hourly averages if the
combustion source’s flow rate varies significantly within an hour.
Example 4-1: Calculation of NOx Emissions Using a CEMS

The following example shows the calculation for 1 hour (60 minutes); the total emissions
during any period (i.e., day, month, quarter, or year) may be calculated as the sum of the
hourly emissions determined by the CEMS. In terms of Equation 4-1, “i” is NOx and the index
for the period in this example is 1, so the result “E” is in tons per hour.
Calculate hourly NOX emissions for a 150 million British thermal units per hour (MMBtu/hr)
unit burning refinery fuel gas given that the following data have been collected:
– The hourly average NOx concentration calculated by the CEMS for this hour is 60 parts
per million by volume (ppmv), wet basis
– The hourly average flow rate calculated by the CEMS for this hour is 500,000 acfm, wet
basis
– The unit operated at standard conditions for the full hour.
Equation 4-1 should be used to calculate the hourly emissions:

 (C ) 
∑  (Q ) × [1 − ( f ) ]× 100%
N
MW NOx T P
E NOx = n H 20 n
NOx n
× × o × n × M N × K 
n =1  MVC Tn Po 
= 500,000 × 1 × (60 ÷ 1,000,000) × 46.01 ÷ 849.5 × 1 × 1 × 60 × 0.0011023
= 0.1075 tons per hour (tons/hr).
Rounded to two significant figures, the hourly emissions are 0.11 tons/hr.
If the combustion source did not operate the entire hour, multiply the hourly emissions by the
fraction of the hour that the unit operated (e.g., if the unit operated for 30 minutes, then
emissions for this hour would be 0.1075 tons/hr × 0.5 = 0.054 tons/hr).
If the combustion source operated steadily and continuously for an entire year and the
emission rate remained perfectly constant over that year, annual emissions would be
0.1075 tons/hr × 8,760 hours per year (hr/yr) = 940 tons/yr.
4-4

Because the emission standards for stationary combustion sources are often provided on a concentration
basis or on a mass per heating–value basis, many stationary combustion sources will have a CEMS for
pollutant concentrations, but may not directly measure exhaust gas flow. In these cases, fairly accurate
exhaust gas flow rates can be calculated from the refinery fuel gas composition and fuel flow
measurements. This method is commonly referred to as the “F factor” method, and procedures for
conducting a fuel analysis and calculating estimated emissions are described in EPA Method 19 (40 CFR
Part 60, Appendix A-7).
EPA Method 19 includes many equations that can be used to calculate emissions depending on moisture
measurements and whether pollutant concentrations and O2 or CO2 content are measured on a wet or dry
basis, and each equation uses one of three different types of F factors. The calculation of the different F
factors is described in Section 12.3.2 of Method 19. In lieu of this calculation, Table 19-2 of EPA Method
19 includes default F factors for some common types of fuel that could be used; however, F factors for
refinery fuel gas are not included in Table 19-2.
Refinery fuel gas generally includes many different components; therefore, one equation that can be used
to calculate the F factor in this situation is an adapted version of Equation 19-16 in EPA Method 19 for
multiple fuels. In Equation 4-2, the dry F factor, Fd, is calculated assuming both the pollutant
concentration and the O2 content are measured on a dry basis:
 n 
 ∑ ( X i × MEVi ) 
Fd = K ×  in=1 
  (Eq. 4-2)
 ∑ ( X i × MHCi ) 
 i =1 
where:
Fd = Volume of combustion components resulting from stoichiometric combustion per unit
of heat content (dry standard cubic feet per million British thermal unit [dscf/MMBtu])
K = Conversion factor, 106 (Btu/MMBtu)
n = Number of fuels or fuel components
i = Index for fuels or fuel components
Xi = Mole or volume fraction of each component in the refinery fuel gas
MEVi = Molar exhaust volume (dry standard cubic feet per mole [dscf/mol])
MHCi = Molar heat content (British thermal units per mole [Btu/mol])
Table 4-2 includes values for the molar exhaust volume and molar heat content for common constituents
of refinery fuel gas. It is important to note that these calculations are based on the higher heating value of
the fuel gas (referred to as gross calorific value in EPA Method 19).
Table 4-2. Molar Exhaust Volumes and Molar Heat Content of Refinery Fuel Gas Constituents
Constituent MEVa (dscf/mol) MHCb (Btu/mol)
Methane (CH4) 7.28 842
Ethane (C2H6) 12.94 1,475
Hydrogen (H2) 1.61 269
Ethene (C2H4) 11.34 1,335
(continued)
4-5
Table 4-2. Molar Exhaust Volumes and Molar Heat Content of Refinery Fuel Gas Constituents
(continued)
Constituent MEVa (dscf/g-mol) MHCb (Btu/g-mol)

Propane (C3H8) 18.61 2,100
Propene (C3H6) 17.01 1,947
Butane (C4H10) 24.28 2,717
Butene (C4H8) 22.67 2,558
Inerts 0.85 0
a MEV = molar exhaust volume (dry standard cubic feet per gram-mole [dscf/g-mol])
b MHC = molar heat content (Btu per gram-mole [Btu/g-mol]); higher heating value basis
The F factor can be used to calculate emissions in one of two ways. In the first option, the F factor, the
amount of fuel combusted, and the heat content of the fuel can be used to calculate the volumetric flow
rate of the exhaust gas, (Q)n, using the following equation.
20.9
Qn = Fd × Q f × HHV × (Eq. 4-3)
(20.9 − %O2d )
where:
Qn = Volumetric flow rate for measurement period “n” (dry standard cubic feet per minute
[dscfm])
Fd = Volume of combustion components per unit of heat content (dscf/MMBtu)
Qf = Volumetric flow rate of fuel (dscfm)
HHV = Higher heating value of fuel (million British thermal units per standard cubic foot
[MMBtu/scf])
%O2d = Concentration of O2 on a dry basis (percent)
Once the volumetric flow rate is known, emissions can be calculated using Equation 4-1 as in
Methodology Rank 1 for stationary combustion sources. It is important to note that when using the F
factor method as indicated here, the exhaust gas flow rate will be in units of dry standard cubic feet per
minute, so the temperature and pressure correction terms are not needed. A moisture correction term is
not needed when the concentration measurement is also made on a dry basis. If the concentration
measurements are made on a wet basis, then they must be corrected to a dry basis by dividing by the
[1−(fH20)n] term.
In the second option, the F factor can be used to calculate an emission factor (Equation 4-4) that can then
be multiplied by the higher heating value of the fuel to estimate emissions (Equation 4-5).
20.9
EFi = C d Fd (Eq. 4-4)
(20.9 − %O2d )
where:
EFi = Emission rate of pollutant (pounds per million British thermal unit [lb/MMBtu])
Cd = Pollutant concentration, dry basis (pounds per standard cubic foot [lb/dscf])
Fd = Volume of combustion components per unit of heat content (dscf/MMBtu)
%O2d = Concentration of O2 on a dry basis (percent)
4-6
Ei = EFi × (60 × Q f )× HHV (Eq. 4-5)
where:
Ei = Emissions of pollutant “i” (pounds per hour [lb/hr])
EFi = Emission rate of pollutant (lb/MMBtu)
Qf = Volumetric flow rate of fuel (dscfm)
60 = Conversion factor, minutes per hour
HHV = Higher heating value of fuel (MMBtu/scf)
Example 4-2 demonstrates the calculation of an F factor based on a fuel analysis for refinery fuel gas.
Examples 4-3 and 4-4 demonstrate the first and second options, respectively, to estimate emissions using
the F factor. As previously mentioned, if the result of a calculation is an hourly emission estimate, then
the calculation process should be repeated for each hour in the year, or 8,760 times. For example, the
hourly pollutant and O2 concentrations from the CEMS should be used in Equation 4-4 to calculate an
emission factor for each hour in the year. That emission factor and the hourly measurements of the fuel
flow rate should then be used in Equation 4-5 to determine the hourly emissions for each hour in the year.
The sum of all of those hourly emission estimates is the annual emission estimate.
Example 4-2: Calculation of Fuel-Specific F Factor

The CEMS measures SO2 and O2 content on a dry basis, so in lieu of using a default F factor,
the facility can use fuel analysis results and Equation 4-2 to calculate an Fd factor
(scf/MMBtu). The Fd factor is calculated at standard conditions of 20°C (68°F) and 29.92
inches of mercury.
The fuel analysis revealed the following mole fractions:
Methane 0.44 Propene 0.03
Ethane 0.04 Butane 0.17
Hydrogen 0.06 Butene 0.01
Ethene 0.01 Inerts 0.04
Propane 0.2
Equation 4-2 is used to calculate Fd as follows:
( ) ( ) ( ) ( ) (
Fd = K × [ X CH 4 × MEVCH 4 + X C 2 H 6 × MEVC 2 H 6 + X H 2 × MEV H 2 + X C 2 H 4 × MEVC 2 H 4 + X C3 H 8 × MEVC3 H 8 + )
(X C3 H 6 ) (
× MEVC3 H 6 + X C 4 H10 × MEVC 4 H10 ) + (XC4 H 8 × MEV ) + ( X × MEV )] ÷
C4 H 8 inerts inerts
[( X CH 4 × MHC CH 4 ) + (XC2 H 6 × MHC C 2 H 6 ) + (X H2 × MHC ) + (X

H2 × MHC ) + (X
C2 H 4 C2 H 4 C3 H 8 )
× MHC C3 H 8 +
(X C3 H 6 × MHC C3 H 6 ) + (X C 4 H 10 × MHC C 4 H10 ) + (X × MHC
C4 H 8 ) + ( X × MHC )]
C4 H 8 inerts inerts
Fd = 106 × [(0.44 × 7.28) + (0.04 × 12.94) + (0.06 × 1.61) + (0.01 × 11.34) + (0.2 × 18.61) + (0.03
× 17.01) + (0.17 × 24.28) + (0.01 × 22.67) + (0.04 × 0.85)] ÷ [(0.44 × 842) + (0.04 × 1475) +
(0.06 × 269) + (0.01 × 1335) + (0.2 × 2100) + (0.03 × 1947) + (0.17 × 2717) + (0.01 × 2558)
+ (0.04 × 0)]
= 106 × 12.55 ÷ 1425
= 8,809 dscf/MMBtu
4-7
Example 4-3: Calculation of Exhaust Flow Rate from F Factor

From Example 4-2, the F factor is 8,809 dscf/MMBtu. The measured SO2 concentration is 20
ppmv (dry basis), the measured O2 concentration is 6% (dry basis), the higher heating value is
1,300 Btu/scf, and the fuel flow rate is 500 dscfm. Equation 4-3 should be used to calculate
the exhaust flow rate from F factor as follows:
20.9
Qn = Fd × Q f × HHV ×
(20.9 − %O2d )
Qn = (8,809) × (500) × (1,300 ÷ 106 [Btu/MMBtu]) × (20.9 ÷ (20.9 – 6))
Qn = 8,031 dscfm
Use this value of Qn in Equation 4-1 without the temperature, pressure, and moisture
correction terms to estimate hourly emissions as follows:
(C i )n
E SO2 = (Q )n × [1 − ( f H 20 )n ] ×
MWi
× × MN ×K
100% MVC
ESO2 = 8,031 × 1 × (20 ÷ 1,000,000) × 64.06 ÷ 849.5 × 60 × 2.2046 = 1.6 lb/hr
Example 4-4: Calculation of Emissions Factor from F Factor

Given the same measurements as Example 4-3, calculate hourly SO2 emissions.
First, the following equation should be used to convert the concentration to the correct units
(from Table 19-1 of EPA Method 19, multiply the SO2 concentration in ppm by 1.660 × 10-7
to convert to lb/scf [see also Equation 4-7 of this section]):
Cd = 20 ppm × (1.660 × 10-7) lb/scf/ppm
Cd = 3.322 × 10-6 lb/scf
Second, Equation 4-4 should be used to calculate the emission factor:
20.9
EFSO2 = C d Fd
(20.9 − %O2 d )
EFSO2 = (3.322 × 10-6) × (8,809) × (20.9 ÷ (20.9 – 6))
EFSO2 = 0.041 lb/MMBtu
Finally, Equation 4-5 should be used to calculate the hourly emissions (the fuel flow rate is
500 dscfm, or 30,000 dscf/hr [500 dscfm × 60 minutes per hour (min/hr)])
E SO2 = EFSO2 × Q f × HHV

ESO2 = (0.041) × (30,000) × (1,300 ÷ 106 [Btu/MMBtu])
ESO2 = 1.6 lb/hr
emission rate remained perfectly constant over that year, then the annual emissions would be
1.6 lb/hr × 8,760 hr/yr ÷ 2,000 lb/ton = 7.0 tons/yr.
4-8
4.3 Methodology Rank 3A for Stationary Combustion Sources

Fuel analysis can be used to predict emissions of pollutants such as SO2 by examining the compounds in
the fuel being combusted. Given a known concentration of a compound, either a pollutant or pollutant
precursor (e.g., hydrogen sulfide [H2S] or other reduced sulfur compounds for purposes of estimating SO2
emissions), emissions of the pollutant can be calculated by assuming that all of that compound is emitted
(e.g., all of the reduced sulfur is oxidized to SO2). Alternatively, if some of the compound is not
combusted (i.e., is in a different physical or chemical state such as ash or unburned hydrocarbons), then
based on the laws of mass conservation, less than 100% of the pollutant is emitted as air emissions (ERG,
2001). In this case, it may be possible to determine the emissions actually resulting from combustion by
considering the uncombusted compounds in a mass balance analysis.
As mentioned in the introduction to Section 4, Methodology Rank 3A for stationary combustion sources
is considered a more accurate method of emission estimation than Methodology Rank 3B for stationary
combustion sources for certain pollutants, namely SO2, particularly if the concentration of the compound
of interest in the fuel and the fuel flow rate are continuously measured. This may also be the case if the
available source tests are not recent or if the combustion unit’s operation varies significantly enough that
there are concerns about the accuracy of applying an emission factor developed at one set of operating
conditions. For example, a test of NOx emissions from a process heater performed at high capacity may
not apply when the process heater is firing at a lower rate. As the heater firing rate decreases, the operator
increases the oxygen content to ensure stable operation, and this increase in oxygen content may lead to
higher concentrations of NOx.
Equation 4-6 is a general equation used to determine a mass emission estimate from fuel analysis data for
gaseous fuels (for fuel analysis of liquid and solid fuels, see the documents referenced in the introduction
to Section 4).
 MWi 
Ei = (Q f )(C f ) 
 MW  (Eq. 4-6)
 f 
where (assuming gaseous fuel):
Ei = Emission estimate of pollutant “i” (lb/hr)
Qf = Fuel flow rate (scf/hr)
Cf = Content of pollutant in fuel (lb/scf)
MWi = Molecular weight of pollutant emitted (pounds per pound mole [lb/lb-mol])
MWf = Molecular weight of pollutant in the fuel (lb/lb-mol)
As previously explained, to calculate yearly emissions, this type of calculation must be repeated for each
hour in the year, or 8,760 times. Also, as described in Example 4-1, if the combustion source did not
operate for part of an hour, then the emission estimates for that hour should be multiplied by the fraction
of the hour that it did operate.
If the compound concentration is measured in parts per million or as a percentage rather than pounds per
cubic feet, then Equation 4-7 will be needed. Equation 4-7 provides a methodology for calculating
compound-specific conversion factors such as the values shown in Table 19-1 of EPA Method 19 for NOx
and SO2.
Cf =
(Cc )× (MWi )
(Eq. 4-7)
K × 385.3
4-9
where:
Cf = Content of pollutant in fuel (lb/scf)
Cc = Volumetric concentration of pollutant in fuel (ppm or percentage)
MWi = Molecular weight of pollutant “i” emitted (lb/lb-mole)
K = Conversion constant: 106 if units of Cc are ppm; 102 if units of Cc are a percentage
385.3 = Conversion constant for ideal gases (standard cubic feet per pound mole [scf/lb-mol])
Example 4-5: Fuel Analysis

Calculate hourly emissions from a combustion source burning refinery fuel gas given:
– Fuel flow rate = 500 dscfm, or 30,000 dscf/hr (500 dscfm × 60 min/hr)
– Sulfur content of the fuel = 160 ppm
– Molecular weight of sulfur = 32 lb/lb-mole; molecular weight of SO2 = 64 lb/lb-mole
Convert the concentration of sulfur in the fuel to the correct units using Equation 4-7:
Cf =
(Cc )× (MWS )
K × 385.3
Cf = (160) × (32) ÷ (106) ÷ (385.3)
Cf = 1.33 × 10-5 lb/scf
Then, use Equation 4-6 to calculate the hourly mass emissions:
 MWSO2 
E SO2 = (Q f )(C f ) 
 MW S 
ESO2 = (30,000) × (1.33 × 10-5) × (64) ÷ (32)
ESO2 = 0.80 lbs SO2/hr
emissions rate remained perfectly constant over that year, then the annual emissions would be
0.80 lbs/hr × 8,760 hr/yr ÷ 2,000 pounds per ton (lb/ton) = 3.5 tons/yr.
4.4 Methodology Rank 3B for Stationary Combustion Sources

Source testing can provide useful data for developing site-specific emission correlations or emission
factors. Source testing provides a measurement of the emissions at a particular point in time, and most
tests are performed at conditions representative of normal operation, in which case the emission
measurement can provide an estimate of emissions at similar operating conditions. Emission factors are
developed by dividing the emission rate by a process parameter such as fuel usage. It is important to note
that this methodology will be less reliable when the unit is operating at conditions other than those tested.
Generally, one source test consisting of three runs is performed at a specific set of conditions, and the
results for each run can be averaged to determine an emission factor that is assumed to apply at all heat
input rates. Example 4-5 demonstrates a sample emission factor calculation based on one test with three
test runs. Section 5.1.2 of this document, Methodology Ranks 3 and 4 for Catalytic Cracking Units, also
provides details on calculating one emission factor based on three test runs. That section also includes a
4-10
detailed description of the ways in which source tests and analyses for PM may need to be handled
differently than other pollutants.
Example 4-6: Site-Specific Emissions Factor

During the most recent source test, three test runs were conducted to determine the NOx
emission rate for a 50 MMBtu/hr process heater. The NOx emissions rate measured during the
source tests were 1.92, 1.51, and 1.76 lbs/hr for tests 1, 2, and 3, respectively. The process
heater was firing at 95% capacity during all three test runs.
First, calculate the appropriate emissions factor for each individual run, and then average the
emissions factors with the following calculations:
– Run 1: Emissions/throughput = 1.92 [lbs/hr] ÷ (50 × 0.95) [MMBtu/hr] = 0.040 lbs
NOx/MMBtu
NOx/MMBtu
NOx/MMBtu
– Average: Emissions/throughput = (0.040 + 0.032 + 0.037) ÷ 3 = 0.036 lbs NOx/MMBtu
A more complex methodology for developing a correlation rather than one emission factor is provided in
Section 2.1.2 of A Customized Approach for Quantifying Emissions for the Electric Power Generation
Sector in Mexico (The United States–Mexico Foundation for Science, 2008). Section 2.1.2 provides a
detailed explanation of a methodology for developing a correlation between NOx emissions and heat input
to the source (fuel use multiplied by heat content of fuel) that is based on EPA’s Optional NOx Emissions
Estimation Protocol for Gas-Fired Peaking Units and Oil-Fired Peaking Units (40 CFR Part 75,
Appendix E). This methodology allows the operator to account for lower mass emissions per heat input
for lower operating loads. The methodology calls for stack measurements at three load levels, ideally
50%, 75%, and 100% firing rates. The test results at each firing rate are used to calculate an emission
factor in units of mass per heat input (or mass per amount of fuel combusted, assuming a constant heat
content of fuel). Following the development of this correlation, hourly emissions can be calculated based
on hourly measurements of fuel combusted and calculations of the heat content of the fuel (or constant
heating value of the fuel). Section 2.1.2 also includes a detailed set of sample calculations in Exhibit 2.1-3
to demonstrate this procedure.

When direct emission monitoring or site-specific emission factors are not available, then default emission
factors may be the only way to estimate emissions. The EPA has developed emission factors for various
types of combustion sources, which are compiled in AP-42. Emission factors for process heaters, boilers,
and other types of external combustion sources are included Chapter 1 of the Compilation of Air Pollutant
Emission Factors. Volume 1: Stationary Point and Area Sources, and emission factors for internal
combustion sources (including engines and combustion turbines) are presented in Chapter 3 of this same
document (U.S. EPA, 1995a). It is important to note that AP-42 does not include emission factors for all
fuels (notably refinery fuel gas and coke). The emission factors in AP-42 are the recommended default
emission factors, and AP-42 should be consulted to obtain the appropriate emission factors for criteria
pollutants such as SO2, NOx, PM, and CO.
In addition, the American Petroleum Institute (API), in conjunction with the Western States Petroleum
Association (WSPA), has conducted emission source tests of combustion sources and has compiled
4-11
emission factors that may be used for refinery combustion sources if AP-42 does not include an
applicable emission factor (Hansell and England, 1998). Separate emission factors were developed for
different combustion sources based on the type of source and fuel used. Appendix D presents the full data
analysis summary for combustion sources of interest as presented in Hansell and England’s Appendices A
and B (1998), including the California Air Resources Board (CARB) emission factor rating; the mean,
median, maximum, and minimum emission factors; number of tests analyzed; standard deviation;
uncertainty; and percent of test run results at or above the detection limit.
The following tables present the recommended emission factors for a variety of metals and organic
compounds. The emission factors in these tables are either the applicable emission factors as presented in
AP-42 or the mean emission factors compiled by Hansell and England (1998). (Additional details on the
rating and quality of these emission factors are located in the original document for AP-42 emission
factors and in Appendix D for the emission factors compiled by Hansell and England (1998).) Emission
factors for boilers and process heaters using various fuels are presented in Table 4-3. Emission factors for
reciprocating internal combustion engines using a variety of fuels are presented in Table 4-4. Emission
factors for combustion turbines using a variety of fuels are presented in Table 4-5.
To use any of these emission factors appropriately, the refinery must monitor or keep records of the
amount of fuel combusted or the heat input.
It is important to note that the bold italic emission factor values in Tables 4-3 through 4-5 from Hansell
and England (1998) are based on method detection limits and are likely biased high. Additional data are
needed to develop accurate emission factors for the compounds that were below the method detection
limits (RTI, 2002). The footnotes to Tables 4-3 through 4-5 explain any differences if a recommended
emission factor is derived from another emission factor.
It is also important to note that this document does not include emission factors for refinery combustion
sources combusting either green or calcined petroleum coke. In addition, while emissions from the
combustion of syngas (gas produced by coke gasification) can be estimated using the methods provided in
this section, this document does not include emission factors for syngas as a fuel. Emission factors for
these fuels will need to be developed from future emission testing and analysis.
4.5.1 Default Emission Factors for Process Heaters

The emissions from boilers and process heaters were dependent on the fuel type: gas (i.e., natural gas or
refinery gas), crude oil/pipeline oil, residual (No. 6) fuel oil, and distillate (No. 2) fuel oil. No distinctions
in HAP emission factors were made for criteria pollutant controls, such as selective catalytic reduction
(SCR), or selective non-catalytic reduction (SNCR). Table 4-3 presents a summary of emission factors
for boilers and process heaters.
4-12
Table 4-3. Summary of Emission Factors for Boilers and Process Heaters Firing Various Fuels
Emission Factor for Fuel Type a
External Combustion, External Combustion, External Combustion, External Combustion,

CAS Natural Gas/Refinery Crude Oil/Pipeline Oil Residual Fuel Oil Distillate Fuel Oil
Number Substance Gas (lb/MMBtu) b (lb/MMBtu) b (No. 6 Oil) (lb/MMBtu) b (No. 2 Oil) (lb/MMBtu) i
83-32-9 Acenaphthene 2.4E-09 1.7E-07 1.4E-07 f --

208-96-8 Acenaphthylene 6.5E-09 2.3E-08 1.7E-09 f --
75-07-0 Acetaldehyde 1.2E-05 1.1E-05 7.0E-06 --
107-02-8 Acrolein 1.7E-05 3.3E-06 -- --
f
120-12-7 Anthracene 4.7E-09 3.7E-08 8.1E-09 --
7440-36-0 Antimony 5.2E-07 -- 3.5E-05 h --
c h
7440-38-2 Arsenic 2.0E-07 6.7E-06 8.8E-06 4E-06
7440-39-3 Barium 4.3E-06 c -- 1.7E-05 h --
d f
71-43-2 Benzene 2.1E-06 4.1E-06 1.4E-06 --
56-55-3 Benzo(a)anthracene 2.2E-08 3.2E-08 2.7E-08 f --
50-32-8 Benzo(a)pyrene 5.7E-08 1.4E-08 1.4E-09 --
205-99-2 Benzo(b)fluoranthene 2.7E-08 5.5E-09 9.9E-09 f, g --
192-97-2 Benzo(e)pyrene -- 3.9E-09 5.8E-09 --
f
191-24-2 Benzo(g,h,i)perylene 1.3E-09 1.9E-08 1.5E-08 --
207-08-9 Benzo(k)fluoranthene 1.7E-08 2.3E-10 9.9E-09 f, g --
h
7440-41-7 Beryllium 1.3E-07 1.9E-06 1.9E-07 3E-06
106-99-0 1,3-Butadiene -- 1.4E-04 -- --
106-97-8 Butane 2.1E-03 d -- -- --
7440-43-9 Cadmium 1.1E-06 c 2.2E-06 2.7E-06 h 3E-06
16887-00-6 Chloride -- -- 2.3E-03 h --
67-66-3 Chloroform -- 6.0E-05 3.4E-05 --
(continued)
4-13
Table 4-3. Summary of Emission Factors for Boilers and Process Heaters Firing Various Fuels (continued)

91-58-7 2-Chloronaphthalene -- 8.2E-08 1.5E-10 --

18540-29-9 Chromium (hexavalent) 2.8E-07 l 1.1E-06 1.7E-06 h --
7440-47-3 Chromium (total) 1.4E-06 c 8.7E-06 5.6E-06 h 3E-06
218-01-9 Chrysene 1.6E-09 7.5E-08 1.6E-08 f --
7440-48-4 Cobalt 8.2E-08 c -- 4.0E-05 h --
c h
7440-50-8 Copper 8.3E-07 9.5E-06 1.2E-05 6E-06
53-70-3 Dibenz(a,h)anthracene 1.2E-09 d 1.2E-08 1.1E-08 f --
d
25321-22-6 Dichlorobenzene 1.2E-06 -- -- --
7,12-Dimethylbenz(a)
57-97-6 1.6E-08 d -- -- --
anthracene
1746-01-6 Dioxin: 4D 2378j -- 3.5E-12 4.3E-12 --
40321-76-4 Dioxin: 5D 12378j -- 1.7E-11 2.5E-12 --
39227-28-6 Dioxin: 6D 123478j -- 1.5E-11 2.5E-12 --
57653-85-7 Dioxin: 6D 123678j -- 2.1E-11 2.5E-12 --
19408-74-3 Dioxin: 6D 123789j -- 3.3E-11 2.5E-12 --
35822-46-9 Dioxin: 7D 1234678j -- 9.3E-11 2.1E-11 --
3268-87-9 Dioxin: 8Dj -- 3.3E-10 2.1E-11 f --
74-84-0 Ethane 3.0E-03 d -- -- --
f
100-41-4 Ethylbenzene 1.6E-05 -- 4.2E-07 --
206-44-0 Fluoranthene 2.9E-09 d 6.9E-08 3.2E-08 f --
86-73-7 Fluorene 2.7E-09 d 2.0E-07 3.0E-08 f --
(continued)
4-14

16984-48-8 Fluoride -- -- 2.5E-04 h --

50-00-0 Formaldehyde 7.4E-05 d 1.1E-05 2.2E-04 f --
51207-31-9 Furan: 4F 2378k -- 6.2E-10 5.5E-12 --
57117-41-6 Furan: 5F 12378k -- 6.0E-11 3.1E-12 --
57117-31-4 Furan: 5F 23478k -- 1.1E-10 3.1E-12 --
70648-26-9 Furan: 6F 123478k -- 1.3E-10 2.5E-12 --
57117-44-9 Furan: 6F 123678k -- 4.3E-11 1.9E-12 --
72918-21-9 Furan: 6F 123789k -- 3.5E-12 2.5E-12 --
60851-34-5 Furan: 6F 234678k -- 6.1E-11 3.7E-12 --
67562-39-4 Furan: 7F 1234678k -- 1.4E-10 9.8E-12 --
55673-89-7 Furan: 7F 1234789k -- 8.3E-12 3.2E-12 --
39001-02-0 Furan: 8Fk -- 7.3E-11 4.9E-11 --
110-54-3 Hexane 1.8E-03 d -- -- --
7783-06-4 Hydrogen sulfide 8.5E-05 -- -- --
f
193-39-5 Indeno(1,2,3-cd)pyrene 7.1E-08 1.9E-08 1.4E-08 --
7439-92-1 Lead 4.9E-07 e 1.9E-06 1.0E-05 h 9E-06
c h
7439-96-5 Manganese 3.7E-07 1.8E-05 2.0E-05 6E-06
7439-97-6 Mercury 2.5E-07 c 1.0E-05 7.5E-07 h 3E-06
56-49-5 3-Methylchloranthrene 1.8E-09 d -- -- --
91-57-6 2-Methylnaphthalene 2.4E-08 d 2.5E-07 7.4E-08 --
7439-98-7 Molybdenum 1.1E-06 c -- 5.2E-06 h --
(continued)
4-15

91-20-3 Naphthalene 6.0E-07 d 5.5E-06 7.5E-06 f --

7440-02-0 Nickel 2.1E-06 c 2.4E-03 5.6E-04 h 3E-06
109-66-0 Pentane 2.5E-03 -- -- --
198-55-0 Perylene -- 5.2E-10 7.4E-10 --
85-01-8 Phenanthrene 1.7E-08 d 1.7E-07 7.0E-08 f --
108-95-2 Phenol 4.0E-06 -- -- --
7723-14-0 Phosphorus 6.4E-07 1.8E-04 6.3E-05 h --
d
74-98-6 Propane 1.6E-03 -- -- --
115-07-1 Propylene 1.5E-04 4.4E-05 -- --
129-00-0 Pyrene 4.9E-09 d 1.2E-07 2.8E-08 f --
7782-49-2 Selenium 8.8E-07 7.9E-06 4.5E-06 h 1.5E-05
71-55-6 1,1,1-Trichloroethane -- -- 1.6E-06 f --
108-88-3 Toluene 3.3E-06 d 3.5E-05 4.1E-05 f --
7440-62-2 Vanadium 2.3E-06 c -- 2.1E-04 h --
f
95-47-6 Xylene (o) -- -- 7.3E-06 --
1330-20-7 Xylene (total) 2.5E-05 2.9E-06 -- --
c h
7440-66-6 Zinc 2.8E-05 4.2E-04 1.9E-04 4E-06
(continued)
4-16
Note: lb/MMBtu = pounds per million British thermal units

a Bold italic values indicate that all test runs were below detection limit; underlined values indicate that 75% or more of the test runs were below the detection
limit.
b Source: Hansell and England, 1998, unless otherwise specified.
c Source: U.S. EPA, 1995a, AP-42 Section 1.4, Table 1.4-4. Emission factors were provided for natural gas combustion sources. Values converted from lb/106
scf to lb/MMBtu by dividing by 1,020 MMBtu/106 scf.

d Source: U.S. EPA, 1995a, AP-42 Section 1.4, Table 1.4-3. Emission factors were provided for natural gas combustion sources. Values converted from lb/106

e Source: U.S. EPA, 1995a, AP-42 Section 1.4, Table 1.4-2. Emission factors were provided for natural gas combustion sources. Values converted from lb/106

f Source: U.S. EPA, 1995a, AP-42 Section 1.3, Table 1.3-9. Emission factors were provided specifically for residual oil (No. 6 oil) fired boilers. Values converted
from lb/103 Gal to lb/MMBtu by dividing by 150 MMBtu/103 Gal.

g Emission factor includes both benzo(b)fluoranthene and benzo(k)fluoranthene.
h Source: U.S. EPA, 1995a, AP-42 Section 1.3, Table 1.3-11. Emission factors were provided specifically for residual oil (No. 6 oil) fired boilers. Values
converted from lb/103 Gal to lb/MMBtu by dividing by 150 MMBtu/103 Gal.

i Source: U.S. EPA, 1995a, AP-42 Section 1.3, Table 1.3-10. Emission factors were provided specifically for distillate oil-fired boilers. Values converted from
lb/1012 Btu to lb/MMBtu by dividing by 106.

j 4D 2378 = 2,3,7,8-Tetrachlorodibenzo-p-dioxin; 5D 12378 = 1,2,3,7,8-Pentachlorodibenzo-p-dioxin; 6D 123478 = 1,2,3,4,7,8-Hexachlorodibenzo-p-dioxin; 6D
123678 = 1,2,3,6,7,8-Hexachlorodibenzo-p-dioxin; 6D 123789 = 1,2,3,7,8,9-Hexachlorodibenzo-p-dioxin; 7D 1234678 = 1,2,3,4,6,7,8-Heptachlorodibenzo-p-

dioxin; 8D = Octachlorodibenzo-p-dioxin.
k 4F 2378 = 2,3,7,8-Tetrachlorodibenzofuran; 5F 12378 = 1,2,3,7,8-Pentachlorodibenzofuran; 5F 23478 = 2,3,4,7,8-Pentachlorodibenzofuran; 6F 123478 =
1,2,3,4,7,8-Hexachlorodibenzofuran; 6F 123678 = 1,2,3,6,7,8-Hexachlorodibenzofuran; 6F 123789 = 1,2,3,7,8,9-Hexachlorodibenzofuran; 6F 234678 =

2,3,4,6,7,8-Hexachlorodibenzofuran; 7F 1234678 = 1,2,3,4,6,7,8-Heptachlorodibenzofuran; 7F 1234789 = 1,2,3,4,7,8,9-Heptachlorodibenzofuran; 8F =
Octachlorodibenzofuran.
l Hexavalent chromium was not detected. Twenty percent of the total chromium emission factor is assumed for the hexavalent chromium emission factor. The
hexavalent chromium emission factors for gas turbines are between 10 and 15 percent of the total chromium emission factor, and the hexavalent chromium
emission factor for boilers firing fuel oil is about 30 percent of the total chromium emission factor (based on detection limits, which are expected to overstate
hexavalent chromium emissions), so 20 percent is a conservative (reasonable high-end) estimate for hexavalent chromium emissions from refinery fuel gas.
4-17
4.5.2 Default Emission Factor for Internal Combustion Engines

The emissions from reciprocating internal combustion engines were dependent on: (1) the type of fuel
used (diesel fuel, or gas [field gas, or natural gas]), (2) the type of engine used (two-stroke versus four-
stroke engines), and (3) the fuel-to-air ratio. As shown in Table 4-4, high air rates (or lean fuel mixtures)
tend to have higher emissions of acetaldehyde and formaldehyde, whereas lean gas mixtures tend to have
higher emissions of benzene, toluene, and xylenes.
4.5.3 Default Emission Factors for Combustion Turbines

The emissions from combustion turbines were dependent on (1) the type of fuel used (e.g., distillate fuel
oil, or gas [refinery fuel gas, liquefied petroleum gas, or natural gas]), and (2) the presence or absence of
duct burners. No distinctions in HAP emission factors were made for criteria pollutant controls, such as
selective catalytic reduction (SCR), or carbon monoxide oxidation catalyst. Table 4-5 presents a
summary of emission factors for combustion turbines.
4-18
Table 4-4. Summary of Emission Factors for Internal Combustion Engines Firing Various Fuels
Emission Factor for Engine and Fuel Type
Natural Gas/Field Natural Natural

Diesel, Diesel, Diesel, Gas, Gas/Field Gas, Gas/Field Gas,
CAS O2 < 13% O2 > 13% O2 not specified Lean (2-Stroke) Lean (4-Stroke) Rich (4-Stroke)
Number Substance (lb/MMBtu) a,b (lb/MMBtu) a,b (lb/MMBtu) c (lb/MMBtu) d (lb/MMBtu) e (lb/MMBtu) j
83-32-9 Acenaphthene 4.5E-06 1.4E-06 < 1.4E-06 f 1.3E-06 g 1.3E-06 h -- i

208-96-8 Acenaphthylene 9.0E-06 5.1E-06 < 5.1E-06 f 3.2E-06 g 5.5E-06 h -- i
75-07-0 Acetaldehyde 2.4E-05 7.6E-04 7.7E-04 7.8E-03 8.4E-03 2.8E-03
107-02-8 Acrolein 7.6E-06 9.4E-05 < 9.3E-05 7.8E-03 5.1E-03 2.6E-03
120-12-7 Anthracene 1.2E-06 1.9E-06 1.9E-06 f 7.2E-07 g 2.4E-07 a,h -- i
100-52-7 Benzaldehyde -- 9.0E-05 -- -- -- --
71-43-2 Benzene 7.1E-04 8.8E-04 9.3E-04 1.9E-03 4.4E-04 1.6E-03
f g a,h
56-55-3 Benzo(a)anthracene 6.1E-07 1.7E-06 1.7E-06 3.4E-07 7.4E-08 -- i
50-32-8 Benzo(a)pyrene 2.5E-07 1.0E-08 < 1.9E-07 f 5.7E-09 g 3.4E-08 a,h -- i
205-99-2 Benzo(b)fluoranthene 1.1E-06 1.9E-07 < 9.9E-08 f 8.5E-09 g 1.7E-07 h -- i
192-97-2 Benzo(e)pyrene -- -- -- 2.3E-08 g 4.2E-07 h -- i
191-24-2 Benzo(g,h,i)perylene 5.4E-07 4.1E-07 < 4.9E-07 f 2.5E-08 g 4.1E-07 h -- i
207-08-9 Benzo(k)fluoranthene 2.1E-07 3.0E-07 < 1.6E-07 f 4.3E-09 g 5.0E-07 a,h -- i
92-52-4 Biphenyl -- -- -- 4.0E-06 2.1E-04 --
106-99-0 1,3-Butadiene -- 3.9E-05 < 3.9E-05 8.2E-04 2.7E-04 6.6E-04
106-97-8 Butane -- -- -- 4.8E-03 5.4E-04 --
Butyr/Isobutyraldehyde -- -- -- 4.4E-04 1.0E-04 4.9E-05
56-23-5 Carbon tetrachloride -- -- -- 6.1E-05 < 3.7E-05 < 1.8E-05
108-90-7 Chlorobenzene -- -- -- 4.4E-05 < 3.0E-05 < 1.3E-05
75-00-3 Chloroethane -- -- -- -- 1.9E-06 --
(continued)
4-19
Table 4-4. Summary of Emission Factors for Internal Combustion Engines Firing Various Fuels (continued)

67-66-3 Chloroform -- -- -- 4.7E-05 < 2.9E-05 < 1.4E-05

f g h
218-01-9 Chrysene 1.5E-06 3.5E-07 3.5E-07 6.7E-07 6.9E-07 -- i
110-82-7 Cyclohexane -- -- -- 3.1E-04 -- --
287-92-3 Cyclopentane -- -- -- 9.5E-05 2.3E-04 --
53-70-3 Dibenz(a,h)anthracene 3.4E-07 4.1E-07 < 5.8E-07 f -- g 1.0E-08 a,h -- i
75-34-3 1,1-Dichloroethane -- -- -- 3.9E-05 < 2.4E-05 < 1.1E-05
107-06-2 1,2-Dichloroethane -- -- -- 4.2E-05 < 2.4E-05 < 1.1E-05
78-87-5 1,2-Dichloropropane -- -- -- 4.5E-05 < 2.7E-05 < 1.3E-05
542-75-6 1,3-Dichloropropene -- -- -- 4.4E-05 < 2.6E-05 < 1.3E-05
74-84-0 Ethane -- -- -- 7.1E-02 1.1E-01 7.0E-02
100-41-4 Ethylbenzene -- -- -- 1.1E-04 4.0E-05 < 2.5E-05
106-93-4 Ethylene dibromide -- -- -- 7.3E-05 < 4.4E-05 < 2.1E-05
206-44-0 Fluoranthene 3.9E-06 7.6E-06 7.6E-06 f 3.6E-07 g 1.1E-06 h -- i
86-73-7 Fluorene 1.2E-05 2.9E-05 2.9E-05 f 1.7E-06 g 5.7E-06 h -- i
50-00-0 Formaldehyde 7.7E-05 1.2E-03 1.2E-03 5.5E-02 5.3E-02 2.1E-02
110-54-3 n-Hexane -- -- -- 4.5E-04 1.1E-03 --
193-39-5 Indeno(1,2,3-cd)pyrene 4.0E-07 2.7E-07 < 3.8E-07 f 9.9E-09 g 1.1E-07 a,h -- i
75-28-5 Isobutane -- -- -- 3.8E-03 -- --
67-56-1 Methanol -- -- -- 2.5E-03 2.5E-03 3.1E-03
108-87-2 Methylcyclohexane -- -- -- 3.4E-04 1.2E-03 --
(continued)
4-20

75-09-2 Methylene chloride -- -- -- 1.5E-04 2.0E-05 4.1E-05

g h
91-57-6 2-Methylnaphthalene -- -- -- 2.1E-05 3.3E-05 -- i
91-20-3 Naphthalene 1.3E-04 8.5E-05 8.5E-05 f 9.6E-05 g 7.4E-05 h < 9.7E-05 i
111-84-2 n-Nonane -- -- -- 3.1E-05 1.1E-04 --
111-65-9 n-Octane -- -- -- 7.4E-05 3.5E-04 --
109-66-0 n-Pentane -- -- -- 1.5E-03 2.6E-03 --
g h
198-55-0 Perylene -- -- -- 5.0E-09 -- -- i
85-01-8 Phenanthrene 4.0E-05 2.9E-05 2.9E-05 f 3.5E-06 g 1.0E-05 h -- i
108-95-2 Phenol -- -- -- 4.2E-05 2.4E-05 --
74-98-6 Propane -- -- -- 2.9E-02 4.2E-02 --
a a
115-07-1 Propylene 2.7E-03 2.6E-03 2.6E-03 2.4E-02 1.7E-02 2.0E-02 a
129-00-0 Pyrene 3.6E-06 4.8E-06 4.8E-06 f 5.8E-07 g 1.4E-06 h -- i
100-42-5 Styrene -- -- -- 5.5E-05 < 2.4E-05 < 1.2E-05
79-34-5 1,1,2,2-Tetrachloroethane -- -- -- 6.6E-05 < 4.0E-05 2.5E-05
108-88-3 Toluene 2.6E-04 4.0E-04 4.1E-04 9.6E-04 4.1E-04 5.6E-04
79-00-5 1,1,2-Trichloroethane -- -- -- 5.3E-05 < 3.2E-05 < 1.5E-05
526-73-8 1,2,3-Trimethylbenzene -- -- -- 3.5E-05 2.3E-05 --
540-84-1 2,2,4-Trimethylpentane -- -- -- 8.5E-04 2.5E-04 --
(continued)
4-21

75-01-4 Vinyl chloride -- -- -- 2.5E-05 1.5E-05 < 7.2E-06

108-38-3,
Xylene (m,p) -- 1.5E-04 -- 5.8E-04 a 1.4E-04 a 4.7E-04 a
106-42-3
95-47-6 Xylene (o) -- 1.5E-04 -- 2.7E-04 a 5.9E-05 a 2.3E-04 a
1330-20-7 Xylene (total) 1.9E-04 2.6E-04 2.9E-04 2.7E-04 1.8E-04 2.0E-04
a Source: Hansell and England, 1998.
b Bold italic values indicate that all test runs were below detection limit.
c Source: U.S. EPA, 1995a, AP-42 Section 3.3, Table 3.3-2. Emission factors are for uncontrolled sources. Emission factors preceded with a less than symbol
are based on method detection limits. Note that these emission factors are for diesel engines up to 600 hp. Emission factors for diesel engines greater than
600 hp can be found in AP-42 Section 3.4.
d Source: U.S. EPA, 1995a, AP-42 Section 3.2, Table 3.2-1. Emission factors are for uncontrolled sources.
e Source: U.S. EPA, 1995a, AP-42 Section 3.2, Table 3.2-2. Emission factors are for uncontrolled sources. Emission factors preceded with a less than symbol
are based on method detection limits.

f
The emission factor for total polycyclic aromatic hydrocarbons (PAH) in U.S. EPA, 1995a, Table 3.3-2 is 1.7E-04 lb/MMBtu.
g The emission factor for total PAH in U.S. EPA, 1995a, Table 3.2-1 is 1.3E-04 lb/MMBtu.
h The emission factor for total PAH in U.S. EPA, 1995a, Table 3.2-2 is 2.7E-05 lb/MMBtu.
i The emission factor for total PAH in U.S. EPA, 1995a, Table 3.2-3 is 1.4E-04 lb/MMBtu.
j Source: U.S. EPA, 1995a, AP-42 Section 3.2, Table 3.2-3. Emission factors are for uncontrolled sources. Emission factors preceded with a less than symbol
are based on method detection limits.
4-22
Table 4-5. Summary of Emission Factors for Combustion Turbines Firing Various Fuelsa
Emission Factors for Combustion Turbine and Fuel Type a, b
Natural Gas/Refinery Natural Gas/Refinery Fuel

Distillate Oil, Type Natural Gas, Type Fuel Gas/Liquefied Gas/Liquefied Petroleum
not specified not specified Petroleum Gas, Duct Gas, No Duct Burners
CAS Number Substance (lb/MMBtu) c (lb/MMBtu) d Burners (lb/MMBtu) (lb/MMBtu)
83-32-9 Acenaphthene -- e -- f 2.2E-08 3.3E-09

208-96-8 Acenaphthylene -- e -- f 1.1E-08 2.9E-09
75-07-0 Acetaldehyde -- 4.0E-05 4.1E-06 2.7E-05
107-02-8 Acrolein -- 6.4E-06 -- 1.7E-05
120-12-7 Anthracene -- e -- f 2.5E-08 3.4E-08
7440-38-2 Arsenic < 1.1E-05 -- -- --
71-43-2 Benzene 5.5E-05 1.2E-05 -- --
e f
56-55-3 Benzo(a)anthracene -- -- 1.5E-08 2.8E-09
50-32-8 Benzo(a)pyrene -- e -- f -- --
e f
205-99-2 Benzo(b)fluoranthene -- -- 2.5E-08 3.3E-09
191-24-2 Benzo(g,h,i)perylene -- e -- f -- 1.9E-09
207-08-9 Benzo(k)fluoranthene -- e -- f -- 2.3E-09
7440-41-7 Beryllium < 3.1E-07 -- -- --
106-99-0 1,3-Butadiene < 1.6E-05 < 4.3E-07 -- --
7440-43-9 Cadmium 4.8E-06 -- 2.9E-06 5.3E-06
18540-29-9 Chromium (hexavalent) 2.2E-06 g -- 7.0E-06 1.5E-06
7440-47-3 Chromium (total) 1.1E-05 -- 5.0E-05 1.3E-05
218-01-9 Chrysene -- e -- f 1.1E-07 4.9E-09
7440-50-8 Copper -- -- 1.2E-05 4.1E-05
53-70-3 Dibenz(a,h)anthracene -- e -- f -- --
(continued)
4-23
Table 4-5. Summary of Emission Factors for Combustion Turbines Firing Various Fuelsa (continued)
Emission Factors for Combustion Turbine and Fuel Type a, b
Natural Gas/Refinery Natural Gas/Refinery Fuel

Distillate Oil, Type Natural Gas, Type Fuel Gas/Liquefied Gas/Liquefied Petroleum
not specified not specified Petroleum Gas, Duct Gas, No Duct Burners
CAS Number Substance (lb/MMBtu) c (lb/MMBtu) d Burners (lb/MMBtu) (lb/MMBtu)
100-41-4 Ethylbenzene -- 3.2E-05 -- --

e f
206-44-0 Fluoranthene -- -- 9.9E-08 1.2E-08
86-73-7 Fluorene -- e -- f 1.8E-07 1.5E-08
50-00-0 Formaldehyde 2.8E-04 7.1E-04 3.1E-03 3.1E-04
7783-06-4 Hydrogen sulfide -- -- -- --
193-39-5 Indeno(1,2,3-cd)pyrene -- e -- f -- 1.8E-09
7439-92-1 Lead 1.4E-05 -- 3.6E-05 2.8E-05
7439-96-5 Manganese 7.9E-04 -- 4.8E-05 1.3E-04
7439-97-6 Mercury 1.2E-06 -- 4.4E-06 1.5E-05
91-20-3 Naphthalene 3.5E-05 e 1.3E-06 f 3.7E-05 7.3E-07
7440-02-0 Nickel < 4.6E-06 -- 7.7E-05 1.7E-04
85-01-8 Phenanthrene -- e -- f 6.4E-07 6.5E-08
108-95-2 Phenol -- -- 2.2E-05 6.7E-06
115-07-1 Propylene -- -- -- 1.6E-03
129-00-0 Pyrene -- e -- f 1.2E-07 2.3E-08
7782-49-2 Selenium < 2.5E-05 -- -- --
108-88-3 Toluene -- 1.3E-04 1.6E-04 3.1E-04
1330-20-7 Xylene (total) -- 6.4E-05 3.7E-04 7.7E-04
7440-66-6 Zinc -- -- 1.2E-04 5.0E-03
(continued)
4-24
Table 4-5. Summary of Emission Factors for Combustion Turbines Firing Various Fuelsa (continued)

a Source: Hansell and England, 1998, Volume 1, unless otherwise specified.
b Bold italic values indicate that all test runs were below detection limit.
c Source: U.S. EPA, 1995a, Section 3.1, Table 3.1-4 and Table 3.1-5. Emission factors are based on an average distillate oil heating value (HHV) of 139
MMBtu/103 gal.
d Source: U.S. EPA, 1995a, Section 3.1, Table 3.1-3. Emission factors are based on an average natural gas heating value (HHV) of 1,020 Btu/scf at 60°F.
e The emission factor for total polycyclic aromatic hydrocarbons (PAH) is 4.0E-05 lb/MMBtu.
f The emission factor for total PAH is 2.2E-06 lb/MMBtu.
g Hexavalent chromium is assumed to be twenty percent of the total chromium emission factor for the hexavalent chromium emission factor. The hexavalent
chromium emission factors for boilers firing fuel oil is about 30 percent of the total chromium emission factor (based on detection limits, which are expected to
overstate hexavalent chromium emissions) and 18 percent of the total chromium for oil-fired utilities (U.S. EPA, 1998c), so 20 percent is expected to be a
reasonable approximation for hexavalent chromium emissions from distillate oil-fired combustion turbines.
4-25
4-26
Version 3 Section 5—Process Vents
5. Process Vents
There are a variety of processes and equipment at petroleum refineries that may release pollutants directly
into the atmosphere through process vents. Many of these process vents may be controlled using a flare,
thermal incinerator, or other air pollution control techniques. This section describes the key process vents
and emissions estimation procedures considering controls. At some facilities, the process gases may be
routed to the refinery’s fuel gas system rather than directly to a control device to the atmosphere. These
gases would then be combusted with the fuel gas in either a stationary combustion unit (process heater or
boiler) or in a flare. As such, emissions from these gases are expected to be included in the emissions
estimated for stationary combustion sources (see Section 4) or for flares (see Section 6). The methods in
this section should be used for process vent gases other than those recovered into the refinery’s fuel gas
system. Particularly, this section provides methods for estimating the emissions associated with catalytic
cracking units (CCUs), coking units, CRUs, sulfur recovery plants, and other process vents (e.g., from
hydrogen (H2) plants, asphalt blowing stills, coke calcining units, blowdown system, and vacuum
generating units).
Table 5-1 summarizes the general hierarchy of process vent emissions measurement or estimation
methods, which are ranked in terms of anticipated accuracy. Within a given measurement method (or
rank), there may be alternative methods for determining the constituent-specific emissions. Methodology
Ranks 1 and 2 for process vents (i.e., the use of CEMS) have been previously described in Section 4,
Stationary Combustion Sources. Although Section 5 will note these methods, where applicable, for
individual process vents, Section 4 should be reviewed for sample calculations and other important
considerations when using CEMS to determine mass emissions rates.
Table 5-1. Summary of Typical Hierarchy of Process Vent Emissions Estimates
Rank Flow Estimate Method Compositional Analysis Data

1 Continuous flow meter Continuous gas composition analyzer
2 Engineering estimates (e.g., F factor) Continuous gas composition analyzer
3 Continuous flow meter or engineering estimates Occasional grab samples
4 Measured process rates Site-specific emissions factor based on source test
5 Measured process rates Default emissions factors
5.1 Catalytic Cracking Units

The CCU catalyst regenerator vent is often the single largest emissions vent at the refinery. The CCU is a
catalytic process used to upgrade (crack) heavy distillates to form lighter, more useful distillates such as
heating oils or gasoline. The CCU system consists of a reactor, a catalyst regenerator (commonly referred
to as the “regenerator”), vent gas process equipment for energy recovery and emissions control, and an
exhaust stack. Nearly all refinery CCU systems operate as fluidized-bed reactors and use air or oil gas
flow to transport the very small catalyst particles between the CCU reactor and regenerator. These
fluidized CCU systems are commonly referred to as fluid catalytic cracking units (FCCU). There are two
or three refineries that use larger catalyst pellets in moving bed-type reactor system, which is commonly
referred to as a thermal catalytic cracking unit (TCCU). Nearly all of the available emissions data are
specific to FCCU, although data for PM, SO2 and NOX emissions are available for one TCCU. The
emissions from the TCCU, when normalized on the basis of the coke burn-off rate, are similar to those of
a similarly operated (uncontrolled) FCCU. Consequently, while the emissions inventory methods
presented in this section are based primarily on FCCU operation and emissions data, these emissions
inventory methods should be used for TCCU systems unless direct emissions data are available.
5-1
The CCU catalyst regenerator vent releases a wide variety of pollutants, including PM, SO2, NOx, carbon
monoxide (CO), VOC, metal HAP, organic HAP, and ammonia. As a by-product of the cracking
reactions, coke is deposited on the catalyst particles. The coke reduces the activity of the catalyst, and the
spent catalyst that is returned from the CCU reactor is regenerated continuously by burning off coke in
the CCU catalyst regenerator. There are two basic types of CCU regenerators: complete combustion
regenerators and partial combustion regenerators. In a complete combustion regenerator, the regenerator
is typically operated at approximately 1,200°F to 1,400°F with excess O2 and low levels (< 500 ppmv) of
CO in the exhaust flue gas. In a partial (or incomplete) combustion regenerator, the regenerator is
typically operated at approximately 1,000°F to 1,200°F under O2-limited conditions and relatively high
levels (1% to 5%) of CO. Prior to exiting the regenerator, catalyst particles entrained with the flue gases
are initially removed by internal cyclone separators and returned to the regenerator catalyst bed for
recirculation to the reactor.
On a routine basis, a small portion of the circulating catalyst, commonly referred to as equilibrium
catalyst (E-cat), is removed from the system and fresh catalyst is added to maintain catalyst activity and
replace catalyst lost or removed from the system. Various other additives may also be included with the
catalyst additions to reduce NOx or SO2 emissions or promote complete combustion (thereby reducing CO
emissions). Although the CCU vent is the primary emissions source associated with the CCU, fugitive
dust emissions associated with handling the fresh catalyst or spent E-cat should also be calculated and
included in the emissions inventory for CCU systems.
After the flue gas exits the CCU regenerator, a variety of energy recovery or emissions control systems
may be used to reduce pollutant releases. For a complete combustion CCU, a waste heat boiler is often
used to recover the latent heat of the flue gas. The energy recovery system also serves to cool the flue gas
prior to PM or other add-on control systems. The most common control systems for CCU are electrostatic
precipitators (ESP) or wet scrubbers. For a partial combustion CCU, the first control system is typically a
CO boiler used to combust the CO in the flue gas to CO2 and recover the heat of combustion and latent
heat of the flue gas (typically by producing steam).
5.1.1 Methodology Ranks 1 and 2 for Catalytic Cracking Units

For SO2, NOx, CO, and VOC, it is anticipated that many CCUs will have CEMS for measuring the
composition of these pollutants in the exhaust gas. Although few CCUs are expected to have PM CEMS,
if PM CEMS are used, these measurements would also qualify as Methodology Rank 1 or 2 for CCUs
even though some “PM augmentation” will be needed as described in Sections 5.1.2.1 and 5.1.2.2.
Similarly, if CEMS are used for specific HAPs, these measurements should be used and reported as
Methodology Rank 1 or 2 for CCUs. Gas flow rate may be directly monitored, but in many cases, the
exhaust flow rate will be calculated based on air blast rates and composition monitors (similar to F factors
based on coke burn-off rates). Equations in 40 CFR Part 63 Subpart UUU can be used to estimate flow
rates based on gas composition analyzers that are typically used to monitor the regenerator combustion
parameters. When CEMS or direct-flow monitors are present, these data should be used to calculate
annual pollutant emissions.
5.1.2 Methodology Ranks 3 and 4 for Catalytic Cracking Units

For VOC, PM, and specific HAP, it is anticipated that many refineries will have performed source tests of
their CCUs. Generally, testing will be conducted infrequently so the testing results represent an emissions
rate at a given point in time and production level. Rather than assuming that the mass emissions rate
measured during the test occurs continuously for all process operating hours, the resulting source test data
are most appropriately used for developing a site-specific emissions factor. This type of emissions factor
is the measured emissions rate (lb/hr) divided by the processing rate (throughput per hr). Generally, the
emissions from the CCU catalyst regenerator vent are best correlated with coke burn-off rates; however, if
CCU feed and operating conditions do not vary significantly, then throughput-based emissions factors
5-2
may be used. If multiple tests have been conducted on the CCU and no significant modifications have
been made on the CCU or its control system, then an arithmetic average of the emission factors should be
used to estimate annual average emissions. If modifications have been made on the CCU or its control
system, the most recently performed source test should be used to develop the site-specific emissions
factor. To estimate maximum hourly emissions, the highest emission factor developed from the individual
runs should be used along with the maximum capacity (or coke burn-off rate) of the CCU.
Example 5-1: Development of Site-Specific Emissions Factor

A source test was performed to determine the PM emissions rate from a CCU. Three test runs
were conducted; the PM emissions rates measured during the three test runs were 20.2, 25.1,
and 17.6 lb/hr for tests 1, 2, and 3, respectively. The processing rates during the three runs were
1,600; 1,700; and 1,500 barrels per hour (bbl/hr) and the coke burn-off rates were 25,000;
29,000; and 23,000 lb/hr for tests 1, 2, and 3, respectively. An appropriate emissions factor for
the CCU can be developed to project annual emissions.
First, the emissions factor for each individual run should be calculated, and then the emissions
factors should be averaged. Emissions factors can be assessed using different normalizing
factors, such as throughput and coke burn-off rates, as follows:
– Run 1: Emissions/throughput = 20.2 (lb/hr) ÷ 1,600 (bbls/hr) = 0.0126 lb/bbl = 12.6 pounds per
thousands of barrels (lb/Mbbl)
Emissions/coke burn-off = 20.2 (lb/hr) ÷ 25 (1,000 lb/hr) = 0.808 lb/1,000 lbs coke burn-off
– Run 2: Emissions/throughput = 25.1 (lb/hr) ÷ 1.7 (Mbbl/hr) = 14.8 lb/Mbbl
– Run 3: Emissions/throughput = 17.6 (lb/hr) ÷ 1.5 (Mbbl/hr) = 11.7 lb/Mbbl
– Average: Emissions/throughput = (12.6 + 14.8 + 11.7) ÷ 3 = 13.0 lb/Mbbl
Emissions/coke burn-off = (0.808 + 0.866 + 0.765) ÷ 3 = 0.813 lb/1,000 lb coke burn-off
There are a variety of ways to determine which emissions factor is most appropriate. One
method is to compare the range of the test runs compared to the three-run average. For the
throughput-based emissions factors, the highest single-run emissions factor is 14% (100% ×
[14.8 − 13] ÷ 13) higher than the average, and the lowest single-run emissions factor is 10%
lower than the average. For the coke burn-off rate emissions factors, the highest single-run
emissions factor is 6.5% (100% × [0.866 − 0.813] ÷ 0.813) higher than the average, and the
lowest single-run emissions factor is 5.9% lower than the average. The smaller range for the
coke burn-off emissions factors (as a percentage of the average) suggests that normalizing the
emissions by coke burn-off accounts for more of the differences in the observed emissions than
does throughput. Consequently, the average coke burn-off emissions factor would be preferred
to the throughput-based emissions factor in this example.
The highest individual run emission factor, 0.866 lb/1,000 lbs coke burn-off in this example,
would be used, along with the maximum hourly coke burn-off rate during the year, to estimate
the maximum hourly emission rate from this CCU.
Although site-specific emissions factors can be developed for PM, VOC, organic HAP species, or metal
HAP species using source test data, there are some unique nomenclature and reporting practices for PM
emissions that deserve additional guidance. Additionally, a quasi Methodology Rank 4 for CCU metal
5-3
HAP emissions estimates based on a site-specific PM emissions rate has been developed and requires
additional explanation. The following subsections discuss additional guidance on providing a complete
PM emissions estimate and on estimating metal HAP emissions from PM source test information.
5.1.2.1 PM Emissions Inventory and Test Method Considerations

PM emissions inventories have their own nomenclature and structure. To correctly report PM emissions,
an understanding of the reporting nomenclature and the PM testing methods is needed. This subsection
provides guidance and background information regarding PM emissions, with a focus on the PM test
methods, particularly those methods commonly used when testing the CCU vent.
PM Emissions Inventory Nomenclature

A complete PM emissions inventory includes the following components:
 PM10-PRI: “Primary” PM emissions that are 10 µm in diameter or less. PM10-PRI = PM10-FIL + PM-CON.
 PM10-FIL: Filterable (or front-half catch) portion of the PM emissions that are 10 µm in diameter or less.
 PM-CON: Condensable PM (or back-half catch). All condensable PM is assumed to be less than 2.5 microns
(µm) in diameter (PM2.5).
 PM2.5-PRI: “Primary” PM emissions that are 2.5 µm in diameter or less. PM2.5-PRI = PM25-FIL + PM-CON.
 PM2.5-FIL: Filterable (or front-half catch) portion of the PM emissions that are 2.5 µm in diameter or less.
Although a complete PM emissions inventory includes PM emissions that are 10 µm in diameter or less, some
measurement methods also collect PM particles that are greater than 10 µm in diameter. The following
nomenclature is used to designate PM emissions that include PM greater than 10 µm in diameter:
 PM-PRI: “Primary” PM emissions of any particle size. PM-PRI = PM-FIL + PM-CON.
 PM-FIL: Filterable (or front-half catch) portion of the PM emissions of any particle size.
EPA Methods 5, 5B, and 5F are the most commonly used test methods for measuring PM emissions from
CCUs. A typical Method 5 sampling train consists of a sampling probe, a heated line and filter, and a
series of impingers that are kept in an ice bath. Method 5, 5B, or 5F sampling measures PM that is
contained in the sampling probe and filter, which is often referred to as the “front-half” or “filterable” PM
catch. PM that condenses in the impinger section of the sampling train is often referred to as the “back
half” catch or the “condensable” PM.
One of the main differences between EPA Methods 5, 5B, and 5F is that the sampling line and filter are
maintained at 250°F for Method 5 compared to 325°F for Methods 5B and 5F. Although generally not
used for the CCU, there are other EPA test methods for PM. EPA Method 17 is similar to Method 5,
except that the filter in the Method 17 sampling probe is within the stack so that the “filterable” PM
content is measured at the stack temperature. EPA Method 201 is similar to Method 17, except that there
is also a cyclone or other sizing device to remove particles greater than 10 µm in diameter prior to the
filter so that Method 201 determines PM10-FIL directly. EPA Method 201A, formerly EPA Other Test
Method 27 (OTM 27), is similar to EPA Method 201 sampling at stack temperature, but uses a series of
cylones to filter out, in succession, particles greater than 10 µm and 2.5 µm in diameter. The particle
catch from the second cyclone can be used to determine the PM emissions that are less than 10 µm in
diameter but greater than 2.5 µm in diameter. The “filterable” catch from EPAMethod 201A is a direct
measure of PM2.5-FIL. Therefore, EPA Method 201A provides a means to directly determine PM-FIL,
PM10-FIL, and PM2.5-FIL. PM CEMS, if used, generally measure only the filterable PM (i.e., PM-FIL).
Any of the Method 5, 17, or 201 variant methods describe only the procedures to determine the front-half
or filterable PM catch. EPA Method 202, formerly EPA Other Test Method 28 (OTM 28), uses a dry
condensing chamber followed by liquid impingers and specifies the procedures to determine the mass of
condensable PM (i.e., PM-CON). Although Method 202 generally references the use of Method 17 (or
201 or 201A) sampling trains, it may also be used in conjunction with EPA Methods 5, 5B, or 5F. As the
5-4
filter temperature in Method 5 is different from

PM fine versus CCU fine
Methods 5B or 5F and also likely different from the
Fine particles that escape from the CCU regenerator
filter temperature when using Method 17, the are often referred to as “fines” or “FCCU fines” by
fraction of PM that is filterable versus condensable, refinery personnel. In this Refinery Emissions
which is a function of the sampling temperature, will Protocol document, fine particles that escape from
also vary depending on the PM sampling method the CCU regenerator are referred to as “CCU fines.”
Please note that CCU fines are not equivalent to
used. Because most of the data available for CCU
“PM fines,” which refers to particles that have a
were collected using EPA Methods 5B or 5F, the diameter of 2.5 µm or less. Generally, we use
condensable PM default emissions factor is most “PM2.5” to denote particles that have a diameter of
suitable to filterable data collected using Methods 2.5 µm or less. As noted in Table 5-2, only a fraction
5B or 5F, but may also be used in conjunction with of the “CCU fines” are “PM fines.”
Method 5 sampling data. If Method 202 is used, then
the “back-half catch” can be considered PM-CON for the purposes of inventory reporting, regardless of
the PM filterable sampling method (i.e., sampling temperature) with which it is paired.
To summarize, PM CEMS and EPA Methods 5 (including 5B or 5F) and 17 provide a measure of PM-
FIL; EPA Method 201 provides a measure of PM-FIL and PM10-FIL; and EPA Method 201A provides a
measure of PM-FIL, PM10-FIL, and PM2.5-FIL. EPA Method 202 provides a measure of PM-CON. One
important caveat to this is that, in Method 5F, ammonia sulfate particulates are determined and subtracted
from the total PM catch to determine the non-sulfate PM emissions. Although the non-sulfate PM value
may be used for compliance purposes (for the Petroleum Refinery NSPS or for Refinery MACT II),
emissions reported for the inventory should include the total filterable PM catch. That is, the total PM
catch before subtracting the sulfate PM emissions should be used to determine PM-FIL for the purposes
of reporting for the emissions inventory. The sulfate particulate PM should be considered to be 100% PM
fine (i.e., PM2.5). Additional guidance on typical CCU particle size distributions is provided in the
following subsection.
5.1.2.2 PM Size Distribution Estimates for Catalytic Cracking Units

AP-42 does not contain PM size distribution data for uncontrolled CCU PM emissions, but it does contain
typical control device default control efficiencies for different types of PM emissions control devices. A
limited number of particle size distribution studies were available in the docket to the Refinery MACT 2
rule (40 CFR Part 63 Subpart UUU; Docket No. A-97-36, Item No. IV-D-19, Attachments 3, 6, and 9).
From these data, the mass fraction of total filterable PM (front-half catch) that is less than 2.5 and 10 µm
in diameter were determined and representative values are provided in Table 5-2. Table 5-2 also provides
PM size distribution default values for controlled CCU. The default distribution factors for controlled
CCU are based on the uncontrolled PM distributions and the control device default control efficiencies
from AP-42. Because the projected size distribution of PM from different control devices was similar, one
single set of distribution factors for controlled CCU are provided for controlled units. Control devices
considered include tertiary cyclones, wet scrubbers, venturi scrubbers, ESP, and fabric filters. All PM-
CON is assumed to be less than 2.5 µm. The size distribution data guidance presented here is specific to
FCCU. Facilities operating a TCCU should use site-specific size distribution data if available; otherwise,
the default factors in Table 5-2 for a CCU with no post-regenerator PM control device should be used to
project PM emissions by particle size.
5-5
Table 5-2. Default Size Distribution for Filterable PM from CCU
Fraction PM-FIL Less Fraction PM-FIL Less

Source Description Than 2.5 µm Than 10 µm
CCU, no post-regenerator PM control device 0.30 0.50
CCU, with post-regenerator PM control device (tertiary 0.80 0.97
cyclone, wet scrubber, ESP, or fabric filter)
Example 5-2: PM Emissions Inventory Calculations

The following information is available for a 50,000 barrels/calendar day (bbl/cd) CCU
controlled with an ESP. On average, the CCU burned 36,000 pounds of coke per hour and
operated for 8,700 hours in the reporting year. During the most recent source test using EPA
Method 5F, the CCU had non-sulfate PM emissions of 0.7 pound per 1,000 pounds of coke
burn-off and PM emissions (including sulfates) of 0.9 pound per 1,000 pounds of coke burn-
off. The “back-half” catch was also analyzed and was determined to be 0.45 pounds per 1,000
pounds of coke burn-off.
PM-FIL = 0.9 (lb/1,000 lb coke burn-off) × 36 (1,000 lb of coke burn-off per hour) × 8,700
(operating hours per year)
= 281,880 lb = 140.9 tons/yr
PM-CON = 0.45 (lb/1,000 lb of coke burn-off) × 36 (1,000 lb of coke burn-off/hr) × 8,700
(operating hours/yr)
= 140,940 lb = 70.5 tons/yr
PM-PRI = PM-FIL + PM-CON = 140.9 + 70.5 = 211.4 tons/yr
For Method 5F, special consideration is made for the sulfate PM, because all of this PM is
expected to be less than 2.5 µm. For Method 5 or 5B, PM10-FIL = (fraction < 10 μm from
Table 5-2) × PM-FIL. For Method 5F, PM10-FIL = (fraction < 10 μm from Table 5-2) ×
PMnon-sulfate + PMsulfate. Similar consideration is needed for the PM2.5-FIL calculation.
Because the CCU is controlled, the appropriate fractions from Table 5-2 are 0.97 for the
PM10-FIL calculation and 0.80 for the PM2.5-FIL calculation.
PMnon-sulfate = 0.7 (lb/1,000 lb coke burn-off) × 36 (1,000 lb coke burn/hr) × 8,700 (hr/yr)
= 219,240 lb = 109.6 tons/yr
PMsulfate = PM-FIL – PMnon-sulfates = 140.9 – 109.6 = 31.3 tons/yr
PM10-FIL = fraction < 10 μm from Table 5-2] × PMnon-sulfate + PMsulfate
= 0.97 × 109.6 + 31.3 = 137.6 tons/yr
PM10-PRI = PM10-FIL + PM-CON = 137.6 + 70.5 = 208.1 tons/yr
PM2.5-FIL = (fraction < 2.5 μm from Table 5-2) × PMnon-sulfate + PMsulfate
= 0.8 × 109.6 + 31.3 = 119.0 tons/yr
PM2.5-PRI = PM2.5-FIL + PM-CON = 119.0 +70.5 = 189.5 tons/yr
5-6
5.1.3 Methodology Rank 5A for CCU Metal HAP Emissions Estimates

During CCU processing, metals deposit on the catalyst particles and slowly deactivate the catalyst. Also
during processing, the catalyst particles may slowly break into finer particles, which can no longer be
recovered by the CCU regenerator internal cyclones. To maintain the desired catalyst activity, a portion of
the recirculating catalyst (i.e., E-cat) is regularly withdrawn from the process and new (fresh) catalyst is
added. New catalyst particles often have irregularities (spurs) that break off during catalyst recirculation.
Consequently, the CCU fines that escape from the CCU regenerator typically have a slightly higher
proportion of fresh catalyst than does the E-cat. This is significant because fresh catalyst will have less
metal deposits than a catalyst that has been recirculating for days or weeks. Available paired observations
of E-cat and CCU fines data indicate that the metal HAP concentration on the CCU fines are consistently
lower than on the E-cat, supporting the hypothesis that the CCU fines contain a higher fraction of fresh
catalyst than does the E-cat.
If metal HAP emissions tests are performed, then site-specific metal HAP emissions factors should be
determined (using the methods provided in Section 5.1.2, Methodology Ranks 3 and 4 for Catalytic
Cracking Units,) and used directly to assess metal HAP emissions rates as a Rank 4 method. However,
few facilities are expected to have performed metal HAP speciation emissions tests for their CCU. On the
other hand, nearly all facilities have their equilibrium catalyst (i.e., the catalyst within the CCU,
commonly referred to as “E-cat”) tested regularly for certain metal contaminants to monitor catalyst
activity and replacement needs. Some facilities also have their CCU fines tested occasionally for metal
HAP concentrations.
If a facility has a site-specific PM emissions factor, then either CCU fines or E-cat analysis can be used as
a means to calculate a site-specific HAP emissions rate using the methodology described in this section.
This Methodology Rank 5A for CCU metal HAP emissions estimates is expected to be more
representative than a default emissions factor, but not as good as a directly measured HAP-specific
emissions factor. This method involves using the site-specific (non-sulfate, if available) PM emissions
rate, and estimating the concentration of metal HAP on the emitted PM based on the concentration of
metal HAP on the CCU fines or E-cat using the procedures provided in this section. This method is not
applicable for mercury (Hg) because it assumes the metals are associated with the catalyst particles and
are not in the vapor phase.
If metal HAP concentrations of CCU fines are available, then these data should be used (preferentially to
E-cat data) to estimate the metal HAP concentration of the emitted PM. It is assumed that the emitted PM
has the same metal HAP concentration as the CCU fines. If metal HAP concentrations for CCU fines are
not available, then the metal HAP composition of the emitted PM can be estimated as 80% of the E-cat
concentration.
It is expected that CCU fine or E-cat concentration data will be available for nickel (Ni) and vanadium
because these are the primary metals that poison CCU catalyst activity, but concentration data may be
available for other metals as well. Table 5-3 lists metal HAP generally present in CCU fines. If the
concentrations of these metal HAP are not determined for the CCU fines or E-cat, then they should be
estimated using the measured Ni concentration and the concentration ratios for these other metal HAP as
provided in Table 5-3.
Equation 5-1 is a basic equation used to calculate annual metal HAP emissions based on the annual PM
emissions rate and CCU fines metal HAP concentrations.
Ci
Ei = PM −FIL × (Eq. 5-1)
1,000,000 mg/kg
5-7
where:
Ei = Emissions rate of metal HAP “i” (tons/yr)
PM-FIL = Filterable PM emissions rate from site-specific PM emissions factor; if EPA Method
5F is used, then use PMnon-sulfate rather than PM-FIL (ton/yr)
Ci = Concentration of metal HAP “i” on the CCU fines (milligrams per kilogram [mg/kg])
= 0.8 × CE-cat,i, if only E-cat concentration are available
= CNickel × (factor from Table 5-3), if other metal HAP concentrations are not available
Note that Ei = ENickel × (factor from Table 5-3) when other metal HAP concentration data are not
available. If vanadium concentrations are available but nickel concentrations are not, the nickel
concentration should be calculated by dividing the vanadium concentration by the default ratio for
vanadium in Table 5-3, and the calculated nickel concentrations and the default ratios in Table 5-3 then
used to project the concentrations of other metal HAPs.
Table 5-3. Default Ratio of Metal HAP Composition of CCU Finesa
Ratio of Metal HAP

Metal HAP to Nickel Concentration
Antimony 0.065
Arsenic 0.010
Beryllium 0.003
Cadmium 0.013
Chromium (total)b 0.25b
Cobalt 0.052
Lead 0.08
Manganese 0.13
Nickel 1.00
Selenium 0.025c
Vanadiumd 1.32
Zincd 0.74
a Concentration ratios are based on average electrostatic precipitator dust
analyses across six refineries.
b Limited data are available to assess the amount for hexavalent chromium that
is present, if any. If site specific data are not available, it is recommended that
a default value of 10 percent of the total chromium emissions be used to
estimate the hexavalent chromium emissions associated with the CCU.
c Selenium was highly variable, being significant for some units and below the
detection limit for several others. Consequently, this value is highly uncertain.
d Vanadium and zinc are not HAPs, but are included here (and in Table 1-1) as
other pollutants of interest.
5-8
Example 5-3: Methodology Rank 5A for Metal HAP Emissions Estimates

Following the previous example, PMnon-sulfate emissions were calculated to be 109.6
tons/yr. If only Ni concentration for E-cat is available, what are the metal HAP emissions if
the average Ni E-cat concentration is 1,200 mg/kg?
The concentration of Ni on the emitted PM is calculated based on the E-cat concentration as
follows:
CNickel = 0.8 × 1,200 = 960 mg/kg
The following equation should be used to calculate the emissions rate of Ni from the PM
emissions rate on estimated Ni concentration:
ENickel = 109.6 tons/yr × (960 mg/kg) ÷ (1,000,000 mg/kg) = 0.105 tons/yr
The following equations should be used to estimate the emissions rate of other metal HAP
from the emissions rate of Ni:
EAntimony = 0.105 tons/yr × 0.065 = 0.0068 tons/yr
EArsenic = 0.105 tons/yr × 0.010 = 0.00105 tons/yr
EBeryllium = 0.105 tons/yr × 0.003 = 0.0003 tons/yr
ECadmium = 0.105 tons/yr × 0.013 = 0.0014 tons/yr
EChromium = 0.105 tons/yr × 0.25 = 0.026 tons/yr
ECobalt = 0.105 tons/yr × 0.052 = 0.0055 tons/yr
ELead = 0.105 tons/yr × 0.08 = 0.0084 tons/yr
EManganese = 0.105 tons/yr × 0.13 = 0.014 tons/yr
ESelenium = 0.105 tons/yr × 0.025 = 0.0026 tons/yr
EVanadium = 0.105 tons/yr × 1.32 = 0.14 tons/yr
EZinc = 0.105 tons/yr × 0.74 = 0.088 tons/yr
Although total chromium emissions have been found above detectable concentrations in the CCU
regenerator vent, very few tests have been performed to determine Cr+6 emissions. The few tests that have
been performed have not found Cr+6 above detection limits, but the detection limits have been comparable
to the detected total chromium emissions. Emissions of Cr+6 from coal-fired power plants indicated that
Cr+6 emissions were 11 percent of the total chromium emissions (U.S. EPA, 1998c). Based on these
limited data, it is recommended that, for the CCU vent, Cr+6 emissions be estimated as 10% of the total
chromium emissions. Thus, in the example calculation provided, Cr+6 emissions would be estimated to be
0.0026 tons/yr.
Emissions of Hg should be estimated for the CCU; however, Hg emissions are not expected to be
correlated with the PM emissions because very little, if any, of the Hg emissions are expected to be
particulate bound. Mercury emissions are more likely correlated with the CCU throughput or coke burn-
off rate. Thus, Hg emissions should be estimated independently of other HAP metals. If Hg emissions
data are available based on source test data for the site, then the source test data can be used, as a
Methodology Rank 4 for metal HAP emissions estimates, to develop a site-specific emissions factor.
Otherwise, Hg emissions should be estimated using a default (Rank 5B) emissions factor rather than
using this Rank 5A method.
5-9
5.1.4 Methodology Rank 5B for Catalytic Cracking Units

For most organic HAP (e.g., formaldehyde, benzene, benzo[a]pyrene, dioxin/furans) and other pollutants
such as hydrogen cyanide and Hg, default emissions factors may be all that are available. When direct
emissions monitoring or site-specific emissions factors are not available, then the default emissions
factors presented in Table 5-4 should be used to calculate the emissions from the CCU regenerator vent.
Note that all CCU are considered “controlled for organics” if they meet the 500 ppmv CO emissions limit.
Table 5-4. Organic HAP Emissions Factors for CCU Catalyst Regenerator Vent
Emissions Factor Emissions Factor
CAS No. Compound (lb/MMbbl)a (lb/klb coke burn-off)b
Volatile Organics
75-07-0 Acetaldehyde 20 0.0013
67-64-1 Acetone 2.4 1.6E-4
107-02-8 Acrolein 1.0 6.6E-5
71-43-2 Benzene 18 1.1E-3
74-83-9 Bromomethane 2.1 1.4E-4
106-99-0 1,3-Butadiene 0.033 2.0E-6
100-41-4 Ethylbenzene 0.24 1.6E-5
50-00-0 Formaldehyde 260 0.016
75-09-2 Methylene chloride 6.7 4.4E-4
108-95-2 Phenol 8.7 5.7E-4
108-88-3 Toluene 3.5 2.1E-4
75-69-4 Trichlorofluoromethane 2.4 1.6E-4
1330-20-7 Xylene 3.2 2.1E-4
Semivolatile and Nonvolatile Organics(excluding dioxin/furans)
83-32-9 Acenaphthene 0.0033 2.2E-7
208-96-8 Acenaphthylene 0.13 7.8E-6
120-12-7 Anthracene 0.10 6.7E-6
56-55-3 Benzo(a)anthracene 0.00052 3.8E-8
50-32-8 Benzo(a)pyrene 0.011 7.1E-7
205-99-2 Benzo(b)fluoranthene 0.0035 2.4E-7
192-97-2 Benzo(e)pyrene 0.00045 3.3E-8
191-24-2 Benzo(g,h,i)perylene 0.0046 3.1E-7
207-08-9 Benzo(k)fluoranthene 0.0026 1.8E-7
218-01-9 Chrysene 0.0033 2.3E-7
53-70-3 Dibenz(a,h)anthracene 0.0042 2.8E-7
206-44-0 Fluoranthene 0.093 6.1E-6
86-73-7 Fluorene 0.037 2.4E-6
193-39-5 Indeno(1,2,3-cd)pyrene 0.0044 3.0E-7
91-57-6 2-Methylnaphthalene 0.026 1.8E-6
91-20-3 Naphthalene 1.0 7.0E-5
85-01-8 Phenanthrene 0.24 1.6E-5
129-00-0 Pyrene 0.0031 2.2E-7
(continued)
5-10
Table 5-4. Organic HAP Emissions Factors for CCU Catalyst

Regenerator Vent (continued)
Emissions Factor Emissions Factor

CAS No. Compound (lb/MMbbl)a (lb/klb coke burn-off)b
Dioxins/Furans
57117-31-4 Pentachlorodibenzofurans 5.5E-07 3.2E-11
57117-44-9 Hexachlorodibenzofuran 1.1E-06 6.3E-11
35822-46-9 Heptachlorodibenzo-p-dioxin 9.4E-07 5.6E-11
Inorganics
7664-41-7 Ammonia 13,000 0.57
75-15-0 Carbon disulfide 0.56 3.7E-5
7647-01-0 Hydrogen chloride 1,800 0.11
74-90-8 Hydrogen cyanide 7,000 0.43
7439-97-6 Mercury 0.094c 6.0E-6c
a Emissions factors for CCU controlled for organics in pounds per million barrels of CCU feed.
b Emissions factors for CCU controlled for organics in pounds per thousand pounds of coke burn-off.
c Total mercury emissions factor. Elemental mercury emissions are approximately 80 percent of total mercury
emissions and divalent (oxidized plus particulate) mercury emissions are approximately 20 percent of total
mercury emissions.
The emission factors presented in Table 5-4 are based primarily on the emission data collected to support
the Refinery MACT 2 rule; however, emissions factors for hydrogen cyanide and mercury were updated
based on source tests conducted in response to EPA’s 2011 ICR for the petroleum refining industry.
While there appear to be slight differences in how outlier and non-detect values were handled, the
emission factors in Table 5-4 agree well with the average emission factors developed by Bertrand and
Siegell (2002).
Based on the lack of data for PAH and furan emissions, the emissions estimates for these compounds
have high uncertainties, likely an order-of-magnitude either high or low. The California Air Resources
Board (CARB), with EPA’s support, conducted an emissions source test at a complete combustion FCCU
(without a post-combustion device). The only dioxin isomer detected in all runs was octachloro-dibenzo-
p-dioxin (OCDD); octachloro-dibenzo-furan (OCDF) and heptachloro-dibenzo-p-dioxin (hepta-CDD)
were detected in one run. All dioxin/furan quantities that were detected were detected at levels below the
method quantitation limit for the analysis. All polychlorinated biphenyls (PCBs) isomers were below
detection limits. This additional source test was not included in the development of the Petroleum
Refinery MACT II emissions factors, but it confirms low emissions of dioxins/furans and PCBs from the
CCU catalyst regenerator vent.
5.2 Fluid Coking Units

Coking units use heat to thermally crack heavy hydrocarbon streams to form lighter, more useful
distillates such as heating oils or gasoline. There are three basic types of coking units: traditional fluid
coking units, flexicoking units, and delayed coking units. Traditional fluid coking units are one of the
largest vent emissions sources at a refinery, being comparable to emissions from the CCU regenerator.
However, there are only a handful of traditional fluid coking units currently in operation in the United
States. Flexicoking units, which are also rare, do not have a direct atmospheric vent. Instead, these units
produce a low heating value syngas that can be subsequently used as fuel in process heaters or boilers.
Emissions from the combustion of flexicoking syngas should be determined using the methods described
5-11
in Section 4, Stationary Combustion Sources. Nearly all new coking units being built at refineries are
delayed coking units, which are discussed further in Section 5.3.
Fluid coking units have several similarities to a CCU. The coking unit contains a burner section and a
reactor section. In the burner section, heat for the coking reaction is supplied by burning a portion of the
coke that is produced. In the reactor, fine coke particles are produced as a result of the thermal cracking
process. Prior to exiting the coking unit, coke particles entrained with the flue gases are removed by
internal cyclone separators and recovered as a product, with a portion of the produced coke recycled to
the burner section. The coke burner is operated with limited air—enough to combust the diverted coke (to
provide heat for the coking reaction)—but limited enough to prevent O2 from getting into the reactor
section. The exhaust gas from the coking unit burner has high levels of CO, which is typically combusted
in a CO boiler to recover the latent heat of the CO. The CO boiler also acts to combust any organics
entrained in the flue gas. The gas may then be further processed to remove PM or SO2 in the flue gas. The
fluid coking unit vent releases a wide variety of pollutants, including PM, SO2, NOx, CO, VOC, metal
HAP, organic HAP, and ammonia.
Special considerations should be taken to estimate the amount of total sulfur compounds present in fuel
gas generated from the fluid coking unit. Coking units have been found to produce appreciable quantities
of methyl mercaptans and other reduced sulfur compounds that are not as efficiently removed from sour
gas as is H2S. Consequently, SO2 emissions from fuel gas combustion devices receiving fuel gas
generated from a fluid coking unit should account for this additional sulfur that is not measured by H2S
monitors (see Section 4, Stationary Combustion Sources, for further details).
Note also that fluid coking units, in particular, may have significant fugitive PM emissions from the
handling of the produced coke. Methodologies for these fugitive PM emissions are described in Section
10, Fugitive Dust Sources.
5.2.1 Methodology Ranks 1 and 2 for Fluid Coking Units

For SO2, NOx, CO, and CO2, it is anticipated that most units will have CEMS for measuring the
composition of the exhaust gas. Gas flow rate may be directly monitored, but in many cases, the exhaust
flow rate will be calculated based on air blast rates and composition monitors (similar to F factors).
5.2.2 Methodology Ranks 3 and 4 for Fluid Coking Units

Source tests should be available for PM (and perhaps SO2 and NOx, if not continuously monitored) so that
site-specific emissions factors should be available. It is anticipated that the pollutant emissions will
primarily be a function of the coke burn rates; however, limited data are available to assess this
hypothesis.
5.2.3 Methodology Rank 5 for Fluid Coking Units

We do not have data to provide default emissions factors for fluid coking units.
5.3 Delayed Coking Units

Coking units use heat to thermally crack heavy hydrocarbon streams to form lighter, more useful
distillates such as heating oils or gasoline. Most of the coking units in operation within the United States
are delayed coking units. Unlike most other refinery operations that are continuous, delayed coking units
are operated in a semi-batch system. Most delayed coking units consist of a large process heater, typically
two coking drums, a single product distillation column, and coke cutting equipment. The process heater
heats the heavy feed oil to near cracking temperatures, and then the oil is fed to one of the coking drums.
As the cracking reactions occur, coke is produced in the drum and begins to fill the drum with sponge-like
solid coke material. Once filled, the feed is diverted to the second coke drum. The full coke drum is
cooled by slowly adding water to the vessel, which quickly turns to steam, and the steam helps to cool
5-12
and purge organics in the coke matrix. After the coke drum is sufficiently cooled, the drum is opened and
the coke is removed from the vessel using high-pressure water. Once the coke is cut out of the drum, the
drum is closed, and prepared to go back online. A single coke drum is typically on-line receiving oil for
14 to 18 hours and then off-line for cooling and decoking for 14 to 18 hours, so a complete cycle time is
approximately 28 to 36 hours.
During the reaction process, the delayed coking unit is a closed system. When the coke drum is taken off-
line, the initial steaming process gas is also recovered through the unit’s product distillation column. As
the steaming cycle continues, the gas is sent to a blowdown system to recover the liquids. The gas is
typically sent to a flare, a flare gas recovery system, or the refinery’s gas plant for further
processing/recovery. Near the end of the steaming process, a vent is opened to allow the remaining steam
and vapors to be released into the atmosphere prior to draining and opening the drum. This atmospheric
vent is commonly referred to as the delayed coking unit steam vent. Emissions from the delayed coking
unit steam vent are primarily CH4 and ethane, but include a variety of volatile and semi-volatile organics.
Although Section 5 primarily addresses the process vent emissions from a given source, fugitive
emissions are also released from the delayed coking unit when the coke drum is drained, opened and the
coke is cut from the drum. During the draining process, hot water (and potentially super-heated water) is
discharged from the coke drum to an open drainage pit, where steam and organic contaminants are
released from the draining water. Some facilities begin draining immediately upon opening the
atmospheric vent while others wait for the steam released during the venting process to subside before
draining the coke drum. Draining the coke drum early in the venting cycle reduces the amount of
emissions that are released from the vent but increases the amount of emissions released in the drainage
area. Because the emissions from the delayed coking unit steam vent is dependent on the timing of water
draining and because a complete emissions inventory is desired, the calculation methodologies and
emission factors presented in this section for organic constituents are intended to account for the
emissions from the entire decoking process operations (venting, draining, drum opening, and coke
cutting). If emissions are estimated based on site-specific measurements or tests conducted only on the
coke drum steam vent, these data should be used to calculate site-specific emissions factors in units of
pounds per 1,000 pounds (lb/1,000 lb) steam for the coker steam vent, as discussed in Section 5.3.2,
Methodology Rank 3 and 4 for Delayed Coking Units. If only pollutant mass emissions rates are available
for the coker steam vent, then a separate analysis should be conducted to estimate the emissions during
the draining and coke cutting processes as discussed in Section 5.3.3, Methodology Rank 5 for Delayed
Coking Units.
PM emissions from the coke cutting operations and subsequent coke storage and handling facilities
should be estimated using the methodologies described in Section 10, Fugitive Dust Sources. The cutting
water will absorb some hydrocarbons, so this water will also become a source of organic emissions;
therefore, organic emissions from impoundments, ponds, or open tanks used to store the cutting liquids
should be estimated using the methodologies described in Section 7.2.5, Equalization Tanks. That is, the
organic emissions methodology presented in this section are expected to include the emissions that occur
during the active draining process as liquid drops from the coke drum vessel and is diverted to the coke
pit. Subsequent emissions from the water stored in the coke pit or other vessels is not considered in the
methodologies presented in this section. The appropriate methods presented in Section 7 of this Refinery
Emissions Protocol document should be used to estimate the emissions from these ancillary sources.
As with fluid coking units, special considerations should be taken to estimate the amount of total sulfur
compounds that are present in fuel gas generated from delayed coking units. Coking units have been
found to produce appreciable quantities of methyl mercaptans and other reduced sulfur compounds that
are not efficiently removed from the sour gas as H2S. Consequently, SO2 emissions from fuel gas
combustion devices receiving fuel gas generated from a delayed coking unit should account for this
5-13
additional sulfur that is not measured by H2S monitors (for more detail, see Section 4, Stationary
Combustion Sources).
5.3.1 Methodology Ranks 1 and 2 for Delayed Coking Units

Delayed coking units vent periodically and only for relatively short periods of time during
depressurization of the unit after the coking cycle. As such, delayed coking units are not expected to use
CEMS. If CEMS are used, Methodology Rank 1 for process vents could be used, but then additional
calculations must be performed to estimate the emissions from the active draining process and coke
cutting process as discussed in Section 5.3.3, Methodology Rank 5 for Delayed Coking Units. The gas
flow rate cannot be estimated using F factors, so Methodology Rank 2 for process vents is not applicable
to the delayed coking unit vent.
5.3.2 Methodology Rank 3 and 4 for Delayed Coking Units

A number of source tests have recently been performed on delayed coking unit vents. Pollutant emissions
from the delayed coking unit steam vent are well correlated with steam generation rates and a heat
balance can be used to estimate the total amount of steam generated (Coburn, 2013). This approach, as
presented in this section, estimates the combined emissions from the delayed coking unit steam vent and
the active draining process. If a refinery begins draining soon after venting is initiated, a larger portion of
the emissions will be released in the drainage area below the delayed coking unit as compared to a
refinery that waits for some time period (30 minutes to several hours) prior to draining. For inventory
purposes, it is not necessary to attempt to differentiate the emissions between the different emissions
release points; one can attribute all of the emissions to the delayed coking unit steam vent stack.
The mass of steam generated and released per decoking cycle can be estimated as follows:
(1 − f ConvLoss )× (M water × C p,water + M coke × C p,coke ) (Toverhead − 212)

M steam = × (Eq. 5-2)
∆H vap 2
Where:
Msteam = Mass of steam generated and released per decoking cycle (pounds per cycle; lbs/cycle).
fConvLoss = fraction of total heat loss that is due to convective heat loss from sides of the coke drum
vessel (unitless). Use the default value of 0.10.
Mwater = Mass of water in the coke drum vessel prior to at the end of the cooling cycle prior to
atmospheric venting (lbs/cycle).
Cp,water = Heat capacity of water (British thermal units per pound per degree Fahrenheit;
Btu/lb/°F). Use the default value of 1.0 Btu/lb/°F.
Mcoke = Typical dry mass of coke in the delayed coking unit vessel at the end of the coking
cycle (lbs/cycle).
Cp,coke = Heat capacity of petroleum coke (Btu/lb/°F). Use the default value of 0.265 Btu/lb/°F.
Toverhead = Temperature of the delayed coking unit vessel overhead line measured as near the coke
drum vessel as practical just prior to venting the atmosphere (°F). If the temperature of
the delayed coking unit vessel overhead line is less than 216 °F, use Toverhead = 216 °F
212 = Temperature at which steam generation is assumed to stop (°F). .
5-14
ΔHvap = Heat of vaporization of water (British thermal units per pound; Btu/lb). Use the default
value of 960 Btu/lb.
The typical dry mass of coke in the coke drum vessel can be calculated based on coke production
quantities per cycle. Alternatively, it can be estimated based on the typical coke drum outage as follows:
 π × D2 
M coke = ρ bulk ×  (H drum − H outage )×
  (Eq. 5-3)
 4 
Where:
ρbulk = Bulk coke bed density pounds per cubic feet; lb/ft3). Use the default value 42.2 lbs/ft3.
Hdrum = Internal height of coking unit vessel (feet; ft).
Houtage = Typical distance from the top of the coking unit vessel and the top of the coke bed (i.e.,
coke drum outage) at the end of the coking cycle (ft).
D = Diameter of coking unit vessel (ft).
The mass of water in the coke drum vessel can be calculated based on the typical water height in the coke
drum while accounting for the volume of coke particles present. Assuming the water completely covers
the coke bed, the mass of water in the coke drum can be estimated as follows:
 π × D 2 M coke 
M water = ρ water ×  (H water ) × − (Eq. 5-4)
 4 ρ particle 

Where:
ρwater = Density of water at the coke drum temperature (lb/ft3). The default value for the density
of water at 220 ºF is 59.5 lbs/ft3
Hwater = Typical distance from the bottom of the coking unit vessel to the top of the water level
at the end of the cooling cycle just prior to atmospheric venting (ft).
ρparticle = Particle density of petroleum coke (lb/ft3). Use the default value 84.3 lbs/ft3.
The annual mass emissions of a pollutant can be estimated using Equation 5-5. Methodology Rank 3 for
delayed coking units is to use site-specific test data for the coker steam vent to determine site-specific
pollutant emissions factors in units of pounds per 1,000 pounds (lbs/1000 lb) steam for use in Equation 5-
5. If site-specific emissions factors are not available, Methodology Rank 4 for delayed coking units is to
use the default emissions factors presented in Table 5-5 (those provided in units of lbs/1000 lb steam).
Typically, all coking drums for a given delayed coking unit will have similar size and operational
characteristics, so Equation 5-5 can be applied once considering all decoking events for the delayed
coking unit. If the coke drums have different sizes or operating characteristics, Equation 5-5 should be
used separately for each coke drum.
E i = M steam × EmFDCU,i × N × 0.001 (Eq. 5-5)
Where:
Ei = Annual emissions of pollutant “i” (pound per year; lb/yr).
5-15
EmFDCU,i = Emissions factor for delayed coking units for pollutant “i” from site-specific test data
(Rank 3) or from Table 5-5 for Methodology Rank 4 for delayed coking units (pounds
per 1,000 pounds (lbs/1000 lb) steam).
N = Cumulative number of decoking (venting) cycles.
0.001 = Conversion factor (number of “1000 pounds” per pound).
Table 5-5. Emissions Factors for Delayed Coking Unit Decoking Operations
Emissions Factor Volume Fraction for Emissions Factor

for Methodology Use in Equation 11-1 for Methodology
Rank 4 for for Coke Drum Rank 5 for
Delayed Coking Opening and Coke Delayed Coking
Units (lb/1,000 lb Cutting (scf/scf of Units
CAS No. Compound steam) coke drum gas) (lb/cycle)a
106-99-0 Methane 7.9 1.8E-2 250
74-84-0 Ethane 1.7 2.0E-3 53
VOC Nonmethane/nonethane VOC 1.7 1.4E-3 b 55
71-43-2 Benzene 0.032 1.5E-5 0.65
108-88-3 Toluene 0.12 4.7E-5 4.9
100-41-4 Ethylbenzene 0.011 3.7E-6 0.085
1330-20-7 Xylene 0.089 3.0E-5 0.67
108-95-2 Phenol 0.0009 3.4E-7 0.046
1319-77-3 Cresols (total) 0.001 3.3E-7 0.077
120-12-7 Anthracene 0.002 4.0E-7 0.072
86-73-7 Fluorene 0.002 4.3E-7 0.081
91-57-6 2-Methylnaphthalene 0.028 7.1E-6 1.3
91-20-3 Naphthalene 0.014 3.9E-6 0.75
85-01-8 Phenanthrene 0.005 1.0E-6 0.21
a Emissions are in pounds per emissions event or per cycle considering a single coke drum.
b Concentration of VOC reported as propane.
5.3.3 Methodology Rank 5 for Delayed Coking Units

When limited data are available regarding delayed coking unit operation, the default emission factors in
Table 5-5 for Methodology Rank 5 for delayed coking units can be used to estimate mass emissions per
cycle (i.e., venting event). The emissions factors for Methodology Rank 5 for delayed coking units
consider only emissions measured from the coker steam vent. In order to estimate emissions from other
decoking operations separate analyses should be conducted to estimate the emissions from the draining
and decoking processes.
The draining emissions can be estimated using the methodologies described in Section 7.2.2, Primary
Weirs (to model the initial drop and cascading of water to the drainage pit); however, the Henry’s law
constants used for the calculation should be adjusted to the temperature of the drain water. The
temperature adjustment input for the weirs model in RWET does not adjust the Henry’s law constant and
appropriate Henry’s law constants for the average temperature of the drained water should be entered into
the site-specific properties section of the Chemical Properties sheet to properly account for the high
temperature of the drain water. Temperature dependent Henry’s law constants can be determined using
the data and methods presented by Sander (1999). Although the precise increase is constituent specific, in
5-16
the absence of temperature-specific Henry’s law coefficients, the Henry’s law constant for compounds in
the drain water can be estimated to be 10 times greater than the Henry’s law constant reported at 25 ºC.
Emissions from opening and decoking the coke drum can be estimated using the methods presented in
Section 11.1, Gaseous Process Vessel Depressurization and Purging provided that the coke drum
contents are sufficiently cooled so that minimal additional steam generation occurs during the coke
cutting process. If significant quantities of steam are generated during the coke cutting process, the
method in Section 11.1 will underestimate the total emissions. Default volume fractions of pollutants in
the coke drum gas for use with Equation 11-1 are provided in Table 5-5.
5.4 Catalytic Reforming Units

The CRU is a series of catalytic reactors that turn naphtha into high-octane gasoline. There are no direct
atmospheric vents from the naphtha-reforming process, but the catalyst activity slowly diminishes with
time and the catalyst must be regenerated. There are three basic types of CRU catalyst regeneration
operations: continuous, cyclic, and semi-regenerative. Continuous CRU catalyst regenerators operate
continuously with a small slip stream of catalyst being re-circulated between the CRU and the
regenerator. In cyclic CRU, there is essentially an extra CRU reactor. When regeneration is needed, one
reactor is cycled offline and regenerated. The regeneration of the offline reactor is a batch process. When
complete, the reactor is returned to service, and the next reactor is cycled offline and regenerated. This
process continues until all reactors are regenerated. In a cyclic CRU, regeneration may occur for 1,000 to
4,000 hours per year. The semi-regenerative CRU operates without regeneration for 8 to 18 months, then
the entire unit is brought offline, and the entire unit is regenerated. The overall regeneration cycle
typically takes 1 to 2 weeks.
During regeneration, there are several potential atmospheric vents. Although the location of the emission
points might vary depending on whether catalyst regeneration is continuous, cyclic, or semi-regenerative,
there are three times or locations during the regeneration process that emissions can occur regardless of
the regenerator type. For continuous regeneration, venting occurs from three distinct vents as follows:
(1) the initial depressurization and purge vent; (2) the coke burn pressure control vent; and (3) the final
catalyst purge vent. For cyclic and semi-regenerative units, the initial depressurization and purge vent is
often a distinct release point, but the coke burn and final catalyst purge emissions are commonly released
at different times during the regeneration cycle from a single atmospheric vent.
The initial depressurization and purge cycle removes the hydrocarbons from the catalyst prior to CRU
catalyst regeneration. The vent gases from this initial purge may have high levels of organic HAP such as
benzene, toluene, xylene, and hexane. The gases generated from the initial depressurization and purge
cycles are typically vented to the refinery fuel gas system or directly to a combustion device (e.g., flare or
process heater). The coke burn cycle is typically the largest (in terms of gas volume) emissions source of
the overall catalyst regeneration cycle. The primary HAP contained in the CRU coke burn vent are
hydrogen chloride (HCl) and chlorine (Cl2), which are produced when the water formed during
combustion leaches chloride atoms from the CRU catalyst. The final purge and reduction cycle removes
O2 and any remaining chlorination agent from the system and reduces the catalyst prior to returning the
CRU catalyst to the reforming process or bringing the unit back online. The vent gases from this final
purge may have low levels of the chlorinating agent (usually an organic HAP such as trichloroethene or
perchloroethene) and residual HCl or Cl2 remaining in the system. The final purge gases are typically
vented into the atmosphere or to the refinery fuel gas system depending on the O2 content of the purge
gases (safety considerations restrict the venting of O2-containing gases to the refinery fuel gas system).
5.4.1 Emissions Estimation Methodology for Catalytic Reforming Units

Few data are available to characterize the emissions from the CRU catalyst regeneration vent because
venting is infrequent, the vent flow rates are slow and usually variable, and the vents have small
5-17
diameters. All of these factors make traditional source testing difficult. Most of the available data are for
HCl and Cl2 emissions from uncontrolled coke burn vents (20 data points are available for HCl emissions;
10 data points are available for Cl2). A few data points were available for a limited number of organic
chemicals. These data are compiled in the Background Information Document (BID) for the proposed
Petroleum Refinery MACT II rule (U.S. EPA, 1998b). During the Petroleum Refinery MACT II project,
CARB, with funding assistance from EPA, conducted a source test of a continuous CRU catalyst
regenerator coke-burn vent for dioxins/furans, PCBs, and PAHs. The results from this source test, which
were not available at the time for inclusion into the Petroleum Refinery MACT II BID, were used to
develop emissions factors for these compounds. Additionally, a total hydrocarbon (THC) 1 emissions
factor was developed based on test data submitted in response to the EPA’s 2011 ICR. The resulting
emissions factors for the coke burn emissions are presented in Table 5-6. These emissions factors are
normalized by the CRU process throughput and are assumed to apply equally for all types of CRU
regenerators.
The most prevalent control device used in association with the coke-burn vent is a wet scrubber. The
dioxin/furan emissions source tests and the volatile organics were performed on a system controlled by a
wet scrubber. Because of the limited solubilities of these chemicals and the scrubbing medium
recirculation rate used for wet scrubbers on this vent stream, the scrubber is assumed to have limited
effectiveness at reducing the emissions of these chemicals. Therefore, the same emissions factor is
recommended for these chemicals for both controlled and uncontrolled CRU. Wet scrubbers are used to
reduce the emissions of HCl and Cl2. The wet scrubbers used for these vents were characterized into two
classes: single-stage scrubbers and multiple-stage scrubbers. Single-stage scrubbers are estimated to
reduce HCl and Cl2 emissions by 92%, and the multiple-stage scrubbers are estimated to reduce HCl and
Cl2 emissions by 97%.
As most emissions from the purge cycles are vented to the refinery fuel gas system or a flare, emissions
from this source were not covered separately in this section; these emissions are presumably included in
the refinery fuel gas combustion sources (process heaters and boilers) or flares emissions estimates. No
data are available to characterize the small portion of venting that occurs directly into the atmosphere
from these purge cycles, but these emissions should be characterized and reported in the inventory.
Table 5-6. Emissions Factors for CRU Catalyst Regeneration Vent
Emissions Factor
CAS No. Chemical Name (lb/1,000 bbl)a
1746-01-6 Dioxin toxic equivalents (TEQ)b 5.7E-09
1336-36-3 Total polychlorinated biphenylsc 2.6E-06
91-20-3 Naphthalene 3.5E-05
91-57-6 2-Methylnaphthalene 1.3E-06
208-96-8 Acenaphthylene 3.0E-08
83-32-9 Acenaphthene 4.3E-08
86-73-7 Fluorene 2.0E-07
85-01-8 Phenanthrene 6.1E-07
120-12-7 Anthracene 9.1E-08
206-44-0 Fluoranthene 1.0E-07
129-00-0 Pyrene 1.5E-08
(continued)
1
Total hydrocarbons may be used as a surrogate for VOC.
5-18
Table 5-6. Emissions Factors for CRU Catalyst Regeneration Vent (continued)
Emissions Factor
CAS No. Chemical Name
(lb/1,000 bbl)a
56-55-3 Benzo(a)anthracene 9.0E-10
218-01-9 Chrysene 2.9E-09
205-99-2 Benzo(b)fluoranthene 1.5E-09
207-08-9 Benzo(k)fluoranthene 7.5E-10
192-97-2 Benzo(e)pyrene 2.9E-09
193-39-5 Indeno(1,2,3-c,d)pyrene 1.7E-09
53-70-3 Dibenzo(a,h)anthracene 7.8E-10
191-24-2 Benzo(g,h,i)perylene 4.0E-09
71-43-2 Benzene 0.004
108-88-3 Toluene 0.0096
1330-20-7 Xylene 0.007
THC Total hydrocarbonsd 0.24
7647-01-0 Hydrogen chloride 4.2e
7782-50-5 Chlorine 0.23e
Notes for Table 5-6.
a Emissions factor in pounds pollutant emitted per 1,000 barrels of catalytic reforming unit process
capacity.
b Dioxin TEQ = toxic equivalents to 2,3,7,8-tetrachloro-dibenzo-p-dioxin used for risk analysis; specific
dioxin/furan isomer emissions data are available.

c Sum total of all chlorinated biphenyl emissions factors; data were available for each class of chlorinated
biphenyls (e.g., mono-, di-, tri-..., decachlorobiphenyl).

d Total hydrocarbons may be used as a surrogate for VOC. Total hydrocarbons are presented on an as
propane basis.
e Emissions factor for uncontrolled coke burn vent; controlled emissions should be estimated based on the
efficiency of the control device for these pollutants. Single-stage scrubbers (including direct caustic
injection) are estimated to reduce HCl and Cl2 emissions by 92%, and the multiple-stage scrubbers are
estimated to reduce HCl and Cl2 emissions by 97%
5.5 Sulfur Recovery Plants

All crude oils contain some sulfur compound impurities. Sulfur compounds in crude oil are converted to
H2S in the cracking and hydrotreating processes of the refinery. The H2S in the generated gas streams is
removed from the process vapors using amine scrubbers. The amine scrubbing solution is subsequently
heated to release the H2S to form an H2S rich “acid gas” that is treated in the sulfur recovery plant to yield
high-purity sulfur that is then sold as product. Most sulfur recovery plants use the Claus reaction and are
commonly referred to as Claus units or Claus sulfur recovery plants. There are a couple of other types of
sulfur recovery plants at smaller refineries, but all larger sulfur recovery plants employ Claus units. The
exhaust gas from the sulfur recovery unit (SRU) is commonly referred to as “tail gas.” The sulfur
recovery plant consists of one or more SRU operated in parallel and may also contain one or more
catalytic tail gas treatment units and/or a thermal oxidizer to combust the tail gas.
The primary HAP components of the final sulfur plant vent are carbonyl sulfide (COS) and carbon
disulfide (CS2). These HAP components are by-products of the SRU and tail gas treatment unit (TGTU)
reactions; COS may also be a product of incomplete combustion from a thermal oxidizer. Unreacted H2S
may also be released during the process. Sulfur recovery plant vents are commonly controlled by a
thermal oxider to oxidize unreacted H2S or H2S in sweep gas from the sulfur pits to SO2. Some sulfur
recovery plants use reducing controls and thus emit H2S rather than SO2. The sulfur plant sour gas feed
5-19
may also contain small amounts of light organics. Therefore, it is important to account for the
hydrocarbons in the sulfur recovery plant feed when estimating emissions from the sulfur recovery plant,
particularly from Lo-Cat® or other sulfur recovery plants that may have atmospheric vents without
thermal destruction.
When the sulfur recovery plant is in operation, the sulfur plant vent flow rate is fairly small so that the
SO2 emissions from the sulfur recovery plant are also relatively small. If the sulfur recovery plant must be
taken offline due to an upset or malfunction, the sour gas may be temporarily directed to a backup sulfur
recovery unit or directed to a flare or the thermal oxidizer. If the sour gas in these cases is sent to a flare
or thermal oxidizer, the SO2 emissions can be very large. As such, it is critical to include accurate
accounting of SO2 emissions during startup, shutdown, or malfunction (SSM) events associated with the
sulfur recovery plant.
5.5.1 Methodology Ranks 1 and 2 for Sulfur Recovery Plants

It is anticipated that most sulfur recovery plants, particularly Claus sulfur recovery plants, will have
continuous SO2 concentration monitors. When continuous flow monitors are also in place, mass
emissions rates can be calculated using the CEMS method previously described in Section 4, Stationary
Combustion Sources. Unlike combustion sources, however, there is not an F factor method to project the
vent flow rate. If the flow rate and H2S concentration of the feed to the sulfur recovery plant, the air or
oxygen feed rates to the sulfur plant burner (used to convert one-third of the H2S to SO2), and the quantity
of sulfur recovered are known, then calculations to determine/quantify emissions can be made, but the
calculations are not trivial. As such, there is not a simple Rank 2 method that can be provided for sulfur
recovery plants. Non-Claus sulfur recovery plants may monitor H2S or reduced sulfur compound
concentrations instead of SO2. Again, H2S emissions estimates should be provided for sulfur recovery
plants.
5.5.2 Methodology Ranks 3 and 4 for Sulfur Recovery Plants

Emissions estimates can be made from inlet flow measurements and assumed recovery efficiencies. For
most Claus units, especially those with tail gas treatment/recovery units, a mass balance approach for
sulfur is expected to yield emissions estimates with high uncertainties (e.g., when two large numbers are
subtracted so that the difference is only a few percentages of the original values, the uncertainties in the
original values may be as large as the difference). A three-stage Claus unit is expected to have sulfur
recovery efficiencies of 95% to 97%. Combining a three-stage Claus unit with a tail gas treatment unit,
sulfur recovery efficiencies are expected to be 99.7% to 99.9%. With these types of recovery efficiencies,
the mass balance approach (i.e., calculating SO2 or other sulfur compound emissions from the difference
of the sulfur feed rate to the SRU and the sulfur produced) will have significant uncertainty.
Consequently, measurement of the mass feed rate of sulfur to the sulfur recovery plant and an assumed
sulfur recovery efficiency (based on the number of Claus reactors in series and the presence or absence of
a tail gas treatment unit) are likely to be more accurate than using a mass balance approach when
calculating sulfur compound emissions from the sulfur recovery plant. A similar approach can be used to
calculate CO2 emissions from the sulfur recovery plant, particularly in systems controlled with a flare or
thermal oxidizer. Because nearly all of the CO2 or hydrocarbons introduced to the sulfur recovery plant
will exit as CO2, measuring inlet C content and flow rate of gases to the sulfur recovery plant can be used
to accurately estimate CO2 emissions.
Given the high uncertainty expected in engineering/mass-balance calculations for sulfur recovery plants,
site-specific emissions factors may be as or more useful than using an engineering/mass-balance approach
for calculating emissions from sulfur recovery plants, particularly sulfur compound emissions. Site-
specific emissions factors are also recommended for reductive tail gas treatment units not followed by an
incinerator and for non-Claus sulfur recovery plants. Evaluation of light hydrocarbons should be
addressed in these systems. For example, Lo-Cat® systems use an oxidation tank in which H2S is
5-20
chemically oxidize to SO2. Any light hydrocarbons entrained in the sour gas are expected to be released in
the vent from the oxidation tank. The emissions from this oxidation tank should be characterized and
reported for the sulfur recovery plant.
5.5.3 Methodology Rank 5 for Sulfur Recovery Plants

The Petroleum Refinery MACT II BID (U.S. EPA, 1998b) presents a range of total sulfur HAP
compounds emissions factors for SRU controlled by an incinerator. Based on the data presented and
additional concentration data submitted by the National Petrochemical and Refiners Association, it was
assumed that 75% of the sulfur HAP emitted was COS and 25% was CS2. The controlled emissions
factors are based on summary data reported for five SRUs. Emissions of uncontrolled sulfur HAP were
estimated from the controlled emissions assuming a control efficiency of 98% for the controlled
emissions (so that uncontrolled emissions are 50 times higher than controlled emissions). The resulting
emissions factors are presented in Table 5-7. Data collected as part of the EPA’s 2011 ICR for petroleum
refineries was reviewed to develop some criteria pollutant emissions factors for sulfur recovery plants.
These factors are also presented in Table 5-7.
Table 5-7. Emissions Factors for Sulfur Recovery Plants
Controlled SRU Controlled SRU Uncontrolled SRU

Emissions Factor Emissions Factor Emissions Factor
CAS No. Compound Name (lb/MMBtu)a (lb/lton)a (lb/lton)a
630-08-0 Carbon monoxide 0.71 1.5
NOx Nitrogen oxides 0.10 0.24
THC Total hydrocarbonsb 0.0014 0.045 2.25c
43-58-1 Carbonyl sulfide 0.12 5.85c
75-15-0 Carbon disulfide 0.040 2.00c
Note: lb/MMBtu = pounds per million British thermal units; lb/lton = pounds per long ton
a Emissions factor in pounds of HAP per long-ton of sulfur produced.
b Total hydrocarbons may be used as a surrogate for VOC. Total hydrocarbons are presented on an as propane
basis.
c Values estimated as 50 times the controlled SRU emissions factor.
5.6 Other Miscellaneous Process Vents

There are many other process vents at a refinery; however, there are limited data for these vents, so
default emissions factors are not available. For these process vents, this section provides a brief
description of the process, the vent source, and the primary pollutants associated with the process vent.
For all of these sources, the general hierarchy of methods as presented in Table 5-1 in the introduction to
this chapter is applicable. For atmospheric vents associated with continuous monitoring systems, the flow
rate and composition of the purged gases should be well characterized and emission estimates can be
developed using Methodology Rank 1 for process vents. For other process vents, measurement data may
not be available. Engineering or model estimates can be used to estimate the uncontrolled emissions. If
the emissions are subsequently controlled, the actual emissions can be estimated based on the efficiency
of the control device. The purpose of this section is to provide default emission factors for Methodology
Rank 5 for these particular process vents.
5.6.1 Hydrogen Plant Vents

Steam CH4 reforming is the primary means by which H2 is produced at a petroleum refinery. Refinery
fuel gas is typically used as the feedstock to the H2 plant. The feedstock is combined with steam in a
reactor at high temperatures (750°C to 800°C) to produce a mixture of H2 and CO. The SMR reaction is
endothermic, meaning the reaction requires heat in order to proceed. Therefore, additional natural gas
5-21
(i.e., natural gas in addition to that used to produce hydrogen) must be supplied to the SMR furnace to
provide the necessary heat for the SMR reaction. A water–gas shift (WGS) reaction is then conducted in a
second series of catalytic reactors to convert the CO and steam to CO2 and H2. Finally, the CO2 is
removed from the H2 product using either a liquid absorption, membrane filtration, or pressure swing
absorption (PSA). There are various H2 plant vents depending on the configuration of the H2 plant. The
primary vent from the H2 plant is from the SMR furnace. The PSA off-gas vent is often fed to the SMR
furnace, so the PSA off-gas will be combusted and emitted along with combustion furnace emissions
from the SMR furnace stack. For units that use liquid adsorption, there is typically a separate high CO2
vent stack from the regeneration of the adsorption liquid (unless the CO2 is captured as a byproduct).
Condensed water from the water-gas shift reactors is often returned to the boiler feed water system after
stripping or deaerating the CO2 from the water. The deaerator vent typically contains methanol and
hydrocarbons; however, no default emission factors are available at this time for the hydrogen plant vent
deaerator vent. The deaerator vent may also be directed to the SMR furnace for hydrocarbon emissions
control. After review of data obtained from the EPA’s 2011 ICR, a NOx emissions factor of
0.081 lb/MMBtu was developed for the SMR furnace.
5.6.2 Asphalt Plant Vents

Asphalt blowing is used for polymerizing and stabilizing asphalt to improve its weathering characteristics.
Air-blown asphalts are used in the production of asphalt roofing products and certain road asphalts. Air
blowing of asphalt may be conducted at petroleum refineries, asphalt processing plants, and asphalt
roofing plants. Asphalt blowing involves the oxidation of asphalt flux by bubbling air through liquid
asphalt flux at 260°C (500°F) for 1 hour to 10 hours. The amount of time depends on the desired
characteristics of the product, such as the softening point and the penetration rate. Shorter periods are
typically used for road asphalt; longer periods are used for roofing asphalt. Asphalt blowing results in an
exothermic reaction that requires cooling. The emissions from a blowing still are primarily organic
particulate with a fairly high concentration of gaseous hydrocarbon and polycyclic organic matter. The
blowing still gas is commonly controlled with a wet scrubber to remove sour gas, entrained oil,
particulates, and condensable organics and/or a thermal oxidizer to combust the hydrocarbons and sour
gas to CO2 and SO2.
Overall PM and TOC emissions factors for asphalt blowing are provided in Section 11.2 of AP-42 (U.S.
EPA, 1995a); these emission factors are also presented here in Table 5-8.
Tumbore (1998) provided an additional summary and evaluation of emissions from controlled asphalt
blowing. Some key findings from this review were:
 The AP-42 emission factors do not address emissions of SO2, which are usually the largest
emissions from the process. For gas fired control systems, H2S resulting from the asphalt blowing
reaction accounted for 70 to 80 % of the SO2 emissions, with the remainder attributable to
entrained or condensed oil.
 The AP-42 emission factors do not address emissions of HCl, which are important when ferric
chloride is used as a catalyst in the process.
5-22
Table 5-8. Emission Factors for Asphalt Blowing (U.S. EPA, 1995a)
Filterable PMa TOCb

Emission Factor (Data Quality Rating)
Operation (lb/ton asphalt processed)
Uncontrolled
Saturant asphalt c 6.6 (E) 1.3 (E)
Coating asphalt d 24 (E) 3.4 (E)
Controlled
Saturant asphalt c 0.27 (D) 0.0043 (D)
Coating asphalt d 0.81 (D) 0.017 (D)
a Filterable (front-half) particulate matter catch using EPA Method 5A sampling at 108°F.
b Total organic compounds measured using EPA Method 25A.
c Saturant blow of 1.5 hours.
d Coating blow of 4.5 hours.
Tumbore (1998) provided an additional summary and evaluation of emissions from controlled asphalt
blowing. Some key findings from this review were:
 The AP-42 emission factors do not address emissions of SO2, which are usually the largest
emissions from the process. For gas fired control systems, H2S resulting from the asphalt blowing
reaction accounted for 70 to 80 % of the SO2 emissions, with the remainder attributable to
entrained or condensed oil.
 The AP-42 emission factors do not address emissions of HCl, which are important when ferric
chloride is used as a catalyst in the process.
The Emission Inventory Guidebook 2006 (EMEP/CORINAIR, 2006) provides the following speciation
profile for organic emissions from asphalt blowing (see Table 5-9). Although not specified, it is assumed
that this speciation profile in Table 5-9 represents uncontrolled asphalt blowing emissions.
Table 5-9. Asphalt Blowing – Nonmethane Volatile Organic Compounds Speciationa
Compound % Weight
Ethane 6.0
Propane 18.8
Butanes 30.5
Pentanes 17.2
Hexanes 8.4
Heptanes 9.8
Octanes 7.4
Cycloparaffins 1.9
Benzene 0.1
a As reported by EMEP/CORINAIR, 2006.
In addition, API and the Western States Petroleum Association (WSPA) conducted emissions source tests
of petroleum refining sources and have compiled emissions factors for controlled asphalt blowing
(Hansell and England, 1998). The average emissions factors are presented in Table 5-10. These factors
are recommended for estimating speciated emissions from controlled asphalt blowing units.
5-23
Table 5-10. Summary of Emissions Factors for Controlled Asphalt Blowing
Mean Emissions Factora
With Blow With Blow Without Blow Without Blow

Cycle and Cycle and Cycle and with Cycle and with
Thermal Thermal Thermal Thermal
Oxidizer Oxidizer Oxidizer Oxidizer
CAS Number Substance (lb/MMcf)b (lb/MMBtu)b (lb/MMcf)b (lb/MMBtu)b
75-07-0 Acetaldehyde 1.8E-03 1.67E-06 4.3E-03 4.1E-06
7440-28-2 Arsenic 1.3E-02 1.2E-05 1.2E-02 1.1E-05
71-43-2 Benzene 3.2E-01 3.0E-04 2.8E-01 2.6E-04
7440-41-7 Beryllium 2.6E-03 2.5E-06 2.3E-03 2.2E-06
7440-43-9 Cadmium 5.3E-03 4.9E-06 4.7E-03 4.4E-06
18540-29-9 Chromium 3.2E-03 3.0E-06 3.3E-03 3.1E-06
(hexavalent)
7440-47-3 Chromium (total) 4.2E-02 3.9E-05 1.4E-02 1.3E-05
7440-50-8 Copper 4.8E-02 4.5E-05 3.8E-02 3.6E-05
100-41-4 Ethylbenzene 8.6E-01 8.1E-04 7.6E-01 7.2E-04
50-00-0 Formaldehyde 3.6E-03 3.3E-06 1.3E-02 1.2E-05
7647-01-0 Hydrogen chloride 2.2E-03 2.1E-06 8.2E-04 7.7E-07
7783-06-4 Hydrogen sulfide 2.1E+00 2.0E-03 1.8E+00 1.7E-03
7439-92-1 Lead 5.3E-02 4.9E-05 4.7E-02 4.4E-05
7439-96-5 Manganese 1.2E-01 1.2E-04 2.1E-01 2.0E-04
7439-97-6 Mercury 9.1E-03 8.5E-06 8.5E-03 8.0E-06
7440-02-0 Nickel 6.7E-02 6.3E-05 6.0E-02 5.7E-05
108-95-2 Phenol 7.6E-02 7.1E-05 4.6E-02 4.4E-05
7782-49-2 Selenium 1.3E-02 1.2E-05 1.2E-02 1.1E-05
1330-20-7 Xylene (total) 8.6E-01 8.1E-04 7.6E-01 7.2E-04
7440-66-6 Zinc 8.4E-01 7.9E-04 5.4E-01 5.0E-04
Note: lb/MMcf = pounds per million cubic feet asphalt fumes, lb/MMBtu = pounds per million British thermal units
a Bold italic values indicate that all test runs were below detection limit.
b Source: Hansell and England, 1998.
5.6.3 Coke Calcining

Coke calcining is one of the processes tested by API and WSPA for which emissions factors were
developed (Hansell and England, 1998). The average emissions factors for a coke calcining unit
controlled with a spray dryer and fabric filter are presented in Table 5-11.
5-24
Table 5-11. Summary of Emissions Factors for Controlled Coke Calcining
CAS Number Substance lb/ton cokeb lb/MMBtub

83-32-9 Acenaphthene 1.5E-08 4.4E-08
208-96-8 Acenaphthylene 1.8E-08 5.6E-08
75-07-0 Acetaldehyde 1.0E-03 3.1E-03
107-02-8 Acrolein 3.4E-04 1.0E-03
120-12-7 Anthracene 1.8E-08 5.4E-08
7440-36-0 Antimony 4.6E-05 1.4E-04
7440-28-2 Arsenic 4.7E-06 1.5E-05
7440-39-3 Barium 2.0E-05 6.1E-05
71-43-2 Benzene 3.2E-04 1.0E-03
56-55-3 Benzo(a)anthracene 8.7E-09 2.6E-08
50-32-8 Benzo(a)pyrene 8.1E-09 2.4E-08
205-99-2 Benzo(b)fluoranthene 8.1E-09 2.4E-08
191-24-2 Benzo(g,h,i)perylene 8.1E-09 2.4E-08
207-08-9 Benzo(k)fluoranthene 8.1E-09 2.4E-08
7440-41-7 Beryllium 1.9E-06 6.0E-06
7440-43-9 Cadmium 9.3E-06 2.9E-05
18540-29-9 Chromium (hexavalent) 6.3E-07 2.1E-06
7440-47-3 Chromium (total) 2.1E-05 6.9E-05
218-01-9 Chrysene 1.3E-08 3.7E-08
7440-50-8 Copper 9.3E-06 2.9E-05
53-70-3 Dibenz(a,h)anthracene 8.1E-09 2.4E-08
1746-01-6 Dioxin:4D 2378 1.1E-11 3.7E-11
Dioxin:4D Other 1.4E-10 4.4E-10
40321-76-4 Dioxin:5D 12378 9.6E-12 2.9E-11
39227-28-6 Dioxin:6D 123478 9.9E-12 3.5E-11
57653-85-7 Dioxin:6D 123678 1.4E-11 4.4E-11
19408-74-3 Dioxin:6D 123789 1.3E-11 4.2E-11
35822-46-9 Dioxin:7D 1234678 1.4E-10 4.2E-10
3268-87-9 Dioxin:8D 1.8E-09 5.3E-09
206-44-0 Fluoranthene 3.6E-08 1.1E-07
86-73-7 Fluorene 5.6E-08 1.7E-07
51207-31-9 Furan:4F 2378 1.3E-11 4.2E-11
(continued)
5-25
Table 5-11. Summary of Emissions Factors for Controlled Coke Calcining (continued)
CAS Number Substance lb/ton cokeb lb/MMBtub

Furan:4F Other 1.4E-10 4.4E-10
57117-41-6 Furan:5F 12378 1.4E-11 4.4E-11
57117-31-4 Furan:5F 23478 1.3E-11 4.1E-11
70648-26-9 Furan:6F 123478 2.4E-11 7.9E-11
57117-44-9 Furan:6F 123678 2.2E-11 7.1E-11
72918-21-9 Furan:6F 123789 9.2E-12 3.0E-11
60851-34-5 Furan:6F 234678 2.0E-11 6.5E-11
67562-39-4 Furan:7F 1234678 1.5E-10 4.8E-10
55673-89-7 Furan:7F 1234789 2.6E-11 8.0E-11
39001-02-0 Furan:8F 1.5E-10 4.1E-10
193-39-5 Indeno(1,2,3-cd)pyrene 8.1E-09 2.4E-08
7439-92-1 Lead 6.2E-05 1.9E-04
7439-96-5 Manganese 4.6E-05 1.4E-04
7439-97-6 Mercury 4.6E-05 1.5E-04
91-20-3 Naphthalene 2.4E-06 7.3E-06
7440-02-0 Nickel 9.1E-05 2.9E-04
85-01-8 Phenanthrene 1.9E-07 5.7E-07
7723-14-0 Phosphorus 4.7E-04 1.5E-03
129-00-0 Pyrene 2.7E-08 7.9E-08
7782-49-2 Selenium 4.7E-06 1.5E-05
108-88-3 Toluene 5.3E-05 1.6E-04
1330-20-7 Xylene (m,p) 3.1E-05 8.9E-05
95-47-6 Xylene (o) 4.5E-05 1.3E-04
7440-66-6 Zinc 1.2E-04 3.7E-04
Note: lb/ton coke = pounds per ton of petroleum coke, lb/MMBtu = pounds per million British thermal units
Abbreviations: 4D 2378 = 2,3,7,8-Tetrachlorodibenzo-p-dioxin; 5D 12378 = 1,2,3,7,8-Pentachlorodibenzo-p-dioxin;
6D 123478 = 1,2,3,4,7,8-Hexachlorodibenzo-p-dioxin; 6D 123678 = 1,2,3,6,7,8-Hexachlorodibenzo-p-dioxin; 6D
123789 = 1,2,3,7,8,9-Hexachlorodibenzo-p-dioxin; 7D 1234678 = 1,2,3,4,6,7,8-Heptachlorodibenzo-p-dioxin; 8D
= Octachlorodibenzo-p-dioxin; 4F 2378 = 2,3,7,8-Tetrachlorodibenzofuran; 5F 12378 = 1,2,3,7,8-
Pentachlorodibenzofuran; 5F 23478 = 2,3,4,7,8-Pentachlorodibenzofuran; 6F 123478 = 1,2,3,4,7,8-
Hexachlorodibenzofuran; 6F 123678 = 1,2,3,6,7,8-Hexachlorodibenzofuran; 6F 123789 = 1,2,3,7,8,9-
Hexachlorodibenzofuran; 6F 234678 = 2,3,4,6,7,8-Hexachlorodibenzofuran; 7F 1234678 = 1,2,3,4,6,7,8-
Heptachlorodibenzofuran; 7F 1234789 = 1,2,3,4,7,8,9-Heptachlorodibenzofuran; 8F = Octachlorodibenzofuran.
a Emissions factors provided are for coke calcining controlled with a spray dryer and fabric filter. Bold italic values
indicate that all test runs were below detection limit; underlined values indicate that more than 75% of the test
runs were below the detection limit.
b Source: Hansell and England, 1998.
5-26
5.6.4 Blowdown Systems

Blowdown systems are used during depressurization processes to recover liquids entrained in a process
gas stream. The remaining uncondensed gases may be compressed and recovered for use as fuel gas or
they may be vented to a thermal destruction device (thermal oxidizer or flare). Either of these scenarios
would be considered a “controlled” blowdown system. In an “uncontrolled” blowdown system, the
uncondensed gases are vented directly to the atmosphere. For blowdown gases that are recovered as fuel
gas, the emissions from these gases would be accounted for in the emissions projected for stationary
combustion devices (see Section 4). For blowdown gases that are vented to a flare, the emissions from
these gases would be accounted for in the emissions projected for flares if Methodology Rank 1, 2, 3, or 4
for flares is used (see Section 6).
AP-42 (U.S. EPA, 1995a) contains default emission factors for uncontrolled blowdown systems as well as
blowdown systems that recover the vapors for destruction in a flare. For systems controlled by a flare, the
measurement methodologies provided in Section 6, Flares, should be used preferentially over these
default emission factors, but these factors may provide an estimate of the portion of the flare’s emissions
originating from blowdown gases. The AP-42 default emission factors for blowdown systems are
provided in Table 5-12. For blowdown gases that are vented to control device other than a flare, use the
methodologies for direct process vent measurement provided in this section, if appropriate data are
available. If appropriate data are not available, use the emission factors for “blowdown systems with
vapor recovery and flaring” in Table 5-12 for blowdown vents controlled with a thermal oxidizer,
catalytic oxidizer, or similar combustion control system; for other control systems, use the uncontrolled
blowdown default THC emissions factor provided in Table 5-12 and adjust for the efficiency of the
control device using the default control efficiencies provided in Table 3-2 (Section 3, Storage Tanks).
Table 5-12. Default Emissions Factors for Blowdown Systems
Emissions Factora
(lb/103 barrel of refinery
Process Description Pollutant feed) Rating
Uncontrolled blowdown Total hydrocarbonsb 580 C
Blowdown System with Carbon monoxide 4.3 C
vapor recovery and flaring
Nitrogen oxides 19 C
Sulfur dioxide 27 C
Total hydrocarbonsb 0.8 C
a Source: U.S. EPA, 1995a.
b
Total hydrocarbon may be estimated as a surrogate for VOC. (Overall, less than 1 weight % of total hydrocarbon
emissions is methane.)
5.6.5 Vacuum Producing Systems

Vacuum producing systems include reciprocating, rotary or centrifugal blowers or compressors, or any jet
ejector or device that takes suction from a pressure below atmospheric and discharges against atmospheric
pressure. Depending on the vacuum producing system design, gases from the system may be released
directly to the atmosphere or may be condensed, to the extent possible, using a condenser. The condensed
liquid is gathered in a “hot well” or “accumulator” and the uncondensed gases are discharged to a control
device or directly to the atmosphere. The vacuum producing system associated with the vacuum
distillation of crude oil is typically the largest vacuum producing system at a petroleum refinery; vacuum
producing systems may also be used to evacuate reactor vessels or vacuum trucks. AP-42 provides a total
hydrocarbon emission factor for the vacuum distillation column condenser, which is provided in
Table 513. Speciation of the gases can be estimated based on the composition in the condensed liquids
5-27
and assuming the gases are saturated (in equilibrium) with the condensed liquid. If the emissions are
controlled, a control device efficiency correction can be used to estimate the controlled emissions.
Table 5-13. Default Emissions Factor for Vacuum Producing Systems
Emissions Factora
(lb/103 barrel
Process Description Pollutant of vacuum feed) Rating
Vacuum producing system Total hydrocarbonsb 50 C
on vacuum distillation
column, uncontrolled
a Source: U.S. EPA, 1995a.
b Total hydrocarbon may be estimated as a surrogate for VOC. (Overall, less than 1 weight % of total
hydrocarbon emissions is methane.)
5-28
Version 3 Section 6—Flares
6. Flares
Flares are point sources used at petroleum refineries to destroy organic compounds in excess refinery fuel
gas, purged products, or waste gases released during startups, shutdowns, and malfunctions. Most flares
have a natural gas pilot flame and use the fuel value of the gas routed to the flare to sustain combustion. If
the heating value of the flare gas falls below certain values (for steam or air assisted flares, typically 300
British thermal units per standard cubic feet [Btu/scf]), then natural gas may be added to the flare gas to
maintain the appropriate heating value for good combustion. Emissions from flares consist of a fraction of
the hydrocarbons in the flare gas (e.g., CH4, CO, VOC, and specific organic HAP) that are not combusted
in the flare; SO2 resulting from the oxidation of sulfur compound impurities, such as H2S, in the gas
stream; and CO2 from the combustion process. Flares are also expected to produce NOx emissions and
may produce PM (soot) if combustion conditions are not adequate. A complete emissions inventory will
include estimates for all these compounds (the specific organic HAP will vary based on the composition
of the gas being flared).
Accurate estimates of emissions from flares are difficult to obtain because they do not lend themselves to
conventional emissions-testing techniques, and only a few attempts have been made to characterize flare
emissions. Therefore, to date, there are limited direct emissions test data for flares. Recent developments
in testing protocols, such as the DIAL technique, provide a direct emissions measurement technique for
flares. However, DIAL measurements provide only a snapshot in time. Unless the flow and composition
of the flare gas is highly stable, inaccuracies build as these measurements are extrapolated to annual
emission rates. Continuous monitoring of the gas stream prior to combustion in the flare is generally the
most accurate means of assessing flare emissions. One difficulty with this approach is that flare gas
composition and flow are highly variable, and the monitors are calibrated to detect compositions or flows
within a certain range or span, so if the flow or composition is outside the instrument’s range, then
inaccuracies in the measurement data result. In addition, because the monitors are evaluating the stream
of gas going to the flare, assumptions must be made regarding flare efficiency to determine the emissions
following combustion. Engineering calculations are another methodology that can be used to assess
certain release events. For example, if a pressure relief valve on a tank opens, then the volume of gas
released can be calculated based on the pressure inside the tank, the pressure outside the tank, the cross-
sectional area of the valve opening, and the duration the valve is open (see Section 12,
Malfunctions/Upsets, in this Refinery Emissions Protocol document). Product knowledge of the tank
liquid composition can be used to calculate the equilibrium vapor space composition, which is assumed to
be the composition of the gas vented. Some emission factors are available, but these have high
uncertainty. EPA concludes that direct measurement methods are best used to develop site- or flare-
specific emission factors or to verify the combustion efficiency of a specific flare under certain
conditions; however, they are not particularly useful in developing an emissions inventory for flares.
Table 6-1 summarizes the hierarchy of flare emissions estimation techniques. Within a given
measurement method (or rank), there may be alternative methods for determining the constituent-specific
emissions; these compositional analysis methods are provided in order of accuracy. Each refinery will
likely use a mixture of different methods. For example, Methodology Rank 1 for flares may be used for
events that are directly monitored and are fairly routine releases, but Methodology Rank 4 for flares may
have to be used to estimate emissions for unusual events if the concentration of a particular pollutant or
the flow rate exceeds the range of the monitoring system employed. The remainder of this section
provides additional detail and guidance regarding the implementation of these methods.
6-1
Table 6-1. Summary of Flare Emissions Estimate Methodologies
Rank Measurement Method Additional Data Needed

1 Continuous composition monitoring (or  Combustion efficiency (based on results of a direct
manual sampling at least once every 3 hours measurement test, if available, or a default
during flaring events) and continuous flow assumption)
rate monitoring of the gas sent to the flare
2 Continuous flow rate monitoring and daily or  Representative sample (grab or integrated)
weekly compositional analysis
 Assumed combustion efficiency
3 Continuous flow rate and heating value  Emission factors based on heating value
monitoring
4 Engineering calculations  Process knowledge of units connected to flare (e.g.,
volume, composition of process streams)
 Temperature and pressure monitoring data or other
process operating data as needed
 Assumed combustion efficiency
5 Emission factors based on energy  Flow estimates (not continuous)
consumption
 Heat value estimates (not continuous)
6 Default emission factors based on refinery or  Refinery or process throughput
process throughput
6.1 Methodology Rank 1 for Flares

Methodology Rank 1 for flares includes continuously measuring composition and flow of gas sent to the
flare. For flares that do not have routine flow, manual sampling of the flare gas at least once every 3 hours
during each flaring event is also considered Methodology Rank 1 for flares. For a flare, the most likely
pollutants to be estimated using Methodology Rank 1 are SO2 and total hydrocarbons (or VOC). For
example, a reduced sulfur or total sulfur monitor can be used to characterize the sulfur content of the gas
being combusted in the flare. Similarly, monitoring of the total hydrocarbon, VOC, or specific organic
HAP content in the flare gas can be used to assess the emissions of these pollutants. However, because
these monitors measure the concentrations in the gas sent to the flare rather than the gas exiting the flare,
the other piece of information needed to estimate emissions from the flare is the flare efficiency (Feff). For
the example of sulfur content above, the portion of the sulfur that is oxidized to SO2 emissions during the
combustion of the flare gas will be dependent on the combustion efficiency. For emissions of compounds
that are emitted as a result of incomplete combustion, such as reduced sulfur compounds and
uncombusted hydrocarbons, one minus the assumed flare efficiency
(1-Feff) is used to estimate the emission rate.
As noted above, direct measurement methods are available and can be used to confirm the efficiency of a
flare. If efficiency information is available for a specific flare at a refinery under certain conditions, that
information should be used to estimate emissions for that flare where appropriate. For other flares,
Section 13.5 of Compilation of Air Pollutant Emission Factors notes that the flare combustion efficiency
for properly operated flares is at least 96.5%, which is equivalent to a destruction efficiency of 98% (U.S.
EPA, 1995a).
Most gas composition monitors measure the concentration on a volume basis (and generally on a dry
basis), although some will provide the concentration on a mass basis. Most flow monitors measure
volumetric flow rate in actual cubic feet. Many flow monitoring systems are equipped with temperature
and pressure monitors to automatically convert the flow to standard conditions. If the gas composition is
determined on a dry basis, which is typical for extractive monitoring systems, then it is important to also
correct the flow to be on a dry basis. The total volumetric flow of the pollutant can be converted to a mass
6-2
emission rate using the molecular weight of the pollutant and/or the ideal gas law, as appropriate. An
emissions estimate can be calculated using Equation 6-1, which is a variation of Equation 4-1 for CEMS:
N 
 (Q )n × [1 − ( f H 20 ) ]×  To   Pn  (Ci,n × PMN i,n ) MWi 
Ei = ∑  n T  ×   × K eff × × × K (Eq. 6-1)

n =1   n   Po  100% MVC 
where:
Ei = Emission rate of pollutant “i” (tons per year [tons/yr] or tons per event [tons/event]).
N = Number of measurement periods per year or per event (e.g., for hourly measurements,
N = 8,760 to calculate annual emissions).
(Q)n = Volume of gas sent to the flare for measurement period “n” (actual cubic feet [acf]). If
the flow rate meter automatically corrects for temperature and pressure, then replace
“(T)o ÷ (T)n × (P)n ÷ (P)o” with “1.” If the pollutant concentration is determined on a
dry basis and the flow rate meter automatically corrects for moisture content, replace
the term [1-(fH20)n] with 1.
(fH2O)n = Moisture content of exhaust gas during measurement period “n,” volumetric basis
(cubic feet water per cubic feet exhaust gas).
To = Temperature at “standard conditions” (520 °R or 528 °R).
Tn = Temperature at which flow is measured during measurement period “n” (°R).
Pn = Average pressure at which flow is measured during measurement period “n” (atm).
Po = Average pressure at “standard conditions” (1 atm).
Keff = Factor to account for the efficiency of the flare. Keff is equal to the flare combustion
efficiency Feff for pollutants created by combustion, such as SO2. Keff is equal to one
minus the flare destruction efficiency (1 – Feff) for pollutants in uncombusted gas, such
as VOC, reduced sulfur, and specific HAP.
Ci,n = Concentration of pollutant “i” or the appropriate precursor to pollutant “i” in the gas
sent to the flare for measurement period “n” (volume %, dry basis). If the pollutant
concentration is determined on a wet basis, then replace the term [1−(fH20)n] with 1.
PMNi = Precursor mole number (use a value of 1 for estimating emissions of pollutants that are
not not fully combusted). For example, for CS2 as a precursor to make SO2, the
pollutant mole number would be 2 because one mole of CS2 will yield 2 moles of SO2
when combusted.
MWi = Molecular weight of pollutant “i” (kilogram per kilogram mole [kg/kg-mol]).
MVC = Molar volume conversion factor (standard cubic feet per kilogram mole [scf/kg-mol])
= 836.6 scf/kg-mol at “standard conditions” of 60°F (520°R) and 1 atmosphere (atm)
= 849.5 scf/kg-mol at “standard conditions” of 68°F (528°R) and 1 atm.
K = Conversion factor = 2.2046/2,000 (tons per kilogram [tons/kg] = 0.0011023 tons/kg.
As explained in Section 4-1, a continuous monitor records multiple measurements per hour, and the
individual measurements can be used to calculate annual emissions in two ways. For a flare, the selection
of an hourly average approach or heavier reliance on individual measurements will depend on the
duration of the flare’s operation (particularly for flares that operate only intermittently) and the process
unit generating the gas being flared.

Methodology Rank 2 for flares includes continuous monitoring of flare gas flow as in Methodology
Rank 1 for flares and routine (daily or weekly) sampling and analysis to determine the composition of the
flare gas. Methodology Rank 2 also includes determining composition by continuously monitoring a
“surrogate” component, and then using sampling or other data to periodically check a correlation between
6-3
the surrogate and the pollutant. For a flare, the most likely pollutants to be estimated using Methodology
Rank 2 for flares are SO2 and specific HAP. A continuous monitor for H2S can be used to estimate SO2
emissions similar to the process described for Methodology Rank 1, but because H2S is only one of
several sulfur compounds potentially present in the flare gas, it is important to periodically characterize
the total sulfur content of the gas so as not to underestimate SO2 emissions from the flare. For example,
periodic sampling of the flare gas can be performed to determine the total sulfur content of the flare gas,
which can be used in conjunction with the H2S monitoring data to determine an average total sulfur-to-
H2S ratio of the flare gas. This average total sulfur-to-H2S ratio, expressed in SO2 equivalence, can be
used to adjust the H2S concentration measured by the continuous monitor to provide a more accurate and
complete assessment of SO2 emissions from the flare. Similarly, sampling can be used to determine the
relative concentration of specific HAP present in the flare gas. These sampling data can be used in
conjunction with total hydrocarbon monitoring data to estimate HAP emissions from the flare. As with
the Methodology Rank 1 for flares, the generation or destruction of pollutants in the flare will depend on
the flare combustion efficiency. As mentioned previously, the current default flare combustion efficiency
for properly operating flares is 96.5 %, which correlates to a destruction efficiency of 98% (U.S. EPA,
1995a).

Methodology Rank 3 for flares includes continuous monitoring of flare gas flow as in Methodology
Rank 1 for flares, continuous monitoring of the heating value of the fuel, and emission factors based on
the heating value. The most likely pollutants estimated using Methodology Rank 3 for flares are CO,
NOx, VOC. A general equation for Methodology Rank 3 for flares is as follows:
Ei = ∑ (Qstd ,n × LHVn × EFi )

N
(Eq. 6-2)
n =1
where:
Ei = Emission rate of pollutant “i” (lbs/yr or lbs/event).
N = Number of measurement periods per year or per event (e.g., for hourly measurements,
N = 8,760 to calculate annual emissions).
Qstd,n = Volume of gas sent to the flare for measurement period “n” (scf). If the flow meter
does not output flow at “standard conditions,” the Qstd,n can be calculated using the first
four terms within the summation of Equation 6-1. Note: the “standard conditions”
used for the volume of gas must match the “standard conditions” at which the heating
value of the gas is determined.
LHVn = Heating value of gas being flared during measurement period “n” on a lower heating
value (LHV) basis (MMBtu/scf). (Heating value for flares is usually determined on a
lower (net) heating value (LHV) basis whereas heating value for process heaters and
boilers are usually determined on a higher heating value basis.)
EFi = Emission factor for pollutant “i” (lb/MMBtu).
Values for “EFi” are presented in Tables 6-2 and 6-3 below. It is important to note that both sets of
emission factors are intended for use with properly operating flares meeting the criteria of 40 CFR 60.18;
therefore, the flare efficiency is already included in these emission factors. It is also important to note that
the number of components for which emission factors are available is limited because these emission
factors are based on general combustion characteristics and limited emission tests. For example, the
emission factors for NOx are generally based on the amount of nitrogen in the air during combustion, so if
there are constituents containing nitrogen in the fuel gas, such as ammonia, additional calculations are
6-4
needed to account for all NOx emissions (based on Methodology Rank 2 for flares, particularly if flow
rate is monitored continuously.)
The emission factors in Table 6-2 are values found in Section 13.5 of EPA’s Compilation of Air Pollutant
Emission Factors (U.S. EPA, 1995a, also known as AP-42) and are based on recent testing of flare
performance using passive Fourier transform infrared spectroscopyspectroscopy, extractive sampling, and
DIAL, as well as considering data from previous EPA testing on flares. These emission factors should be
used with LHV heat content measurements.
Table 6-2. Flare Energy Consumption-Based Emission Factorsa
Emission Factora
Pollutant (lb/106 Btu, LHV basis) Representativenessb
Volatile organic compounds 0.57 Poorly

Carbon monoxide (CAS No. 630-08-0) 0.31 Poorly
Nitrogen oxides 0.068
a Emissions factors for Section 13.5 of EPA’s AP-42 (U.S. EPA, 1995a)
b The nitrogen oxides emissions factor precedes the 2013 draft final recommended procedures for development of
emissions factors. As such, there is no representativeness assigned to it. Under the old rating letter-based
rating scale, this factor has a “B” rating.
Additional emission factors for soot (i.e., PM) are also provided AP-42; however, the soot factors
provided in units of concentration in the flare exhaust stream. These factors have been converted to
heating value-based factors to allow calculation of soot (PM) emissions using Equation 6-2. To calculate
the soot (PM) emissions from flares, each measurement period would be assigned a flare operation
category based on the amount of smoke generated by the flare during that measurement period, so the
appropriate emission factor could be applied. It should be assumed that all flare soot is fine PM
(PM2.5FIL).
Table 6-3. Emission Factors for Soot from Flares
Emission Factora Emission Factorb

Flare Operation (µg/L in exhaust) (lb/MMBtu, LHV basis)
Nonsmoking flares 0 0
Lightly smoking flares 40 0.027
Average smoking flares 177 0.12
Heavily smoking flares 274 0.19
a Source: U.S. EPA, 1995a; reported as micrograms per liter (µg/L) in flare exhaust.
b Calculated from concentration using F-factor method on a dry basis, assuming 3% O2 in exhaust
gas stream.
6-5
Example 6-1: Calculation of CO Emissions from a Flare with Continuous Monitors for
Flow Rate and Heating Value
Calculate hourly emissions from a refinery flare given:
– The flare is used to control routine emissions from one source at the refinery, so emissions
can be estimated using the hourly average flow rate and heat content measurements
– The average flow rate to the flare rate during a certain hour is 250 scfm
– The lower heating value of the flow to the flare during that hour is 900 Btu/scf, or 9E-04
MMBtu/scf (900 Btu/scf × 1E-06Btu/MMBtu)
Calculate hourly emissions using Equation 6-2:
∑ (Q × LHVn × EFCO )
N
ECO = std , n
n =1
Qstd,n = 250 scfm × 60 min (for 1 hr period) = 15,000 scf

ECO = (9E-04) × (15,000) × (0.34)
ECO = 4.59 lbs/hr
Annual emissions for the flare would be calculated as the sum of emissions estimates for all
hours the flare received the vent gas. If the process unit being controlled operated 8,400 hours
during the year and the flow rate and composition of the flare vent gas remained constant, the
annual flare emissions of carbon monoxide would be: 4.59 × 8,400 = 38,600 lbs.

Methodology Rank 4 for flares includes a variety of estimation methods based on engineering
calculations. Estimates of nearly all pollutants from flares can be calculated using this rank based on
process knowledge of units connected to the flare, including process unit volume, process stream
compositions, temperature, and pressure. Any of this information used in conjunction with periodic
sampling also constitutes Methodology Rank 4. The specific methods used will vary based on the ways in
which the flare is used and the units the flare serves. Estimating emissions from flares with routine flow,
such as those used for controlling blowdown from delayed coking units or other routine discharges, may
rely on key process operating parameters that can be defined in advance. For process upsets or
malfunctions, the estimation methods used may be very event-focused and are likely to rely on
engineering judgment as much as actual monitoring data. Estimation methods may also be needed from
flares with monitoring systems when an upset or malfunction causes the concentration or flow to exceed
the calibrated span of the monitor. A few examples are described in this section, but many others are
possible, given the complexity of refineries and differences between them. When emissions are estimated
using Methodology Rank 4 for flares, a description of the methodology, assumptions, and specific pieces
of data should be recorded and kept for future reference, both to document the emissions calculation
methodology and to ensure consistency from one estimate to the next for a particular process.
For routine or planned releases from flares not equipped with a continuous monitor, the refiner should
know the quantity and composition of the gas being released. If the gas is excess fuel gas that has been
amine-treated and mixed in a central drum, then the H2S concentration of the gas being flared should be
the same as the concentration going to any fuel gas combustion device (e.g., process heater, boiler) at the
refinery. (As noted previously, it is important to note that H2S may not be the only reduced sulfur
compound contributing to SO2 emissions.) Alternatively, if a fuel gas combustion device has an
6-6
instrument for measuring SO2 emissions, then the SO2 emissions from the flare can be calculated based on
the quantities of gas being combusted in the fuel gas combustion device and the flare. If the efficiency of
that fuel gas combustion device and the flare are not the same, then an efficiency correction may be
needed. (This method would account for all sulfur from the flare since SO2 emissions would be measured
directly.) As another example, if a refiner is emptying a tank and combusting the vapors, VOC and HAP
emissions can be calculated based on the composition of the tank contents and the size of the tank.
For emergency flares and flares with flare gas recovery systems (or otherwise equipped with water seals
intended to prevent gases from going to the flare during normal operations), an alternative method of
estimating flow rate for non-routine flaring events is to monitor the pressure drop across the flare water
seal drum. A limited number of flow measurements can be made at different flare gas and water seal
differential pressures (or water heights) to develop a flow rate correlation with the measured pressure
drop. The water seal differential pressure or water height is routinely measured to determine the flow to
the flare. Alternatively, the pressure in the flare line can also be monitored using a direct pressure monitor
installed at the final liquid knock-out drum prior to the water seal. Concentrations of specific compounds
are estimated by sampling and analysis of the flare gases (preferred, if possible) or calculations based on
the composition of the gases in the process unit(s) contributing to the flaring event. Emissions are then
calculated using those concentrations, the calculated flow rate, and an assumed flare efficiency. Some
refiners that use this method for flares with a flare gas recovery system also monitor the recovery system
to ensure that gases are not sent to the flare during normal operation.
6.5 Methodology Ranks 5 and 6 for Flares

As previously mentioned, flares are difficult to test, so only a few emission factors have been developed
for flares. Available emission factors for flares were provided in Tables 6-2 and 6-3 for use with
continuous flow and heat content monitoring. Methodology Rank 5 for flares consists of using these same
emission factors, but with estimates of the flow rate and heating content of the flare gas rather than direct
measurement of these values. The total VOC emission factor could be used with knowledge of the
components in the flare gas to estimate HAP emissions.
Methodology Rank 6 for flares includes emission factors for HAP that were developed for the Petroleum
Refinery Source Characterization and Emission Model for Residual Risk Assessment (RTI, 2002) based
on the State of Louisiana’s title V permit applications data. The original emission factors presented in
RTI’s document (2002) were based on arithmetic averages. However, the arithmetic average is
occasionally skewed by one high estimate. The log-mean average emission factors from the State of
Louisiana’s title V permit applications data were calculated and are provided in Table 6-4. These
emission factors yield a cumulative emission estimate from all flares (not per flare) for the refinery based
on the total refinery crude capacity in barrels per calendar day (bbl/cd). It is uncertain how the permit
application estimates were developed and whether these estimates include SSM events, but it is likely that
these factors would not include emissions released during significant SSM events. As such, these
emission factors are very general and are not recommended unless no other information is available.
Refineries are expected to have more source-specific information for estimating flare emissions, but these
factors may be used for the purposes of modeling or for general emissions estimates.
The application of these emission factors is straightforward, using Equation 6-3.
Ei = Cap × EFi (Eq. 6-3)

where:
Ei = Mass emissions per year of pollutant “i” (tons/yr)
Cap = Crude capacity of refinery (bbl/cd)
6-7
EFi = Emission factor for pollutant “i” (tons/yr/bbl/cd)
Table 6-4. Flare General Emission Factorsa
Emission Factor
CAS Number Component (tons/yr/bbl/cd)
71-43-2 Benzene 9 E-06

108-88-3 Toluene 7 E-06
1330-20-7 Xylene 6 E-06
1634-04-4 Methyl tertiary-butyl ether 3 E-06
110-54-3 Hexane 1 E-05
50-00-0 Formaldehyde 1 E-06
100-41-4 Ethylbenzene 2 E-07
106-99-0 1,3-Butadiene 7 E-06
a Source: Log-mean average from State of Louisiana’s title V permit
applications data
Example 6-2 shows the calculation of emissions from flares at a refinery with a given crude capacity.
Example 6-2: Emissions Factor

Given that a refinery has a crude capacity of 100,000 bbl/cd crude capacity, the annual
emissions of benzene (C6H6) emissions can be calculated from flares at the refinery as follows:
E C6 H 6 = Cap × EFC6 H 6
EC6H6 = (100,000) × (9E-06)
EC6H6 = 0.9 tons/yr
6-8
Version 3 Section 7—Wastewater Collection and Treatment Systems
7. Wastewater Collection and Treatment Systems

Industrial wastewater collection and treatment operations range from simple pre-treatment operations that
discharge to publicly owned treatment works (POTW) to full-scale wastewater treatment systems. Full-
scale treatment systems include the collection of process and/or storm water that is treated to a quality
that is acceptable for discharge to a receiving water body or for re-use.
Wastewater collection systems differ among facilities, but they generally include drains, manholes,
trenches, lift stations, sumps, junction boxes, and weirs. As water passes through each of these system
components, emissions may occur by volatilization of organic compounds at the water/air interface.
The overall objectives of wastewater treatment at refineries are to: (1) equalize flow and pollutant load by
buffering flow surges in large tanks, (2) separate free and emulsified oils and solids from the wastewater
by oil/water separators and flotation unit operations (e.g., dissolved air flotation [DAF], induced air
flotation [IAF], dissolved nitrogen flotation [DNF]), and (3) oxidize organic molecules and remove or
transform nutrients through biodegradation. Figure 7-1 shows a typical full-scale refinery wastewater
treatment scheme.
Collection Oil-Water Dissolved Air Equalization

Aeration Tank
System Separator Flotation Unit Tank
Return
Sludge Clarifier
Process
Wastewater Slop Oil Tank Sludge to Coker,
In Land Treatment,
or Disposal
Return to
Crude Polishing Pond
Processing
Discharge
Figure 7-1. Typical refinery wastewater treatment system.
Tanks or treatment units upstream of any benzene removal or destruction systems (e.g., stream strippers,
enhanced biodegradation units) at most petroleum refineries are regulated under BWON (Benzene Waste
Operations NESHAP) and thus should be covered and vented to appropriate control devices. Equalization
tanks, oil/water separators, and flotation units typically fall into this category; however, in the event these
units are not covered because they do not contain high benzene quantities, emission estimation methods
are presented in this Refinery Emissions Protocol document. Biological treatment units are typically the
first uncovered process in the wastewater treatment system and thus the first air emission source.
Table 7-1 lists the emission estimation methods for wastewater treatment systems. The methods are
ranked according to anticipated reliability. There are three primary estimation methods: (1) direct
measurement, (2) predictive modeling, and (3) engineering estimates. Direct measurements can only be
7-1
taken from process units that are covered and vented to a control device. Predictive modeling can be
accomplished using analytical equations, such as those presented in AP-42, by using computer-based fate
and transport models (e.g., WATER9, TOXCHEM), or by using the simplified refinery wastewater
emission tool (RWET) presented in this Refinery Emissions Protocol document. Engineering estimates
can be made using emission factors based on crude throughput or wastewater load.
Table 7-1. Summary of Wastewater Treatment Emission Estimates
Rank Measurement Method Application Data Requirements

1 Direct measurement Covered and  Constituent load and speciation of collected
vented units gas samples
2a Predictive modeling with site- Uncovered units  Constituent load and speciation of process
specific factors and wastewaters
biodegradation rates followed by  Site-specific biodegradation rates
validation  Model validation by a direct measurement
method
2b Predictive modeling with site- Uncovered units  Constituent load and speciation of process
specific factors and wastewaters
biodegradation rates  Site-specific biodegradation rates
2c Predictive modeling with site- Uncovered units  Constituent load and speciation of process
specific factors wastewaters
3a Engineering estimates based on Uncovered units  Constituent load and speciation of process
wastewater treatment plant load wastewaters
3b Engineering estimates based on Uncovered units  Crude throughput
crude throughput
7.1 Methodology Rank 1 for Wastewater Treatment Units

Emissions from covered and vented wastewater treatment units and drainage system components can be
directly measured using CEMS or periodic sampling; emission inventory estimates for these vented
sources can be developed using the methods summarized in Section 4, Stationary Combustion Sources,
and Section 5, Process Vents, of this Refinery Emissions Protocol document.
Although there are direct measurement methods for uncovered units (i.e., off gas collectors and open path
optical methods such as DIAL), these methods do not provide continuous monitoring data. Therefore,
these methods are not recommended as primary techniques for emission estimations, but are reasonable
for predictive modeling validation.
7.2 Methodology Rank 2 for Wastewater Treatment Units

Air emission estimation modeling techniques use mathematical equations to predict the fate and transport
of specific constituents in wastewater. This section describes some key wastewater treatment units and the
applicable methods for estimating the emissions from these units. Wastewater collection systems are
considered an initial wastewater treatment unit to account for the emissions.
AP-42 presents a variety of emissions equations for estimating air emissions from various waste water
treatment systems (U.S. EPA, 1995a). There are various computer-based fate and transport models (e.g.,
WATER9 and TOXCHEM), that implement these equations for specific wastewater treatment units and
that support the development of system-wide wastewater collection and treatment emissions estimates.
Although successful use of the available predictive models has been demonstrated, reporting facilities
have expressed concerns regarding complexity, user friendliness, and accuracy; therefore, a simplistic
refinery wastewater emission tool (RWET) has been developed as part of this document to address these
issues and provide more accurate comparisons among refineries. RWET was designed to help the user
7-2
identify critical model inputs for each process unit, recognize variables and constants that may be site-
specific, and calculate air emissions and constituent effluent concentrations. Additionally, constituent
effluent concentrations from a treatment unit can be used as inputs for the next downstream unit when
processes are linked in a series (e.g., oil-water separator effluent data equals input data for the DAF unit).
The “Critical Inputs” section of the refinery wastewater emission tool includes data that is required to
estimate air emissions from the specific process unit. The “Variables with Default Values” section
includes inputs that may be site-specific or unknown to the user. Default values are taken from AP-42, but
they can be changed if more accurate data are known or determined. The Chemical Properties sheet
contains the best available chemical, physical, and biodegradation information on the 30 organic HAP
listed in the Petroleum Refinery MACT I. However, if more accurate or site-specific data are available,
then these data can be inserted in the Site-Specific Table on the Chemical Properties sheet, and the model
will override the default values. Detailed instructions for use and application are provided in the refinery
wastewater emission tool.
Since reliable variables to estimate emissions for all volatile organic compounds (VOC) are not available,
air emission estimates for compounds not listed in the MACT I can be made using surrogate compounds.
Butane is the surrogate compound used for C2 – C4 VOC and octane for C4 – C9 in the refinery
wastewater emission tool. Air emission estimates for compounds not listed and considered non-volatile
(e.g., > C9 and Henry’s law constant > 10-3 atm-m3/gmol) can be made using the Compound A, B, or C
rows in the tool. The following list of chemical, physical, and biological degradation properties for each
compound are required to reliably estimate air emissions:
1. Molecular weight
2. Vapor pressure
3. Henry’s law constant
4. Diffusivity in water
5. Diffusivity in air
6. Octanol-water partitioning coefficient
7. Maximum biorate constant
8. Half saturation biorate constant
The calculations used in RWET and variable definitions are based on those presented in AP-42 and on
peer-reviewed journals and are presented in detail in Appendix B, Wastewater Treatment System
Equations, of this Refinery Emissions Protocol document. RWET is available for free at the following
EPA website: http://www.epa.gov/ttn/chief/efpac/esttools.html.
7.2.1 Wastewater Collection Systems

Wastewater collection systems are complex networks of various components that drain large surface
areas. These complexities make accurately estimating air emissions a very challenging task. RWET
utilizes process drainage areas (PDAs), shown schematically in Figure 7-2, to more-easily estimate air
emissions. Rather than summing all of the drains and other components within the processing area, model
drainage systems are developed for each PDA. The model drainage components are based on the typical
(or reasonable worst-case) components that an individual waste stream will experience. In the top
drainage area of Figure 7-2, an individual waste stream would experience typically two drains (or, for a
worst-case assessment, three drains), one junction box, one manhole cover, and a lift station. For the
lower drainage area of Figure 7-2, an individual waste stream in this area would experience typically two
drains (or, for a worst-case assessment, three drains), a length of open trench (from either the middle or
7-3
furthest most drain for the typical or worst-case assessment, respectively), one sump, and a lift station.
Each of the model drainage system components may be identified as either controlled or uncontrolled.
Components meeting the requirements for controlled drainage components in BWON are considered
controlled components. The drainage components do not necessarily have to be subject to BWON to be
considered controlled, only that the specified controls are used (e.g., water seals for drains, gasketed
manhole covers, or drainage systems vented to a control device).
For each PDA, the wastewater generation rate and pollutant concentration at the point of generation
(POG) are needed to determine the pollutant load to the system. Model drainage components, constituent
load, and control efficiencies are then utilized within RWET to estimate air emissions for a given PDA.
The emissions from each PDA are then combined to obtain the total for the wastewater collection system.
More details of this method are presented in Appendix B, Wastewater Treatment System Equations.
Table 7-2 lists the critical inputs, variables with default values, and chemical properties specific to
wastewater collection system PDAs.
As previously stated, RWET utilizes model drainage system components to ease the complexity of
estimating air emissions from collection systems. However, if system components (e.g., lift stations,
junction boxes, and sumps) vary greatly from those described in the Best Available Control Technology/
Lowest Achievable Emission Rate (BACT/LAER), site-specific air emissions can be determined from the
procedures described in Appendix B, Wastewater Treatment System Equations.
Process Drainage Area
Manholes
Branch
Lateral
Junction Sewer
Sewers Drain Risers
Boxes
Trunk
Sewer
Sump
Open Trench
Lift
Station
To Wastewater
Treatment
Figure 7-2. Typical refinery wastewater collection system process drainage areas.
7-4
Table 7-2. Critical Inputs and Chemical Properties Specific to Wastewater Collection System
PDA Air Emission Calculations
Critical Inputs Variables with Default Values Chemical Properties

Number of drains Controlled / Uncontrolled Henry’s Law Constant (H)
Linear meters of open trench Control Efficiency
Number of manholes
Number of junction boxes
Number of lift stations
Number of sumps
Wastewater flow rate (Q)
Constituent concentration at POG (C0)
7.2.2 Primary Weirs

Weirs can serve as open-channel dams and utilized in settling basins to discharge cleaner effluent. As
previously stated, these components of the wastewater collection system are typically covered and vented;
therefore, open-air emission estimations are not required. However, if the treatment units are uncovered,
then the refinery wastewater emission tool can be used to estimate air emissions. Table 7-3 lists the
critical inputs, variables with default values, and chemical properties specific to weirs.
Table 7-3. Critical Inputs, Variables, and Chemical Properties Specific
to Primary Weir Air Emission Calculations

Wastewater flow rate (Q) Diffusivity of oxygen in water (DO2,w) Diffusivity of constituent in water (Dw)
Constituent influent concentration (C0) Henry’s Law Constant (H)
Weir height (h)
7.2.3 Oil-Water Separators

An oil-water separator is a treatment unit designed to separate oil and suspended solids from wastewater.
As previously mentioned, these treatment units are typically covered and vented; therefore, open-air
emission estimations are not required. However, if the treatment units are uncovered, then the refinery
wastewater emission tool can be used to estimate air emissions. Table 7-4 lists the critical inputs,
variables with default values, and chemical properties specific to oil-water separators.
If an oil-water separator is covered except for the effluent weir, air emissions can be calculated by using
the methods described in Section 7.2.3 (primary weirs) and the associated sheet in RWET.
to Oil-Water Separators Air Emission Calculations

Surface area (A) Flow of oil (Qoil) Diffusivity of constituent in air (Da)
Oil layer thickness (Olayer) Density of oil (ρoil) Octanol–water partitioning coefficient
(Kow)
Total pressure (P0) Molecular weight of oil (MW oil) Vapor pressure of constituent (P*)
Wastewater flow rate (Q) Density of air (ρair)
Constituent influent concentration (C0) Molecular weight of air (MW air)
Wind speed (U10) Viscosity of air (µa)
Fraction of volume that is oil (FO)
7-5
7.2.4 Dissolved Air Flotation Units

DAF is a wastewater treatment unit that uses bubble flotation to remove suspended oil and solids from
water. As previously mentioned, these treatment units are typically covered and vented to control devices.
However, if the treatment units are uncovered, then the refinery wastewater emission tool can be used to
estimate air emissions. Table 7-5 lists the critical inputs, variables with default values, and chemical
properties specific to DAF units.

to DAF Air Emission Calculations

Surface area (A) Viscosity of water (µL) Diffusivity of constituent in air (Da)
Temperature (T) Density of water (ρL) Diffusivity of constituent in water (Dw)
Total pressure (P0) Molecular weight of water (MWL) Henry’s law constant (H)
Wastewater flow rate (Q) Density of air (ρair)
Constituent influent concentration (C0) Molecular weight of air (MW air)
Diffused air flow rate (Qa) Viscosity of air (µa)
Wind speed (U10)
7.2.5 Equalization Tanks

Equalization tanks dampen variations in wastewater flow rate and pollutant load to lessen negative
impacts on downstream processes. As previously mentioned, these tanks are commonly covered and
vented to control devices; however, if the treatment units are uncovered, then the refinery wastewater
emission tool can be used to estimate air emissions. Table 7-6 lists the critical inputs, variables with
default values, and chemical properties specific to equalization tanks.

to Equalization Tank Air Emission Calculations

Surface area (A) Density of water (ρL) Diffusivity of constituent in air (Da)
Temperature (T) Viscosity of water (µL) Diffusivity of constituent in water
(Dw)
Total pressure (P0) Density of air (ρair) Henry’s law constant
Wastewater flow rate (Q) Molecular weight of air (MW air)
Constituent influent concentration Viscosity of air (µa)
(C0)
Wastewater depth (D) Rotational speed of impeller (w)
Total power to aerators (Ptot) Quiescent surface area (AQ)
Number of aerators (NI) Oxygen transfer correction factor
(Ot)
Turbulent surface area (AT) Oxygen transfer rating to surface
area (J)
Wind speed (U10) Impeller diameter (d)
7.2.6 Biological Treatment Units

Biological treatment is an effective process to reduce, remove, or transform organic constituents and
nutrients typically found in refinery wastewater to an acceptable form or concentration prior to discharge
or re-use. However, biological treatment units are complex and vary greatly in design, operation, and
7-6
treatment efficiency, resulting in units that are difficult to characterize. Multiple fate and transport
mechanisms are often involved in the ultimate removal of a specific compound. As seen in Figure 7-3,
biodegradation, volatilization, adsorption, hydrolysis, and photo degradation are common mechanisms in
wastewater treatment that may compete against each other. Additionally, biological systems are dynamic
in nature, resulting in shifts in the dominant fate mechanism. Therefore, it is important to obtain and use
site-specific variables when estimating emissions to obtain accurate results.
Constituent Out
(Air)
Biological
Constituent Out
Treatment Unit (Wastewater)
Constituent In (Wastewater) Reactions and

Absorption
Constituent Out
(Sludge)
Figure 7-3. Simplified drawing of a constituent mass balance in a biological treatment unit.
The two primary classes of biological treatment units are suspended-growth and attached-growth systems.
Suspended-growth systems maintain the biomass (flocs) in suspension by mechanical or aeration devices
where biodegradation occurs. Examples of suspended-growth systems include activated sludge, aerated
lagoons, sequencing batch reactors, and membrane bioreactors. Attached-growth systems establish
biofilms on fixed surfaces where biochemical reactions occur. Examples of attached-growth systems
include trickling filters, rotating biological contactors, and fluidized-bed reactors. Regardless of the type
of biological treatment used, the biochemical reactions are generally the same, with the oxidation of
organic compounds and ammonia forming new cells, CO2, and water.
To maintain the health of the microorganisms and ensure adequate biologically mediated oxidation of the
organic compounds and ammonia, O2 must be introduced into the system. There are two basic methods
for aerating wastewater: (1) introducing air or pure O2 into the wastewater through diffusers or other
devices, and (2) agitating the water to increase surface area and promote the gaseous exchange with the
atmosphere. However, the mechanisms that increase dissolved oxygen concentration in the water also
increase the emission rate of organic constituents through increased volatilization. The volatilization rate
can be affected by the wastewater surface area, turbulence, and temperature; hydraulic retention; water
depth; concentration and physical properties of the organic constituents (e.g., volatility, diffusivity,
inhibitory mechanisms); and atmospheric conditions.
As previously mentioned, the biological treatment unit will most likely be the first uncovered process in
the wastewater treatment system and potentially the greatest source of air emissions. Table 7-7 lists the
critical inputs, variables with default values, and chemical properties specific to biological treatment unit
air emission calculations.
7-7

to Biological Treatment Unit Air Emission Calculations

Surface area (A) Density of water (ρL) Diffusivity of constituent in air (Da)
Temperature (T) Viscosity of water (μL) Diffusivity of constituent in water
(Dw)
Total pressure (P0) Density of air (ρair) Henry’s law constant
Wastewater flow rate (Q) Molecular weight of air (MW air) Maximum biodegradation rate
constant (Kmax)
Constituent influent concentration Viscosity of air (µa) Half saturation biorate constant (Ks)
(C0,X)
Wastewater depth (D) Weight fraction of the total carbon in Octanol-water partitioning
biomass (fOC) coefficient (Kow)
Total power to aerators (Ptot) Quiescent surface area (AQ)
Number of aerators (NI) Oxygen transfer correction factor
(Ot)
Mixed liquor volatile suspended Oxygen transfer rating to surface
solids (CMLVSS) area (J)
Influent BOD concentration (C0,BOD) Impeller diameter (d)
Wind speed (U10) Rotational speed of impeller (w)
Turbulent surface area (AT) Biomass volatile suspended solids
yield from influent biochemical
oxygen demand (Y)
Wasted sludge flow rate (Qw)
7.2.7 Polishing Ponds

Although air emissions downstream of biological treatment should be minimal, ineffective treatment,
constituent slugs, or hydraulic surges may lead to HAP contamination of tertiary treatment processes (i.e.,
polishing ponds). If air emission estimates are needed for polishing ponds, then the equalization tank
sheet of the refinery wastewater emission tool can be used. Although polishing ponds may function as a
plug-flow reactor, the assumption can be made that volatilization is the major constituent fate mechanism.
If the polishing pond is aerated and has significant active biomass (> 40 milligrams per liter [mg/L]
volatile suspended solids), the biological treatment sheet of the model may be used.
7.2.8 Site-Specific Factors

There has been tremendous effort to compile default values for specific variables to estimate air
emissions; however, site-specific data provide the most accurate results. The factors that can have the
most dramatic impact on air emissions from a biological treatment unit are the ones impacting
biodegradation. Compound-specific biodegradation rate constants (i.e., k0 and k1) and the half-saturation
concentration (i.e., KS) can be determined by using the aerated reactor test (BOX test). The empirically
derived values can then be used in a predictive model for more accurate results than can be developed
when default biodegradation rate constant values are use. The methods used to determine the fraction of
organic constituent biodegraded are provided in 40 CFR Part 63, Appendix C (also available at:
http://www.epa.gov/EPA-AIR/2004/June/Day-30/a14826.htm).
Although these methods have been successfully used to estimate biodegradation rates, there are concerns.
Specifically, it is argued that dosing the bioreactor with only one constituent of interest yields inaccurate
biodegradation rates compared to those dosed with a mixture of compounds commonly encountered by
the microorganisms. Preferential biodegradation, degradation by-products, and co-metabolism are
7-8
important factors to consider when determining biodegradation rates that are not addressed in the
recommended methods. Yerushalmi and Guiot (1998) reported biodegradation rates of benzene and
toluene that were 62.9 and 16.4 times greater, respectively, when used as the sole substrate, versus a
mixture of organic compounds. It is recommended that the constituent of interest be dosed in the
appropriate ratio with organic compounds found in the real-world system (as described in Appendix C).
Additionally, the methods do not distinguish between biodegradation, hydrolysis, and adsorption. Rather,
the results of the BOX test are a summation of the three primary fate mechanisms common to aerated
bioreactors. This fact is important to consider when assessing all the fate mechanisms involved in
compound degradation.
7.2.9 Model Validation

To ensure that predictive model outputs are accurate, validation studies can be conducted to support the
results. This task is accomplished by secondary direct or in-direct measurement techniques such as offgas
collectors, DIAL, concentration-profile methods. Direct measurements are taken of a modeled process
unit, and the results are compared. Favorable comparisons are indicative of accurate predictive modeling,
whereas poor comparisons could be the result of incorrect assumptions or errors in the model. If
corrective actions are necessary, a review of the constants and site-specific variables should be conducted.
7.3 Methodology Rank 3 for Uncovered Units
7.3.1 Engineering Estimates Based on Wastewater Treatment Plant Load

If a facility only has benzene wastewater concentration data, then this method can be used to estimate the
unknown HAP concentrations for use in the predictive modeling methods in Section 7.2, Methodology
Rank 2 for Uncovered Units. The wastewater generation rates, benzene concentrations, and total HAP
concentrations presented in Table 7-8 (U.S. EPA, 1998a) were combined with the total U.S. refinery
processing and production capacities reported by EIA (2009) to calculate an average wastewater mixture
to determine an average benzene and total HAP concentration of refinery wastewater. These data were
evaluated using the average petroleum stream composition provided in Appendix B, Wastewater
Treatment System Equations, of this Refinery Emissions Protocol document and an evaluation of the log
octanol water partition coefficient to help speciate the total HAP concentration before and after oil
removal. The results of this analysis (see Table 7-9) are relative concentrations of key pollutants as a ratio
of the benzene concentration. If the benzene concentration of the refinery wastewater is known, then the
concentration of other organics in the wastewater can be estimated using the ratios in Table 7-9.
7.3.2 Engineering Estimates Based on Process Capacities

Facilities that do not have any of the required data to estimate air emissions (e.g., for proposed units) can
use the factors in Table 7-8 to estimate wastewater generation rates and benzene concentrations to the
wastewater treatment system. Either the factors in Table 7-8 can be used to estimate the total organic
HAP concentration, or the factors in Table 7-9 can be used to estimate the specific HAP (and total HAP)
concentrations of the wastewater. The generated data can then be used to estimate specific constituent
wastewater loads (see Section 7.3.1, Engineering Estimates Based on Wastewater Treatment Plant Load)
and used in the predictive modeling methods in Section 7.2, Methodology Rank 2 for Uncovered Units. If
the wastewater treatment system design is unknown, then the default emission factors in Table 7-10 can
be used to estimate the fraction of the total constituent load to the wastewater treatment system
(calculated using the factors in Tables 7-8 and 7-9) that is expected to be released into the atmosphere.
7-9
Table 7-8. Model Process Unit Characteristics for Petroleum Refinery Wastewatera
Average Flow Average Benzene Average Organic

Factorb Concentrationc HAP Concentration
Process Unit (gal/bbl) (ppmw) (ppmw)
Crude distillation 2.9 21 140
Alkylation unit 6 3 6.9
Catalytic reforming 1.5 106 238
Hydrocracking unit 2.6 14 72
Hydrotreating/hydrorefining 2.6 6.3 32
Catalytic cracking 2.4 13 165
Thermal cracking/coking 5.9 40 75
Thermal cracking/visbreaking 7.1 40 75
Hydrogen plant 80d 62 278
Asphalt plant 8.6 40 75
Product blending 2.9 24 1,810
Sulfur plant 9.7e 0.8 3.4
Vacuum distillation 3 12 53
Full range distillation 4.5 12 65
Isomerization 1.5 33 117
Polymerization 3.5 0.01 0.04
MEK dewaxing units 0.011 0.1 27
Lube oil/specialty processing unit 2.5 40 75
Tank drawdown 0.02 188 840
Note: gal/bbl = gallons of wastewater per barrel of capacity at a given process unit, ppmw = parts per million by
weight
a Source: U.S. EPA, 1998.
b All flow factors were derived from Clean Air Act Section 114 questionnaire responses
c Average concentration in the wastewater
d This flow factor is given in gallons per million cubic feet (gal/MM ft3) of gas production
e This flow factor is given in gal/ton of sulfur
Table 7-9. Refinery Wastewater Contaminant Concentrations as a Ratio to Benzene
Mass Concentration Ratio of Compounds

to the Concentration of Benzene
Inlet to Biological
Inlet to Oil-Water Treatment Unit (After Oil-
CAS No. HAP Separator or DAF Unit Water Separator)
540841 2,2,4-Trimethylpentane 1.97 0.022
71432 Benzene 1.00 1.00
92524 Biphenyl 0.034 0.0005
1319773 Cresols 0.25 0.38
98828 Cumene 0.37 0.013
100414 Ethylbenzene 0.88 0.086
110543 Hexane 3.50 0.047
(continued)
7-10
Table 7-9. Refinery Wastewater Contaminant Concentrations as a Ratio to Benzene (continued)
Mass Concentration Ratio of Compounds

to the Concentration of Benzene
Inlet to Biological
Inlet to Oil-Water Treatment Unit (After Oil-
CAS No. HAP Separator or DAF Unit Water Separator)
1634044 Methyl tertiary-butyl ether 0.58 0.98
91203 Naphthalene 0.29 0.02
108952 Phenol 0.18 0.80
100425 Styrene 0.58 0.09
108883 Toluene 3.3 0.80
1330207 Xylene 3.6 0.33
106990 1,3-Butadiene 0.0006 0.0027
VOC Total VOCs 81 17
Table 7-10. Default Mass Emission Factors for Refinery Wastewater Collection
and Treatment Systems
Mass Fraction of Compound Emitted from

Wastewater Collection and Treatment System Based
on Total Wastewater Loading
Open Wastewater BWON-Compliant
Collection and Wastewater Collection
CAS No. HAP Treatment System and Treatment System
106990 1,3-Butadiene 0.91 0.75
540841 2,2,4-Trimethylpentane 0.95 0.55
71432 Benzene 0.65 0.25
92524 Biphenyl 0.34 0.031
1319773 Cresols 0.002 0.000
98828 Cumene 0.68 0.24
100414 Ethylbenzene 0.66 0.22
110543 Hexane 0.97 0.55
1634044 Methyl tertiary-butyl ether 0.45 0.091
91203 Naphthalene 0.41 0.098
108952 Phenol 0.000 0.000
100425 Styrene 0.81 0.64
108883 Toluene 0.66 0.19
1330207 Xylene 0.64 0.21
VOC Total VOCs (using butane) 0.94 0.60
7-11
Example 7-1. Methodology Rank 3 for Wastewater

A refinery has the following processing and production rates. The average annual emissions of
hexane, benzene, toluene, and xylene can be estimated from the wastewater collection and treatment
system for a BWON-compliant refinery.
Process Unit or Operation Average Throughput (bbl/cd)

Atmospheric distillation 100,000
Vacuum distillation 50,000
Catalytic reforming 20,000
Catalytic cracking 40,000
Alkylation 10,000
Product blending 60,000
Total crude and product tank throughput 180,000
Use the factors in Table 7-8 and the density of water (ρwater = 8.34 lb/gal) to estimate the total
wastewater flow rate and benzene loading rate.
(A) (B) (C) = A×B (D)
Average Wastewater Wastewater Benzene (E) = C×ρ×D/106
Process Unit Throughput Flow Factor Flow Concentration Benzene Mass
or Operation (bbl/cd) (gal/bbl) (gal/cd) (ppmw) Flow (lb/cd)
Atmospheric 100,000 2.9 290,000 21 50.8
distillation
Vacuum 50,000 3 150,000 12 15.0
distillation
Catalytic 20,000 1.5 30,000 106 26.5
reforming
Catalytic 40,000 2.4 96,000 13 10.4
cracking
Alkylation 10,000 6 60,000 3 1.5
Product 60,000 2.9 174,000 24 34.8
blending
Tank 180,000 0.02 3,600 188 5.6
drawdown
Totals 803,600 144.7
Use the factors in Table 7-9 to estimate the mass loading rate of other compounds and the
BWON-compliant emission factors in Table 7-10 to estimate the annual emissions.
(A) (C) (D) = B×C×365/2000
Mass Ratio from (B) = A×144.7 Emission Factor Annual Emissions
Compound Table 7-9 Mass Flow (lb/cd) from Table 7-10 (tons/yr)
Hexane 3.5 506.5 0.55 51
Benzene 1.0 144.7 0.25 6.6
Toluene 3.34 483.3 0.19 17
Xylene 3.57 516.6 0.21 20
7-12
Version 3 Section 8—Cooling Towers
8. Cooling Towers
Cooling water is used in refineries in heat exchange systems and condensers to cool or condense various
product streams. The cooling water for a closed-loop recirculation system may be sent to a cooling tower,
where it is cooled to near ambient temperature for reuse, and then is returned to the process or to
refrigeration units for additional cooling. The cooling water for a once-through system may be cooled and
sent to a receiving waterbody. This section focuses on cooling towers because these are used
predominately in the petroleum refining industry. Some refineries use once-through cooling systems
(emission estimates for once-through cooling systems can be estimated using essentially the same
techniques used for cooling towers).
Organic HAP and VOC are picked up by cooling water when leaks occur in heat exchangers or
condensers. Product on the high-pressure side leaks through cracks in the exchanger and contaminates the
water. Organic HAP, VOC, PM10, and chlorine are subsequently emitted from the water into the
atmosphere due to stripping (i.e., active air and water contact) in the cooling tower and drift loss. Large
emissions on the order of tons per year can occur for low levels of contamination or concentration
because refineries use large volumes of cooling water. Generally, CH4 or other GHGs are not significant
pollutants for cooling towers; however, if the cooling water is used on a high CH4 process stream, then
CH4 emissions can be calculated using the same methods presented in this section for calculating volatile
hydrocarbon emissions from cooling towers. The cooling tower emission estimates require information on
the cooling water flow rate and the composition of the process stream in the heat exchanger or in the
condenser or refrigeration units. These emission estimates are also dependent on the type of heat
exchange monitoring system used.
Table 8-1 summarizes the hierarchy of cooling tower emission estimation techniques. The methods are
ranked in terms of anticipated reliability. Within a given measurement method (or rank), there may be
alternative methods for determining the constituent-specific emissions; these compositional analysis
methods are also provided in order of reliability. Multiple methods are provided for volatile organic HAP
and VOC, while one method is provided for PM10, chlorine, and heavy organic HAP. Cooling towers may
service more than one process and have multiple cooling water return lines. The monitoring performed
on different cooling water return lines may differ for return lines leading to a single cooling tower. The
methods described in this section can be applied to individual return lines (if the flow rate of each line is
known) so that the total emissions from a cooling tower may be calculated as the sum of the emissions
estimated for each return line. Different return lines to a single cooling tower may use different rank
methodologies based on difference in the monitoring conducted on the individual return lines.
The remainder of this section provides additional detail and guidance regarding the implementation of
these methods. In general, annual emissions from cooling towers should account for the full operating
time of the cooling tower on an annual basis. Annual cooling tower emissions can be estimated using the
measured concentrations for each monitoring period (whether the measured concentration constitutes a
leak or not) and summing emissions over all monitoring periods or can be estimated using emission
factors for the full annual operating time of the cooling tower. Hourly emissions should be estimated
based on the highest hourly emission rate for the cooling tower; when default emission factors are used,
the hourly emissions should be determined based on the highest recirculation rate experienced for that
year.
8-1
Table 8-1. Summary of Cooling Tower Emissions Estimation Methodologies
Rank Measurement Method or Emission Factor Compositional Analysis Data
1  Air stripping simulation using Appendix P,  Speciation of collected gas samples with EPA TO-
Modified El Paso method (speciated VOC) a 14 or TO-15 methods, with EPA Method 18, or
 Water recirculation rate with a portable (not handheld) gas
chromatograph/flame ionization detector
2  Air stripping simulation using Appendix P,  Process-specific, service-specific concentrations

Modified El Paso methoda, using a flame  Process-specific average concentrations
ionization detection analyzer (total VOC)
 Site-specific refinery average stream
 Water recirculation rate concentrations
 Default process compositions
3  Direct water measurement by EPA Method  Speciation of collected water samples

8260Bb before and after exposure to the
atmosphere (e.g., at the cooling tower return
line and at the outlet of cooling tower)
 Water recirculation rate
4  Direct water measurement by EPA Method  Speciation of collected water samples

8260Bb before exposure to the atmosphere
(e.g., at the heat exchanger exit line, or at the
cooling tower return line)
 Water recirculation rate
5  AP-42 emission factor for VOC, PM10, and  Process-specific, service-specific concentrations
chlorine  Process-specific average concentrations
Water recirculation rate  Site-specific refinery average stream
OR concentrations
 Material balance  Default process compositions
a Source: TCEQ, 2003.
b Source: U.S. EPA, 1996.
8.1 Methodology Ranks 1 and 2 for Cooling Towers

The El Paso method (TCEQ, 2003) uses a flow-through system for air stripping a sample of water and
analyzing the stripped gases for VOC using a flame ionization detection (FID) analyzer. This method
measures the quantity of easily strippable components from the cooling water that have boiling points
below 140°F. In this method, a continuous stream of cooling water is transferred via a pipe or flexible
tubing to an air-stripping column apparatus. Air flowing countercurrent to the cooling water in the
apparatus strips VOC and organic HAP from the water. Concentrations of pollutants in the air exiting the
stripping column are measured, along with the air and water flow rates to the apparatus, to allow
estimation of the concentrations of strippable VOC in the cooling water. At the apparatus air outlet, the
concentrations may be measured using an on-site FID analyzer to determine the total strippable VOC
(Methodology Rank 2 for cooling towers) or an on-site portable gas chromatograph (GC); (Methodology
Rank 1 for cooling towers), or the concentrations may be measured by collecting samples in sample
canisters for off-site laboratory analysis for speciation of air contaminants (Compendium Method TO-
14A, Compendium Method TO-15, or Method 18 [which are all Rank 1 methods]).
The data necessary to implement Methodology Ranks 1 and 2 for cooling towers include speciated air-
stripped compounds concentration, operating parameters of the air-stripping apparatus, and the cooling
water flow rate. Table 8-2 summarizes the data requirements for Methodology Ranks 1 and 2.
8-2
Table 8-2. Data Requirements for VOC or Speciated VOC Emissions, Methodology Ranks 1 and 2
Rank Data That Must Be Collected The Way in Which Data Are Obtained
Rank 1—  Speciated air-strippable VOC concentration  Periodic Appendix P testing results using
speciated VOC measured in the stripped air (ppmv) an air-stripping column apparatus, followed
emissions by TO-14A, TO-15, EPA Method 18, or
GC/FID
 Sample water flow rate of the apparatus,  Measurements during Appendix P testing
stripping air flow rate of the apparatus,
pressure in the air-stripping column
apparatus chamber, and the temperature of
stripping chamber
 Cooling water flow recirculation rates  Continuous measurements from pump flow
(gal/min) rate curves, rotameters, or similar methods
 Length of time of the monitoring period  Assume the measured concentration has
(hours) occurred for half of the time period since
the last sampling date; if a leak occurs,
then add the time period it takes to repair
the leak
Rank 2—VOC  Air strippable VOC concentration measured  Periodic Appendix P testing results using
emissions in the stripped air (ppmv) an air stripping column apparatus, followed
by FID
 Sample water flow rate of the apparatus,  Measurements during Appendix P testing
stripping air flow rate of the apparatus,
pressure in the air-stripping column
apparatus chamber, and temperature of
stripping chamber
 Cooling water flow recirculation rates  Continuous measurements from pump flow
(gal/min) rate curves, rotameters, or similar methods
 Length of time of the monitoring period  From dates of monitoring events
(hours)
For cooling towers servicing process streams with multiple components, it is best to determine speciated
compounds for air-strippable compounds from the cooling water. For cooling towers servicing process
streams with single components, using the FID analyzer is likely adequate. Equation 8-1 should be used
to calculate the concentration of the air-strippable components in the cooling water from the concentration
detected in the stripped air outlet:
C Air ,i * MW * P * PC * bstripairflow
CWater ,i = (Eq. 8-1)
R * (T + 273) * a samplewaterflow * ρWater
where:
CWater,i = Concentration of air strippable compound “i” in the water matrix (parts per million by
weight [ppmw])
CAir,i = Concentration of compound “i” in the stripped air (parts per million by volume [ppmv])
MW = Molecular weight of the compound (grams per mole [g/mol])
P = Pressure in the air-stripping column apparatus chamber (inches [in.] of Hg)
PC = Pressure conversion factor (0.03342 atmospheres per inch [atm/in.] of Hg)
bstripairflow = Stripping air flow rate of apparatus (milliliters per minute [mL/min])
R = Universal gas constant (82.054 milliliters-atmospheres per mole-Kelvin [mL-atm/mol-K])
T = Stripping chamber temperature (°C)
8-3
asamplewaterflow = Sample water flow rate of apparatus (mL/min)

ρWater= Density of the sample cooling water (g/mL).
When using the El Paso method (TCEQ, 2003), the calculated concentration of the compounds in the
water matrix represents the concentration that was stripped and does not represent the total concentration
of the compounds in the water matrix prior to stripping. As such, this calculated water concentration is
referred to as the “strippable pollutant concentration” in the cooling water. In general, when estimating
the total VOC based on the FID analyzer, use the molecular weight (MW) of the calibration gas.
Typically, CH4 is used for the calibration, so the MW for CH4 (16 g/mol) would be used to estimate the
total VOC concentration in the water and the emission rate (reported “as CH4”). Even when a different
calibration gas is used, one may need to use the MW for CH4 in the event a leak threshold is defined in
units of ppmw or ppmv as CH4.
In general, when the process stream is a single component, then the concentration on a CH4 basis from the
FID can be converted to specific compounds that are expected to be leaking by accounting for C. For a
multi-component process stream, it is better to determine speciated compounds using a portable GC/FID,
TO-14A, TO-15, or Method 18. For portable GC results, in addition to calibration with a standard
compound (CH4 or propane), the response factor of the unit should be determined for typical compounds
expected in the air-stripped stream. For speciated compound results from an off-site laboratory, the
concentration in the water for specific compounds can be estimated using the MW for the respective
compound (e.g., benzene MW is 78 g/mol). These speciated concentrations (calculated using the
compound-specific MW) should be used when estimating emissions for the purposes of an emissions
inventory. However, it is important to note that for specific regulatory purposes, these concentrations
must be converted to a CH4 basis (using the MW of CH4) for comparison to a leak definition threshold
that is on an “as CH4” basis, as shown below in Example 8-1.
Example 8-1: How to Compare Speciated Concentration Results with Leak Definitions
That Are Defined as Methane
A facility conducts El Paso monitoring that provides speciated results, but, for regulatory
purposes, these concentrations must be converted to a CH4 basis for comparison to a leak
definition for the stripping air that is on a CH4 basis.
Given: Speciated El Paso monitoring found hexane concentration in the stripped air of 5
ppmv. What is the concentration “as methane?”
In this example, hexane concentration in the stripped air is 5 ppmv. The following equation
should be used to convert this concentration to CH4 basis:
 5 mole C 6 H 14 6 mole C 1 mole CH 4 

5 ppmv Hexane =  6
* *  = 30 ppmv as CH 4
 10 mole air 1 mole C 6 H 14 1 mole C 
If the leak definition is 6.2 ppmv as CH4, then a leak has been detected because the hexane
concentration is equivalent to 30 ppmv as CH4.
8-4
Equation 8-2 should be used to estimate emissions for air strippable total VOC from the cooling water:
CWater ,i min lb 1 ton
Ei = 6
× FlowCoolingWater × 60 × 8.34 × H Year × (Eq. 8-2)
10 ppm hr gal 2000 lb
where:
Ei = Emissions of air strippable pollutant “i” from the cooling water (tons)
FlowCoolingWater = Flow rate of cooling water (gallons per minute [gal/min])
60 = Unit conversion for minutes per year (min/yr)
8.34 = Density of water in pounds per gallon (lb/gal)
1/2000 = Unit conversion for tons per pound (tons/lb)
HYear = Length of time of the monitoring period or the length of time a leak occurs (hr)
When Cair,i is based on total VOC results from an FID analyzer (Methodology Rank 2 for cooling towers),
pollutant “i” is simply VOC.
When Cair,i is based on GC analysis results for individual species (Methodology Rank 1 for cooling
towers) or based on site-specific or default species composition data (Methodology Rank 2 for cooling
towers), pollutant “i” will be the individual pollutant.
If monitoring or sampling events occur regularly or periodically (i.e., quarterly or monthly), then
concentration measurements for consecutive monitoring events can be used to estimate the average
emission rate during the intervening period by assigning each concentration measurement to half of the
time period between monitoring/sampling events. If significant concentrations of organics are detected in
the cooling water (i.e., a “leak” is detected), then the measured concentration is also attributed to the time
span from when the leak was discovered until the time when it was repaired. This method is analogous to
the “midpoint” method used to annualize equipment leak emissions when periodic monitoring is used. As
with equipment leaks, the “modified trapezoid” or “average period” methods could also be used to
annualize emissions from cooling towers when periodic monitoring of the cooling water is performed (see
Section 2.2.2, Calculating Annual Average Equipment Leak Emissions, in this Refinery Emissions
Protocol for additional details).
When monitoring or sampling is not conducted regularly or periodically, any measured concentration
should be used to calculate the average emission rate from the cooling tower, assuming the concentration
is constant for the entire year. In this situation, if the concentration measurement suggests a leak that is
subsequently repaired, then use the initially determined “leak” concentration for the entire reporting year
up to the time when the leak was repaired.
8-5
Example 8-2: Calculation for Methodology Rank 1 for Cooling Towers

Given: A cooling tower with a water recirculation rate of 32,000 gal/min is monitored
quarterly using speciated El Paso monitoring. A hexane concentration in the stripped air of 5.0
ppmv is detected during the second quarterly monitoring event, indicating a “leak.” The
properties of the El Paso stripping column during this monitoring event were as follows:
pressure =29.9 in. of Hg; strippable air flow in the apparatus = 2,500 mL/min; temperature =
32°C; sample water flow rate in the apparatus = 125 mL/min. The leak repair took 45 days
from the monitoring event to complete. Assuming all other monitoring events found no
detectible concentrations of hexane, what are the annual emissions of hexane?
Solution: First use Equation 8-1 to calculate the concentration of air strippable compound in
the water as follows.
g atm ml
5.0 ppmv * 86.17 * 29.9in.Hg * 0.03342 * 2500
mole in.Hg min
CWaterSpecies = = 0.34 ppmw Hexane
ml − atm ml g
82.054 * (32 + 273) * 125 *1
mole − K min ml
Thus, the 5 ppmv hexane concentration in the air stream translates to 0.34 ppmw “strippable”
hexane in the cooling water.
Next, the emissions must be estimated for the period between the first and second quarterly
monitoring events. The exact timing of the monitoring events should be recorded and used,
but from the information given (quarterly monitoring), we assume the monitoring events
occur 91 days (2,184 hours) apart. One can use either the midpoint, modified trapezoid, or
average period emission rate. Using the modified trapezoid or average period method and
noting the effective water concentration of the first monitoring event is 0 ppmv hexane, the
emissions between the first and second monitoring event is:
Ei ,1 =
(0.34 − 0) 2 * 32,000 gal * 60
min
* 8.34
lb
* 2,184 hr *
1 ton
= 2.97 ton Hexane
6
10 min hr gal 2000 lb
Similarly, the emissions for the period from the second quarterly monitoring event and the
leak repair is estimated. This period is given as 45 days (or 1,080 hours).
0.34 gal min lb 1 ton
Ei , 2 = 6
* 32,000 * 60 * 8.34 * 1,080 hr * = 2.94 ton Hexane
As these are the only measurable emissions for the year, the annual emissions would be
calculated as Ei = Ei,1 + Ei,2 = 2.97+2.94 = 5.91 tons, which is 5.9 tons when rounded to two
significant figures.
Note: The advantage of the midpoint method is that it uses the same leak concentration for the
period between the leak being detected and repaired as the “half” period prior to the
monitoring event. Consequently, the time period from the monitoring event and leak repair
can be added to the “half” period prior to the monitoring event (91 days÷2 = 1,092 hrs) to
calculate a total duration of the leak and calculate the cumulative emissions in one step as
follows:
* (1,092 + 1,080 ) hr *
0.34 gal min lb 1 ton
Ei = 6
* 32,000 * 60 * 8.34 = 5.9 ton Hexane
8-6
Example 8-3: Calculation for VOC Emissions Using Methodology Rank 2 for Cooling
Towers
Given: The cooling tower with a water recirculation rate of 25,000 gal/min is monitored
quarterly using El Paso stripping column (same stripping column properties as in Example
82) and a VOC concentration in the stripped air of 14 ppmv (as CH4) is measured on
February 1, 2009. The leak was repaired on March 4, 2009 (31 days later). All previous and
subsequent monitoring events found VOC concentration in the stripped air of 1 ppmv. What
is the annual emissions of VOC from this cooling tower in 2009? If the process fluid contains
10 wt% benzene and a total light (strippable) organic content of 80%, what is the annual
emissions of benzene from this cooling tower in 2009?
Solution: Equation 8-1 is used to calculate the concentration of air strippable VOC in the
water as CH4. The water concentration corresponding to the stripping air concentration of 14
ppmv is:
g atm ml
14 ppmv * 16.04 * 29.9 in.Hg * 0.03342 * 2500
mole in.Hg min
CWater ,VOC = = 0.179 ppmw VOC as CH 4
ml − atm ml g
82.054 * (32 + 273) * 125 *1
mole − K min ml
In the same manner, the water concentration corresponding to the stripping air concentration
of 1 ppmv is: 0.0128 ppmw as CH4 (0.179/14).
Because the leak was identified during the first quarterly monitoring event, the method used to
estimate emissions between monitoring intervals (i.e., the midpoint, modified trapezoid, or
average period method) will yield slightly different results for 2009, although the total
emissions across both 2008 and 2009 will be identical. The modified trapezoid method is the
most complicated, requiring interpolation to determine the effective leak rate on January 1,
2009, before the emissions between Januaruy 1 and February 1 can be calculated. We
generally recommend the midpoint method due to its ease of use, but the same method should
be used for all cooling towers and all inventory years.
Using the midpoint method, the water concentration of 0.179 ppmw as CH4 is present for 62
days (or 1,488 hour) in 2009 (from Jan. 1 to March 4, 2009). The emissions during this period
are:
0.179 gal min lb 1ton
EVOC ,1 = 6
* 25,000 * 60 * 8.34 * 1,488hr * = 1.67 ton VOC
For the remainder of the year (8760 – 1488 = 7,272 hrs), the emissions are:
0.0128 gal min lb 1ton
EVOC , 2 = 6
* 25,000 * 60 * 8.34 * 7,272hr * = 0.58 ton VOC
Therefore, the 2009 VOC emissions from this cooling tower are 1.67 + 0.58 = 2.25 tons VOC.
8-7
Example 8-4: Calculation for Speciated HAP Emissions using Methodology Rank 2 for
Cooling Towers
Given: The cooling tower in Example 83 serves a process unit whose process fluid contains
10 wt% benzene and a total light (strippable) organic content of 80%. What is the annual
emissions of benzene from this cooling tower in 2009?
Solution: In Example 83, theVOC emissions were calculated to be 2.25 tons in 2009. To
determine the mass emissions of benzene, the mass of benzene per mass of VOC stripped in
the El Paso column must be calculated. In this example, 80 percent of the process fluid is
expected to be strippable organics. The remaining 20 percent may be either inorganic
compounds, like H2 or N2 that are stripped but are not detected by the organic analyzer, or
heavier organics that do not strip appreciably in the stripping column. While more detailed
calculations could be performed for compounds that may marginally strip, for the purposes of
speciating volatile emissions, it is assumed these non-volatile compounds are not stripped.
Benzene represents 10% of the total mass of fluid leaking into the cooling water. Since the
total strippable organic content of fluid leaking into the cooling water is 80 percent, the mass
fraction of benzene per mass of strippable VOC is 0.1/0.8 or 0.125.
Therefore, benzene is expected to contribute 12.5 percent of the total strippable VOC
concentration or 0.28 tons benzene (0.125 tons benzene per ton VOC × 2.25 tons VOC).
8.2 Methodology Rank 3 for Cooling Towers

Methodology Rank 3 is used to estimate emissions from cooling towers and involves using a mass
balance approach based on sampling the cooling water before and after the cooling tower. Although
Methodology Rank 3 has the potential to be as appropriately reliable as Methodology Ranks 1 or 2 for
large leakers, Rank 3 suffers from detection limit issues with water sampling techniques. Consequently,
this monitoring approach will not detect the smaller leaks that could significantly contribute to VOC or
HAP emissions.
Cooling water typically has low concentrations of components and large volumes of flow rate, so direct
water sampling typically requires analysis for low-level concentrations. A common method used for
cooling water sampling is Method 8260B. In this method, the sample is introduced into a gas
chromatography/mass spectrometry (GC/MS) system, the GC column is temperature programmed to
separate the components, and the components are detected with MS. The results from the method are the
concentrations of components in the cooling water. Typically, two cooling water samples are taken to
estimate the emissions: one sample before and one sample after the cooling water has been exposed to the
atmosphere. The change in the component concentration of the cooling water multiplied by the cooling
water recirculation rate provides the quantity emitted into the atmosphere for the duration of the leak.
The data necessary to implement Methodology Rank 3 for cooling towers include speciated compounds
concentration in the cooling water and the cooling water flow rate. Table 8-3 summarizes the data
requirements of Methodology Rank 3.
8-8
Table 8-3. Data Requirements for Speciated Compound Emissions, Methodology Rank 3
Rank 3—  Speciated compounds concentration  Periodic sampling results using Method 8260B,
speciated measured in the inlet cooling tower return including a measurement prior to exposure to the
emissions line (ppmw) atmosphere
 Speciated compounds concentration  Periodic sampling results using Method 8260B,
measured in the outlet cooling water including a measurement after exposure to the
(ppmw) atmosphere
 Cooling water flow recirculation rates  Continuous measurements from pump flow rate
(gal/min) curves, rotameters, or similar methods
(hours) occurred for half of the time period since the last
sampling date; if a leak occurs, then add the time
period it takes to repair the leak
The change in the component concentration in the cooling water before and after being exposed to the
atmosphere is used, along with the cooling water recirculation rate, to estimate the emissions from the
cooling water. The following equation (Equation 8-3) can be used to calculate the emissions of
components from the cooling water. It is important to note that it is assumed that the water make-up rates
offset any drift loss and blowdown loss; therefore, flow “in” to a cooling tower is the same as flow “out”
and can be determined based on the water recirculation flow rate, regardless of where this measurement
takes place.
(C − C i ,Out ) lb min 1 ton

Ei = × FlowCoolingWater × 8.34 × 60 × H Year ×
i , In
6
(Eq. 8-3)
10 gal hr 2000 lb
where:
Ci,In = Concentration of component “i” in the cooling water prior to exposure to the
atmosphere (ppmw)
Ci,Out = Concentration of component “i” in the cooling water after exposure to the atmosphere
(ppmw)
As shown in Equation 8-4, the total emissions from the cooling water are the sum of all component
species emitted and can be summed specifically for VOC or organic HAP:
n
ETotal = ∑ Ei (Eq. 8-4)
i =1
8-9

Given: Concentrations for total xylenes in the cooling water are determined semi-annually.
During the first measurement event the concentration of xylene in the cooling water return
line before exposure to the atmosphere was 0.220 ppmw and 0.080 ppmw in the cooling water
flow to the heat exchanger after exposure to the atmosphere. Sampling during the second
monitoring event showed xylene concentrations of 0.320 ppmw before and 0.100 ppmw after
exposure to the atmosphere in the cooling tower. The cooling water recirculation rate averages
55,000 gal/min with a maximum recirculation rate of 65,000 gal/min. The cooling tower
operated for 8,000 hours during the year. What are the annual and maximum hourly
emissions of xylene from the cooling tower?
Solution: For monitoring that occurs semi-annually or less often attributing the emissions
between monitoring events and the calendar year can be cumbersome. As no leak repair was
conducted, it can be assumed that these two measurement are equally representative for the
year. The average hourly emission rate of xylene is calculated using Equation 8-3 for each of
the monitoring events as follows:
E i ,1 =
(0.22 − 0.08) × 55,000 gal × 8.34
lb
× 60
min
= 3.85 lb/hr
6
10 min gal hr
Ei,2 =
(0.32 − 0.10) × 55,000 gal × 8.34
lb
× 60
min
= 6.05 lb/hr
6
10 min gal hr
The average hourly emissions rate during the year can be calculated as the arithmetic average
of these two hourly estimates, which is 4.95 lb/hr [(3.85+6.05)/2]. The annual emissions are
then calculated based on the operating hours as: 4.95 lb/hr × 8,000 hours × 1 ton/2,000 lbs =
19.8 tons of xylene.
The maximum hourly emissions of xylene are calculated using they highest net measurement
concentration and the maximum recirculation rate as follows.
E i ,max =
(0.32 − 0.10) × 65,000 gal × 8.34
lb
× 60
min
= 7.16 lb/hr
6
10 min gal hr

Methodology Rank 4, similar to Methodology Rank 3, uses direct cooling water samples for analysis of
VOC and HAP contained in the cooling water, and a water sample is taken and sent for laboratory
analysis. Under Methodology Rank 4 for cooling towers, only one sample is taken of the cooling water
prior to being exposed to the atmosphere to estimate emissions. In this method, it is assumed that 100% of
the compound is emitted at the cooling tower. As such, this methodology will likely overestimate the
cooling water emissions, and would be a conservative (i.e., resulting in greatest emissions) estimate.
The data necessary to implement Methodology Rank 4 for cooling towers include speciated compounds
concentration in the cooling water and the cooling water flow rate. Table 8-4 summarizes the data
requirements of Methodology Rank 4.
8-10
Table 8-4. Data Requirements for Speciated Compound Emissions, Methodology Rank 4
Rank 4—  Speciated compounds concentration  Periodic sampling results using Method
speciated measured in the inlet cooling tower return 8260B, including a measurement prior to
emissions line (ppmw) exposure to the atmosphere
 Cooling water flow recirculation rates  Continuous measurements from pump
(gal/min) flow rate curves, rotameters, or similar
methods
(hours) occurred for half of the time period since
the last sampling date; if a leak occurs,
then add the time period it takes to repair
the leak
The component concentration in the cooling water before being exposed to the atmosphere is used with
the cooling water flow rate to estimate the emissions from the cooling water. The following equation
(Equation 8-5) should be used to calculate the emissions of components from the cooling water:
(C ) lb min 1 ton
Ei = × FlowCoolingWater × 8.34 × 60 × H Year ×
i , In
6
(Eq. 8-5)
10 gal hr 2000 lb
where:
Ci,In = Concentration of component “i” in the cooling water prior to exposure to the
atmosphere (ppmw)
Equation 8-5 is equivalent to Equation 8-3 with Ci,Out set equal to zero. Thus, the calculations performed
for Methodology Rank 4 for cooling towers is essentially identical to those for Methodology Rank 3 for
cooling towers. Equation 8-4 can be used to calculate the total VOC or total organic HAP emissions from
the cooling water.

Methodology Rank 5 for cooling towers uses emission factors from AP-42 (U.S. EPA, 1995a; Sections
5.1 and 13.4). The AP-42 emission factors approved for use in Methodology Rank 5 for cooling towers
are summarized in Table 8-5. The remainder of this section describes the use of these emission factors
for different classes of pollutants.
Table 8-5. Methodology Rank 5 Default Emission Factors
VOC Emission Factor Drift Factor PM10 Emission Factor

Type of Cooling Tower (lbs/MMgal) a (lbs/MMgal) a (lbs/MMgal) b
Induced draft, counter flow 6.0 1,700 31.5
Induced draft, cross flow 6.0 1,700 40.8
Unspecified draft or flow type 6.0 1,700 35.0
Natural draft 6.0 73 No data
a Source: U.S. EPA, 1995a; presented in units of pounds per million gallons (lb/MMgal) of recirculation.
b Source: U.S. EPA, 1995a; calculated as the product of the drift factor from Table 13.4-1 and the total dissolved
solids concentration in Table 13.4-2.
8-11
8.4.1 VOC and Volatile Organic HAP

Methodology Rank 5 for volatile organics from cooling towers uses the uncontrolled VOC emission
factor given in AP-42 (U.S. EPA, 1995a; Section 5.1) of 6 pounds of total VOC per million gallons of
water (MM gal) (equivalent to a concentration in the water of 0.7 ppmw). The controlled emission factor
provided in AP-42 should not be used. The controlled emission factor only applies to refineries that
directly monitor for hydrocarbons (and organic HAP) and repair leaks when they occur, in which case the
refineries would have monitoring data to conduct emission estimates under Methodology Ranks 1, 2, 3, or
4 for cooling towers.
To estimate HAP emissions, refineries should use site-specific information on the composition of process
streams cooled in heat exchangers and condensers if available. Process-specific concentrations, process-
specific average concentrations, or site-specific refinery average stream concentrations should be used
when available. If these data are not available, then the default average refinery process stream
compositions given in Appendix A, Average Stream Compositions, Table A-1, may be used to speciate
the VOC and generate HAP emissions estimates.
The data necessary to implement the Methodology Rank 5 for cooling towers include the AP-42
uncontrolled VOC emission factor, the cooling water flow rate, and the process stream composition.
Table 8-6 summarizes the data requirements of Methodology Rank 5.
Table 8-6. Data Requirements for VOC or Speciated VOC Emissions, Methodology Rank 5
Rank 5—estimated  AP-42 uncontrolled emission factor for  Not applicable
speciated cooling towers of 6 pounds of VOC per
emissions million gallons of water
 Process stream composition data  Measurement/sampling of process stream
composition, engineering knowledge of
process composition, or default average
composition for the industry
 Cooling water flow recirculation rates  Continuous measurements from pump
(gal/min) flow rate curves, rotameters, or similar
methods
 Operating hours (hours)  Use operating hours of cooling tower or
assume continuous operation (8,760
hr/yr, non-leap year)
The following equation (Equation 8-6) should be used to estimate VOC emissions:
min 1 ton
EVOC = EFUnc × FlowCoolingWater × 60 × H Year × (Eq. 8-6)
hr 2000 lb
where:
EFUnc = Uncontrolled emission factor for cooling towers from AP-42 (Section 5.1)
= 6.0 pounds of VOC per 106 gallons of cooling water (lb/MMgal).
The following equation (Equation 8-7) should be used to estimate the emissions for a particular species
component:
min 1 ton
Ei = EFUnc × WtFraci × FlowCoolingWater × 60 × H Year × (Eq. 8-7)
hr 2000 lb
8-12
where:
WtFraci = Approximate weight fraction of component i in the cooling water

Given: An unmonitored cooling tower with a water recirculation rate of 25,000 gal/min is
servicing a heat exchanger cooling a reformulated gasoline stream. Estimate the annual VOC
and HAP emissions from this cooling tower.
Solution: Equation 8-6 should be used to calculate the annual emissions of VOC, EVOC:
6 lb VOC gal min 1 ton
EVOC = 6
× 25,000 × 60 × 8,760hr × = 39.4 ton VOC
10 gal min hr 2000 lb
Without site-specific stream composition, use the default concentrations for reformulated
gasoline in Appendix A, Average Stream Compositions, and use Equation 8-7 to estimate the
emissions for each HAP component in the gasoline stream. For example, to calculate the
annual emissions of benzene, the liquid composition for benzene is 0.7 wt%. Using this
concentration in Equation 8-7 yields:
6 lb VOC gal min 1 ton
Ei = 6
× 0.007 × 25,000 × 60 × 8,760hr × = 0.28 ton Benzene
Similarly, emissions of other consistuents are estimated as follows:
 Cumene of 0.19% by weight; emissions are 0.075 ton
 Ethylbenzene of 1.26% by weight; emissions are 0.50 ton
 Hexane of 1.36% by weight; emissions are 0.54 ton
 Naphthalene of 0.21% by weight; emissions are 0.083 ton
 Toluene of 7.0% by weight; emissions are 2.8 ton
 2,2,4-Trimethylpentane of 1.1% by weight; emissions are 0.43 ton
 Xylene of 7.3% by weight; emissions are 2.9 ton
The total HAP emissions for this cooling tower are calculated by summing the emissions of
the individual HAP; the total annual HAP emissions for this cooling tower are 7.6 tons.
8.4.2 Particulate Matter Emissions

Methodology Rank 5 for cooling towers uses the total liquid drift emission factors given in AP-42 (U.S.
EPA, 1995a; Section 13.4), which are also provided for convenience in Table 5-5, and the total dissolved
solids (TDS) weight fraction to estimate PM10 emissions. It is conservatively assumed that all TDS are in
the PM10 size range.
Site-specific TDS concentration in the cooling water should be used when available (Methodology Rank
5A for PM from cooling towers). When direct TDS monitoring is not conducted, AP-42 (U.S. EPA,
1995a) recommends that the TDS be estimated from TDS concentration of the makeup water times the
“cycles of concentration ratio” of a monitored parameter (Methodology Rank 5B for PM from cooling
towers). The cycles of concentration ratio is the concentration of a monitored parameter (such as
conductivity, calcium, chlorides, or phosphate) in the cooling tower recirculating water divided by the
concentration of that measured parameter in the makeup water. Alternatively, Methodology Rank 5B for
PM from cooling towers can use a direct correlationof a monitored parameter to TDS concentration. For
8-13
example, it is commonly suggested that the Equation 8-8 be used to estimate TDS concentration from
conductivity measurements.
TDS (ppmw) = CFTDS × Conductivity ( µ mho/cm)) (Eq. 8-8)

where:
TDS = Total dissolved solids concentration of the cooling water (ppmw)
Conductivity = Conductivity of cooling water (micromho per centimeter [µmho/cm]).
CFTDS = Correlation factor to convert conductivity to TDS concentration (ppmw per µmho/cm)
= 0.5 to 1.0; default is 0.67 (see http://www.appslabs.com.au/salinity.htm)
The correlation factor is dependent on the specific electrolytes present and the total concentration of
electrolytes. At low conductivity (<1,000 µmho/cm), CFTDS = 0.5. The correction factor increases with
increasing TDS concentrations. A site-specific correction factor or curve should be used based on paired
measurements over the range of typical measured conductivity values. Otherwise, the default correction
factor of 0.67 should be used.
Once the the TDS weight fraction is determented, the PM10 emissions corresponding to the TDS
measurement value is calculated using Equation 8-9.
min 1 ton
E PM 10 = EFDrift × WtFracTDS × FlowCoolingWater × 60 × H Period × (Eq. 8-9)
hr 2000 lb
where:
EPM10 = Emissions of PM10 during the measurement period (tons or tons per period)
WtFracTDS = Weight fraction of TDS (Note, 1,000 ppmw = 0.001 wt fraction)
HPeriod = Number of hours associated with the monitored TDS concentration (hrs).
If no data are available to estimate TDS concentration in the cooling water, then the default average PM10
emission factors provided in Table 8-5 (which are based in the TDS fraction from AP-42, Table 13.4-2)
should be used. These default PM10 emission factors represent Methodology Rank 5C for PM from
cooling towers. Emissions of PM10 are calculated using Equation 8-10, which is essentially identical to
the calculation for total VOC emissions.
min 1 ton
E PM 10 = EFPM 10 × FlowCoolingWater × 60 × H Year × (Eq. 8-10)
hr 2000 lb
where:
EFPM10 = Default emission factor for PM10 from cooling towers from Table 8-5.
8-14
Example 8-7: Calculation for Annual PM Emissions Using Methodology Rank 5B for
PM from Cooling Towers
Given: A cooling tower with a water recirculation rate of 25,000 gal/min is sampled monthly
for TDS. What are the annual emissions of PM10? If the highest recirculation rate is 30,000
gal/min, what is the maximum hourly PM10 emissions rate?
Solution: Using the site-specific TDS fraction, the operating hours associated with each TDS
concentration period can be determined using the midpoint method as shown in the following
table (Column 3). Equation 8-9 is applied to each concentration period to calculate the
emissions of PM10 for that period (Column 4). The sum of the emissions in Column 4 yields a
total annual emissions of 22.7 tons (or 23 tons rounded to 2 significant digits).
2 3 4
1 TDS Concentration Hours Assigned to Emissions (tons for
Date (ppmw) TDS Concentration concentration period)
Jan 10 (startup Jan 1) 360 240+300=540 0.25
February 4 520 300+336=636 0.33
March 4 780 336+372=708 0.70
April 4 1,100 372+360=732 1.03
May 4 1,260 360+372=732 1.18
June 4 2,300 372+360=732 2.15
July 4 3,500 360+372=732 3.27
August 4 5,500 372+372=744 5.22
September 4 4,600 372+360=732 4.29
October 4 1,700 360+372=732 1.59
November 4 2,100 372+624=996 2.67
December (shutdown Dec 1 0 744 0.00
- not operating in December)
Total 8,760 22.7 ton
The maximum TDS concentration (5,500 ppmw) and the maximum cooling water recirculation
rate of 30,000 gal/min is used in Equation 8-9 to calculate the maximum hourly PM10 emissions
rate as follows:
1,700 lb drift 5,500 lb TDS gal min
E PM 10 = 6
× 6 × 25,000 × 60 = 14 lbs/hr PM 10
10 gal 10 lb water min hr
8-15
Example 8-8: Calculation Using Methodology Rank 5C for PM from Cooling Towers
Given: A forced draft, cross-flow cooling tower with a water recirculation rate of 25,000
gal/min in service all year. No TDS concentration data are available. What are the annual
emissions of PM10 emissions?
Solution: Since no TDS measurement data are available, use the default PM10 emission
factors from Table 8-5. The appropriate emission factor for a forced draft, cross-flow
cooling tower is 40.8 lb/MMgal. The cooling tower was operated all year (8760 hours). Using
Equation 8-10, the annual emissions of PM10, EPM10, are calculated as follows:
40.8 lb PM 10 gal min 1 ton
E PM 10 = 6
× 25,000 × 60 × 8,760 hr × = 268 ton PM 10
8.4.3 Non-volatile Organic HAP Emissions

Organic compounds with low volatility, such as most POM compounds, will not be effectively stripped
from the cooling water. These compounds will tend to accumulate in recirculated cooling water, and will
primarily be released as drift from the cooling tower. The emissions of these compounds can be
estimated from their concentration in the cooling water and the liquid drift factor in the same manner used
to calculate PM10 emissions using Equation 8-9.
8.4.4 Chlorine Emissions

Site-specific material balance may be one method of estimating chlorine emissions. Otherwise, limited
information is available regarding chlorine emission factors for cooling towers. EPA is currently
reviewing available data to develop guidance for estimating chlorine emissions from cooling towers. To
aid in developing that guidance, EPA is seeking additional data and suggestions for emission estimation
methodologies.
8-16
Version 3 Section 9—Loading Operations
9. Loading Operations
Product or commodity loading emissions occur when vapor is displaced by the product or commodity
when it is loaded into tank trucks, rail cars, and marine vessels (which include both ships and barges). The
vapor may contain constituents from the material previously transported and from the product currently
being loaded. HAP, VOC, and GHG are emitted from loading operations. Emission from loading
operations should be estimated for all materials loaded into tank trucks, rail cars, and marine vessels such
as final products, intermediates, additives, or wastes. Loading operations also include loading materials
into drums or containers, such as LPG container filling operations or loading waste into drums for off-site
disposal.
Loading emission estimates require information on the quantity of material loaded annually, the
component composition of the material loaded and the vapor pressures of the components, the loading
procedure, the type of vessel being loaded, and the effectiveness of the capture system and the controls
used. Emissions from gasoline loading racks are regulated under the Petroleum Refinery MACT I (40
CFR Part 63, Subpart CC) and are limited to 10 milligrams of TOCs per liter of gasoline (Subpart CC
references Subpart R, which references some requirements for 40 CFR Part 60, Subpart XX). Emissions
from marine vessel loading operations for crude, gasoline, and other products are also regulated under the
Petroleum Refinery MACT I (Subpart CC references the Marine Vessel Loading MACT in 40 CFR Part
63, Subpart Y) and are subject to a 97% emission-reduction requirement for existing sources and 98%
emission-reduction requirement for new sources.
Emissions from loading operations may be reduced by using submerged loading or bottom loading.
Additionally, emissions from loading operations may be captured and controlled using a vapor collection
system and an add-on air pollution control device. Vapor balancing may be used as an alternative to add-
on controls to reduce emissions from loading operations. If an add-on air pollution control device or a
vapor balance system is used to capture and control emissions, then capture and control efficiencies
should be included in the emission estimate.
Control devices used for loading operations typically include thermal or catalytic incinerators, adsorption
systems, scrubbers, and flares. Vapor balancing systems are allowed as an option or an alternative for
marine vessel loading operations under Subpart CC of the Petroleum Refinery MACT I. For loading racks
subject to gasoline loading requirements, typically an EPA Method 25A or 25B performance test is
conducted on the control device to collect outlet TOC concentration data and determine an outlet
emission factor based on the volume of gasoline loaded (i.e., meet the less than 10 milligrams of total
hydrocarbon per liter gasoline loaded [mg/L]). For marine vessel loading subject to requirements under
Subpart CC MACT I, typically, a Method 25 performance test is conducted on the control devices to
collect inlet and outlet TOC concentration data and determine a control efficiency.
Table 9-1 summarizes the hierarchy of loading operation emission estimation techniques. Within a given
measurement method (or rank), there may be alternative methods for determining the constituent-specific
emissions; these compositional analysis methods are also provided in order of accuracy. Methodology
Ranks 1, 2, and 3 for loading operations only apply loading operations that are controlled; uncontrolled
loading operations must estimate their emissions using Methodology Ranks 4 for loading operations. The
remainder of this section provides additional details and guidance regarding the ways in which to
implement these methods.
9-1
Table 9-1. Summary of Loading Operations Emission Estimates
Rank Measurement Method or Emission Factor Compositional Analysis Data

1A Direct measurement (CEMS) for both flow rate  Pressure, temperature, and moisture content
and gas composition (depending on the monitoring system)
1B Direct measurement (CEMS) for both flow rate  Process-specific, service-specific concentrations
and THC based on measurement data
2 Direct measurement by EPA Method 18 (site-  Not applicable
specific emission factor) and loading rate
3 Direct measurement by EPA Method 25,  Process-specific, service-specific concentrations
Method 25A, or Method 25B (site-specific  Process-specific average concentrations
emission factor) and loading rates  Site-specific refinery average stream concentrations
4 AP-42 emission factor (default emission  Process-specific, service-specific concentrations
factor) and loading rates  Process-specific average concentrations
 Site-specific refinery average stream concentrations
9.1 Data Available on Product Composition and Properties

To speciate the emissions into VOC or HAP, facilities should use site-specific information on the
composition of the material loaded if available (for more information, see Section 9.3.2, Estimate
Controlled or Uncontrolled Emissions and Speciate). Data are often available on the product or
commodity composition, sometimes for both the liquid and vapor phases and sometimes only for the
liquid phase. Material-specific or site-specific concentrations should be used when available. If material-
specific or site-specific data are not available, then the default average compositions for refinery process
and product streams presented in Table A-1 of Appendix A, Average Stream Compositions, may be used
to speciate and generate VOC and HAP emission factors.
Other properties such as the MW of the liquid, the MW of the vapor, and the vapor pressure of a
multicomponent product or commodity are also often available. When these data are not readily available
at the refinery, data from various references can be used or can be calculated using basic principles. For
instance, property data are available in AP-42, Section 7.1 (Organic Liquid Storage Tanks), in Table
7.1-2 for various petroleum fuels and in Table 7.1-3 for various petrochemicals (U.S. EPA, 1995a).
If the weight composition of a commodity is available but other property data are not, estimates can be
made. Appendix A presents equations that may be used for calculating estimates of the liquid vapor
pressure, the molecular weight of the vapor, and the weight fraction of the vapor given a known liquid
component composition.
9.2 Methodology Rank 1 for Loading Operations

For pollutants such as THC, it is anticipated that some loading operations may have CEMS for measuring
the composition and the gas flow rate of the emissions from loading operations. If such monitors are
present, the methods in Section 4.1, Methodology Rank 1 for Stationary Combustion Sources, should be
used to calculate emissions. Although it is unlikely that CEMS are used that can provide emissions for
speciated organic compounds, if a facility monitors THC and has sampled the organic vapors to determine
specific organic compound concentrations, then speciation profile developed from the sampling data can
be used with the CEMS data for THC. The THC emissions would be reported as Methodology Rank 1A
for loading operations and the speciated organic emissions would be reported as Methodology Rank 1B
for loading operations.
9-2
9.3 Methodology Ranks 2 and 3 for Loading operations

Performance testing conducted at the refinery may be used to develop site-specific emission factors with
emission data correlated to throughput. These site-specific emission factors may be used with throughput
data to estimate the annual loading emissions.
For gasoline loading activities regulated under Petroleum Refinery MACT I (40 CFR Part 63, Subpart CC
and Subpart R), refineries have conducted periodic EPA Method 18, Method 25A, or Method 25B testing
for speciated hydrocarbons or for outlet nonmethane organic compounds (NMOC); determined the flow
rate based on EPA Method 2A or 2B testing; and measured and recorded the gasoline loading throughput
during the testing. The outlet test results on a mass basis are correlated with the gasoline throughput
during the test to demonstrate that the refinery is meeting the 10 mg/L gasoline emission limitation. The
test data could also be used to estimate the speciated hydrocarbons, total hydrocarbon, CH4, NMOC, or
VOC loading emissions into the atmosphere based on the mass emissions per volume loaded factor
developed during the performance test and the annual quantity of that product loaded. It is important to
note that when NMOC is used, a CH4 emission factor may also be developed. Additional information
about speciation of the gasoline or product either as site-specific information or as defaults may be used
to estimate individual species component emissions.
For marine vessel loading regulated under the Marine Vessel Loading MACT (40 CFR Part 63, Subpart
CC and Subpart Y), refineries have conducted EPA Method 2, 2A, 2C, or 2D flow rate testing and have
conducted EPA Method 25 or 25A testing on the inlet and outlet streams of the control device to
determine control device efficiency. The Marine Vessel Loading MACT requires initial performance
testing of the control device, but it does not require ongoing or periodic testing. The outlet test data, along
with the recorded product throughput during the testing, could also be used to estimate the total
hydrocarbon, CH4, NMOC, or VOC loading emissions per volume loaded factor. This factor and the
annual loading data of the product can be used to estimate emissions. It is important to note that when
NMOC is used, a CH4 emission factor may also be developed. Additional information on speciation of the
gasoline or product either as site-specific information or as defaults may be used to estimate individual
species component emissions.
The data necessary to implement Methodology Ranks 2 and 3 for loading operations include the emission
rate or emission factor (either speciated by component or as total hydrocarbon, CH4, NMOC, or VOC)
and throughput data of the products loaded. Table 9-2 summarizes the data requirements of Methodology
Ranks 2 and 3.
Table 9-2. Data Requirements for VOC or Speciated Emissions, Methodology Rank 1 or 2
Rank 2—  Speciated component emission factor  Periodic EPA Method 18 testing results
speciated for a product (milligrams per liter [mg/L]
emissions or lb/gal)
 Throughput data by product over the  Volume flow rate measurements determined by
time period of interest (gallons or liters) flow meter or pump rates, gallons, or liters
Rank 3—  Total hydrocarbon or VOC emission  Periodic EPA Method 25 or 25A testing results
VOC or total factor for a product (mg/liter or lb/gal)
hydrocarbon
 Spectiaion profile from measurement data or from
process knowledge
emissions
 Throughput data by product over the  Volume flow rate measurements determined by
time period of interest (gallons or liters) flow meter or pump rates, gallons, or liters
9-3
Equation 9-1 should be used to estimate the emissions of each product using emission rates or emission
factors developed from testing.
E Li = ERLi × Thruput (Eq. 9-1)

where:
ELi = Emissions of component species “i” (lbs)
ERLi = Emission rate of component species “i” (lb/gal)
Thruput = Throughput of product (gallons)
When ERLi is based on total hydrocarbon, NMOC, or VOC, pollutant “i” is simply the total hydrocarbon,
NMOC, or VOC. When ERLi is based on single pollutant, pollutant “i” will be the individual pollutant.
When ERLi is based on total hydrocarbon, NMOC, or VOC (i.e., Method 25, 25A, or 25B testing),
additional calculations are necessary to estimate individual pollutant emissions. To estimate the emissions
of each pollutant, multiply Equation 9-1 for ELi by the vapor weight fraction of the component “i”, as
shown in Equation 9-2.
E LSpecies = ERLi × Thruput × Wtfrac Species (Eq. 9-2)
where:
ELSpecies = Emissions of component species “i” (lbs)
WtfracSpecies = Vapor weight fraction of component species “i” (fraction by weight)
If the vapor weight fraction of the components are not know, then these concentrations can be calculated
from the product knowledge, i.e., the composition of compounds in the liquid. Appendix A, Average
Stream Compositions, presents the methodology for calculating the vapor phase concentration based on
the liquid-phase composition and the system temperature and pressure. It is important to characterize even
low levels of the most volatile compounds in the liquid, because even low levels of the most volatile
compounds can significantly affect the vapor phase concentration. An example calculation is provided in
Example A-1 of Appendix A.
9.4 Methodology Rank 4 for Loading operations

The emission estimating methodology for loading operations is given in AP-42 (U.S. EPA, 1995a); these
methodologies are for loading operations that are vented directly to the atmosphere. Equations are
provided in this section for estimating controlled emissions from the “uncontrolled” emissions based on
the capture and control efficiency. In section 5.2 of AP-42, commodities are categorized into crude
loading, gasoline loading, and what is termed “other products” (i.e., not crude or gasoline). In the AP-42
methodology, there is one equation that is used for tanker trucks and rail cars for crude, gasoline, and
other commodity products and for marine vessels for other commodity products (but not for crude or
gasoline). In addition, Table 5.2-2 in AP-42 presents other emission factors that are used for marine
vessels for gasoline. These equations provide the emissions for the “product” or “total hydrocarbon,”
whether that product has multiple component species (e.g., gasoline) or whether that product is a single
species (e.g., toluene). Depending on the commodity being loaded, the emission factor (EFL) could be
used for gasoline emissions, crude emissions, or other product emissions. The AP-42 methodology for
gasoline and other products is summarized in the following sections. It is not expected that refineries will
be loading crude for transport, but if so, the refiner should consult AP-42 for the emission estimation
methodology.
9-4
9.4.1 AP-42 Emission Factors for “Product” or “Total Hydrocarbon” Emissions
9.4.1.1 Tanker Trucks and Railcars (Any Commodity) and Marine Vessels (Products
Other than Gasoline and Crude Oil)
Equation 9-3 should be used to estimate the uncontrolled emission factor or the emission rate for the
product loaded, in terms of total hydrocarbon or the product:
S × P × MWVap
ERL = 12.46 × (Eq. 9-3)
T
where:
ERL = Loading emission factor or emission rate for the product (or total hydrocarbon) loaded
(in lb per 103 gallons of liquid loaded)
S = Saturation factor (see Table 9-3)
P = True vapor pressure of liquid loaded (pounds per square inch absolute [psia]) 1
MWVap = Molecular weight of vapors (lb/mol)
T = Temperature of bulk liquid loaded (°R; °F + 460)
Table 9-3. Saturation Factorsa
Type of Tanker Loaded Loading Scenario Saturation Factor

 Tank trucks  Submerged loading 0.60
 Rail cars  Dedicated normal service
 Any commodity  Submerged loading 1.00
 Dedicated vapor balance service
 Splash loading 1.45
 Dedicated normal service
 Splash loading 1.00
 Dedicated vapor balance service
 Marine vessels  Submerged loading 0.2
 Other products (only)  Ships
 Submerged loading 0.5
 Barges
a Source: U.S. EPA, 1995.
9.4.1.2 Marine Vessels (Gasoline)

Section 5.2, Table 5.2-2, of AP-42 provides VOC emission factors for marine vessels loading gasoline at
marine terminals based on measurements of gasoline loading losses from ships and barges. Those
emission factors are presented in Table 9-4. (Table 5.2-2 of AP-42 also includes emission factors in units
of mg/L.) AP-42 does not provide ratings for these emission factors.
1
True vapor pressure is defined as the equilibrium partial pressure exerted by a volatile organic liquid as a function
of temperature. It differs from the Reid vapor pressure, which is defined as the absolute vapor pressure exerted by a
liquid at 100°F and is a common measure of the volatility of petroleum liquids. For a pure component at 100°F, the
true vapor pressure is equivalent to the Reid vapor pressure.
9-5
Table 9-4. VOC Emission Factors for Marine Vessel Loading of Gasoline at Marine Terminalsa
Ships and Ocean Barges

Vessel Tank Condition Previous Cargo (lb/103 gal) Barges (lb/103 gal)
Uncleaned Volatilec 2.6 3.9
Ballasted Volatile 1.7 Not applicable
Cleaned Volatile 1.5 No data
Gas-freed Volatile 0.7 No data
Any condition Nonvolatile 0.7 No data
Gas-freed Any cargo No data 2.0
Typical overall situationd Any cargo 1.8 3.4
a Source: U.S. EPA, 1995a. The emission factors for VOC (which excludes methane and ethane) may be used
to estimate total organic emissions (i.e., including methane and ethane) because methane and ethane
constitute a negligible fraction of the components in gasoline.
b Ocean barges (tank compartment depth about 40 ft) exhibit emission levels similar to tank ships. Shallow draft
barges (compartment depth 10 to 12 ft) exhibit higher emission levels.
c Volatile cargoes are those with a true vapor pressure greater than 10 kilopascals (kPa) (1.5 psia).
d Based on observation that 41% of tested ship compartments were uncleaned, 11% ballasted, 24% cleaned,
and 24% gas-freed. For barges, 76% were uncleaned.
9.4.2 Estimate Uncontrolled Emissions and Speciate

Using the emission factor developed or identified from the methods previously discussed, the emissions
can be estimated based on Equation 9-4:
E L = ERL × Thruput (Eq. 9-4)
where:
EL = Uncontrolled emissions (lb) of commodity
ERL = Loading emission rate (lb/103 gal of liquid loaded)
Thruput = Throughput of commodity (gallons)
As previously discussed, depending on the commodity being loaded, the emission rate ERL could be for
gasoline, crude, or other product. Also as previously discussed, to speciate the emissions for specific
components, the composition or concentration data available (i.e., product-specific, product-specific
average, site-specific product, or default averages) can be used as shown in Equation 9-5.
E LSpecies = ERL × Thruput × WtfracSpecies (Eq. 9-5)
where:
ELSpecies = Emissions (lb) of the component species “i”
WtfracSpecies = Weight fraction of component species “i” in the vapor (fraction)
9.4.3 Capture Efficiency

If an add-on air pollution control device or vapor balance system is used to capture and control emissions,
then capture and control efficiencies should be included in the emission estimate. The capture efficiency
is the portion of the total emissions that is captured by the vapor collection system. The effectiveness of
the vapor collection system, or capture system, is dependent on the leak tightness of the truck tank, rail
car, or marine vessel; the condition of the connection from the tanker or vessel collection header
equipment to the vapor collection system; the leak tightness of the terminal’s vapor collection system; and
the control device. For gasoline loading racks subject to Subpart CC, refineries are required to confirm
9-6
and document the vapor tightness of the vehicles being loaded. For loading racks subject to gasoline
loading requirements, an annual pressure and vacuum certification is required on the tanker truck (e.g.,
using EPA Method 27 in 40 CFR Part 60), a leak test using EPA Method 21 from 40 CFR Part 60 may be
performed on the tanker truck, and a leak check using EPA Method 21 may be conducted on the
terminal’s vapor collection system. In addition, the type of loading conducted (i.e., atmospheric,
pressurized) and type of connection from the tank headspace to the vapor collection system and to the
control device are factors. Typical capture efficiencies assumed for vapor collection procedures and
systems are shown in Table 9-5. Capture efficiency for the vapor collection system can be applied based
on the leak check conducted for the tanker truck, rail car, and marine vessel.
Table 9-5. Capture Efficiencies for Vapor Collection Systemsa
Loading Characteristics and Leak Check Frequency for Tankers Capture Efficiencies
 No leak check on tanker 65%
 No leak check on tanker 85%
 Maintain minimum positive pressure below +3 to +5 inches of water
 Annual leak check on tanker per 40 CFR Part 60, Subpart XX (nongasoline) 95%
 Semi-annual leak check on tanker per 40 CFR Part 60, Subpart XX (nongasoline) 97.5%
 Annual leak check on tanker per 40 CFR Part 60, Subpart XX (gasoline) 98.7%
 Vacuum loading, maintaining vacuum less than −1.5 inches of water 100%
 Hard-piped bolted, flanged connection from tanker to the vapor collection system
 Pressure tank that is U.S. Department of Transportation certified 100%
 Hard-piped bolted, flanged connection from tanker to the vapor collection system
Note: Use latest available version if updates to this document have occurred since the cited version.
a Source: TCEQ, 2000.
9.4.4 Overall Control Efficiency

The overall control efficiency represents the portion of the total captured emissions that are removed from
the emission stream by the add-on air pollution control device. The control efficiency determined for the
control device should be applied based on performance testing, if available, or based on design efficiency
and operating parameters.
To estimate controlled emissions from loading, use Equations 9-6 and 9-7 to apply the capture efficiency
and the control efficiency to the uncontrolled emissions:
ECont = ERL × Thruput × [(1 − Capture) + (Capture × (1 − CE ) )] (9-6)
EContSpecies = ERL × Thruput × Wtfrac Species × ( 1 − Capture ) × ( 1 − CE ) (9-7)
where:
Capture = Capture efficiency of the vapor collection system (fraction)
CE = Control efficiency of the add-on air pollution control device (fraction)
9-7
Example 9-1: Calculation for Rank 4—When Property Data Are Not Available
Calculate the VOC and HAP emissions for submerged loading of reformulated gasoline into
dedicated-service gasoline transport tank trucks (each with a volume of 8,000 gallons) based
on the following information:
The refinery loads 210 tank trucks per year, with 100 tank trucks loading gasoline at a bulk
temperature of 90°F, 55 loading at a bulk temperature of 77°F, and 55 loading at a bulk
temperature of 40°F. Liquid composition or vapor-phase property data are not available,so the
property data for liquid composition of reformulated gasoline in Appendix A are used (see
Tables A-1 and Table A-2).
First, the vapor pressure of reformulated gasoline and MWVAP must be calculated. In this
example, the weight fraction of the vapor is estimated separately for each loading temperature
following the calculation methodology presented in Appendix A (see Example A-1 and Table
A-2). The vapor pressure for each constituent is determined at the three loading temperatures
using Antoine’s equation and Antoine’s coefficients, such as those provided in Table 7.1-5 of
AP-42 (U.S. EPA, 1995a).
From Example A-1 in Appendix A, at 77°F, the total system vapor pressure is 4.77 psia and
the MW of the vapor, MWGasolineVAP, is 67.49 lb/lb-mol. Similar calculations for loading
gasoline with bulk temperatures of 90°F and 40°F reveal the following (see Appendix A for
further details):
Temperature Vapor Pressure MWGasolineVAP
90°F 6.13 psia 68.08 lb/lb-mol
9-8
Example 9-1 (continued): Calculation for Rank 4—When Property Data Are Not
Available
The emissions rates at each temperature are calculated using Equation 9-3:
lb
0.60 × 6.13psia × 68.08
ERGasoline 90 = 12.46 lb − mole = 5.673 lb/10 3 gal
(90°F + 460)°R
lb
0.60 × 4.77psia × 67.49
(77°F + 460)°R
lb
0.60 × 2.17psia × 65.79
(40°F + 460)°R
The emissions at each temperature are then calculated using Equation 9-4:
5.673 lb  gal 
E Gasoline 90 = ×  8,000 × 100 trucks  = 4,538 lb gasoline
1,000 gal  truck 
4.482 lb  gal 
EGasoline 77 = ×  8,000 × 55 trucks  = 1,972 lb gasoline
2.135 lb  gal 
EGasoline 40 = ×  8,000 × 55 trucks  = 939 lb gasoline
The individual VOC or HAP species at each bulk temperature can be estimated using the
calculated gasoline vapor weight fraction for each species in Equation 9-5:
E LSpecies = EFL × Thruput × Wtfrac Species = E gasoline × Wtfrac Species
See Table 9-6 for the emission estimate summary for individual HAP species.
Finally, total emissions are estimated by summing the emission estimates for each of the three
bulk temperatures. Total gasoline emissions (VOC) are 7,449 lb (4,538 lb + 1,972 lb + 939
lb), or 3.7 tons (7,449 lb ÷ 2,000 lb/ton).
Similarly, total emissions of each of the individual HAP species are estimated by summing the
emission estimates for each of the three bulk temperatures. The results are presented in
Table 9-6.
9-9
Table 9-6. Sample Calculation Methodology Rank 4—Summary of Emissions (When Property Data Are Not Available)
At 40°F At 77°F At 90°F

Gasoline, At 40°F Gasoline, At 77°F Gasoline, At 90°F Sum Total
CAS Vapor, EGasoline = 939 lb Vapor, EGasoline = 1,972 lb Vapor, EGasoline = 4,538 lb Annual
Number Component “i” wt% ELSpecies “i” (lb) wt% ELSpecies “i” (lb) wt% ELSpecies “i” (lb) Emissions (lb)
71-43-2 Benzene 0.33 3.14 0.42 8.24 0.45 20.22 31.60
110-54-3 Hexane 0.96 9.05 1.13 22.30 1.18 53.64 84.99
108-88-3 Toluene 0.77 7.23 1.09 21.55 1.21 55.01 83.80
1330-20-7 Xylene 0.22 2.02 0.35 6.93 0.41 18.46 27.41
100-41-4 Ethylbenzene 0.04 0.38 0.07 1.30 0.08 3.45 5.13
540-84-1 2,2,4-Trimethylpentane 0.23 2.15 0.30 5.87 0.32 14.58 22.59
98-82-8 Cumene 0.0027 0.025 0.0048 0.094 0.0057 0.26 0.38
91-20-3 Naphthalene 0.0001 0.001 0.0003 0.006 0.0004 0.018 0.025
Total HAP 24.0 66.3 165.6 255.9
9-10
Version 3 Section 10—Fugitive Dust Sources
10. Fugitive Dust Sources

There are three main sources of fugitive dust (or PM) at a petroleum refinery: roads (paved and unpaved);
fluid catalytic cracking catalyst handling; and petroleum coke storage and handling. Petroleum refineries
may also operate a landfill or land application unit, which would also be a source of fugitive PM
emissions. There are no direct emission measurement methodologies commonly employed for fugitive
dust sources. Fugitive dust sources are described in Chapter 13 of AP-42 (U.S. EPA, 1995a).
Unfortunately, none of the default factors for the fugitive dust emission correlations are specific to
petroleum refineries. Table 10-1 provides recommended default values for the correlations based on a
review of the available data. The methodologies are detailed in Chapter 13 of AP-42 and basic guidance
for applying the methodologies are provided below.
 For paved roads, Equation 1 of Section 13.2.1.3 of AP-42 (U.S. EPA, 1995a) should be used with
the silt loading recommended in Table 10-1; values for all other equation variables are provided
in Section 13.2.1.3.
 For unpaved roads, Equation 1a and 1b of Section 13.2.2.2 of AP-42 (U.S. EPA, 1995a) should
be used with the silt content recommended in Table 10-1; values for all other equation variables
are provided in Section 13.2.2.2.
 For handling of FCCU catalyst or petroleum coke, use Equation 1 of Section 13.2.4.3 of AP-42
(U.S. EPA, 1995a) with the moisture content recommended in Table 10-1; values for all other
equation variables are provided in Section 13.2.4.3.
Table 10-1. Default Values for Fugitive Dust Emission Estimates
Potential Range Recommended

Source/Variable Description Value Value
Paved road : silt loading, g/m2 0.4a – 120b 10
Unpaved road : silt content % 1.8 – 25.2c 7
FCCU catalyst “drops”: silt content % 50d
FCCU catalyst “drops”: moisture content % 8e
FCU or calcined coke “drops”: silt content % 5e
FCU coke or calcined “drops”: moisture content % 8e
Delayed coking unit coke “drops”: silt content % 5e
Delayed coking unit coke “drops”: moisture content % 10f
Flexicoking or petroleum coke ash: silt content % 13g
Flexicoking or petroleum coke ash: moisture content % 7g
a Based on low average daily traffic (ADT) public road factor in Table 13.2.1-3 in AP-42.
b Based in asphalt batching industrial facility factor in Table 13.2.1-4 in AP-42.
c Based on range for industrial sites in Table 13.2.2-3 in AP-42.
d Assumed based on median particle size of FCCU catalyst and equilibrium catalyst (E-cat) of 65 µm.
e Values for coke breeze assumed to apply.
f Value suggested as typical by industry representatives (API/NPRA, 2010).
g Values for flue dust assumed to apply.
The methodologies outlined in Chapter 13 provide means of estimating PM10 and PM2.5 emissions. These
emissions are all filterable PM so that PM10-PRI = PM10-FIL and PM25-PRI = PM25-FIL. For FCCU
catalyst handling emissions and for petroleum coke handling activities, emission estimates should also be
provided for metal HAP emissions. Concentrations of specific metal HAP are commonly determined for
the FCCU equilibrium catalyst; these concentrations can be used times the PM10 emission rate to
10-1
Version 3 Section 10—Fugitive Dust Sources
determine the metal HAP emission rate from FCCU catalyst and E-cat handling (external from the
FCCU). Table 5-3 provides concentration ratios of various metal HAP to nickel on CCU fines. These
concentration ratios can be used to augment the FCCU catalyst and E-cat metal HAP emission estimates
when only nickel concentrations are routinely determined. Compositional analysis of petroleum coke
material or ash should also be determined so that the metal HAP emissions from petroleum coke or ash
handling can be reported.
10-2
Version 3 Section 11—Startup and Shutdown
11. Startup and Shutdown

Except for the CEMS-based measurements (i.e., Methodology Rank 1), most of the emission estimation
methods provided in this Refinery Emissions Protocol document characterize the emissions during
normal operating conditions. However, during startup and shutdown, the process operations are not
normal. During startup, the reactor temperatures may not be sufficient for the necessary reactions to occur
or the process flow rate may be well below typical operating conditions. Under these conditions,
processes that do not typically have atmospheric emissions may have significant emissions. While most
control devices will operate well at lower vent rates, some control devices, such as venturi scrubbers,
require a minimum flow for effective operation. During startup and shutdown, the emission
characteristics of the processes may be significantly different than those during normal operations. When
the emission estimation methodology relies on normal operating conditions (e.g., most site-specific and
default emission factors), it is important to specifically estimate the emissions during startup and
shutdown and include these emissions in the facility’s annual emission totals. During shutdown, the
process vessels need to be degassed and purged.
Process unit startup and shutdown of one process unit may also affect the emission characteristics of other
units, particularly with respect to the fuel gas system. When a hydrotreater or hydrocracking unit is
shutdown, this often leads to excess hydrogen being added to the fuel gas. This leads to higher NOx
emissions from process heaters. When certain energy intensive units, such as catalytic reforming units,
are shutdown, the refinery may have excess fuel gas, which is subsequently flared. Depending on the
methods used to estimate emissions from combustion sources and flares, these fuel gas fluctuations may
already be fully accounted within the inventories for combustion sources and flares. However, if default
estimates (or site-specific estimates based on source tests conducted under typical operating conditions)
are used, then emissions during these significant fuel gas fluctuation events should be estimated and the
emissions included in the annual emission values.
It is beyond the scope of this protocol document to provide methods of estimating emissions during all
startup or shutdown events. However, the primary emission event during startup or shutdown is expected
to be vessel depressurization and purging (or degassing). Methods are provided in this section to account
for these events are described in this section.
11.1 Gaseous Process Vessel Depressurization and Purging

Process vessel depressurization for gaseous processes can be estimated using Equation 11-1. This
equation assumes ideal gas law; for gases that do not follow ideal gas law, compressibility factors can be
included. The equation includes a void fraction term to account for inert material, such as packing
material, distillation trays, and certain catalyst particles, that reduce the gaseous space. This equation may
underestimate the emissions if the solid material in the vessel may adsorb certain gas constituents (e.g.,
petroleum coke or activated carbon), that can desorb at lower pressures and temperatures.
 (P + 14.7 ) 528°R 
× [VV × f void ]×
MWi
Ei =  V × × MFi  (Eq. 11-1)
 14.7 TV MVC 
where:
Ei = Emissions of pollutant “i” during depressurization event (lbs/event).
PV = Gauge pressure of the vessel when depressurization gases are first routed to the
atmosphere (pounds per square inch gauge, psig).
14.7 = Assumed atmospheric pressure (pounds per square inch, psi).
TV = Absolute temperature of the vessel when depressurization gases are first routed to the
atmosphere (degrees Rankine, °R).
11-1
528 = “Standard” temperature used in the calculation (°R).

VV = Volume of the vessel (cubic feet, ft3).
fvoid = Volumetric void fraction of vessel. Use 1 unless packing material or trays are present
in the vessel.
MWi = Molecular weight of pollutant “i” (lb/lb-mole).
MVC = Molar volume conversion factor (385.3 scf/ lb-mole).
MFi = Volume fraction of pollutant “i” in vessel gas (scf pollutant i/scf of gas).
Typically, depressurization and purging occurs initially to the refinery fuel gas system or a flare or similar
control device. Combustion of gases purged to the refinery fuel gas system (either directly or via a flare
gas recovery system) should be accounted for under stationary combustion source emissions regardless of
the methodology used for stationary combustion sources. Depending on the methodology used to estimate
the emissions from the flare, these gas flows may or may not be included in the flare emission estimates.
If the flow rate and composition or heating value of the flare gas is measured (Methodology Ranks 1, 2 or
3 for flares), special calculations are not needed because episodic releases to the flare, such as those from
depressurization and purging events, are adequately accounted for. If the flare emission methodology
does not account for these flows (e.g., default factors in Methodology Rank 6 for flares), Equation 11-1
can be used to estimate the mass of gas sent to the flare or other control device (using initial
depressurization temperatures and pressures) and these mass rates can be adjusted by the flare or control
device efficiency to estimate the emissions during the controlled venting period. The use of Equation 11-1
is illustrated in Example 11-1, Gaseous Depressurization Calculation.
11.2 Liquid Process Vessel Depressurization and Purging

For vessels that contain liquids, depressurization and purging emissions can be estimated based on the
“heel” or fraction of the liquid that remains in the vessel prior to purging. This method assumes that the
mass of liquid remaining in the tank will overwhelm the quantity of contaminants in the headspace. For
systems operating with very volatile material (such as butane in gasoline), it may be necessary to consider
the initial gaseous emissions for the most volatile components (using Equation 11-1) as well as the liquid
heel. A heel fraction may be between 0.1 and 1 percent, depending on the viscosity of the liquid, the
surface area of internals within the vessel, and the location of the drainage port or pump line (higher
viscosities, higher internal surface areas, and higher drainage port locations will increase the heel
fraction). Using the heel method, it is assumed the entire liquid heel is eventually volatilized. Knowledge
of the process stream can be used to speciate the emissions. Again, some of the heel may be purged to a
flare or other control device and a control efficiency for this portion is warranted, but some fraction will
also be purged directly to the atmosphere. Example 11-2, Liquid Heel Emissions Calculation, illustrates
this methodology.
11-2
Example 11-1: Gaseous Depressurization Calculation

A catalytic reforming unit is depressurized and purged to a flare. After the purge sequence, no
liquid materials are assumed to be present. The reactor is then depressurized and purged to the
atmosphere. The total volume of the reactor vessels are 20,000 ft3 and the catalyst occupies
40% of the total reactor volume. The temperature and pressure of the vessel when first purged
to the atmosphere are 250°C and 10 psig, respectively. The composition of the gas is in the
reactor vessel after the flare purge is:
Benzene = 1.2 vol%
Hexane = 0.5 vol%
Toluene = 1.4 vol%
Xylenes = 0.8 vol%
Other VOC = 2.1 vol%
Nitrogen = 94 vol%
For use in Equation 10-1, the reactor vessel temperature must be converted to °R: 250°C =
250×9/5 + 32 = 482°F = 482+460 = 942°R.
The void fraction is the volume fraction not occupied by catalyst particles: 1-0.4 = 0.6.
The molecular weight of the constituents listed range from 78 to 106 lb/lb-mol. It is assumed
the “other VOC” has similar molecular weight so a molecular weight for this unspecified
material is estimated from, the weighted average molecular weight of the known organic
constituents: MolWt “other VOC” = (1.2×78+0.5×86+1.4×92+0.8×106)/4.9 = 90 lb/lb-mol.
Using benzene as the example, substituting into Equation 11-1 yields:
 (10 + 14.7 ) 528°R 

× [20000 × 0.6]×
78
Ei =  × × 0.01 = 27.5 lbs = 0.014 tons
 14.7 942 385.3 
Similarly, the emissions of other organic constituents are estimated to be:
Benzene = 27.5 lbs/event
Hexane = 12.6 lbs/event
Toluene = 37.8 lbs/event
Xylenes = 24.9 lbs/event
Other VOC = 55.4 lbs/event
The total VOC is the sum of these emissions, which is 158 lbs/event or 0.079 tons/event. If
this event occurred twice during the year, the annual emission would be 0.16 tons VOC/yr.
11-3
Example 11-2: Liquid Heel Emissions Calculation

Problem: Using the same catalytic reforming unit as in Example 11-1, estimate the emissions
from depressurizing and purging the reactor vessels using the “heel” approach. Assume 90
percent of the heel is purged to the flare.
Solution: As there is high internal surface area, the heel is assumed to be 0.5% of the liquid
volume. The maximum liquid volume is 20,000 ft3 × (1-0.4) = 12,000 ft3, so the volume of
liquid in the reactor is estimated to be 60 ft3 (12,000 × 0.005).
Appendix A contains average composition of various process fluids. Reformate composition is
given as:
Benzene = 4.6 wt%
Hexane = 3.9 wt%
Toluene = 14.5 wt%
Xylenes = 13.8 wt%
Other VOC = 63.2 wt%
To calculate the potential mass emissions, a mass of liquid remaining must be calculated.
Using a specific gravity of 0.8 for reformate, the mass of liquid remaining is estimated to be:
60 ft3 × 0.8 × 62.4 lb/ft3 (density of water) = 3,000 lbs.
Given 90 percent of the heel liquid is removed while purging to the flare, then 2,700 lbs of
material is sent to the flare and 300 lbs are purged directly to the atmosphere. Assuming a 98
percent destruction efficiency for the flare, the uncombusted emissions from the flare are
estimated to be 54 lbs [2700×(1-0.98)].
Therefore, a total of 354 lbs of organic material (VOC) is estimated to be emitted considering
emissions from both the flare and atmospheric purging.
Emissions of specific constituents are calculated based on the mass composition in the liquid:
Benzene = 354 ×0.046 = 16 lbs/event
Hexane = 354 ×0.039 = 14 lbs/event
Toluene = 354 ×0.145 = 51 lbs/event
Xylenes = 354 ×0.138 = 49 lbs/event
Other VOC = 354 ×0.632 = 224 lbs/event.
11-4
Version 3 Section 12—Malfunctions/Upsets
12. Malfunctions/Upsets
Malfunctions or upsets may occur either within the process unit or to a control device used to reduce a
source’s emissions. During malfunction/upset events (hereafter simply referred to as malfunction events),
emissions may be significantly higher than the emissions that occur under normal operating conditions.
Except for the CEMS-based measurements (i.e., Methodology Rank 1), most of the emissions estimation
methods provided in this Refinery Emissions Protocol document characterize the emissions during
normal operating conditions. As such, emissions during malfunction events must be accounted for
separately for each malfunction event and these malfunction/upset emissions must be added to the normal
process emissions to accurately estimate annual emissions.
Because of the myriad of potential malfunction events that could occur, it is impossible to provide
specific guidance for all possible malfunction scenarios. However, because malfunction events are
important to both annual and short-term emissions, the duration and emissions associated with each
malfunction event should be recorded, and these emissions should be included in the annual emissions
reported in response to the ICR.
The following list provides specific events for which malfunction/upset emissions estimates should be
made to accurately account for these emissions. Again, this is not intended to be an exhaustive list, but
rather representative examples. In some malfunction events, such as pressure relief valve opening to a
flare during a system over-pressurization event, the emissions during the malfunction event may be
adequately accounted for by the flare emission methodology (i.e., when using continuous monitoring of
gases is sent to a flare), but in other cases it may not (i.e., when using default factors). As such, emissions
inventory developers will need to evaluate the malfunction events in light of their inventory methods to
determine if it is appropriate to use special malfunction emission calculations. Emission estimates should
be provided for the following malfunction events:
1. Any instance when a control device is bypassed or is not functioning properly.
2. Any instance when the amine scrubbing system and/or sulfur recovery plant is offline or not
operating at normal efficiencies (generally affecting SO2 emission estimates from combustion
sources, flares, and/or sulfur recovery plants).
3. Instances of over-steaming a flare (steam-to-gas ratios exceeding 4) or instances where the flare
operating conditions do not satisfy 40 CFR 60.18 (i.e., inadequate BTU content, exit velocities
exceeding limits). Eighty percent flare destruction efficiencies should be used during periods of
over-steaming; 93 percent flare destruction efficiencies should be used during periods when the
flare operating conditions do not satisfy 40 CFR 60.18.
4. Any instance when a spill or similar emergency release occurs.
Leaks identified as part of a LDAR work practice (for equipment leaks or cooling towers) are not
considered malfunction events, and these emissions are covered using the methodologies presented in
Section 2, Equipment Leaks, and Section 8, Cooling Towers, of this Refinery Emissions Protocol
document. Note, however, the methodology for spills presented in this section can be used for estimating
emissions from equipment leaks that result in liquid puddles. As noted in Section 3, Storage Tanks, of this
Refinery Emissions Protocol document, specific calculations should be made for tank roof landings. The
emissions from tank roof landings are independent of the cause of the landing (intentional or
unintentional), so there is no need to distinguish between landings that occur as a result of a “malfunction
event” or for another reason.
There are three types of malfunction events considered in this section: control device malfunctions;
process vessel overpressurization; and spills.
12-1
12.1 Control Device Malfunctions

When a control device is used to reduce the emissions from a particular source, the site-specific emission
factors or other methodologies provided in this Refinery Emissions Protocol document account for these
controls when the equipment is operating normally. However, when control systems are not operating
normally or or bypassed the emissions may be orders of magnitude more than when these emissions are
controlled. If a control device that normally achieves 99% control efficiency is offline for 3 days, then the
emissions during these 3 days would equal the annual emissions projected for the controlled emission
source.
When there is a malfunction with a control device, the emissions from the source can be estimated based
on the uncontrolled emission factor or by adjusting the controlled emissions based on the control device’s
efficiency. It is important to note that the control device efficiency is often different for different
pollutants. A wet scrubber on an FCCU may be 95% efficient at reducing SO2 emissions, 90% efficient
for PM and metal HAPs, but has no real impact on VOC or NOX emissions. Similarly, oversteaming a
flare that causes poor combustion efficiency will increase VOC and reduced sulfur emissions but reduce
NOX and SO2 emissions. Table 12-1 provides default control efficiencies and correction factors for
certain control device malfunctions. If site-specific emission data are available for the controlled
emissions, then the “uncontrolled” emissions during the control device malfunction or bypass can be
estimated using Equation 12-1.
Eunc ,i = E cont ,i × EM cont ,i × t (Eq. 12-1)
where:
Eunc,i = Uncontrolled emissions estimate from control device malfunction or bypass for
pollutant “i” (kg/event)
Econt,i = Controlled emission rate of pollutant “i” from measurement data or site-specific
emission test data (kg/hr)
EMcont,i = Controlled emission multiplier for pollutant “i” based on the source on control device
from Table 12-1
t = Duration of the event (hr/event).
Table 12-1. Control Device Efficiency and Multiplier Factors for Control Device Malfunctions
Controlled
Source/Control Device Control Device Emission
Descriptiona Pollutant Classb Efficiencyc (%) Multiplierd
FCCU or FCU/wet scrubber PM, metal HAP 92 12.5
SO2 95 20
FCCU or FCU/ESP PM, Metal HAP 92 12.5
SO2, NOX, VOC, organic 0 1
HAP, CO
FCCU or FCU/cyclone PM, Metal HAP 85 6.7
FCCU or FCU/CO boiler CO, VOC, most organic HAP 98% 50
NOX, PAH, Formaldehyde -100% 0.5
FCCU or FCU/SCR NOX 92 12.5
FCCU or FCU/SNCR NOX 60% 1.7
(continued)
12-2
Table 12-1. Control Device Efficiency and Multiplier Factors for Control Device Malfunctions
(continued)
Controlled
Source/Control Device Control Device Emission
Descriptiona Pollutant Classb Efficiencyc (%) Multiplierd
Sulfur plant malfunction with sour SO2 96 – 99.8% 25 - 500
gas sent to flare
H2S 0 – 98% 1– 50
Tail gas treatment malfunction with SO2 95% 20
tail gas sent to flare
H2S 0 – 95% 1 – 20
a Abbreviations:
FCCU = fluid catalytic cracking unit
FCU = fluid coking unit
ESP = electrostatic precipitator
CO boiler = carbon monoxide boiler
SCR = selective catalytic reduction
SNCR = selective non-catalytic reduction
b Pollutant class. Only pollutants affected by the control device are listed. For other pollutants, assume the
control device efficiency is 0% and the controlled emission multiplier is 1.
c Control device efficiency. A negative number indicates a control device that may increase the emissions of a
particular pollutant.
d Controlled emission multiplier. Factor used to escalate the controlled emission factor to account for periods of
control device malfunction = 1/(1-control device efficiency).
12.2 Vessel Overpressurization

Process malfunctions often result in temperature or pressure excursions that must be released from the
vessel to prevent a catastrophic failure. Generally, these emergency releases will be sent to a flare.
Depending on the methodology used to estimate the emissions from the flare, these gas flows may be
included in the flare emission estimates (e.g., when measuring flare gas volumes and composition). If the
flare emission methodology does not account for these flows (e.g., default factors) or if the flow rate
during the event exceeds the range of the flow meter installed on the flare gas line, then the methods
provided in this section should be used.
For discharges through a pressure relief valve or similar discharge of a compressible fluid, the flow
velocity through the valve, pipe, or other restriction is limited to the speed of sound or Mach 1.
Discharges that are limited by the speed of sound are termed “choked” or “sonic,” and discharges that are
less than Mach 1 are termed “unchoked” or “subsonic.” For a pressure relief valve, the outlet pressure is
typically known, being either atmospheric pressure (14.7 psia) or atmospheric pressure plus the back-
pressure of the flare system or other control device. To determine if the flow is choked, first use Equation
12-2 to determine the critical vessel pressure for sonic flow conditions.
k / ( k −1)
 k +1
Pvessel ° = Pout   (Eq. 12-2)
 2 
where:
Pvessel° = Critical vessel pressure for sonic flow conditions (atm)
Pout = Outlet pressure of vent or discharge piping (atm)
k = Ratio of the specific heat at constant pressure to the specific heat at constant volume,
Cp/Cv (dimensionless); see Table 12-2 for values of k for different gases.
12-3
Table 12-2. Values of k for Various Gasesa
Compound k Compound k
Methane 1.30 Air 1.40
Natural gas (methane/ethane) 1.27 Hydrogen 1.40
Ethane 1.22 Nitrogen 1.40
Ethylene 1.20 Oxygen 1.40
Propane 1.14 Carbon monoxide 1.40
n-Butane or Iso-butane 1.11 Carbon dioxide 1.28
Pentane 1.09 Hydrogen sulfide 1.32
Hexane or Cyclohexane 1.08 Sulfur dioxide 1.26
Benzene 1.10
a k = Cp/Cv = ratio of the specific heat at constant pressure to the specific heat at constant volume
If the actual vessel pressure is less than the critical vessel pressure, Pvessel°, calculated from Equation 12-2,
then the flow will be subsonic (unchoked). For subsonic flow, first calculate the Mach number of the
discharge flow using Equation 123.
  k −1 
  
 2    Pvessel 
− 1
k 
M=  ×  (Eq. 12-3)
 k − 1    Pout 

 
where:
M = Mach number of the discharge flow
Pout = Pressure at the discharge outlet, typically atmospheric (atm).
The mass emissions rate is then calculated using Equation 12-4.
k × MWt
(
Ei = C i × A × 6.8087 × 10 4 Pvessel × ) R × Tvessel
×
M
k +1
(Eq. 12-4)
 M (k − 1) 
2 2 ( k −1)
1 + 
 2 
where:
Ei = Emission rate of pollutant “i” (lb/sec)
Ci, = Concentration of pollutant “i” in the discharged gas (weight fraction)
6.8087×104 = Conversion factor (lb/(ft sec2) per atm)
A = Cross-sectional area of the vent outlet (ft2)
MWt = Molecular weight of gas discharged (lb/lb-mol)
R = Ideal gas law constant = 4.968×104 (lb ft2)/(sec2 °R lb-mol)
Tvessel = Temperature of gas in vessel (°R).
If the vessel pressure is greater than Pvessel°, then the flow will be choked (i.e., limited to sonic flow).
Therefore, M is set to 1, and Equation 12-4 can be used (with M=1) to calculate the mass emission rate
during the discharge.
12-4
If the releases are discharged to a flare or other control device (and these discharges are not otherwise
measured), the calculated mass release rates determined using Equation 12-4 must then be adjusted by the
control efficiency to estimate the emissions during the emergency release event.
Equation 12-4 is expressed as an instantaneous emission rate. If the vessel pressure and temperature
remain fairly constant (less than 5% change) during the event, the instantaneous rate can be multiplied by
the duration of the release to determine the total emissions for the event. If the vessel pressure and
temperature vary, Equation 12-4 should be estimated for discrete time intervals and the emissions for each
time interval determined and summed to calculate the total emissions during the event.
Example 12-1: Calculation for Emissions for Vessel Overpressurization

Given: A hydrocracking unit has a 4 inch diameter pressure relief valve set to open at 2,000 psig.
The gas phase is 50 vol% H2 and 50 vol% light (C2 through C4) hydrocarbons. If the pressure
relief valve opened for 30 minutes, what are the VOC emissions from the unit during the event?
Assume the vessel discharges to a flare with an average back pressure of 10 psig. The average
temperature and pressure of the vessel during the event were 400 °C and 2,050 psig, respectively.
Solution: Pout = 10+14.7 = 24.7 psia. On a mass basis, most of the flow will be the hydrocarbons.
Therefore, the k value for propane of 1.14 is selected from Table 2-2 as representative of the gas
stream. To determine if the flow is choked, first use Equation 12-2 to determine the critical vessel
pressure for sonic flow conditions.
k / ( k −1) 1.14 / (0.14 )
 k +1  1.14 + 1 
Pvessel ° = Pout   = 24.7   = 42.9 psia
 2   2 
The pressure of the vessel is far greater than this, so the flow is choked. Equation 2-4 is used to
calculate the discharge rate by setting the Mach number is set to 1. Equation 2-4 requires the
following variables in specific units.
• MWt: Using propane’s molecular weight of 44 lb/lb-mol for the hydrocarbon portion, the
average molecular weight of the discharge gas is: 0.5×2 + 0.5×44 = 23 lb/lb-mol.
• Ci: Using propane as the surrogate for the VOC, the weight fration of VOC is 0.5×44/23
= 0.9565 weight fraction VOC.
• A: Valve diameter = 4”/12 = 0.3333 ft. Area = (π/4)×(0.3333)2 = 0.08725 ft2
• Pvessel: (2050+14.7)/14.7 = 140.5 atm
• Tvessel: (400 + 273.15)×1.8 = 1,212 °R
Plugging these values into Equation 2-4 yields:
1.14 × 23
(
Ei = 0.9565 × 0.08725 × 6.8087 × 10 4 × 140.5 × ) ×
1
4.968 × 10 4 × 1,212 2.14
 0.14  2(0.14 )
1 + 
 2 
Ei = 798,400 × 4.355 × 10 −7 × 0.5962 = 314 lb/sec

The event occurred for 1,800 sec (30 min ×60 sec/min), so the total release was 565,200 lb or
283 tons. With a flare efficiency of 98%, the emissions to the atmosphere related to this event was
5.66 tons VOC [283 × (1-0.98)].
12-5
12.3 Spills
Generally, it can be assumed that 100% of the compounds spilled are emitted into the atmosphere. For
heavier liquids, the mass transfer correlations for an oil layer (Koil) as provided in Appendix B,
Wastewater Treatment System Equations, Section B.2.1, Oil Water Separators, can be used. Equation 12-
5 can be used to estimate the emissions based on the quantity of material spilled, the area of the spill, and
the time period from when the initial spill occurred to the final clean up:
(
Ei = ρVC i ,o 1 − e − K oil At / V ) (Eq. 12-5)
where:
Ei =
Emission estimate of pollutant “i” (kg/spill event)
ρ Density of spilled liquid (kg/m3)
=
V Volume of spilled material (m3)
=
Ci,o =
Concentration of pollutant “i” in spilled fluid prior to the spill event (mass fraction;
kg/kg)
Koil = Mass transfer coefficient for an oil (organic) liquid (m/sec)
t = Duration of the spill (sec)
A = Area covered by the spill (m2).
12-6
Version 3 Section 13—References
13. References
API (American Petroleum Institute). 2002. Refinery Stream Speciation. API Publication No. 4723.
November 1.
API (American Petroleum Institute). 2009. Compendium of Greenhouse Gas Emissions Methodologies
for the Oil and Natural Gas Industry. August 2009. Available at:
http://www.api.org/ehs/climate/new/upload/2009_GHG_COMPENDIUM.pdf.
API (American Petroleum Institute) and NPRA (National Petrochemical and Refiners Association). 2010.
Review of the Emission Estimation Protocol for Petroleum Refineries. Letter to Ms. Brenda
Shine, U.S. EPA. March 31.
Bertrand R. R., and J.H. Siegell. 2002. Emission of trace compounds from catalytic cracking regenerators,
Environmental Progress, 21(3)163-167. October.
Bertrand R. R., and J.H. Siegell. 2003. Emission of trace compounds from catalytic reforming units,
Environmental Progress, 22(1)74-77. April.
Coburn, J. 2013. Memorandum from J. Coburn, RTI, to Brenda Shine, EPA/OAQPS/SPPD. Impacts
Estimates for Delayed Cokin Units. September 12
Coburn, J. and M. Icenhour. 2008. Memorandum from J. Coburn and M. Icenhour, RTI, to Brenda Shine,
EPA/SPPD. Preliminary Analysis of Short-term Variability in Storage Vessels Emissions.
December 29.
EIA (U.S. Energy Information Administration). 2009. Refinery Capacity Report 2009. Prepared by the
Energy Information Administration, Washington, DC. June 25.
EIIP (Emission Inventory Improvement Program). 2007. Methods for Estimating Air Emissions from
Chemical Manufacturing Facilities. Volume 2, Chapter 16. Prepared by Mitchell Scientific, Inc.
and RTI International. August.
EMEP/CORINAIR (European Monitoring and Evaluation Programme/COoRdination of Information on

AIR emissions). 2006. Atmospheric Emission Inventory Guidebook 2006. Technical Report No.
11/2006. Available at: http://www.eea.europa.eu/publications/EMEPCORINAIR4.
ERG (Eastern Research Group, Inc). 2001. Emission Inventory Improvement Program Technical Report
Series. Volume II, Chapter 2: “Preferred and Alternative Methods for Estimating Air Emissions
from Boilers.” Prepared for Point Sources Committee Emission Inventory Improvement Program.
January.
Hansell, D., and G. England. 1998. Air Toxic Emission Factors for Combustion Sources Using Petroleum
Based Fuels, Volumes 1 and 2. API Publication 348. Prepared by EER for S. Folwarkow,
Western States Petroleum Association, Concord, CA, and K. Ritter, American Petroleum
Institute, Washington, DC.
Levelton Consultants Ltd. 2005. Emission Monitoring and Reporting Strategy – Summary and
Background. Prepared for Canadian Council of Ministers of the Environment and NFPRER
Monitoring and Reporting Team.
13-1
Lev-On, M., Epperson, D., Siegell, J., and Ritter, K. (2007). Derivation of new emission factors for
quantification of mass emissions when using optical gas imaging for detecting leaks. J Air Waste
Manag Assoc 57( 9):1061-1070
Lucas, B. 2007. Memorandum from B. Lucas, EPA/SPPD, to Project Docket File (EPA Docket No. EPA-
HQ-OAR-2003-0146). Storage Vessels: Control Options and Impact Estimates. August 3, 2007.
Docket Item No. EPA-HQ-OAR-2003-0146-0014.
NYSERDA (New York State Energy Research and Development Authority). 2005. Hydrogen Fact Sheet.
Hydrogen Production–Steam Methane Reforming (SMR). Available at:
http://www.getenergysmart.org/Files/HydrogenEducation/6HydrogenProductionSteamMethaneR
eforming.pdf.
Potter, T., and K. Simmons. 1998. Composition of Petroleum Mixtures, Volume 2. Total Petroleum
Hydrocarbon Criteria Working Group Series. University of Massachusetts. Amherst,
Massachusetts.
RTI (Research Triangle Institute). 2002. Petroleum Refinery Source Characterization and Emission
Model for Residual Risk Assessment. Prepared for U.S. Environmental Protection Agency, Office
of Air Quality Planning and Standards, Research Triangle Park, NC. EPA Contract No. 68-D6-
0014. July.
Sander, R. 1999. Compilationof Henry’s Law Constants for Inorganic and Organic Species of Potential
Importand in Environmental Chemistry. Version 3. Max-Planck Institute of Chemistry, Mainz,
Germany. April 8.
SCAQMD (South Coast Air Quality Management District). 2004a. Source Test Report 03-194 Conducted
at Chevron/Texaco Refinery, El Segundo, California—Volatile Organic Compound (VOC),
Carbon Monoxide (CO), and Particulate Matter (PM) Emissions from a Coke Drum Steam Vent.
May 14.
SCAQMD (South Coast Air Quality Management District). 2004b. Source Test Report 03-197 Conducted
at Conoco-Phillips Refinery, Carson, California—Volatile Organic Compound (VOC), Carbon
Monoxide (CO), and Particulate Matter (PM) Emissions From a Coke Drum Steam Vent. July 23.
SCAQMD (South Coast Air Quality Management District). 2004c. Source Test Report 03-198 Conducted
at Exxon Mobil Refinery, Torrance, California—Volatile Organic Compound (VOC), Carbon
Monoxide (CO), and Particulate Matter (PM) Emissions From a Coke Drum Steam Vent.
March 4.
SCAQMD (South Coast Air Quality Management District). 2004d. Source Test Report 03-200 Conducted
at Shell Oil Refinery, Wilmington, California—Volatile Organic Compound (VOC), Carbon
Monoxide (CO), and Particulate Matter (PM) Emissions From a Coke Drum Steam Vent. July 1.
TCEQ (Texas Commission on Environmental Quality). 2000. Air Permit Technical Guidance for
Chemical Sources: Loading Operations.
TCEQ (Texas Commission on Environmental Quality). 2003. Source Sampling Procedures Manual,
Appendix P, Cooling Tower Monitoring, Air Stripping Method (Modified El Paso Method) for
Determination of Volatile Organic Compound Emissions from Water Sources. January 31.
TCEQ (Texas Commission on Environmental Quality). Air Quality Division. 2009. 2008 Emissions
Inventory Guidelines. RG-360A/08. January.
13-2
The United States – Mexico Foundation for Science. 2008. A Customized Approach for Quantifying
Emissions for the Electric Power Generation Sector in Mexico. Prepared for: SEMARNAT and
CFE. Revision 2, Revised by RTI International for U.S. Environmental Protection Agency Clean
Air Markets Division and Climate Change Division. December.
Tumbore, DC. 1998. The magnitude and source of air emissions from asphalt blowing operations.
Environmental Progress 17(1):53–59.
U.S. EPA (Environmental Protection Agency). 1995a. Compilation of Air Pollutant Emission Factors.
Volume 1: Stationary Point and Area Sources. AP-42, Fifth Edition. Office of Air Quality
Planning and Standards, Research Triangle Park, NC.
U.S. EPA (Environmental Protection Agency). 1995b. Protocol for Equipment Leak Emission Estimates.
EPA-453/R-95-017. Office of Air Quality Planning and Standards, Research Triangle Park, NC.
U.S. EPA (Environmental Protection Agency). 1996. Method 8260B, Volatile Organic Compounds by
Gas Chromatography/Mass Spectrometry. Revision 2, December. SW-846.
U.S. EPA (Environmental Protection Agency). 1998a. Locating and Estimating Air Emissions from
Sources of Benzene. EPA-454/R-98-011. Office of Air Quality Planning and Standards, Research
Triangle Park, NC.
U.S. EPA (Environmental Protection Agency). 1998b. Petroleum Refineries—Background Information

for Proposed Standards, Catalytic Cracking (Fluid and Other) Units, Catalytic Reforming Units,
and Sulfur Recovery Units. EPA-453/R-98-003. Washington, DC: Government Printing Office.
U.S. EPA (Environmental Protection Agency). 1998c. Study of Hazardous Air Pollutant Emissions from
Electric Utility Steam Generating Boilers—Final Report to Congress. Volume 1. EPA-453/R-98-
004a. Office of Air Quality Planning and Standards, Research Triangle Park, NC.
U.S. EPA (Environmental Protection Agency). 2006. TANKS Emissions Estimation Software, Version
4.09D. Office of Air Quality Planning and Standards, Technology Transfer Network,
Clearinghouse for Inventories and Emissions Factors, Available at
http://www.epa.gov/ttn/chief/software/tanks/index.html.
U.S. EPA (Environmental Protection Agency). 2009. Mandatory Reporting of Greenhouse Gases (Final
Rule). Federal Register: 74 FR 56260. October 30.
URS Corporation. 2008. Source Test Report of the Coker Steam Vent. Prepared for Hovensa, LLC, St.
Croix, U.S. Virgin Islands. September 8.
Yerushalmi, L., and S.R. Guiot. 1998. Kinetics of biodegradation of gasoline and its hydrocarbon
constituents. Appl Microbiol Biotechnol 49:475−481.
13-3
13-4
Version 3 Appendix A—Average Stream Compositions
Appendix A
Average Stream Compositions
A-1
A-2
Table A-1. Refinery Average Stream Hazardous Air Pollutant Compositions—Default Values
Average Weight Percent of Compound in Process Unit Stream a
2,2,4-Trimethyl
Xylenes (total)
1,2,4-trimethyl
1,3-Butadiene
Ethylbenzene
Naphthalene
1330-20-7
n-Hexane
benzene b
110-54-3
Biphenyl
Benzene
106-99-0
540-84-1
108-88-3
100-41-4
100-42-5
Cumene
pentane
Toluene
Styrene
71-43-2
98-82-8
95-63-6
91-20-3
92-52-4
Process Units
Atmospheric Distillation 0.01 4.28 0.05 0.86 1.7 1.97 0.63 0.12 0.63 0.25 0.06 0.001
Vacuum Distillation - 0.007 0.004 0.004 0.043 0.020 0.024 0.001 0.020 0.12 0.09 -
Coking Unit 0.04 2.49 0.75 0.42 1.30 1.76 0.70 0.12 0.70 0.28 0.001 0.19
Cat Hydrocracker 0.03 1.86 1.04 1.27 2.72 2.67 1.02 0.09 1.33 0.20 0.001 0.19
FCCU 0.01 0.99 0.29 1.03 3.28 4.90 1.07 0.10 1.92 0.72 0.43 0.14
CRU 0.009 2.75 0.25 6.34 17.44 17.61 3.88 0.42 5.93 0.87 0.001 0.11
Hydrodesulfurization - 1.89 0.0001 0.37 1.72 1.94 0.37 0.07 0.40 0.25 0.22 -
Alkylation 0.22 1.57 25.24 0.03 2.05 0.08 5.0E-5 5.0E-5 5.0E-5 5.0E-5 5.0E-5 0.02
Isomerization - 3.22 0.01 0.51 0.64 0.15 - - - - - -
Polymerization - 0.54 0.73 1.24 1.76 1.18 0.09 - - - -
Products
Aromatics Extract -
Benzene - - - 99.99 0.009 0.002 - - - - - -
Aromatics Extract -
Toluene - - - 0.296 95.61 3.49 2.82 - 0.3 - - -
Aromatics Extract -
Xylenes - - - - - 77.89 16.08 0.59 - - - -
Aromatics Extract -
Heavy Aromatics - - - 5.7 19.36 25.03 3.56 0.965 4.94 0.314 - -
Commercial Jet Fuel - 1.68 0.0006 0.577 0.965 1.22 0.257 0.106 0.815 0.627 0.409 -
Conventional Gasoline 0.0009 2.22 2.94 1.4 c 7.75 6.56 1.490 0.219 3.002 0.415 - 0.0767
Aviation Gasoline - 2.30 - 1.01 7.69 3.37 1.2 0.103 0.5 0.695 - -
Reformulated Gasoline - 1.36 1.10 0.8b 7 7.31 1.26 0.19 2.52 0.21 - -
Diesel Fuel - 0.030 0.007 0.004 0.081 0.241 0.053 0.056 0.349 0.285 0.117 -
Home Heating Oil - - - - 0.108 0.300 0.095 - 0.429 0.453 0.181 -
Solvent-refined Lubes - - - - - - - - - - - -
9.81E-
Crude Oil 05 1.341 0.185 0.312 0.542 0.811 0.186 0.108 0.377 0.139 0.030
a
Based on Petroleum Environmental Research Forum (PERF) refinery process stream speciation study (API, 2002),
unless otherwise specified.
b
1,2,4-Trimethylbenzene is not a HAP, but it is a prevalent VOC in many refinery streams.
c
Average benzene content of conventional and reformulated gasoline based on Fuel Trends Resport: Gasoline
1995-2005 (U.S. EPA. Report No. EPA420-R-08-002. January 2008), corrected to weight percent using the
density correction factor of 0.876/0.739 (i.e., 1.16vol%×0.876/0.739 = 1.4wt%; 0.67vol%×0.876/0.739 = 0.8wt%).
A-3
Property Calculations
If the weight composition of a commodity is available but other property data are not, estimates can be
made. For example, if the liquid component mixture and the weight fraction for gasoline are known, then
the vapor pressure of the liquid, the MW of vapor, and the weight fraction of the vapor can be estimated.
The following steps and equations may be used for calculating these estimates:
1. From the weight fraction values in the gasoline (liquid), Equation A-1 should be used to estimate
the number of moles of component “i.”
Wtfraci
ni = (Eq. A-1)
MWi
2. From moles of component “i” (liquid), Equation A-2 should be used to estimate the mole fraction
of component “i” (liquid).
ni
xi = (Eq. A-2)
nT
3. Equation A-3 should be used to calculate the partial pressure of each component “i” in liquid.
P(i) = P × xi (Eq. A-3)

where:
P(i) = Partial pressure of component “i” (psia)
P = Vapor pressure of pure component “i” at the liquid temperature (psia)
xi = Liquid mole fraction
4. To estimate the total pressure of the liquid, Equation A-4 should be used to estimate the average
vapor pressure for the gasoline mixture by summing the partial pressure of each component “i”
(liquid).
n
PtotalVP = ∑ Pi (Eq. A-4)
1
5. From mole fraction of component “i” (liquid), Equation A-5 should be used to estimate the mole
fraction of component “i” in the vapor using Raoult’s law.
Pi
yi = (Eq. A-5)
PtotalVP
6. Equation A-6 should be used to estimate the MW of the gasoline vapor, MWVAP, using the mole
fraction of vapor (yi) and the MW of the pure component “i”(MWi).
n
MWVap = ∑ y i × MWi (Eq. A-6)
1
where:
MWVAP = Molecular weight of the vapor (lb/mol)
MWi = Molecular weight of component “i” (lb/mol)
A-4
7. Equation A-7 should be used to estimate the gasoline vapor weight fractions using the gasoline
vapor mole fractions and the MWVAP.
y i × MWi
Wtfracvap = (Eq. A-7)
MWVap
Example A-1: Calculation of Vapor Pressure and Molecular Weight for Reformulated
Gasoline Loading Example
Calculate the vapor pressure of gasoline and MWVAP for submerged loading of reformulated
gasoline into dedicated-service gasoline transport tank trucks (each with a volume of 8,000
gallons) based on the following information:
The refinery loads 100 tank trucks per year at a bulk temperature of 77°F. Property data are
available for liquid composition of the gasoline only and are provided in Table A-2.
First, the vapor pressure for each constituent is determined at the loading temperature using
Antoine’s equation and Antoine’s coefficients, such as those provided in Table 7.1-5 of AP-42
(U.S. EPA, 1995a).
Using Steps 1 through 7 from this Appendix A, the gasoline liquid mole fraction can be
calculated from the weight fraction or weight percent component composition, the partial
pressure of each component in liquid, and the gasoline vapor mole fraction using Raoult’s law
(see each calculation “step” in the columns in Table A-2).
Wtfraci
 Step 1. ni =
MWi
ni
 Step 2. xi =
nT
 Step 3. P (i ) = P × xi
n
 Step 4. PtotalVP = ∑P 1
i
Pi
 Step 5. y i =
PtotalVP
n
 Step 6. MWVap = ∑y
1
i × MWi
y i × MWi
 Step 7. wtfrac vap =
MWVap
See Table A-2 for the calculations of thee variables. At 77°F, the total system vapor pressure is
4.77 psia and the MW of the vapor, MWGasolineVAP, is 67.4 lb/lb-mol.
A-5
Table A-2. Sample Calculation of Average Molecular Weight and Vapor Weight Fraction a
Step 1 Step 2 Step 3 Step 4 Step 5 Step 6 Step 7

G=H/(ΣH)*
A B C D=C/A E=D/(ΣD) F=E*B G=F/(ΣF) H=G*A 100%
Gasoline Gasoline Gasoline Gasoline MolWt
Gasoline Liquid Liquid Partial Vapor Contribution
CAS Pvap at Liquid (moles/100 (mole Pressure (mole Gasoline Gasoline
Number Compound MolWt 77°F (psia) (wt%) g) fraction) (psia) fraction) Vapor (g/mol) Vapor (wt%)
71-43-2 Benzene 78.11 1.840 0.80 1.02E-04 9.36E-03 1.72E-02 3.61E-03 2.82E-01 0.42
110-54-3 Hexane 86.18 2.928 1.36 1.58E-04 1.44E-02 4.22E-02 8.85E-03 7.63E-01 1.13
108-88-3 Toluene 92.14 0.550 7.0 7.60E-04 6.94E-02 3.82E-02 8.01E-03 7.38E-01 1.09
1330-20-7 Xylene 106.14 0.169 7.31 6.89E-04 6.30E-02 1.07E-02 2.23E-03 2.37E-01 0.35
100-41-4 Ethylbenzene 106.17 0.184 1.26 1.19E-04 1.08E-02 1.99E-03 4.18E-04 4.44E-02 0.07
540-84-1 2,2,4-Trimethylpentane 114.23 0.954 1.098 9.61E-05 8.79E-03 8.38E-03 1.76E-03 2.01E-01 0.30
98-82-8 Cumene 120.19 0.089 0.19 1.58E-05 1.44E-03 1.28E-04 2.68E-05 3.23E-03 0.005
91-20-3 Naphthalene 128.17 0.005 0.21 1.64E-05 1.50E-03 7.63E-06 1.60E-06 2.05E-04 0.0003
106-97-8 n-Butane 58.12 35.220 4.7 8.09E-04 7.39E-02 2.60E+00 5.46E-01 3.17E+01 46.99
72-28-5 Isobutane 58.12 13.294 1.7 2.92E-04 2.67E-02 3.55E-01 7.45E-02 4.33E+00 6.42
109-66-0 Pentane 72.15 10.285 3.9 5.41E-04 4.94E-02 5.08E-01 1.07E-01 7.69E+00 11.39
78-78-4 Isopentane 72.15 3.670 7.9 1.09E-03 1.00E-01 3.67E-01 7.70E-02 5.56E+00 8.23
C6 surrogate 86.18 4.094 10.08 1.17E-03 1.07E-01 4.38E-01 9.18E-02 7.91E+00 11.72
C7 surrogate 100.2 1.274 7.21 7.20E-04 6.58E-02 8.38E-02 1.76E-02 1.76E+00 2.61
C8 surrogate 114.23 0.522 3.21 2.81E-04 2.57E-02 1.34E-02 2.81E-03 3.21E-01 0.48
C9 surrogate 120.19 0.039 7.29 6.07E-04 5.54E-02 2.18E-03 4.57E-04 5.49E-02 0.08
Other VOC 100.20 0.884 34.78 3.47E-03 3.17E-01 2.80E-01 5.88E-02 5.89E+00 8.73
nT PtotalVP MW vap
Total 100 0.0109 1.00 4.77 1.00 67.49 100
a
Weight percents for HAP based on composition provided in Table A-1; weight percent VOC from n-butane to C9 surrogate are based on Composition of
Petroleum Mixtures, Volume 2 of the Total Hydrocabon Criteria Working Group Series. Prepared by T.L. Potter and K.E. Simmons. May 1998.
ISBN 1844940-19-6; n-Heptane was used as the surrogate for the remaining uncharacterized fraction of the liquid gasoline (“other VOC”). Facilities should
use facility-specific data for light VOC (C4 to C6 compounds) concentrations because these significantly impact the calculated vapor-phase composition.
A-6
Version 3 Appendix B—Wastewater Treatment System Equations
Appendix B
Wastewater Treatment System Equations
B-1
B-2
B. Wastewater Treatment System Equations

This appendix provides a brief presentation of the key equations needed to calculate air emissions from
typical wastewater treatment units. The purpose of this appendix is to document the equations used in the
simplified refinery wastewater emission tool and to correct errors found in the presentation of some of
these equations as provided in Chapter 4.3 of AP-42 (EPA, 1995).
B.1 Mass Transfer Rate Equations

The overall mass transfer coefficient that determines the rate of volatilization is determined based on a
two-resistance module: a liquid phase mass transfer resistance and a gas phase mass transfer resistance.
The liquid and gas phase mass transfer resistances are very different for turbulent surfaces compared to
quiescent (laminar flow) surfaces. Therefore, the overall mass transfer coefficient is a composite of the
overall mass transfer coefficient for the turbulent surface area and the overall mass transfer coefficient for
the quiescent surface area based on an area weighted average as follows:
K OL ,t × At + K OL ,q × Aq
K OL = (Eq. B-1)
A
where:
KOL = Overall mass transfer coefficient (m/s)
KOL,t = Overall mass transfer coefficient for turbulent surface areas (m/s)
At = Turbulent surface area = faer A, m2
faer = Fraction of total surface area affected by aeration
KOL,q = Overall mass transfer coefficient for quiescent surface areas (m/s)
Aq = Quiescent surface area = (1-faer) A, m2 (Note: At + Aq must equal A)
A = Total surface area (m2)
The overall mass transfer coefficient for turbulent surface areas based on the two resistance module is:
−1
 1 1 
K OL ,t = +  (Eq. B-2)
k 
 l ,t H ′ × k g ,t 
where:
kl,t = liquid phase mass transfer coefficient for turbulent surface areas (m/s)
H' = dimensionless Henry's law constant = H/RTH
H = Henry's law constant (atm-m3/mol)
R = ideal gas law constant = 0.00008205 (atm-m3/mol-K)
TH = temperature at which Henry's law constant was evaluated = 298 K.
kg,t = gas phase mass transfer coefficient for turbulent surface areas (m/s)
Similarly, the overall mass transfer coefficient for quiescent surface areas is
−1
 1 1 
K OL ,q = +  (Eq. B-3)
k 
 l ,q H ′ × k g ,q 
where:
kl,q = Liquid phase mass transfer coefficient for quiescent surface areas (m/s)
kg,q = Gas phase mass transfer coefficient for quiescent surface areas (m/s).
B-3
The mass transfer correlations used in this module to estimate the individual mass transfer coefficients are
the same as those used in the WATER9 emission model developed by EPA. The documentation of these
mass transfer correlations can be accessed from EPA’s Web site at
http://www.epa.gov/ttn/chief/software/water/index.html. Only the basic equations are provided here. For a
more detailed discussion of these mass transfer correlations, the reader is referred to Chapter 5 of the Air
Emissions Models for Waste and Wastewater report (U.S. EPA, 1994).
B.1.1 Liquid Phase Mass Transfer Coefficient for Turbulent Surfaces

The liquid phase, turbulent surface mass transfer coefficient is calculated as follows:
 8.22 × 10 −3 × J × Ptot × 1.024 (T − 20 ) × Ocf × MWl

0.5
  Di ,l 
k l ,t =   (Eq. B-4)
 10.76 × At × ρ l   DO 2,l 
  
where:
J = Oxygen transfer factor (lb/h/hp)
Ptot = Total power to the impellers (hp)
T = Water temperature in (°C)
Ocf = Oxygen correction factor
MWl = Molecular weight of liquid (water) (g/mol)
ρl = Density of liquid (water) (g/cm3 = Mg/m3)
Di,l = Diffusivity in liquid (water) (cm2/s)
DO2,l = Diffusivity of oxygen in liquid (water) (cm2/s)
B.1.2 Gas Phase Mass Transfer Coefficient for Turbulent Surfaces

The gas phase, turbulent surface mass transfer coefficient is calculated as follows:
−1
k g ,t = 1.35 × 10 −7 × Re g × p 0.4 × Sc g × Fr −0.21 × Di.a × MWa × d imp
1.42 0.5
(Eq. B-5)
where:
Reg = Gas phase Reynolds number = (dimp2 w ρg)/μg
ρg = Density of gas (air) (g/cm3)
μg = Viscosity of gas (air) (g/cm-s)
p = Power number = 0.85 (550 Ptot/Naer) gc,2 / [(62.428ρl )w3 (dimp/30.48)5 ]
gc,2 = Gravitational constant = 32.17 lbm-ft/s2-lbf = 0.03283 gc
Naer = Number of aerators
w = Rotational speed (rad/s)
Scg = Gas phase Schmidt number = μg/(ρg Di,a)
Fr = Froud number = [w2 (dimp/30.48) ]/ gc,2
Di,a = Diffusivity of constituent in air (cm2/s)
MWa = Molecular weight of air (g/mol)
dimp = Impeller diameter (cm)
gc = Gravitational constant = 980 cm/s2
B-4
B.1.3 Liquid Phase Mass Transfer Coefficient for Quiescent Surfaces

The appropriate correlation to use to estimate the liquid phase mass transfer coefficient is dependent on
the wind speed and the fetch-to-depth ratio of the impoundment. The fetch is the linear distance across the
treatment unit, and it is calculated from the unit’s surface area, assuming a circular shape. That is,
0.5
 4A 
F =  (Eq. B-6)
 π 
where:
F = Fetch (m)
For wind speeds less than 3.25 m/s, the following correlation is used regardless of the fetch-to-depth ratio
(F/dliq):
2
 Di ,l 3
k l ,q = 2.78 × 10 −6   (Eq. B-7)
 Dether 
where:
kl,q = Liquid phase, quiescent surface mass transfer coefficient (m/s)
Di,l = Diffusivity of constituent in liquid (water) (cm2/s)
Dether = Diffusivity of ether in water = 8.5 × 10-6 cm2/s
For windspeeds greater than or equal to 3.25 m/s, the appropriate correlation is dependent on the fetch-to-
depth ratio as follows:
For
F
d liq
< 14, (
k l ,q = 1.0 × 10 −6 + a × 10 − 4 U * )( ) b
Scliq
− 0.5
(Eq. B-8)
where:
a = Equation constant, a = 34.1 for U* > 0.3 m/s; a = 144 for U* < 0.3 m/s
U* = Friction velocity, m/s = 0.01U (6.1 + 0.63U)0.5
U = Wind speed at 10 m above the liquid surface
b = Equation constant, b = 1 for U* > 0.3 m/s; b = 2.2 for U* < 0.3 m/s
Scliq = Liquid phase Schmidt number = μl/(ρl Di,l)
μl = Viscosity of water (g/cm-s)
ρl = Density of water (g/cm3)
2
F   F    D 3
For 14 ≤ ≤ 51.2, k l ,q = 2.605 × 10 −9   + 1.277 × 10 −7  U 2  i ,l  (Eq. B-9)
d  D
d liq   liq    ether 
2
F  Di ,l 3
For > 51.2, k l ,q = 2.611 × 10 −7 U 2   (Eq. B-10)
d liq  Dether 
B-5
B.1.4 Gas Phase Mass Transfer Coefficient for Quiescent Surfaces

The gas phase mass transfer coefficient for quiescent surface areas is estimated as follows:
−0.67
k g ,q = 4.82 × 10 −3 × U 0.78 × Sc g × F −0.11 (Eq. B-11)
B.2 Collection Systems and Process Units

The fraction emitted (fe) values for the collection system components are based on the models described
in the EPA document titled Industrial Wastewater Volatile Organic Compound Emissions (U.S. EPA,
1990). Briefly, the correlation between the fraction emitted (fe) and Henry’s law constant were
determined. The correlation equations were then used to determine the emissions for each compound and
uncontrolled collection system component.
Table B-1. Model Collection System Components and fe–Henry’s Law Correlation
Collection System Component Correlation

Drains fe=0.035*ln(HLC)+0.4079
Trenches fe=0.005*ln(HLC)+0.0658
Manholes fe=0.009*ln(HLC)+0.1036
Junction Boxes fe=0.0105*ln(HLC)+0.1416
Lift Stations fe=0.0312*ln(HLC)+0.4163
Sumps fe=0.0004*ln(HLC)+0.007
B.2.1 Junction Boxes and Lift Stations

Junction boxes and lift stations collect and equalize wastewater prior to pumping to the wastewater
treatment system. As previously mentioned, these components should be covered the vented; however, if
uncovered, the following procedure can be used to estimate air emissions:
 Step 1: Calculate the liquid phase mass transfer coefficient for turbulent surfaces (kl,t) with
Equation B-4. Alternatively, a simplified version specifically for junction boxes and lift stations
that is based on default values presented in AP-42 yields:
 0.184V 1.024 (T − 20 ) (Di ,l )0.5 
k l ,t =  (Eq. B-12)
 At 
where:
V = Unit volume (m3)
Default values:
J = 3 (lb/h/hp)
Ocf = 0.83
MWl = 18 (g/mol)
ρl = 1 (g/m3)
Ptot = 0.0264 (hp/m3) ×V
DO2,l = 2.4×10-5 (cm2/s)
 Step 2: Calculate the gas phase mass transfer coefficient for quiescent surfaces (kg,q) as described
in Section B.1.4, Gas Phase Mass Transfer Coefficient for Quiescent Surfaces.
B-6
 Step 3: Calculate the overall mass transfer coefficient (K) with the following equation:
(k H ' k g ,q )
K OL =
l ,t
(H ' k g ,q + k l ,t )
(Eq. B-13)
 Step 4: Calculate the tank effluent concentration (CL) with the following equation:
QC 0
CL (g / m3 ) = (Eq. B-14)
(K OL A + Q )
 Step 5: Calculate the air emissions (N) with the following equation:
N ( g / s ) = K OL C L A (Eq. B-15)
B.2.2 Sumps
Sumps typically collect and equalize wastewater from various collection systems prior to treatment. As
previously mentioned, these components should be covered the vented; however, if uncovered, the
following procedure can be used to estimate air emissions:
 Step 1: Calculate the liquid phase mass transfer coefficient for quiescent surfaces (kl,q) as
described in Section B.1.3, Liquid Phase Mass Transfer Coefficient for Quiescent Surfaces.
 Step 3: Calculate the overall mass transfer coefficient (K) with the following equation:
(k H ' k g ,q )
K OL =
l ,q
(H ' k g ,q + k l ,q )
(Eq. B-16)
 Step 4: Calculate the tank effluent concentration (CL) with using Equation B-14:
 Step 5: Calculate the air emissions (N) with using Equation B-15:
B.2.3 Weirs
Weirs are typically used in wastewater collection and treatment units as dams, allowing solids to settle in
the quiescent areas. The liquid phase mass transfer coefficients (kl) is determined by Equations B-17a and
17b. The gas phase mass transfer coefficients (kg) is determined using Equation B-18. These liquid and
gas phase mass transfer coefficient equations are based on the work presented by Nakasone (1987) and
Pincince (1991). Air emissions (N) for each pollutant are calculated using equation B-20.
ln(r ) = 0.042h 0.872 q 0.509 (Eq. B-17a)

0.66
qD 
kl =  lv  ln(r ) (Eq. B-17b)
z  Dlo 
0.66
D 
k g = 0.05 gv  (Eq. B-18)
D 
 go 
B-7
−1
1 1 
Ko =  + (Eq. B-19)
k 
 l kg K 
g  K Z
N   = 1 − exp − o QCo (Eq. B-20)
s  q 
where:
Q = Volumetric flow rate (m3/s)

Co = Initial constituent concentration (g/m3)
h = Weir height (ft)
r = oxygen deficit ratio
q = hydraulic weir loading [m3/(m∙h)]
Z = tailwater depth (m)
kl = liquid-phase mass transfer coefficient (m/s)
Dlv = diffusion coefficient for the organic in water (cm2/s)
Dlo = diffusion coefficient for air in water (cm2/s)
kg = gas-phase mass transfer coefficient (m/s)
Dgv = diffusion coefficient of the organic in air (cm2/s)
Dgo = diffusion coefficient of the reference material in air (cm2/s)
Ko = overall mass transfer coefficient (m/s)
K = partition coefficient (atm-m3/mol)
N = air emissions (g/s).
B.2.4 Oil-Water Separators (API)

An oil-water separator is a treatment unit designed to separate oil and suspended solids from wastewater.
As previously mentioned, these units are typically covered and vented and thus open-air emission
estimations are not required. However, if the treatment units are uncovered, then the following procedure
can be used to estimate air emissions.
First, calculate the gas phase mass transfer coefficient for quiescent surfaces (kg,q) with Equation B-11.
Next, calculate the overall mass transfer coefficient for oil (Koil) with the following equations:
K oil = k g ,q Keqoil (Eq. B-21)
where:
P * ρ a MWoil
Keq oil = (Eq. B-22)
ρ oil MWa Po
Air emissions are determined from the following equations:
N ( g / s ) = K oil C L ,oil A (Eq. B-23)
where:
( )
C L ,oil g / m 3 =
Qoil Cooil
K oil A + Qoil
(Eq. B-24)
B-8
If the oil layer is > 1cm, then use this equation:

K owCo
Cooil = (Eq. B-25)
[1 − f oil + f oil K ow ]
If the oil layer is < 1cm, then use this equation:
Co
Cooil = (Eq. B-26)
f oil
Qoil = f oil Q (Eq. B-27)
where:
P* = Constituent vapor pressure (mm Hg)
MWoil = Molecular weight of oil (g/mol)
ρoil = Density of oil (g/m3)
P0 = Total pressure (atm)
Qoil = Volumetric flow rate of oil (m3/s)
Cooil = Initial concentration of constituent in oil phase (g/m3)
Kow = Octanol-water partition coefficient
foil = Fraction of volume that is oil
B.2.5 Dissolved Air Flotation

DAF is a wastewater treatment unit that clarifies water by removing suspended oil and solids by bubble
flotation. As previously mentioned, these treatment units are typically covered and vented to control
devices. However, if the treatment units are uncovered, then the following procedure can be followed to
estimate air emissions.
First, the liquid phase mass transfer coefficient for quiescent surfaces (kl,q) can be determined by the
method described in Section B.1.3, Liquid Phase Mass Transfer Coefficient for Quiescent Surfaces.
Second, the gas phase mass transfer coefficient for quiescent surface (kg,q) can be calculated by the
method described in Section B.1.4. Third, the overall mass transfer coefficient (KOL) can be determined
by Equation B-1. Finally, the air emissions (N) can be calculated with the following equations:
N ( g / s ) = (K OL A + Qa H ′)C L (Eq. B-28)

where:
QC 0
CL (g / m3 ) = (Eq. B-29)
(K OL A + Q + Qa H ′)
where:
Qa = Volumetric flow rate of diffused air (m3/s), at conditions of 1 atmosphere and bulk
liquid temperature
B.2.6 Equalization Tanks (Mixed Tank with No Biodegradation)

Equalization tanks dampen variations in wastewater flow rate and pollutant load to lessen negative
impacts on downstream processes. As previously mentioned, these tanks should be covered the vented;
however, if these tanks are uncovered, then the following procedure can be used to estimate air emissions:
B-9
 Step 1: Calculate the liquid phase mass transfer coefficient for turbulent surfaces (kl,t) with
Equation B-4.
 Step 2: Calculate the gas phase mass transfer coefficient for turbulent surfaces (kg,t) with Equation
B-5.
 Step 3: Calculate the liquid phase mass transfer coefficient for quiescent surfaces (kl,q) as
described in Section B.1.3, Liquid Phase Mass Transfer Coefficient for Quiescent Surfaces.
 Step 5: Calculate the overall mass transfer coefficient for turbulent surfaces (KOL,t) with Equation
B-2.
 Step 6: Calculate the overall mass transfer coefficient for quiescent surfaces (KOL,q) with Equation
B-3.
 Step 7: Calculate the overall mass transfer coefficient (KOL) with Equation B-1.
 Step 8: Calculate the tank effluent concentration (CL) with the following equation:
QC 0
CL (g / m3 ) = (Eq. B-30)
(K OL A + Q )
 Step 9: Calculate the air emissions (N) with the following equation:
N ( g / s ) = K OL C L A (Eq. B-31)
B.2.7 Biological Treatment Unit (with a Known Wasting Rate)

The determination of air emissions from biological treatment units is similar to equalization tanks with
the addition of a biodegradation factor. Additionally, whereas turbulent and quiescent surface areas
impact air emissions for mechanically aerated systems, the air volumetric flow rate impacts diffused
aeration systems. Consequently, two slightly different methods are used to estimate air emissions.
For mechanically aerated systems, follow Steps 1 through 7 of Section B.2.5 to determine the overall
mass transfer coefficient (KOL).
For diffused aeration systems, first calculate the average diffused air flow rate Qa,avg (m3/s):
 Qa P0 
 + Qa 
 P0 + Pg 
Qa ,avg = (Eq. B-32)
2
where:
Qa = Volumetric flow rate of diffused air (m3/s), at conditions of 1 atmosphere and bulk
liquid temperature
Pg = Blower gauge pressure (lb/in2, gauge)
P0 = Atmospheric pressure (atm) × 14.696 (lb/in2atm)=(lb/in2)
Monteith et al. (1996) reported the effect of turbulence, induced by diffused bubble aeration, on mass
transfer at the water surface. Correlations between the mass transfer of ammonia and VOCs were
established and used to determine KL[NH4]a, kl,ta and kg,ta with the following equations:
 3600 × Qa ,avg
0.4

K L[ NH 4 ] a = 0.08 + 0.012  (Eq. B-33)
 At 
B-10
where:
KL[NH4]a = Ammonium volumetric mass transfer coefficient (hr-1)
3600 = Conversion factor (sec/hr)
At = Turbulent surface area (m2)
0.5
 Di ,l 
k l ,t a = K L[ NH 4 ] a ×  

(Eq. B-34)
 Dl , NH 4 
where:
Dl,NH4 = Diffusivity of Ammonium in water (6.93×10-5 cm2/s)
kl,ta = Volumetric liquid phase mass transfer coefficient (hr-1)
The kl,t values must then be calculated from the kl,ta value with the following equation:
k l ,t a
k l ,t = (Eq. B-35)
3600 × a
where:
kl,t = Liquid phase mass transfer coefficient (m/s)
a = Specific interfacial area based on liquid volume (m-1) = A/V
V = Aeration basin volume (m3)
Hsieh et al. (1992) reported the ratio of kg/kl ranged from 2.2 to 3.6 for diffused aeration systems.
Therefore, kg can be approximated by the following equation:
k g ,t = 3k l ,t (Eq. B-36)
where:
kg,t = Gas phase mass transfer coefficient (m/s)
The KOL,t, KOL,q, and KOL are then calculated by using steps by using equations B-2, B-3, and B-1.
Next, calculate the tank effluent concentration (CL) and the emission rate (N) with the following
equations.
For mechanically aerated systems:

N ( g / s ) = K OL C L A (Eq. B-37)
For diffused aeration systems:
N ( g / s ) = (K OL A + Qa H ′)C L (Eq. B-38)
where:
C (g / m ) =
L
[− b + (b
3
2
− 4ac )
0.5
] (Eq. B-39)
2a
B-11

K OL A + Qw k abs CWAS ,VSS + Qe
a= (Eq. B-40)
Qi
b=
Ks
(K OL A + Qw k abs CWAS ,VSS + Qe ) + Vmax C MLVSSV − C0 (Eq. B-41)
Qi Qi
For diffused aerated systems:
K OL A + Qa ,avg H ′ + Qw k abs CWAS ,VSS + Qe

a= (Eq. B-42)
Qi
b=
Ks
(K OL A + Qa,avg H ′ + Qw k abs CWAS ,VSS + Qe ) + Vmax C MLVSSV − C0 (Eq. B-43)
Qi Qi
c = − K S C0 (Eq. B-44)
where:
Qw = Wasted activated sludge flow rate (m3/s)
Qe = Effluent flow rate (m3/s)
Qi = Influent flow rate (m3/s)
KS = Half saturation biorate constant (g/m3)
Vmax = Maximum biorate constant (g/s-g biomass)
CMLVSS = Mixed-liquor volatile suspended solids (g/m3)
CWAS,VSS = Wasted sludge volatile suspended solids (g/m3)
kabs = Partition coefficient for organic contaminant (m3/g)
where:
k abs = 1 × 10 −6 f OC K OC (Eq. B-45)
fOC = Weight fraction of carbon in biomass (g C/g) (default value 0.33)
KOC = Organic carbon partition coefficient (g/MgC/(g/ m-3)
1×10-6 = unit conversion (megagrams per gram, Mg/g)
where:
log K OC = log K OW − 0.32 (Eq. B-46)
B.2.8 Biological Treatment Unit (with an Unknown Wasting Rate)

Air emission estimate calculations from aerated lagoons are very similar to those for activated sludge
units. The difference is sludge is not wasted from the lagoon, rather it settles and accumulates on the
lagoon bottom until dredging. The assumptions made in this document are: (1) at steady-state conditions
the amount of sludge that settles and thus removed from the unit is equal to the biomass production rate,
and (2) once the sludge and associated contaminants settle, they are considered removed from the system.
Follow the procedure in Section B.2.6 to determine the overall mass transfer coefficient (KOL), the unit
effluent concentration (CL), and the emission rate (N). However, variables a, b, and c are determined by
the following equations
B-12

K OL A + 0.67C BODi Qi k abs + Qe
a= (Eq. B-47)
Qi
b=
Ks
(K OL A + 0.67C BODi Qi k abs + Qe ) + Vmax C MLVSSV − C 0 (Eq. B-48)
Qi Qi
For diffused aerated system:
K OL A + Qa ,avg H ′ + 0.67C BODi Qi k abs

a= (Eq. B-49)
Qi
b=
Ks
(K OL A + Qa,avg H ′ + 0.67C BODi Qi k abs ) + Vmax C MLVSSV − C0 (Eq. B-50)
Qi Qi
c = − K S C0 (Eq. B-51)
B.3 References
Hsieh, C., R. Babcock, and M. Stenstrom. 1992. Estimating semi-volatile organic compound emission
rates and oxygen transfer coefficients in diffused aeration. Presented at the 65th Water
Environment Federation Conference, New Orleans, LA, September 20–24.
Monteith, H., W. Parker, and H. Melcer. 1996. Impact of bubble-induced surface turbulence on gas-liquid
mass transfer in diffused aeration systems. Presented at the Water Environment Federation’s
Annual Technical Exhibition and Conference, Dallas, TX, October 5–9.
Nakasone, H. 1987. Study of aeration at weirs and cascades. J. Environ. Eng.113:64.
Pincince, A.B. 1991. Transfer of oxygen and emissions of volatile organic compounds at clarifier weirs.
Res. J. Water Pollut. Control Fed., 63, 114-119.
U.S. EPA (Environmental Protection Agency). 1990. Industrial Wastewater Volatile Organic Compound
Emissions—Background Information for BACT/LAER Determinations. EPA 450/3-90-004. Office
of Air Quality Planning and Standards, Research Triangle Park, NC. January.
U.S. EPA (Environmental Protection Agency). 1994. Air Emissions Models for Waste and Wastewater.
EPA-453/R-94-080A. Office of Air Quality Planning and Standards, Research Triangle Park, NC.
U.S. EPA (Environmental Protection Agency). 1995. Compilation of Air Pollutant Emission Factors.
Volume 1: Stationary Point and Area Sources. AP-42, Fifth Edition. Office of Air Quality
Planning and Standards, Research Triangle Park, NC.
B-13
B-14
Version 3 Appendix C—Primer for TANKS Model Users
Appendix C
Primer for TANKS Model Users
C-1
C-2
C. Primer for TANKS Model Users

In general, EPA recommends that the emission estimation procedures detailed in Chapter 7.1 of AP-42
(U.S. EPA, 1995a) be used to calculate air pollutant emissions from organic liquid storage vessels. There
are many tools available, such as TANKS v4.09D emission estimation software that can be used to
perform the necessary calculations detailed in Chapter 7.1 of AP-42 (U.S. EPA, 1995a). TANKS v4.09D
software can be downloaded for free at http://www.epa.gov/ttn/chief/software/tanks/index.html (under the
How to Get TANKS 4.09D link). Because TANKS v4.09D is widely used, this appendix is included to
provide tips and insights on using the TANKS program. This appendix will also highlight some potential
enhancements to the modeling of storage tank emissions that may be implemented when using the AP-42
equations directly (or potentially other software packages) that are not currently afforded by TANKS
v4.09D.
The calculations in the TANKS program are specific to the type of tank, and the results will be highly
dependent on the type of liquid stored. Each individual storage tank should be modeled separately. There
may be instances in which a set of tanks have identical properties and materials and very similar
throughputs so that the emissions from a single storage tank can be modeled and used as the emissions for
each tank in the set. However, this type of “model tank” analysis should be used and generally be limited
to situations in which limited data are available (i.e., if throughput is only measured by product and not by
individual tank, then modeling a group of identical tanks used to store that product may be necessary).
For inventories developed by the petroleum refinery or storage tank operators, it is recommended that
individual tank data be entered into the TANKS program (or other similar software), and the emission
results for individual tanks be reported in the emission inventory. For some uses of the emission inventory
data, specific emission location and source characteristics data are needed. These data become lost or less
accurate when emissions from a group of tanks (e.g., those that would be assigned to the same source
classification code [SCC]) are aggregated when reporting to the emission inventory.
To enter a new tank, first select the tank type from the drop-down menu labeled “Create a New Tank
Record.” There are five basic tank types: horizontal fixed roof tanks; vertical fixed roof tanks; external
floating roof tanks; internal floating roof tanks; and domed external floating roof tanks. Next, fill out the
information in the Identification, Physical Characteristics, Site Selection, Tank Contents, and Monthly
Calculations tabs with the appropriate information. Tables C-1 and C-2 show the required input by tank
orientation. These inputs may either be from a drop-down list based on the TANKS database for location,
tank color, and other tank-specific information or taken directly from process data. In TANKS, selecting
the city will set the meteorological data that will be used in the calculations. After entering all of the
requested data, the tank information can be saved by pressing the Save button at the bottom of the gray
window.
C-3
Table C-1. TANKS Inputs by Tank Configuration-Tank Dimensions and Characteristics
Domed
Vertical Internal External External
Horizontal Fixed-Roof Floating Floating Floating
Tank Tank Roof Roof Roof
Identification number     
Description     
State     
City     
Company     
Dimensions and Characteristics
Shell height (feet [ft])  
Diameter (shell diameter, ft)   
Volume  
Maximum liquid height  
Average liquid height  
Working volume (gallons)  
Turnovers per year    
Net throughput (gallons per year
  
[gal/yr])
Is tank heated?  
Is tank underground? 
Self-supporting roof 
Number of columns 
Effective column diameter 
Roof type  
Roof fitting category  
Tank construction  
C-4
Table C-2. TANKS Inputs by Tank Configuration-Specific Attributes, Site Selection, and Contents
Domed
Vertical Internal External External
Horizontal Fixed-Roof Floating Floating Floating
Tank Tank Roof Roof Roof
Tank/Shell/Roof Characteristics
(External) shell/paint color/shade     
(External) shell/paint shell condition     
Internal shell condition 
Rim Seal
Primary seal   
Secondary seal   
Deck Characteristics
Deck type 
Deck fitting category 
Breather Vent Settings
Vacuum setting (psig)  
Pressure setting (psig)  
Roof Characteristics
Color/shade  
Condition 
Type 
Height (ft) 
Roof paint condition 
Site Selection
Nearest major city     
Daily average ambient temperature     
Annual average maximum temperature     
Annual average minimum temperature     
Average wind speed     
Annual average solar insulation factor
    
(Btu/(ft × ft ×day)
Atmospheric pressure (pounds of force
    
per square inch absolute [psia])
Tank Contents
Chemical category of liquid     
Single or multi-component liquid     
Mixture properties, if applicable     
Throughput by month     
C-5
For the internal floating roof tank, the external floating roof tank, and the domed external floating roof
tank, it is possible to define the fitting type to represent specific characteristics of the tanks at your
facility. Using facility-specific tank information will ensure that the calculated emission estimates are as
accurate as possible. The fittings information can be entered under the Physical Characteristics tab by
selecting the View/Add Fittings button and providing the fitting type and status. Table C-3 contains a list
of the different fitting types and status options. Using the information provided in Tables C-1 through
C-3, the tanks program estimates the anticipated emissions specifically for the particular location, tank
type, and process characteristics.
Once a tank is entered, you can use or import that tank as a starting point for adding new tanks. To copy
data from an existing tank to a new tank, select DATA, then TANKS, and then New Record from the
TANKS 4.0 main menu or by selecting the appropriate acronym from the toolbar on the main screen.
Click on the Copy button at the bottom of the tank pop-up screen. Then click on the tank that you want to
copy from a dropdown menu, and then select whether you want to copy just the Tank Data, just the
Liquid Data, or both (“All Data”). Simply opening an existing tank and changing the name does not
create a new tank record, it only renames and replaces the old record.
Because TANKS uses preloaded meteorological data, year-to-year variations in emissions will only result
from differing throughput or tank service (storing different material than in previous years). If
accommodated in other software packages, meteorological data that are specific to the reporting year
should be used because this information will provide more accurate emission values. Studies have
indicated that emissions may vary by up to 25% from year to year depending on the meteorological
conditions for that year (Coburn and Icenhour, 2008). Also, if accommodated in other software packages,
directly measured liquid temperatures or monthly average ambient temperatures should be used for the
bulk liquid temperatures, and these values should be used to calculate monthly average emission rates. In
TANKS, even when using the monthly calculations option, unfortunately the annual average liquid
temperature is still used. As such, unless there are extreme differences in throughputs by month, the
monthly calculation procedures used in TANKS are unlikely to significantly alter the annual emission
estimates. Based on limited model analyses, monthly variations in liquid temperature may be significant
and may influence the annual average emissions if monthly emissions are estimated more rigorously (i.e.,
by accounting for variations in bulk liquid temperature) (Coburn and Icenhour, 2008).
The TANKS program typically provides total hydrocarbon emission estimates. TANKS is also designed
to calculate the individual component emissions from known mixtures and to estimate emissions from
typical refined petroleum products. More than 100 organic liquids are included in the TANKS chemical
database, but it is also designed to analyze other substances based on their chemical properties and the
parameters shown in Table C-4. If you store a component or mixture in the storage tanks and have
available property data (i.e. chemical database inputs), then these can be added to the TANKS program.
In addition to the analysis of individual chemicals, the TANKS program also has the capability to
calculate the characteristics of defined mixtures. When the “Multiple” option is selected for the “Single-
or Multi-Component Liquid” option, and “Full Speciation” is selected for the “Speciation” option, then
the full chemical profile can be created using the “View/Add Components” button. Any defined chemical,
including those defined using the criteria in Table C-4, can be included in the mixture as a component as
long as the percentage of total liquid weight or relative weight is known. With the chemical components
and their relative percentages and/or weights entered, the menu can be closed, and the mixture properties
can be calculated by clicking on the Calculate Mixture Properties button.
C-6
Table C-3. Fitting Type and Status Option Inputs for TANKS Program
Fitting Type Status

Bolted cover, gasketed
Access hatch (24 inches in diameter) Unbolted cover, gasketed
Unbolted cover, ungasketed
Bolted cover, gasketed
Automatic gauge float well Unbolted cover, gasketed
Unbolted cover, ungasketed
Built-up column, sliding cover, gasketed
Built-up column, sliding cover, ungasketed
Column well (24 inches in diameter) Pipe column, flexible fabric sleeve seal
Pipe column, sliding cover, gasketed
Pipe column, sliding cover, ungasketed
Weighted mechanical actuation, gasketed
Gauge-hatch/sample well (8 inches in diameter)
Weighted mechanical actuation, ungasketed
Sliding cover, gasketed
Ladder well (36 inches in diameter)
Sliding cover, ungasketed
Rim vent (6 inches In diameter)
Open
Roof drain (3 inches in diameter)
90% closed
Adjustable, pontoon area, ungasketed
Adjustable, center area, ungasketed
Adjustable, double-deck roofs
Fixed
Roof leg (3 inches In diameter)
Adjustable, pontoon area, gasketed
Adjustable, pontoon area, sock
Adjustable, center area, gasketed
Adjustable, center area, sock
Adjustable
Roof leg or hanger well
Fixed
Slotted pipe, sliding cover, gasketed
Sample pipe or well (24 inches In diameter) Slotted pipe, sliding cover, ungasketed
Slit fabric seal 10% open
Sliding cover, without float, ungasketed
Sliding cover, with float, ungasketed
Sliding cover, without float, gasketed
Sliding cover, with float, gasketed
Slotted guide-pole/sample well Sliding cover, with pole wiper, gasketed
Sliding cover, with pole sleeve, gasketed
Sliding cover, with float, wiper, gasketed
Sliding cover, with float, sleeve, wiper, gasketed
Sliding cover, with pole sleeve, wiper, gasketed
Stub drain (1 inch in diameter) None listed
Sliding cover, ungasketed
Sliding cover, gasketed
Unslotted guide-pole well Sliding cover, with sleeve ungasketed
Sliding cover, with sleeve, gasketed
Sliding cover, with wiper, gasketed
Vacuum breaker (10 inches in diameter)
C-7
Table C-4. Chemical Database Inputs for TANKS

Chemical name
CAS number (Chemical Abstracts Service number)
Category
Liquid molecular weight
Liquid density (lb/gal; ~60°F)
Vapor molecular weight
Vapor pressure at 40°F, 50°F, 60°F, 70°F, 80°F, 90°F, 100°F
Constants for Antoine's equation using °C (A, B, and C)
Constants for Antoine's equation using °K (A and B)
Reid vapor pressure (psia): Distillates and crude oil
ASTM slope: distillates only
C-8
Version 3 Appendix D—Emission Factors for Combustion Sources
Appendix D
Emission Factors from Air Toxic Emission Factors for
Combustion Sources Using Petroleum Based Fuels –
Final Report, Volume 2 – Development of Emission
Factors Using CARB Approach
D-1
D-2
D. Emission Factors from Air Toxic Emission Factors for Combustion

Sources Using Petroleum Based Fuels – Final Report, Volume 2 –
Development of Emission Factors Using CARB Approach
 Detect Ratio: Ratio of detected values to the sum of detected and nondetected values.
 RSD: 100 times the standard deviation divided by the arithmetic average.
 Uncertainty: 100 times the 95% confidence interval divided by the arithmetic average.
Dioxin Acronyms
4D 2378 2,3,7,8-Tetrachlorodibenzo-p-dioxin.
4D Other Tetrachlorodibenzo-p-dioxin other.
4D Total Tetrachlorodibenzo-p-dioxin total.
5D 12378 1,2,3,7,8-Pentachlorodibenzo-p-dioxin.
5D Other Pentachlorodibenzo-p-dioxin other.
5D Total Pentachlorodibenzo-p-dioxin total.
6D 123478 1,2,3,4,7,8-Hexachlorodibenzo-p-dioxin.
6D Other Hexachlorodibenzo-p-dioxin other.
6D Total Hexachlorodibenzo-p-dioxin.
7D 1234678 1,2,3,4,7,8-Heptachlorodibenzo-p-dioxin.
7D Other Heptachlorodibenzo-p-dioxin other.
7D Total Heptachlorodibenzo-p-dioxin total.
8D Octachlorodibenzo-p-dioxin.
Furan Acronyms
4F 2378 2,3,7,8-Tetrachlorodibenzofuran.
4F Other Tetrachlorodibenzofuran other.
4F Total Tetrachlorodibenzofuran total.
5F 12378 1,2,3,7,8-Pentachlorodibenzofuran.
5F 23478 2,3,4,7,8-Pentachlorodibenzofuran.
5F Other Pentachlorodibenzofuran other.
5F Total Pentachlorodibenzofuran total.
6F 123478 1,2,3,4,7,8-Hexachlorodibenzofuran.
6F Other Hexachlorodibenzofuran other.
6F Total Hexachlorodibenzofuran total.
7F 1234678 1,2,3,4,6,7,8-Heptachlorodibenzofuran.
7F 1234789 1,2,3,4,7,8,9-Heptachlorodibenzofuran.
7F Other Heptachlorodibenzofuran other.
7F Total Heptachlorodibenzofuran total.
8F Octachlorodibenzofuran.
D-3
Table D-1a. Summary of Data for Emission Factor Development –

Asphalt Blowing with Blow Cycle and a Thermal Oxidizer
Emission Factor (lb/MMcf) Uncer-

CARB tainty, Detect
Substance Rating Mean Median Maximum Minimum Tests RSD, % % Ratio
Acetaldehyde C3-v- 1.78E-03 1.78E-03 1.78E-03 1.78E-03 1 0.00 0.00 1.00
Arsenic D3-v- 1.31E-02 1.31E-02 1.31E-02 1.31E-02 1 0.00 0.00 0.00
Benzene E3-v- 3.16E-01 3.16E-01 3.16E-01 3.16E-01 1 0.00 0.00 0.00
Beryllium D3-v- 2.63E-03 2.63E-03 2.63E-03 2.63E-03 1 0.00 0.00 0.00
Cadmium D3-v- 5.25E-03 5.25E-03 5.25E-03 5.25E-03 1 0.00 0.00 0.00
Chromium (Hex) C3-v- 3.17E-03 3.17E-03 3.17E-03 3.17E-03 1 0.00 0.00 1.00
Chromium (Total) C3-v- 4.18E-02 4.18E-02 4.18E-02 4.18E-02 1 0.00 0.00 1.00
Copper D3-v- 4.75E-02 4.75E-02 4.75E-02 4.75E-02 1 0.00 0.00 1.00
Ethylbenzene E3-v- 8.61E-01 8.61E-01 8.61E-01 8.61E-01 1 0.00 0.00 0.00
Formaldehyde C3-v- 3.55E-03 3.55E-03 3.55E-03 3.55E-03 1 0.00 0.00 1.00
HCl C3-v- 2.21E-03 2.21E-03 2.21E-03 2.21E-03 1 0.00 0.00 1.00
Hydrogen Sulfide A3-v- 2.07E+00 2.07E+00 2.07E+00 2.07E+00 1 0.00 0.00 0.00
Lead D3-v- 5.25E-02 5.25E-02 5.25E-02 5.25E-02 1 0.00 0.00 0.00
Manganese D3-v- 1.23E-01 1.23E-01 1.23E-01 1.23E-01 1 0.00 0.00 1.00
Mercury A3-v- 9.07E-03 9.07E-03 9.07E-03 9.07E-03 1 0.00 0.00 1.00
Nickel D3-v- 6.65E-02 6.65E-02 6.65E-02 6.65E-02 1 0.00 0.00 0.00
Phenol C3-v- 7.57E-02 7.57E-02 7.57E-02 7.57E-02 1 0.00 0.00 1.00
Selenium D3-v- 1.31E-02 1.31E-02 1.31E-02 1.31E-02 1 0.00 0.00 0.00
Xylene (Total) E3-v- 8.61E-01 8.61E-01 8.61E-01 8.61E-01 1 0.00 0.00 0.00
Zinc D3-v- 8.41E-01 8.41E-01 8.41E-01 8.41E-01 1 0.00 0.00 1.00
D-4
Table D-1b. Summary of Data for Emission Factor Development –

Asphalt Blowing with Blow Cycle and a Thermal Oxidizer
Emission Factor (lb/MMBtu) Uncer-

CARB tainty, Detect
Arsenic D3-v- 1.23E-05 1.23E-05 1.23E-05 1.23E-05 1 0.00 0.00 0.00
Benzene E3-v- 2.98E-04 2.98E-04 2.98E-04 2.98E-04 1 0.00 0.00 0.00
Cadmium D3-v- 4.94E-06 4.94E-06 4.94E-06 4.94E-06 1 0.00 0.00 0.00
Copper D3-v- 4.47E-05 4.47E-05 4.47E-05 4.47E-05 1 0.00 0.00 1.00
HCl C3-v- 2.08E-06 2.08E-06 2.08E-06 2.08E-06 1 0.00 0.00 1.00
Hydrogen Sulfide A3-v- 1.95E-03 1.95E-03 1.95E-03 1.95E-03 1 0.00 0.00 0.00
Lead D3-v- 4.94E-05 4.94E-05 4.94E-05 4.94E-05 1 0.00 0.00 0.00
Mercury A3-v- 8.53E-06 8.53E-06 8.53E-06 8.53E-06 1 0.00 0.00 1.00
Nickel D3-v- 6.26E-05 6.26E-05 6.26E-05 6.26E-05 1 0.00 0.00 0.00
Phenol C3-v- 7.12E-05 7.12E-05 7.12E-05 7.12E-05 1 0.00 0.00 1.00
Selenium D3-v- 1.23E-05 1.23E-05 1.23E-05 1.23E-05 1 0.00 0.00 0.00
Zinc D3-v- 7.91E-04 7.91E-04 7.91E-04 7.91E-04 1 0.00 0.00 1.00
D-5

Asphalt Blowing with No Blow Cycle and a Thermal Oxidizer

CARB tainty, Detect
Arsenic D3-v- 1.16E-02 1.16E-02 1.16E-02 1.16E-02 1 0.00 0.00 0.00
Benzene E3-v- 2.80E-01 2.80E-01 2.80E-01 2.80E-01 1 0.00 0.00 0.00
Cadmium D3-v- 4.65E-03 4.65E-03 4.65E-03 4.65E-03 1 0.00 0.00 0.00
Copper D3-v- 3.79E-02 3.79E-02 3.79E-02 3.79E-02 1 0.00 0.00 1.00
HCl C3-v- 8.22E-04 8.22E-04 8.22E-04 8.22E-04 1 0.00 0.00 1.00
Hydrogen Sulfide A3-v- 1.83E+00 1.83E+00 1.83E+00 1.83E+00 1 0.00 0.00 0.00
Lead D3-v- 4.65E-02 4.65E-02 4.65E-02 4.65E-02 1 0.00 0.00 0.00
Mercury A3-v- 8.53E-03 8.53E-03 8.53E-03 8.53E-03 1 0.00 0.00 1.00
Nickel D3-v- 6.01E-02 6.01E-02 6.01E-02 6.01E-02 1 0.00 0.00 0.00
Phenol C3-v- 4.64E-02 4.64E-02 4.64E-02 4.64E-02 1 0.00 0.00 1.00
Selenium D3-v- 1.16E-02 1.16E-02 1.16E-02 1.16E-02 1 0.00 0.00 0.00
Zinc D3-v- 5.35E-01 5.35E-01 5.35E-01 5.35E-01 1 0.00 0.00 1.00
D-6

Asphalt Blowing with No Blow Cycle and a Thermal Oxidizer

CARB tainty, Detect
Arsenic D3-v- 1.09E-05 1.09E-05 1.09E-05 1.09E-05 1 0.00 0.00 0.00
Benzene E3-v- 2.64E-04 2.64E-04 2.64E-04 2.64E-04 1 0.00 0.00 0.00
Cadmium D3-v- 4.37E-06 4.37E-06 4.37E-06 4.37E-06 1 0.00 0.00 0.00
Copper D3-v- 3.56E-05 3.56E-05 3.56E-05 3.56E-05 1 0.00 0.00 1.00
HCl C3-v- 7.74E-07 7.74E-07 7.74E-07 7.74E-07 1 0.00 0.00 1.00
Hydrogen Sulfide A3-v- 1.73E-03 1.73E-03 1.73E-03 1.73E-03 1 0.00 0.00 0.00
Lead D3-v- 4.38E-05 4.38E-05 4.38E-05 4.38E-05 1 0.00 0.00 0.00
Mercury A3-v- 8.03E-06 8.03E-06 8.03E-06 8.03E-06 1 0.00 0.00 1.00
Nickel D3-v- 5.65E-05 5.65E-05 5.65E-05 5.65E-05 1 0.00 0.00 0.00
Phenol C3-v- 4.37E-05 4.37E-05 4.37E-05 4.37E-05 1 0.00 0.00 1.00
Selenium D3-v- 1.09E-05 1.09E-05 1.09E-05 1.09E-05 1 0.00 0.00 0.00
Zinc D3-v- 5.04E-04 5.04E-04 5.04E-04 5.04E-04 1 0.00 0.00 1.00
D-7
Table D-3a. Summary of Data for Emission Factor Development – Boilers Firing Fuel Oil
Emission Factor (lb/Mgal) Uncer-

CARB tainty, Detect
Acenaphthene B3-v0 8.93E-07 8.50E-07 1.56E-06 3.34E-07 2 67.31 53.86 0.87
Acenaphthylene B3-v0 3.17E-07 3.28E-07 3.58E-07 2.34E-07 2 14.29 11.43 0.00
Acetaldehyde A3-v1 1.04E-03 4.98E-04 3.35E-03 2.80E-04 2 116.00 92.82 0.75
Acrolein A3-v0 2.17E-03 1.75E-03 3.35E-03 1.43E-03 1 47.31 53.53 0.00
Anthracene B3-v1 3.22E-07 2.84E-07 7.42E-07 1.90E-08 2 98.99 79.20 0.94
Arsenic B3-v0 1.44E-03 1.44E-03 1.68E-03 1.19E-03 1 17.00 19.24 0.00
Benzene A3-v0 4.63E-03 4.67E-03 4.76E-03 4.47E-03 1 3.26 3.69 0.00
Benzo(a)anthracene B3-v1 2.04E-07 2.21E-07 3.58E-07 2.98E-08 2 78.58 62.87 0.00
Benzo(a)pyrene B3-v0 2.09E-07 2.24E-07 3.58E-07 4.61E-08 2 73.55 58.55 0.00
Benzo(b)fluoranthene B3-v0 1.21E-06 1.02E-06 2.36E-06 3.34E-07 2 73.99 59.20 0.40
Benzo(e)pyrene C3-v0 8.53E-07 1.03E-06 1.12E-06 4.11E-07 1 45.15 51.09 0.00
Benzo(g,h,i)perylene B3-v0 9.99E-07 8.47E-07 2.20E-06 3.34E-07 2 77.58 62.08 0.00
Benzo(k)fluoranthene B3-v1 2.08E-07 2.37E-07 3.58E-07 2.98E-08 2 76.76 61.42 0.00
Beryllium B3-v0 6.29E-05 6.35E-05 7.03E-05 5.50E-05 1 12.22 13.83 0.00
1,3-Butadiene B3-v1 6.17E-03 5.97E-03 1.18E-02 8.95E-04 2 93.74 75.01 0.00
Cadmium B3-v0 8.56E-04 6.58E-04 1.26E-03 6.47E-04 1 41.15 46.57 0.00
Chloroform A3-v0 4.69E-03 5.00E-03 5.10E-03 4.78E-03 1 3.26 3.69 0.00
2-Chloronaphthalene C3-v0 2.27E-08 2.33E-08 2.86E-08 1.63E-08 1 27.13 30.70 0.00
Chromium (Hex) A3-v0 1.15E-03 1.15E-03 1.21E-03 1.09E-03 1 4.95 5.60 0.00
Chromium (Total) A3-v0 4.83E-03 5.18E-03 5.85E-03 3.45E-03 1 25.70 29.08 1.00
Chrysene B3-v0 3.88E-06 3.26E-06 8.57E-06 1.16E-06 2 78.87 63.11 1.00
Copper B3-v0 3.86E-03 2.02E-03 8.67E-03 8.92E-04 1 108.86 123.19 1.00
Dibenz(a,h)anthracene B3-v1 7.06E-07 3.43E-07 2.68E-06 2.07E-07 2 136.88 109.33 0.00
Dioxin: 4D 2378 C3-v0 6.33E-10 5.71E-10 7.83E-10 5.46E-10 1 20.50 23.20 0.00
Dioxin: 5D 12378 C3-v0 3.68E-10 2.73E-10 5.71E-10 2.61E-10 1 47.67 53.94 0.00
Dioxin: 6D 123478 C3-v0 3.68E-10 2.73E-10 5.71E-10 2.61E-10 1 47.67 53.94 0.00
Dioxin: 6D 123678 C3-v0 3.68E-10 2.73E-10 5.71E-10 2.61E-10 1 47.67 53.94 0.25
Dioxin: 6D 123789 C3-v0 3.68E-10 2.73E-10 5.71E-10 2.61E-10 1 47.67 53.94 0.25
Dioxin: 7D 1234678 C3-v0 3.12E-09 2.86E-09 5.19E-09 1.30E-09 1 62.77 71.03 1.00
Dioxin: 8D C3-v0 7.50E-08 8.28E-08 1.23E-07 1.93E-08 1 69.66 78.83 1.00
Fluoranthene B3-v1 7.78E-06 6.18E-06 1.65E-05 9.31E-07 2 87.83 70.27 1.00
Fluorene B3-v0 4.65E-06 4.60E-06 8.48E-06 8.93E-07 2 83.31 66.66 1.00
Formaldehyde A3-v1 6.72E-03 7.14E-03 1.67E-02 2.83E-04 2 91.09 72.89 0.18
Furan: 4F 2378 C3-v0 8.16E-10 7.83E-10 1.09E-09 5.71E-10 1 32.18 36.42 1.00
Furan: 5F 12378 C3-v0 4.59E-10 5.46E-10 5.71E-10 2.61E-10 1 37.52 42.46 0.00
Furan: 5F 23478 C3-v0 4.59E-10 5.46E-10 5.71E-10 2.61E-10 1 37.52 42.46 0.00
Furan: 6F 123478 C3-v0 3.64E-10 2.86E-10 5.46E-10 2.61E-10 1 43.44 49.15 0.50
Furan: 6F 123678 C3-v0 2.73E-10 2.73E-10 2.86E-10 2.61E-10 1 4.51 5.10 0.33
(continued)
D-8
Table D-3a. Summary of Data for Emission Factor Development – Boilers Firing Fuel Oil
(continued)

CARB tainty, Detect
Furan: 6F 123789 C3-v0 3.68E-10 2.73E-10 5.71E-10 2.61E-10 1 47.67 53.94 0.00
Furan: 6F 234678 C3-v0 5.51E-10 5.71E-10 8.20E-10 2.61E-10 1 50.85 57.54 0.50
Furan: 7F 1234678 C3-v0 1.44E-09 1.14E-09 2.09E-09 1.09E-09 1 38.89 44.01 1.00
Furan: 7F 1234789 C3-v0 4.64E-10 2.73E-10 8.57E-10 2.61E-10 1 73.43 83.10 0.00
Furan: 8F C3-v0 7.15E-09 7.14E-09 1.04E-08 3.91E-09 1 45.27 51.23 1.00
Indeno(1,2,3-cd) pyrene B3-v0 4.59E-07 3.55E-07 7.82E-07 2.90E-07 2 43.75 35.01 0.00
Lead B3-v0 1.56E-03 8.59E-04 3.32E-03 4.89E-04 1 98.91 111.92 0.00
Manganese B3-v0 5.81E-03 2.82E-03 1.22E-02 2.40E-03 1 95.34 107.89 1.00
Mercury B3-v0 1.03E-05 9.04E-06 1.47E-05 7.19E-06 1 38.06 43.07 0.00
2-Methylnaphthalene C3-v0 1.09E-05 1.07E-05 1.22E-05 9.76E-06 1 11.52 13.04 1.00
Naphthalene B3-v0 6.07E-05 6.22E-05 8.12E-05 4.25E-05 2 25.60 20.48 1.00
Nickel B3-v0 3.34E-01 3.05E-01 4.06E-01 2.90E-01 1 18.95 21.44 1.00
Perylene C3-v0 1.09E-07 5.95E-08 2.20E-07 4.88E-08 1 87.56 99.09 0.00
Phenanthrene B3-v0 1.10E-05 1.10E-05 1.99E-05 3.29E-06 2 57.86 46.30 1.00
Propylene A3-v0 2.19E-02 2.21E-02 2.25E-02 2.11E-02 1 3.26 3.69 0.00
Pyrene B3-v1 4.01E-06 2.35E-06 1.08E-05 3.53E-07 2 110.05 88.06 0.97
Selenium B3-v0 2.88E-03 3.59E-03 3.61E-03 1.45E-03 1 43.13 48.81 0.17
Toluene A3-v0 5.75E-03 5.79E-03 5.91E-03 5.55E-03 1 3.26 3.69 0.00
Xylene (Total) A3-v0 1.10E-02 1.11E-02 1.14E-02 1.06E-02 1 3.26 3.69 0.00
Zinc B3-v0 1.60E-02 8.61E-03 3.09E-02 8.56E-03 1 80.41 90.99 1.00
D-9
Table D-3b. Summary of Data for Emission Factor Development – Boilers Firing Fuel Oil

CARB tainty, Detect
Acenaphthene B3-v0 6.00E-09 5.71E-09 1.06E-08 2.14E-09 2 69.26 55.42 0.87
Acenaphthylene B3-v0 2.09E-09 2.16E-09 2.29E-09 1.59E-09 2 12.32 9.86 0.00
Acrolein A3-v0 1.48E-05 1.19E-05 2.28E-05 9.69E-06 1 47.31 53.53 0.00
Anthracene B3-v1 2.07E-09 1.83E-09 4.76E-09 1.29E-10 2 98.24 78.60 0.94
Arsenic B3-v0 9.76E-06 9.76E-06 1.14E-05 8.10E-06 1 17.00 19.24 0.00
Benzene A3-v0 3.15E-05 3.17E-05 3.24E-05 3.04E-05 1 3.26 3.69 0.00
Benzo(a)anthracene B3-v1 1.32E-09 1.44E-09 2.29E-09 2.03E-10 2 77.09 61.68 0.00
Benzo(a)pyrene B3-v0 1.36E-09 1.46E-09 2.29E-09 3.13E-10 2 71.84 57.48 0.00
Benzo(b)fluoranthene B3-v0 8.11E-09 6.77E-09 1.60E-08 2.14E-09 2 75.81 60.66 0.40
Benzo(g,h,i)perylene B3-v0 6.72E-09 5.69E-09 1.50E-08 2.14E-09 2 79.36 63.50 0.00
Benzo(k)fluoranthene B3-v1 1.34E-09 1.55E-09 2.29E-09 2.03E-10 2 75.32 60.26 0.00
1,3-Butadiene B3-v1 4.18E-05 4.04E-05 8.00E-05 5.74E-06 2 94.58 75.68 0.00
Cadmium B3-v0 5.82E-06 4.47E-06 8.58E-06 4.40E-06 1 41.15 46.57 0.00
Chrysene B3-v0 2.62E-08 2.19E-08 5.83E-08 7.46E-09 2 80.59 64.48 1.00
Copper B3-v0 2.62E-05 1.37E-05 5.89E-05 6.07E-06 1 108.86 123.19 1.00
Dibenz(a,h)anthracene B3-v1 4.73E-09 2.20E-09 1.82E-08 1.41E-09 2 139.45 111.58 0.00
Dioxin:4D 2378 C3-v0 4.31E-12 3.88E-12 5.32E-12 3.71E-12 1 20.50 23.20 0.00
Dioxin:5D 12378 C3-v0 2.50E-12 1.86E-12 3.88E-12 1.77E-12 1 47.67 53.94 0.00
Dioxin:6D 123478 C3-v0 2.50E-12 1.86E-12 3.88E-12 1.77E-12 1 47.67 53.94 0.00
Dioxin:6D 123678 C3-v0 2.50E-12 1.86E-12 3.88E-12 1.77E-12 1 47.67 53.94 0.25
Dioxin:6D 123789 C3-v0 2.50E-12 1.86E-12 3.88E-12 1.77E-12 1 47.67 53.94 0.25
Dioxin:7D 1234678 C3-v0 2.12E-11 1.94E-11 3.53E-11 8.87E-12 1 62.77 71.03 1.00
Dioxin:8D C3-v0 5.10E-10 5.63E-10 8.36E-10 1.31E-10 1 69.66 78.83 1.00
Fluoranthene B3-v1 5.25E-08 4.14E-08 1.12E-07 5.97E-09 2 89.14 71.33 1.00
Fluorene B3-v1 3.14E-08 3.10E-08 5.76E-08 5.73E-09 2 84.62 67.71 1.00
Furan:4F 2378 C3-v0 5.54E-12 5.32E-12 7.43E-12 3.88E-12 1 32.18 36.42 1.00
Furan:5F 12378 C3-v0 3.12E-12 3.71E-12 3.88E-12 1.77E-12 1 37.52 42.46 0.00
Furan:5F 23478 C3-v0 3.12E-12 3.71E-12 3.88E-12 1.77E-12 1 37.52 42.46 0.00
Furan:6F 123478 C3-v0 2.48E-12 1.94E-12 3.71E-12 1.77E-12 1 43.44 49.15 0.50
Furan:6F 123678 C3-v0 1.86E-12 1.86E-12 1.94E-12 1.77E-12 1 4.51 5.10 0.33
(continued)
D-10
Table D-3b. Summary of Data for Emission Factor Development – Boilers Firing Fuel Oil
(continued)

CARB tainty, Detect
Furan:6F 123789 C3-v0 2.50E-12 1.86E-12 3.88E-12 1.77E-12 1 47.67 53.94 0.00
Furan:6F 234678 C3-v0 3.74E-12 3.88E-12 5.57E-12 1.77E-12 1 50.85 57.54 0.50
Furan:7F 1234678 C3-v0 9.79E-12 7.76E-12 1.42E-11 7.43E-12 1 38.89 44.01 1.00
Furan:7F 1234789 C3-v0 3.15E-12 1.86E-12 5.82E-12 1.77E-12 1 73.43 83.10 0.00
Furan:8F C3-v0 4.86E-11 4.85E-11 7.06E-11 2.66E-11 1 45.27 51.23 1.00
Indeno(1,2,3-cd)pyrene B3-v0 3.05E-09 2.28E-09 5.32E-09 1.97E-09 2 46.19 36.96 0.00
Lead B3-v0 1.06E-05 5.84E-06 2.26E-05 3.32E-06 1 98.91 111.92 0.00
Mercury B3-v0 7.02E-08 6.15E-08 1.00E-07 4.89E-08 1 38.06 43.07 0.00
Naphthalene B3-v0 4.04E-07 4.11E-07 5.52E-07 2.73E-07 2 28.11 22.49 1.00
Nickel B3-v0 2.27E-03 2.08E-03 2.76E-03 1.97E-03 1 18.95 21.44 1.00
Perylene C3-v0 7.44E-10 4.05E-10 1.49E-09 3.32E-10 1 87.56 99.09 0.00
Phenanthrene B3-v0 7.39E-08 7.35E-08 1.35E-07 2.11E-08 2 60.02 48.02 1.00
Pyrene B3-v1 2.72E-08 1.58E-08 7.37E-08 2.26E-09 2 110.83 88.68 0.97
Selenium B3-v0 1.96E-05 2.44E-05 2.45E-05 9.83E-06 1 43.13 48.81 0.17
Toluene A3-v0 3.91E-05 3.94E-05 4.02E-05 3.77E-05 1 3.26 3.69 0.00
Zinc B3-v0 1.09E-04 5.86E-05 2.10E-04 5.82E-05 1 80.41 90.99 1.00
D-11

Boilers Firing Refinery Fuel Gas

CARB tainty, Detect
Acenaphthene A3-v0 5.88E-06 5.75E-06 6.47E-06 5.44E-06 1 8.97 10.15 0.37
Acenaphthylene A3-v0 2.56E-06 2.59E-06 2.59E-06 2.50E-06 1 2.04 2.31 0.00
Acetaldehyde C1-v3 3.97E-03 2.43E-03 1.21E-02 4.92E-06 5 100.80 51.01 1.00
Anthracene A3-v0 2.28E-05 2.50E-05 3.89E-05 4.40E-06 1 76.23 86.26 1.00
Arsenic D3-v0 7.04E-04 7.65E-04 1.13E-03 2.11E-04 1 66.10 74.80 1.00
Benzene C1-v2 2.06E-01 6.24E-02 1.40E+00 3.39E-03 5 180.67 88.53 0.80
Benzo(a)anthracene A3-v0 1.83E-05 1.79E-05 2.46E-05 1.25E-05 1 33.08 37.44 1.00
Benzo(a)pyrene A3-v0 3.42E-06 3.37E-06 4.40E-06 2.50E-06 1 27.72 31.36 0.76
Benzo(b)fluoranthene A3-v0 6.76E-06 7.24E-06 8.28E-06 4.75E-06 1 26.90 30.44 1.00
Benzo(g,h,i)perylene A3-v0 3.85E-06 3.89E-06 5.17E-06 2.50E-06 1 34.64 39.20 0.78
Benzo(k)fluoranthene A3-v0 2.56E-06 2.59E-06 2.59E-06 2.50E-06 1 2.04 2.31 0.00
Beryllium D3-v0 1.55E-04 1.55E-04 1.56E-04 1.53E-04 1 1.35 1.87 1.00
Cadmium D3-v0 2.38E-03 2.05E-03 3.13E-03 1.98E-03 1 27.01 30.57 1.00
Chromium (Hex) C3-v0 7.70E-03 7.56E-03 1.09E-02 4.67E-03 1 40.17 45.45 0.00
Chromium (Total) C3-v1 1.28E-02 5.43E-03 3.08E-02 2.16E-03 1 122.51 138.63 1.00
Chrysene A3-v0 3.42E-06 2.59E-06 5.17E-06 2.50E-06 1 44.30 50.13 0.50
Copper D3-v0 6.30E-03 6.30E-03 7.71E-03 4.89E-03 1 31.59 43.78 1.00
Dibenz(a,h)anthracene A3-v0 2.56E-06 2.59E-06 2.59E-06 2.50E-06 1 2.04 2.31 0.00
Fluoranthene A3-v0 4.25E-05 3.10E-05 7.24E-05 2.40E-05 1 61.63 69.74 1.00
Fluorene A3-v0 9.78E-06 5.75E-06 1.82E-05 5.44E-06 1 74.23 84.00 1.00
Formaldehyde C1-v1 1.60E-02 1.41E-02 4.14E-02 3.24E-03 5 62.29 31.52 1.00
Hydrogen Sulfide A1-v1 2.74E-01 2.37E-01 7.15E-01 5.94E-02 5 89.81 44.00 0.00
Indeno(1,2,3-cd)pyrene A3-v0 2.56E-06 2.59E-06 2.59E-06 2.50E-06 1 2.04 2.31 0.00
Lead D3-v0 2.42E-03 2.42E-03 2.49E-03 2.36E-03 1 3.87 5.36 1.00
Manganese D3-v0 2.39E-03 2.39E-03 3.14E-03 1.64E-03 1 44.52 61.70 1.00
Mercury D3-v0 3.23E-04 3.18E-04 3.82E-04 2.70E-04 1 17.33 19.60 0.00
Naphthalene A3-v0 2.06E-04 1.94E-04 2.40E-04 1.85E-04 1 14.51 16.42 1.00
Nickel D3-v0 5.59E-03 5.59E-03 7.03E-03 4.16E-03 1 36.33 50.35 1.00
Phenanthrene A3-v0 5.64E-05 4.49E-05 8.80E-05 3.62E-05 1 49.24 55.72 1.00
Phenol C2-v0 2.18E-03 8.64E-04 5.66E-03 7.28E-04 4 98.89 54.25 0.85
Pyrene A3-v0 5.98E-05 5.17E-05 8.28E-05 4.49E-05 1 33.82 38.27 1.00
Selenium D3-v0 2.06E-03 2.36E-03 2.83E-03 9.79E-04 1 46.79 52.94 0.16
Toluene E2-v2 8.40E-01 8.35E-02 5.00E+00 4.14E-02 3 189.87 124.04 0.97
Zinc D3-v2 3.42E+00 3.82E-01 9.78E+00 9.27E-02 1 161.24 182.45 1.00
D-12

Boilers Firing Refinery Fuel Gas

CARB tainty, Detect
Arsenic D3-v0 5.88E-07 6.46E-07 9.40E-07 1.78E-07 1 65.36 73.95 1.00
Benzene C1-v2 1.74E-04 5.03E-05 1.22E-03 2.86E-06 5 186.11 91.19 0.80
Cadmium D3-v0 2.00E-06 1.70E-06 2.64E-06 1.67E-06 1 27.60 31.23 1.00
Chrysene A3-v0 2.86E-09 2.19E-09 4.36E-09 2.02E-09 1 45.70 51.71 0.50
Copper D3-v0 5.32E-06 5.32E-06 6.51E-06 4.13E-06 1 31.59 43.78 1.00
Fluorene A3-v0 8.19E-09 4.65E-09 1.53E-08 4.59E-09 1 75.53 85.46 1.00
Lead D3-v0 2.05E-06 2.05E-06 2.10E-06 1.99E-06 1 3.87 5.36 1.00
Mercury D3-v0 2.72E-07 2.69E-07 3.22E-07 2.24E-07 1 18.19 20.58 0.00
Nickel D3-v0 4.72E-06 4.72E-06 5.94E-06 3.51E-06 1 36.33 50.35 1.00
Phenol C2-v0 1.83E-06 7.04E-07 4.91E-06 5.45E-07 4 99.84 56.49 0.85
Pyrene A3-v0 5.00E-08 4.36E-08 6.99E-08 3.63E-08 1 35.35 40.00 1.00
Selenium D3-v0 1.73E-06 1.99E-06 2.39E-06 8.11E-07 1 47.44 53.68 0.16
Toluene E2-v2 7.23E-04 7.25E-05 4.37E-03 3.59E-05 3 193.00 126.09 0.97
Zinc D3-v2 2.83E-03 3.22E-04 8.10E-03 7.83E-05 1 161.02 182.20 1.00
D-13
Table D-5a. Summary of Data for Emission Factor Development – Boilers Firing Refinery
Fuel Gas Controlled with Selective Catalytic Reduction (NOx)

CARB tainty, Detect
Arsenic D3-v0 7.04E-04 7.65E-04 1.13E-03 2.11E-04 1 66.10 74.80 1.00
Benzene C1-v2 2.06E-01 6.24E-02 1.40E+00 3.39E-03 5 180.67 88.53 0.80
Cadmium D3-v0 2.38E-03 2.05E-03 3.13E-03 1.98E-03 1 27.01 30.57 1.00
Chrysene A3-v0 3.42E-06 2.59E-06 5.17E-06 2.50E-06 1 44.30 50.13 0.50
Copper D3-v0 6.30E-03 6.30E-03 7.71E-03 4.89E-03 1 31.59 43.78 1.00
Fluorene A3-v0 9.78E-06 5.75E-06 1.82E-05 5.44E-06 1 74.23 84.00 1.00
Lead D3-v0 2.42E-03 2.42E-03 2.49E-03 2.36E-03 1 3.87 5.36 1.00
Mercury D3-v0 3.23E-04 3.18E-04 3.82E-04 2.70E-04 1 17.33 19.60 0.00
Nickel D3-v0 5.59E-03 5.59E-03 7.03E-03 4.16E-03 1 36.33 50.35 1.00
Phenol C2-v0 2.18E-03 8.64E-04 5.66E-03 7.28E-04 4 95.89 54.25 0.85
Pyrene A3-v0 5.98E-05 5.17E-05 8.28E-05 4.49E-05 1 33.82 38.27 1.00
Selenium D3-v0 2.06E-03 2.36E-03 2.83E-03 9.79E-04 1 46.79 52.94 0.16
Toluene E2-v2 8.40E-01 8.35E-02 5.00E+00 4.14E-02 3 189.87 124.04 0.97
Zinc D3-v2 3.42E+00 3.82E-01 9.78E+00 9.27E-02 1 161.24 182.45 1.00
D-14
Table D-5b. Summary of Data for Emission Factor Development – Boilers Firing Refinery
Fuel Gas Controlled with Selective Catalytic Reduction (NOx)

CARB tainty, Detect
Arsenic D3-v0 5.88E-07 6.46E-07 9.40E-07 1.78E-07 1 65.36 73.95 1.00
Benzene C1-v2 1.74E-04 5.03E-05 1.22E-03 2.86E-06 5 186.11 91.19 0.80
Cadmium D3-v0 2.00E-06 1.70E-06 2.64E-06 1.67E-06 1 27.60 31.23 1.00
Chrysene A3-v0 2.86E-09 2.19E-09 4.36E-09 2.02E-09 1 45.70 51.71 0.50
Copper D3-v0 5.32E-06 5.32E-06 6.51E-06 4.13E-06 1 31.59 43.78 1.00
Fluorene A3-v0 8.19E-09 4.65E-09 1.53E-08 4.59E-09 1 75.53 85.46 1.00
Lead D3-v0 2.05E-06 2.05E-06 2.10E-06 1.99E-06 1 3.87 5.36 1.00
Mercury D3-v0 2.72E-07 2.69E-07 3.22E-07 2.24E-07 1 18.19 20.58 0.00
Nickel D3-v0 4.72E-06 4.72E-06 5.94E-06 3.51E-06 1 36.33 50.35 1.00
Phenol C2-v0 1.83E-06 7.04E-07 4.91E-06 5.45E-07 4 99.84 56.49 0.85
Pyrene A3-v0 5.00E-08 4.36E-08 6.99E-08 3.63E-08 1 35.35 40.00 1.00
Selenium D3-v0 1.73E-06 1.99E-06 2.39E-06 8.11E-07 1 47.44 53.68 0.16
Toluene E2-v2 7.23E-04 7.25E-05 4.37E-03 3.59E-05 3 193.00 126.09 0.97
Zinc D3-v2 2.83E-03 3.22E-04 8.10E-03 7.83E-05 1 161.02 182.20 1.00
D-15

Coke Calcining Controlled by Spray Dryer and Fabric Filter
Emission Factor (lb/ton coke) Uncer-

CARB tainty, Detect
Acrolein A3-v0 3.36E-04 3.41E-04 3.60E-04 3.08E-04 1 7.79 8.82 0.00
Antimony C3-v0 4.61E-05 4.64E-05 4.79E-05 4.41E-05 1 4.04 4.57 0.34
Arsenic C3-v0 4.66E-06 4.64E-06 4.92E-06 4.41E-06 1 5.42 6.13 0.00
Barium C3-v0 1.95E-05 1.99E-05 2.46E-05 1.39E-05 1 27.51 31.13 1.00
Benzene C3-v0 3.24E-04 4.06E-04 4.50E-04 1.15E-04 1 56.20 63.60 1.00
Beryllium C3-v0 1.93E-06 1.72E-06 2.43E-06 1.63E-06 1 22.77 25.76 0.42
Cadmium C3-v0 9.32E-06 9.27E-06 9.84E-06 8.84E-06 1 5.34 6.04 0.00
Chrysene A3-v0 1.28E-08 1.18E-08 1.84E-08 8.25E-09 1 40.02 45.28 0.79
Copper C3-v0 9.32E-06 9.27E-06 9.84E-06 8.84E-06 1 5.34 6.04 0.00
Dioxin:4D 2378 A3-v0 1.14E-11 1.18E-11 1.32E-11 9.05E-12 1 18.56 21.00 0.00
Dioxin:4D Other A3-v0 1.44E-10 1.01E-10 2.99E-10 3.17E-11 1 96.53 109.23 1.00
Dioxin:5D 12378 A3-v0 9.56E-12 9.20E-12 1.35E-11 6.01E-12 1 39.05 44.19 0.00
Dioxin:6D 123478 A3-v0 9.93E-12 9.24E-12 1.57E-11 4.83E-12 1 55.10 62.35 0.31
Dioxin:6D 123678 A3-v0 1.38E-11 1.27E-11 2.04E-11 8.30E-12 1 44.48 50.33 0.69
Dioxin:6D 123789 A3-v0 1.29E-11 1.45E-11 1.50E-11 9.20E-12 1 24.83 28.10 0.39
Dioxin:7D 1234678 A3-v0 1.38E-10 1.18E-10 2.18E-10 7.96E-11 1 51.62 58.41 1.00
Dioxin:8D A3-v0 1.77E-09 1.79E-09 2.76E-09 7.60E-10 1 56.40 63.82 1.00
Fluorene A3-v0 5.64E-08 5.21E-08 6.61E-08 5.10E-08 1 14.90 16.86 1.00
Furan:4F 2378 A3-v0 1.33E-11 1.19E-11 1.63E-11 1.18E-11 1 19.13 21.65 0.70
Furan:4F Other A3-v0 1.39E-10 1.10E-10 2.15E-10 9.21E-11 1 47.70 53.98 1.00
Furan:5F 12378 A3-v0 1.36E-11 1.43E-11 1.50E-11 1.16E-11 1 13.01 14.72 0.65
(continued)
D-16

Coke Calcining Controlled by Spray Dryer and Fabric Filter (continued)
Emission Factor (lb/ton coke) Uncer-

CARB tainty, Detect
Furan:5F 23478 A3-v0 1.25E-11 1.38E-11 1.46E-11 9.10E-12 1 23.87 27.01 0.63
Furan:6F 123478 A3-v0 2.44E-11 2.49E-11 2.71E-11 2.12E-11 1 12.27 13.89 1.00
Furan:6F 123678 A3-v0 2.21E-11 2.19E-11 3.13E-11 1.31E-11 1 41.02 46.42 1.00
Furan:6F 123789 A3-v0 9.19E-12 8.96E-12 1.20E-11 6.64E-12 1 29.03 32.85 0.67
Furan:6F 234678 A3-v0 2.01E-11 2.04E-11 2.04E-11 1.96E-11 1 2.55 2.88 1.00
Furan:7F 1234678 A3-v0 1.50E-10 1.42E-10 1.76E-10 1.32E-10 1 15.51 17.55 1.00
Furan:7F 1234789 A3-v0 2.56E-11 2.56E-11 2.99E-11 2.12E-11 1 17.09 19.33 0.72
Furan:8F A3-v1 1.46E-10 1.36E-10 2.86E-10 1.61E-11 1 92.60 104.78 1.00
Lead C3-v0 6.20E-05 4.92E-05 9.27E-05 4.41E-05 1 43.08 48.75 0.50
Manganese C3-v0 4.56E-05 4.41E-05 7.63E-05 1.63E-05 1 65.87 74.54 0.88
Mercury C3-v1 4.63E-05 1.63E-05 1.12E-04 1.08E-05 1 122.58 138.71 1.00
Nickel C3-v0 9.06E-05 4.92E-05 1.76E-04 4.64E-05 1 81.90 92.67 0.65
Phosphorus C3-v0 4.66E-04 4.64E-04 4.92E-04 4.41E-04 1 5.42 6.13 0.00
Pyrene A3-v0 2.67E-08 2.64E-08 3.23E-08 2.15E-08 1 20.27 22.94 1.00
Selenium C3-v0 4.66E-06 4.64E-06 4.92E-06 4.41E-06 1 5.42 6.13 0.00
Silver C3-v0 1.63E-05 1.63E-05 1.72E-05 1.54E-05 1 5.45 6.17 0.00
Thallium C3-v0 6.99E-05 6.96E-05 7.38E-05 6.64E-05 1 5.31 6.01 0.00
Toluene C3-v0 5.34E-05 4.46E-05 7.17E-05 4.39E-05 1 29.73 33.64 1.00
Xylene (m,p) C3-v0 3.09E-05 3.44E-05 3.79E-05 2.04E-05 1 29.89 33.82 0.22
Xylene (o) C3-v0 4.49E-05 4.34E-05 4.79E-05 4.34E-05 1 5.72 6.47 0.00
Zinc C3-v0 1.17E-04 1.04E-04 1.63E-04 8.37E-05 1 35.15 39.78 1.00
D-17

Coke Calcining Controlled by Spray Dryer and Fabric Filter

CARB tainty, Detect
Acrolein A3-v0 1.04E-03 1.06E-03 1.14E-03 9.25E-04 1 10.62 12.02 0.00
Antimony C3-v0 1.44E-04 1.42E-04 1.52E-04 1.37E-04 1 5.19 5.87 0.34
Arsenic C3-v0 1.45E-05 1.42E-05 1.56E-05 1.37E-05 1 6.73 7.62 0.00
Barium C3-v0 6.10E-05 6.40E-05 7.79E-05 4.11E-05 1 30.51 34.52 1.00
Benzene C3-v0 1.03E-03 1.39E-03 1.45E-03 2.62E-04 1 64.73 73.25 1.00
Cadmium C3-v0 2.90E-05 2.85E-05 3.12E-05 2.74E-05 1 6.71 7.59 0.00
Chrysene A3-v0 3.73E-08 3.48E-08 4.98E-08 2.72E-08 1 30.84 34.90 0.79
Copper C3-v0 2.90E-05 2.85E-05 3.12E-05 2.74E-05 1 6.71 7.59 0.00
Dioxin:4D 2378 A3-v0 3.68E-11 3.66E-11 4.86E-11 2.51E-11 1 31.88 36.08 0.00
Dioxin:5D 12378 A3-v0 2.92E-11 2.56E-11 3.74E-11 2.47E-11 1 24.24 27.43 0.00
Dioxin:6D 123478 A3-v0 3.45E-11 2.57E-11 6.45E-11 1.34E-11 1 77.21 87.37 0.31
Dioxin:6D 123678 A3-v0 4.40E-11 5.23E-11 5.68E-11 2.30E-11 1 41.59 47.07 0.69
Dioxin:6D 123789 A3-v0 4.22E-11 4.16E-11 5.95E-11 2.56E-11 1 40.22 45.52 0.39
Dioxin:7D 1234678 A3-v0 4.19E-10 3.27E-10 6.05E-10 3.27E-10 1 38.30 43.34 1.00
Dioxin:8D A3-v0 5.25E-09 4.98E-09 7.65E-09 3.12E-09 1 43.40 49.11 1.00
Fluorene A3-v0 1.70E-07 1.50E-07 2.18E-07 1.41E-07 1 24.78 28.04 1.00
Furan:4F 2378 A3-v0 4.23E-11 4.53E-11 4.85E-11 3.31E-11 1 19.19 21.72 0.70
(continued)
D-18

Coke Calcining Controlled by Spray Dryer and Fabric Filter (continued)

CARB tainty, Detect
Furan:5F 12378 A3-v0 4.42E-11 4.16E-11 5.87E-11 3.23E-11 1 30.36 34.35 0.65
Furan:5F 23478 A3-v0 4.09E-11 4.07E-11 5.66E-11 2.53E-11 1 38.39 43.44 0.63
Furan:6F 123478 A3-v0 7.88E-11 7.53E-11 1.02E-10 5.88E-11 1 27.77 31.43 1.00
Furan:6F 123678 A3-v0 7.11E-11 8.68E-11 9.00E-11 3.65E-11 1 42.25 47.81 1.00
Furan:6F 123789 A3-v0 2.95E-11 3.32E-11 3.68E-11 1.84E-11 1 33.01 37.36 0.67
Furan:6F 234678 A3-v0 6.50E-11 5.68E-11 8.39E-11 5.43E-11 1 25.28 28.60 1.00
Furan:7F 1234678 A3-v0 4.75E-10 4.90E-10 5.42E-10 3.94E-10 1 15.81 17.89 1.00
Furan:7F 1234789 A3-v0 8.04E-11 8.31E-11 8.70E-11 7.12E-11 1 10.23 11.58 0.72
Furan:8F A3-v1 4.13E-10 3.77E-10 7.94E-10 6.63E-11 1 88.52 100.16 1.00
Lead C3-v0 1.91E-04 1.56E-04 2.74E-04 1.42E-04 1 37.94 42.93 0.50
Mercury C3-v1 1.48E-04 4.81E-05 3.60E-04 3.43E-05 1 124.89 141.33 1.00
Nickel C3-v0 2.87E-04 1.56E-04 5.68E-04 1.37E-04 1 84.93 96.10 0.65
Pyrene A3-v0 7.86E-08 7.75E-08 8.76E-08 7.09E-08 1 10.68 12.09 1.00
Selenium C3-v0 1.45E-05 1.42E-05 1.56E-05 1.37E-05 1 6.73 7.62 0.00
Silver C3-v0 5.07E-05 4.97E-05 5.45E-05 4.81E-05 1 6.52 7.37 0.00
Thallium C3-v0 2.18E-04 2.14E-04 2.34E-04 2.06E-04 1 6.65 7.53 0.00
Toluene C3-v0 1.63E-04 1.44E-04 2.45E-04 9.97E-05 1 45.82 51.85 1.00
Xylene (m,p) C3-v0 8.90E-05 8.62E-05 1.11E-04 6.99E-05 1 23.14 26.18 0.22
Xylene (o) C3-v0 1.32E-04 1.40E-04 1.48E-04 1.09E-04 1 15.78 17.86 0.00
Zinc C3-v0 3.66E-04 3.08E-04 5.25E-04 2.65E-04 1 38.08 43.10 1.00
D-19
Table D-7a. Summary of Data for Emission Factor Development – Heaters Firing Natural Gas

CARB tainty, Detect
Acrolein A3-v0 4.51E-03 4.56E-03 4.64E-03 4.32E-03 1 3.66 4.14 0.00
Benzene A3-v0 2.34E-03 1.65E-03 3.71E-03 1.65E-03 1 50.72 57.40 0.53
Chrysene A3-v0 1.39E-06 1.18E-06 1.83E-06 1.15E-06 1 27.63 31.27 0.44
Ethylbenzene A3-v0 2.25E-03 2.25E-03 2.25E-03 2.25E-03 1 0.00 0.00 0.00
Fluorene A3-v0 4.59E-06 4.50E-06 5.82E-06 3.46E-06 1 25.79 29.19 1.00
Pyrene A3-v0 5.60E-06 2.84E-06 1.16E-05 2.31E-06 1 93.70 106.03 1.00
Toluene A3-v0 3.23E-02 1.38E-02 7.47E-02 8.37E-03 1 114.08 129.09 1.00
D-20
Table D-7b. Summary of Data for Emission Factor Development – Heaters Firing Natural Gas

CARB tainty, Detect
Acrolein A3-v0 4.40E-06 4.46E-06 4.53E-06 4.22E-06 1 3.66 4.14 0.00
Benzene A3-v0 2.28E-06 1.61E-06 3.62E-06 1.61E-06 1 50.72 57.40 0.53
Chrysene A3-v0 1.36E-09 1.16E-09 1.79E-09 1.13E-09 1 27.63 31.27 0.44
Fluorene A3-v0 4.49E-09 4.39E-09 5.69E-09 3.38E-09 1 25.79 29.19 1.00
Pyrene A3-v0 5.47E-09 2.77E-09 1.14E-08 2.25E-09 1 93.70 106.03 1.00
Toluene A3-v0 3.15E-05 1.35E-05 7.29E-05 8.17E-06 1 114.08 129.09 1.00
D-21

Heaters Firing Natural Gas and Refinery Fuel Gas

CARB tainty, Detect
Acrolein C3-v0 2.29E-03 2.10E-03 2.84E-03 1.93E-03 1 21.23 24.03 0.00
Benzene A3-v0 2.12E-02 1.99E-02 2.44E-02 1.94E-02 1 13.02 14.73 0.00
Chrysene A3-v0 7.91E-07 6.38E-07 1.24E-06 4.91E-07 1 50.32 56.94 0.52
Fluorene A3-v1 6.48E-04 1.50E-04 1.69E-03 1.01E-04 1 139.64 158.01 1.00
Indeno(1,2,3-
A3-v0 4.56E-07 4.04E-07 6.67E-07 2.96E-07 1 41.87 47.38 0.49
cd)pyrene
Phenol A3-v0 1.72E-03 1.82E-03 2.08E-03 1.26E-03 1 24.34 27.54 1.00
Pyrene A3-v0 1.25E-05 7.14E-06 2.62E-05 4.04E-06 1 96.40 109.08 1.00
Toluene A3-v0 2.63E-02 2.46E-02 3.03E-02 2.41E-02 1 13.02 14.73 0.00
D-22

Heaters Firing Natural Gas and Refinery Fuel Gas

CARB tainty, Detect
Acrolein C3-v0 1.08E-06 1.10E-06 1.12E-06 1.03E-06 1 4.51 5.11 0.00
Benzene A3-v0 1.01E-05 1.02E-05 1.06E-05 9.61E-06 1 4.81 5.44 0.00
Chrysene A3-v0 4.27E-10 3.62E-10 6.67E-10 2.52E-10 1 50.33 56.95 0.52
Fluorene A3-v1 3.49E-07 8.52E-08 9.09E-07 5.20E-08 1 139.21 157.53 1.00
Indeno(1,2,3-
A3-v0 2.45E-10 2.08E-10 3.58E-10 1.68E-10 1 41.08 46.49 0.49
cd)pyrene
Phenol A3-v0 9.22E-07 9.35E-07 1.12E-06 7.14E-07 1 21.82 24.69 1.00
Pyrene A3-v0 6.74E-09 4.05E-09 1.41E-08 2.08E-09 1 95.66 108.24 1.00
Toluene A3-v0 1.26E-05 1.26E-05 1.31E-05 1.19E-05 1 4.81 5.44 0.00
D-23
Table D-9a. Summary of Data for Emission Factor Development – Heaters Firing Fuel Oil

CARB tainty, Detect
Acenaphthene C3-v1 1.75E-06 2.20E-06 2.99E-06 6.81E-08 1 86.27 97.62 1.00
Acenaphthylene C3-v0 8.02E-08 5.33E-08 1.37E-07 5.03E-08 1 61.23 69.29 1.00
Acrolein C3-v0 5.98E-04 5.99E-04 6.03E-04 5.91E-04 1 0.95 1.07 0.00
Anthracene C3-v0 6.62E-08 6.81E-08 7.41E-08 5.65E-08 1 13.48 15.25 1.00
Arsenic D3-v0 8.34E-04 8.28E-04 8.62E-04 8.13E-04 1 3.01 3.41 0.00
Benzene A3-v0 8.47E-03 8.49E-03 8.74E-03 8.19E-03 1 3.22 3.65 0.00
Benzo(a)anthracene C3-v1 3.84E-06 2.05E-07 1.12E-05 1.33E-07 1 165.56 187.35 1.00
Benzo(a)pyrene C3-v0 9.80E-08 8.63E-08 1.84E-07 2.40E-08 1 82.16 92.97 1.00
Benzo(b)fluoranthene C3-v0 7.95E-07 6.37E-07 1.15E-06 5.98E-07 1 38.74 43.84 1.00
Benzo(g,h,i)perylene C3-v1 2.12E-06 5.50E-07 5.57E-06 2.55E-07 1 140.55 159.04 1.00
Benzo(k)fluoranthene C3-v0 3.36E-08 2.08E-08 6.81E-08 1.18E-08 1 90.02 101.86 1.00
1,3-Butadiene A3-v0 1.95E-02 1.96E-02 2.01E-02 1.89E-02 1 3.14 3.55 0.00
Chrysene C3-v1 1.12E-05 2.80E-06 2.92E-05 1.54E-06 1 139.73 158.12 1.00
Copper D3-v0 2.62E-03 1.64E-03 4.58E-03 1.63E-03 1 64.94 73.48 1.00
Dibenz(a,h)anthracene C3-v1 1.76E-06 1.18E-07 5.09E-06 6.54E-08 1 164.18 185.78 1.00
Dioxin:4D 2378 C3-v0 4.96E-10 5.94E-10 5.97E-10 2.97E-10 1 34.70 39.26 0.00
Dioxin:5D 12378 C3-v1 2.49E-09 5.94E-10 6.57E-09 2.97E-10 1 142.30 161.02 0.00
Dioxin:6D 123478 C3-v1 2.19E-09 5.94E-10 5.68E-09 2.97E-10 1 138.10 156.28 0.00
Dioxin:6D 123678 C3-v1 2.99E-09 5.94E-10 8.07E-09 2.97E-10 1 147.43 166.83 0.07
Dioxin:6D 123789 C3-v1 4.78E-09 5.94E-10 1.34E-08 2.97E-10 1 157.08 177.75 0.04
Dioxin:7D 1234678 C3-v1 1.33E-08 4.45E-09 3.29E-08 2.68E-09 1 127.05 143.76 1.00
Dioxin:8D C3-v0 4.68E-08 5.64E-08 5.97E-08 2.41E-08 1 42.12 47.67 1.00
Fluoranthene C3-v0 1.97E-06 2.28E-06 2.48E-06 1.13E-06 1 37.09 41.97 1.00
Fluorene C3-v0 7.48E-05 3.25E-05 1.67E-04 2.52E-05 1 106.53 120.55 1.00
Furan:4F 2378 C3-v2 8.93E-08 1.48E-09 2.66E-07 5.95E-10 1 171.19 193.71 1.00
Furan:5F 12378 C3-v1 8.56E-09 5.94E-10 2.48E-08 2.97E-10 1 164.20 185.80 0.00
Furan:5F 23478 C3-v2 1.52E-08 5.94E-10 4.48E-08 2.97E-10 1 168.14 190.26 0.00
Furan:6F 123478 C3-v2 1.92E-08 5.94E-10 5.68E-08 1.49E-10 1 169.85 192.20 0.01
Furan:6F 123678 C3-v2 6.12E-09 2.97E-10 1.79E-08 1.49E-10 1 166.91 188.87 0.02
Furan:5F 23478 C3-v2 1.52E-08 5.94E-10 4.48E-08 2.97E-10 1 168.14 190.26 0.00
Furan:6F 123478 C3-v2 1.92E-08 5.94E-10 5.68E-08 1.49E-10 1 169.85 192.20 0.01
(continued)
D-24
Table D-9a. Summary of Data for Emission Factor Development – Heaters Firing Fuel Oil
(continued)

CARB tainty, Detect
Furan:6F 123678 C3-v2 6.12E-09 2.97E-10 1.79E-08 1.49E-10 1 166.91 188.87 0.02
Furan:6F 123789 C3-v0 4.96E-10 5.94E-10 5.97E-10 2.97E-10 1 34.70 39.26 0.00
Furan:6F 234678 C3-v1 8.76E-09 8.91E-10 2.48E-08 5.95E-10 1 158.52 179.38 0.03
Furan:7F 1234678 C3-v1 1.95E-08 1.19E-09 5.68E-08 5.95E-10 1 165.30 187.05 1.00
Furan:7F 1234789 C3-v0 1.19E-09 5.94E-10 2.69E-09 2.97E-10 1 109.22 123.59 0.00
Furan:8F C3-v1 1.04E-08 4.75E-09 2.54E-08 1.19E-09 1 125.10 141.57 1.00
Indeno(1,2,3-cd)pyrene C3-v1 1.81E-06 1.55E-07 5.12E-06 1.48E-07 1 158.71 179.59 1.00
Lead D3-v0 2.96E-04 1.79E-04 5.48E-04 1.62E-04 1 73.61 83.30 0.62
Mercury D3-v0 1.72E-05 1.29E-05 2.83E-05 1.04E-05 1 56.41 63.84 0.00
Naphthalene C3-v0 8.46E-04 1.04E-03 1.11E-03 3.88E-04 1 47.07 53.26 1.00
Nickel D3-v0 3.46E-01 3.47E-01 4.09E-01 2.81E-01 1 18.57 21.01 1.00
Perylene C3-v0 7.41E-08 3.57E-08 1.66E-07 2.07E-08 1 107.75 121.93 1.00
Phenanthrene C3-v1 2.49E-05 1.18E-05 6.02E-05 2.83E-06 1 123.75 140.04 1.00
Pyrene C3-v0 1.32E-06 1.19E-06 2.14E-06 6.28E-07 1 57.94 65.56 1.00
Selenium D3-v0 4.63E-03 3.96E-03 6.59E-03 3.33E-03 1 37.32 42.23 0.00
Toluene A3-v0 9.99E-03 1.00E-02 1.03E-02 9.67E-03 1 3.06 3.46 0.00
Zinc D3-v0 8.93E-03 8.35E-03 1.22E-02 6.27E-03 1 33.43 37.83 1.00
D-25
Table D-9b. Summary of Data for Emission Factor Development – Heaters Firing Fuel Oil

CARB tainty, Detect
Acrolein C3-v0 4.17E-06 4.18E-06 4.21E-06 4.13E-06 1 0.95 1.07 0.00
Arsenic D3-v0 5.82E-06 5.78E-06 6.02E-06 5.67E-06 1 3.01 3.41 0.00
Benzene A3-v0 5.91E-05 5.92E-05 6.10E-05 5.72E-05 1 3.22 3.65 0.00
1,3-Butadiene A3-v0 1.36E-04 1.37E-04 1.40E-04 1.32E-04 1 3.14 3.55 0.00
Cadmium D3-v1 5.73E-06 8.28E-06 8.60E-06 2.99E-07 1 82.13 92.94 1.00
Chrysene C3-v1 7.81E-08 1.95E-08 2.04E-07 1.07E-08 1 139.73 158.12 1.00
Copper D3-v0 1.83E-05 1.15E-05 3.20E-05 1.14E-05 1 64.94 73.48 1.00
Dioxin:4D 2378 C3-v0 3.46E-12 4.15E-12 4.17E-12 2.08E-12 1 34.70 39.26 0.00
Dioxin:5D 12378 C3-v1 1.74E-11 4.15E-12 4.59E-11 2.08E-12 1 142.30 161.02 0.00
Dioxin:6D 123478 C3-v1 1.53E-11 4.15E-12 3.96E-11 2.08E-12 1 138.10 156.28 0.00
Dioxin:6D 123678 C3-v1 2.08E-11 4.15E-12 5.63E-11 2.08E-12 1 147.43 166.83 0.07
Dioxin:6D 123789 C3-v1 3.33E-11 4.15E-12 9.38E-11 2.08E-12 1 157.08 177.75 0.04
Dioxin:7D 1234678 C3-v1 9.30E-11 3.11E-11 2.29E-10 1.87E-11 1 127.05 143.76 1.00
Dioxin:8D C3-v0 3.26E-10 3.94E-10 4.17E-10 1.68E-10 1 42.12 47.67 1.00
Fluorene C3-v0 5.22E-07 2.26E-07 1.16E-06 1.76E-07 1 106.53 120.55 1.00
Furan:4F 2378 C3-v2 6.23E-10 1.04E-11 1.86E-09 4.15E-12 1 171.19 193.71 1.00
Furan:5F 12378 C3-v1 5.97E-11 4.15E-12 1.73E-10 2.08E-12 1 164.20 185.80 0.00
Furan:5F 23478 C3-v2 1.06E-10 4.15E-12 3.13E-10 2.08E-12 1 168.14 190.26 0.00
Furan:6F 123478 C3-v2 1.34E-10 4.15E-12 3.96E-10 1.04E-12 1 169.85 192.20 0.01
Furan:6F 123678 C3-v2 4.27E-11 2.07E-12 1.25E-10 1.04E-12 1 166.91 188.87 0.02
(continued)
D-26
Table D-9b. Summary of Data for Emission Factor Development – Heaters Firing Fuel Oil
(continued)

CARB tainty, Detect
Furan:6F 123789 C3-v0 3.46E-12 4.15E-12 4.17E-12 2.08E-12 1 34.70 39.26 0.00
Furan:6F 234678 C3-v1 6.11E-11 6.22E-12 1.73E-10 4.15E-12 1 158.52 179.38 0.03
Furan:7F 1234678 C3-v1 1.36E-10 8.29E-12 3.96E-10 4.15E-12 1 165.30 187.05 1.00
Furan:7F 1234789 C3-v0 8.33E-12 4.15E-12 1.88E-11 2.08E-12 1 109.22 123.59 0.00
Furan:8F C3-v1 7.29E-11 3.32E-11 1.77E-10 8.30E-12 1 125.10 141.57 1.00
Lead D3-v0 2.07E-06 1.25E-06 3.83E-06 1.13E-06 1 73.61 83.30 0.62
Mercury D3-v0 1.20E-07 8.99E-08 1.98E-07 7.28E-08 1 56.41 63.84 0.00
Nickel D3-v0 2.41E-03 2.42E-03 2.86E-03 1.96E-03 1 18.57 21.01 1.00
Perylene C3-v0 5.17E-10 2.49E-10 1.16E-09 1.45E-10 1 107.75 121.93 1.00
Pyrene C3-v0 9.22E-09 8.32E-09 1.49E-08 4.38E-09 1 57.94 65.56 1.00
Selenium D3-v0 3.23E-05 2.77E-05 4.60E-05 2.33E-05 1 37.32 42.23 0.00
Toluene A3-v0 6.97E-05 6.99E-05 7.18E-05 6.75E-05 1 3.06 3.46 0.00
Zinc D3-v0 6.23E-05 5.83E-05 8.49E-05 4.38E-05 1 33.43 37.83 1.00
D-27
Table D-10a. Summary of Data for Emission Factor Development – Heaters Firing Refinery
Fuel Gas Controlled with Selective Non-Catalytic Reduction (NOx)

CARB tainty, Detect
Acetaldehyde B1-v3 1.95E-02 1.08E-02 1.12E-01 1.10E-04 8 129.02 51.62 0.88
Antimony C3-v0 5.81E-04 6.56E-04 8.51E-04 2.36E-04 1 54.13 61.25 1.00
Arsenic C3-v0 9.54E-04 1.11E-03 1.43E-03 3.19E-04 1 60.10 68.00 1.00
Barium C3-v0 6.49E-03 6.49E-03 6.64E-03 6.33E-03 1 2.45 2.78 0.00
Benzene B1-v2 8.42E-02 6.76E-02 2.76E-01 2.35E-03 11 93.71 31.97 0.02
Cadmium C3-v0 1.11E-03 1.08E-03 1.33E-03 9.15E-04 1 18.76 21.23 1.00
Chrysene A2-v0 1.66E-06 1.29E-06 5.37E-06 9.83E-07 4 76.71 45.33 0.63
Copper C3-v0 4.73E-03 2.16E-03 1.05E-02 1.48E-03 1 106.76 120.81 1.00
Fluorene A2-v0 1.06E-05 8.55E-06 2.64E-05 3.33E-06 4 67.66 39.98 1.00
Formaldehyde B1-v3 1.51E-01 3.03E-02 1.76E+00 7.92E-04 7 254.14 108.70 1.00
Hydrogen Sulfide A1-v1 4.05E-01 3.27E-01 1.20E+00 1.82E-02 7 84.49 36.13 0.00
Lead C3-v0 5.49E-03 4.43E-03 8.43E-03 3.61E-03 1 47.03 53.22 1.00
Mercury C3-v0 2.02E-04 1.97E-04 2.17E-04 1.92E-04 1 6.49 7.34 0.36
Nickel C3-v1 1.06E-02 1.48E-03 2.88E-02 1.44E-03 1 149.30 168.95 0.95
Phenol C1-v1 6.96E-03 4.70E-03 2.68E-02 2.93E-04 7 103.61 44.32 0.97
Pyrene A2-v0 2.84E-06 2.63E-06 5.09E-06 1.78E-06 4 33.95 20.06 1.00
Selenium C3-v0 2.20E-05 2.28E-05 2.85E-05 1.48E-05 1 31.23 35.34 0.78
(continued)
D-28
Fuel Gas Controlled with Selective Non-Catalytic Reduction (NOx) (continued)

CARB tainty, Detect
Silver C3-v1 1.81E-03 1.48E-03 3.61E-03 3.53E-04 1 91.23 103.24 0.94
Thallium C3-v0 6.49E-03 6.49E-03 6.64E-03 6.33E-03 1 2.45 2.78 0.00
Toluene D1-v2 1.37E-01 7.81E-02 1.21E+00 3.71E-03 11 154.31 52.65 0.55
Zinc C3-v0 2.34E-02 2.89E-02 3.18E-02 9.52E-03 1 51.72 58.53 1.00
D-29
Table D-10b. Summary of Data for Emission Factor Development – Heaters Firing Refinery
Fuel Gas Controlled with Selective Non-Catalytic Reduction (NOx)

CARB tainty, Detect
Antimony C3-v0 5.17E-07 5.84E-07 7.58E-07 2.10E-07 1 54.13 61.25 1.00
Arsenic C3-v0 8.50E-07 9.90E-07 1.28E-06 2.84E-07 1 60.10 68.00 1.00
Barium C3-v0 5.78E-06 5.78E-06 5.92E-06 5.63E-06 1 2.45 2.78 0.00
Benzene B1-v1 6.47E-05 5.49E-05 1.85E-04 2.54E-06 11 87.67 29.91 0.02
Cadmium C3-v0 9.88E-07 9.65E-07 1.18E-06 8.15E-07 1 18.76 21.23 1.00
Chrysene A2-v0 1.63E-09 1.23E-09 4.79E-09 1.02E-09 4 66.84 39.50 0.63
Copper C3-v0 4.21E-06 1.93E-06 9.39E-06 1.31E-06 1 106.76 120.81 1.00
Fluorene A2-v0 1.08E-08 8.77E-09 2.74E-08 2.96E-09 4 70.62 41.74 1.00
Formaldehyde B1-v3 1.11E-04 1.90E-05 1.34E-03 7.67E-07 7 262.94 112.46 1.00
Lead C3-v0 4.89E-06 3.94E-06 7.51E-06 3.21E-06 1 47.03 53.22 1.00
Mercury C3-v0 1.80E-07 1.75E-07 1.93E-07 1.71E-07 1 6.49 7.34 0.36
Nickel C3-v1 9.42E-06 1.31E-06 2.57E-05 1.29E-06 1 149.30 168.95 0.95
Phenol C1-v1 5.63E-06 3.14E-06 2.54E-05 2.84E-07 7 114.62 49.02 0.97
Pyrene A2-v0 2.84E-09 2.72E-09 4.53E-09 1.87E-09 4 28.87 17.06 1.00
Selenium C3-v0 1.96E-08 2.03E-08 2.54E-08 1.32E-08 1 31.23 35.34 0.78
(continued)
D-30
Fuel Gas Controlled with Selective Non-Catalytic Reduction (NOx) (continued)

CARB tainty, Detect
Silver C3-v1 1.61E-06 1.31E-06 3.21E-06 3.14E-07 1 91.23 103.24 0.94
Thallium C3-v0 5.78E-06 5.78E-06 5.92E-06 5.63E-06 1 2.45 2.78 0.00
Toluene D1-v2 1.07E-04 7.00E-05 9.19E-04 4.04E-06 11 148.57 50.69 0.55
Zinc C3-v0 2.08E-05 2.58E-05 2.83E-05 8.48E-06 1 51.72 58.53 1.00
D-31

Heaters Firing Refinery Fuel Gas

CARB tainty, Detect
Antimony C3-v0 5.81E-04 6.56E-04 8.51E-04 2.36E-04 1 54.13 61.25 1.00
Arsenic C3-v0 9.54E-04 1.11E-03 1.43E-03 3.19E-04 1 60.10 68.00 1.00
Barium C3-v0 6.49E-03 6.49E-03 6.64E-03 6.33E-03 1 2.45 2.78 0.00
Benzene B1-v2 8.42E-02 6.76E-02 2.76E-01 2.35E-03 11 93.71 31.97 0.02
Cadmium C3-v0 1.11E-03 1.08E-03 1.33E-03 9.15E-04 1 18.76 21.23 1.00
Chrysene A2-v0 1.66E-06 1.29E-06 5.37E-06 9.83E-07 4 76.71 45.33 0.63
Copper C3-v0 4.73E-03 2.16E-03 1.05E-02 1.48E-03 1 106.76 120.81 1.00
Fluorene A2-v0 1.06E-05 8.55E-06 2.64E-05 3.33E-06 4 67.66 39.98 1.00
Lead C3-v0 5.49E-03 4.43E-03 8.43E-03 3.61E-03 1 47.03 53.22 1.00
Mercury C3-v0 2.02E-04 1.97E-04 2.17E-04 1.92E-04 1 6.49 7.34 0.36
Nickel C3-v1 1.06E-02 1.48E-03 2.88E-02 1.44E-03 1 149.30 168.95 0.95
Phenol C1-v1 6.96E-03 4.70E-03 2.68E-02 2.93E-04 7 103.61 44.32 0.97
Pyrene A2-v0 2.84E-06 2.63E-06 5.09E-06 1.78E-06 4 33.95 20.06 1.00
Selenium C3-v0 2.20E-05 2.28E-05 2.85E-05 1.48E-05 1 31.23 35.34 0.78
(continued)
D-32

Heaters Firing Refinery Fuel Gas (continued)

CARB tainty, Detect
Silver C3-v1 1.81E-03 1.48E-03 3.61E-03 3.53E-04 1 91.23 103.24 0.94
Thallium C3-v0 6.49E-03 6.49E-03 6.64E-03 6.33E-03 1 2.45 2.78 0.00
Toluene D1-v2 1.37E-01 7.81E-02 1.21E+00 3.71E-03 11 154.31 52.65 0.55
Zinc C3-v0 2.34E-02 2.89E-02 3.18E-02 9.52E-03 1 51.72 58.53 1.00
D-33

Heaters Firing Refinery Fuel Gas

CARB tainty, Detect
Antimony C3-v0 5.17E-07 5.84E-07 7.58E-07 2.10E-07 1 54.13 61.25 1.00
Arsenic C3-v0 8.50E-07 9.90E-07 1.28E-06 2.84E-07 1 60.10 68.00 1.00
Barium C3-v0 5.78E-06 5.78E-06 5.92E-06 5.63E-06 1 2.45 2.78 0.00
Benzene B1-v1 6.47E-05 5.49E-05 1.85E-04 2.54E-06 11 87.67 29.91 0.02
Cadmium C3-v0 9.88E-07 9.65E-07 1.18E-06 8.15E-07 1 18.76 21.23 1.00
Chrysene A2-v0 1.63E-09 1.23E-09 4.79E-09 1.02E-09 4 66.84 39.50 0.63
Copper C3-v0 4.21E-06 1.93E-06 9.39E-06 1.31E-06 1 106.76 120.81 1.00
Fluorene A2-v0 1.08E-08 8.77E-09 2.74E-08 2.96E-09 4 70.62 41.74 1.00
Lead C3-v0 4.89E-06 3.94E-06 7.51E-06 3.21E-06 1 47.03 53.22 1.00
Mercury C3-v0 1.80E-07 1.75E-07 1.93E-07 1.71E-07 1 6.49 7.34 0.36
Nickel C3-v1 9.42E-06 1.31E-06 2.57E-05 1.29E-06 1 149.30 168.95 0.95
Phenol C1-v1 5.63E-06 3.14E-06 2.54E-05 2.84E-07 7 114.62 49.02 0.97
Pyrene A2-v0 2.84E-09 2.72E-09 4.53E-09 1.87E-09 4 28.87 17.06 1.00
Selenium C3-v0 1.96E-08 2.03E-08 2.54E-08 1.32E-08 1 31.23 35.34 0.78
(continued)
D-34

Heaters Firing Refinery Fuel Gas (continued)

CARB tainty, Detect
Silver C3-v1 1.61E-06 1.31E-06 3.21E-06 3.14E-07 1 91.23 103.24 0.94
Thallium C3-v0 5.78E-06 5.78E-06 5.92E-06 5.63E-06 1 2.45 2.78 0.00
Toluene D1-v2 1.07E-04 7.00E-05 9.19E-04 4.04E-06 11 148.57 50.69 0.55
Zinc C3-v0 2.08E-05 2.58E-05 2.83E-05 8.48E-06 1 51.72 58.53 1.00
D-35
Fuel Gas Controlled with Selective Catalytic Reduction

CARB tainty, Detect
Antimony C3-v0 5.81E-04 6.56E-04 8.51E-04 2.36E-04 1 54.13 61.25 1.00
Arsenic C3-v0 9.54E-04 1.11E-03 1.43E-03 3.19E-04 1 60.10 68.00 1.00
Barium C3-v0 6.49E-03 6.49E-03 6.64E-03 6.33E-03 1 2.45 2.78 0.00
Benzene B1-v2 8.42E-02 6.76E-02 2.76E-01 2.35E-03 11 93.71 31.97 0.02
Cadmium C3-v0 1.11E-03 1.08E-03 1.33E-03 9.15E-04 1 18.76 21.23 1.00
Chrysene A2-v0 1.66E-06 1.29E-06 5.37E-06 9.83E-07 4 76.71 45.33 0.63
Copper C3-v0 4.73E-03 2.16E-03 1.05E-02 1.48E-03 1 106.76 120.81 1.00
Fluorene A2-v0 1.06E-05 8.55E-06 2.64E-05 3.33E-06 4 67.66 39.98 1.00
Lead C3-v0 5.49E-03 4.43E-03 8.43E-03 3.61E-03 1 47.03 53.22 1.00
Mercury C3-v0 2.02E-04 1.97E-04 2.17E-04 1.92E-04 1 6.49 7.34 0.36
Nickel C3-v1 1.06E-02 1.48E-03 2.88E-02 1.44E-03 1 149.30 168.95 0.95
Phenol C1-v1 6.96E-03 4.70E-03 2.68E-02 2.93E-04 7 103.61 44.32 0.97
Pyrene A2-v0 2.84E-06 2.63E-06 5.09E-06 1.78E-06 4 33.95 20.06 1.00
Selenium C3-v0 2.20E-05 2.28E-05 2.85E-05 1.48E-05 1 31.23 35.34 0.78
(continued)
D-36
Fuel Gas Controlled with Selective Catalytic Reduction (continued)

CARB tainty, Detect
Silver C3-v1 1.81E-03 1.48E-03 3.61E-03 3.53E-04 1 91.23 103.24 0.94
Thallium C3-v0 6.49E-03 6.49E-03 6.64E-03 6.33E-03 1 2.45 2.78 0.00
Toluene D1-v2 1.37E-01 7.81E-02 1.21E+00 3.71E-03 11 154.31 52.65 0.55
Zinc C3-v0 2.34E-02 2.89E-02 3.18E-02 9.52E-03 1 51.72 58.53 1.00
D-37
Fuel Gas Controlled with Selective Catalytic Reduction

CARB tainty, Detect
Antimony C3-v0 5.17E-07 5.84E-07 7.58E-07 2.10E-07 1 54.13 61.25 1.00
Arsenic C3-v0 8.50E-07 9.90E-07 1.28E-06 2.84E-07 1 60.10 68.00 1.00
Barium C3-v0 5.78E-06 5.78E-06 5.92E-06 5.63E-06 1 2.45 2.78 0.00
Benzene B1-v1 6.47E-05 5.49E-05 1.85E-04 2.54E-06 11 87.67 29.91 0.02
Cadmium C3-v0 9.88E-07 9.65E-07 1.18E-06 8.15E-07 1 18.76 21.23 1.00
Chrysene A2-v0 1.63E-09 1.23E-09 4.79E-09 1.02E-09 4 66.84 39.50 0.63
Copper C3-v0 4.21E-06 1.93E-06 9.39E-06 1.31E-06 1 106.76 120.81 1.00
Fluorene A2-v0 1.08E-08 8.77E-09 2.74E-08 2.96E-09 4 70.62 41.74 1.00
Indeno(1,2,3-
A1-v3 1.03E-07 1.75E-09 1.42E-06 1.02E-09 9 343.01 131.85 0.99
cd)pyrene
Lead C3-v0 4.89E-06 3.94E-06 7.51E-06 3.21E-06 1 47.03 53.22 1.00
Mercury C3-v0 1.80E-07 1.75E-07 1.93E-07 1.71E-07 1 6.49 7.34 0.36
Nickel C3-v1 9.42E-06 1.31E-06 2.57E-05 1.29E-06 1 149.30 168.95 0.95
Phenol C1-v1 5.63E-06 3.14E-06 2.54E-05 2.84E-07 7 114.62 49.02 0.97
Pyrene A2-v0 2.84E-09 2.72E-09 4.53E-09 1.87E-09 4 28.87 17.06 1.00
(continued)
D-38
Fuel Gas Controlled with Selective Catalytic Reduction (continued)

CARB tainty, Detect
Selenium C3-v0 1.96E-08 2.03E-08 2.54E-08 1.32E-08 1 31.23 35.34 0.78
Silver C3-v1 1.61E-06 1.31E-06 3.21E-06 3.14E-07 1 91.23 103.24 0.94
Thallium C3-v0 5.78E-06 5.78E-06 5.92E-06 5.63E-06 1 2.45 2.78 0.00
Toluene D1-v2 1.07E-04 7.00E-05 9.19E-04 4.04E-06 11 148.57 50.69 0.55
Zinc C3-v0 2.08E-05 2.58E-05 2.83E-05 8.48E-06 1 51.72 58.53 1.00
D-39

Internal Combustion Engines Firing Diesel Fuel, O2 < 13%

CARB tainty, Detect
Acrolein A3-v0 1.07E-03 7.15E-04 1.79E-03 7.15E-04 1 57.74 65.33 0.56
Benzene A3-v0 1.01E-01 9.93E-02 1.04E-01 9.93E-02 1 2.71 3.06 1.00
Chrysene C3-v0 2.11E-04 2.21E-04 2.28E-04 1.84E-04 1 11.03 12.48 1.00
Fluorene C3-v0 1.77E-03 1.76E-03 1.81E-03 1.72E-03 1 2.62 2.96 1.00
Pyrene C3-v0 5.11E-04 4.89E-04 5.60E-04 4.86E-04 1 8.16 9.24 1.00
Toluene A3-v0 3.74E-02 3.70E-02 3.87E-02 3.64E-02 1 3.27 3.70 1.00
D-40

Internal Combustion Engines Firing Diesel Fuel, O2 < 13%

CARB tainty, Detect
Acrolein A3-v0 7.57E-06 5.04E-06 1.26E-05 5.04E-06 1 57.74 65.33 0.56
Benzene A3-v0 7.11E-04 7.00E-04 7.33E-04 7.00E-04 1 2.71 3.06 1.00
Chrysene C3-v0 1.49E-06 1.56E-06 1.60E-06 1.30E-06 1 11.03 12.48 1.00
Fluorene C3-v0 1.25E-05 1.24E-05 1.28E-05 1.21E-05 1 2.62 2.96 1.00
Pyrene C3-v0 3.61E-06 3.45E-06 3.95E-06 3.43E-06 1 8.16 9.24 1.00
Toluene A3-v0 2.63E-04 2.61E-04 2.73E-04 2.57E-04 1 3.27 3.70 1.00
D-41

Internal Combustion Engines Firing Diesel Fuel, O2 > 13%

CARB tainty, Detect
Acrolein A3-v0 1.30E-02 8.43E-03 3.16E-02 6.62E-03 2 74.07 59.27 0.82
Benzaldehyde A3-v0 1.26E-02 1.24E-02 1.35E-02 1.20E-02 1 6.29 7.12 0.68
Benzene B3-v0 1.22E-01 1.13E-01 1.91E-01 6.57E-02 2 48.35 38.69 1.00
Benzo(b+k)fluoranthene C3-v0 1.44E-06 1.45E-06 1.46E-06 1.41E-06 1 1.90 2.14 0.00
1,3-Butadiene C3-v0 5.41E-03 5.41E-03 5.41E-03 5.41E-03 1 0.00 0.00 0.00
Chrysene C3-v0 4.90E-05 5.23E-05 6.75E-05 2.72E-05 2 33.96 27.17 1.00
Fluorene C3-v1 4.06E-03 3.96E-03 7.57E-03 2.10E-04 2 87.57 70.07 0.99
Propylene B3-v0 3.58E-01 3.38E-01 5.83E-01 1.45E-01 2 56.61 45.30 1.00
Pyrene C3-v0 6.66E-04 6.10E-04 1.07E-03 1.17E-04 2 51.99 41.60 1.00
Toluene B3-v0 5.50E-02 5.49E-02 7.56E-02 3.44E-02 2 38.06 30.45 1.00
Xylene (m,p) C3-v0 2.16E-02 2.09E-02 2.40E-02 1.98E-02 1 10.07 11.40 1.00
Xylene (o) C3-v0 2.09E-02 2.09E-02 2.09E-02 2.09E-02 1 0.00 0.00 0.00
D-42

Internal Combustion Engines Firing Diesel Fuel, O2 > 13%

CARB tainty, Detect
Acrolein A3-v0 9.37E-05 6.05E-05 2.29E-04 4.72E-05 2 74.63 59.71 0.82
Benzaldehyde A3-v0 9.01E-05 8.84E-05 9.64E-05 8.54E-05 1 6.29 7.12 0.68
Benzene B3-v0 8.81E-04 8.17E-04 1.38E-03 4.69E-04 2 48.96 39.17 1.00
Benzo(b+k)fluoranthene C3-v0 1.03E-08 1.03E-08 1.04E-08 1.00E-08 1 1.90 2.14 0.00
1,3-Butadiene C3-v0 3.86E-05 3.86E-05 3.86E-05 3.86E-05 1 0.00 0.00 0.00
Chrysene C3-v0 3.52E-07 3.76E-07 4.81E-07 1.97E-07 2 33.43 26.75 1.00
Fluorene C3-v1 2.90E-05 2.83E-05 5.40E-05 1.52E-06 2 87.31 69.86 0.99
Propylene B3-v0 2.58E-03 2.44E-03 4.22E-03 1.04E-03 2 57.16 45.74 1.00
Pyrene C3-v0 4.78E-06 4.38E-06 7.62E-06 8.43E-07 2 52.00 41.61 1.00
Toluene B3-v0 3.96E-04 3.95E-04 5.47E-04 2.46E-04 2 38.72 30.98 1.00
Xylene (m,p) C3-v0 1.54E-04 1.49E-04 1.71E-04 1.42E-04 1 10.07 11.40 1.00
Xylene (o) C3-v0 1.49E-04 1.49E-04 1.49E-04 1.49E-04 1 0.00 0.00 0.00
D-43

Two-Stroke Internal Combustion Engines Firing Field Gas, Lean

CARB tainty, Detect
Acetaldehyde A3-v0 8.77E+00 9.42E+00 1.14E+01 5.53E+00 1 33.90 38.37 1.00
Acrolein A3-v1 2.00E+00 1.44E+00 4.36E+00 2.00E-01 1 106.79 120.84 1.00
Benzene A3-v0 7.85E+00 4.60E+00 1.36E+01 4.23E+00 2 60.68 53.19 1.00
Chrysene A3-v0 1.69E-03 1.69E-03 1.80E-03 1.58E-03 1 9.43 13.07 1.00
Fluorene A3-v0 2.44E-03 2.44E-03 3.61E-03 1.27E-03 1 67.90 94.10 0.00
Formaldehyde A3-v1 5.09E+01 6.79E+01 8.53E+01 4.21E+00 2 69.80 55.85 1.00
Propylene A3-v0 2.49E+01 1.77E+01 4.02E+01 1.14E+01 2 56.74 49.73 1.00
Pyrene A3-v0 2.39E-04 2.39E-04 3.06E-04 1.71E-04 1 40.10 55.57 0.64
Toluene A3-v1 2.86E+00 2.36E+00 5.74E+00 2.23E-01 2 72.83 63.84 1.00
Xylene (m,p) A3-v0 6.04E-01 3.17E-01 1.25E+00 1.89E-01 2 80.76 70.79 1.00
Xylene (o) A3-v0 2.88E-01 1.36E-01 5.91E-01 6.80E-02 2 88.41 77.49 1.00
D-44

Two-Stroke Internal Combustion Engines Firing Field Gas, Lean

CARB tainty, Detect
Acrolein A3-v1 1.90E-03 1.37E-03 4.15E-03 1.90E-04 1 106.79 120.84 1.00
Benzene A3-v0 7.48E-03 4.38E-03 1.30E-02 4.03E-03 2 60.68 53.19 1.00
Chrysene A3-v0 1.61E-06 1.61E-06 1.72E-06 1.50E-06 1 9.43 13.07 1.00
Fluorene A3-v0 2.32E-06 2.32E-06 3.43E-06 1.21E-06 1 67.90 94.10 0.00
Pyrene A3-v0 2.27E-07 2.27E-07 2.92E-07 1.63E-07 1 40.10 55.57 0.64
Toluene A3-v1 2.72E-03 2.25E-03 5.47E-03 2.12E-04 2 72.83 63.84 1.00
Xylene (m,p) A3-v0 5.75E-04 3.02E-04 1.19E-03 1.80E-04 2 80.76 70.79 1.00
Xylene (o) A3-v0 2.74E-04 1.30E-04 5.63E-04 6.48E-05 2 88.41 77.49 1.00
D-45

Four-Stroke Internal Combustion Engines Firing Field Gas, Lean

CARB tainty, Detect
Benzene A2-v0 1.72E+00 2.07E+00 2.24E+00 9.25E-01 3 35.58 28.47 1.00
Toluene A2-v0 7.68E-01 9.38E-01 1.17E+00 2.69E-01 3 51.12 40.90 1.00
Xylene (m,p) A2-v1 3.02E-01 1.90E-01 9.03E-01 7.74E-02 3 99.71 79.78 1.00
Xylene (o) A2-v0 8.97E-02 9.57E-02 1.22E-01 6.13E-02 3 25.60 20.49 1.00

Four-Stroke Internal Combustion Engines Firing Field Gas, Lean

CARB tainty, Detect
Benzene A2-v0 1.64E-03 1.97E-03 2.14E-03 8.81E-04 3 35.58 28.47 1.00
Toluene A2-v0 7.31E-04 8.94E-04 1.11E-03 2.56E-04 3 51.12 40.90 1.00
Xylene (m,p) A2-v1 2.87E-04 1.81E-04 8.60E-04 7.37E-05 3 99.71 79.78 1.00
Xylene (o) A2-v0 8.55E-05 9.11E-05 1.16E-04 5.84E-05 3 25.60 20.49 1.00
D-46

Four-Stroke Internal Combustion Engines Firing Field Gas, Rich

CARB tainty, Detect
Benzene A3-v0 1.10E+01 1.10E+01 1.11E+01 1.09E+01 1 1.15 1.59 1.00
Toluene A3-v0 3.44E+00 3.44E+00 3.55E+00 3.33E+00 1 4.56 6.32 1.00
Xylene (m,p) A3-v0 5.37E-01 5.37E-01 5.62E-01 5.11E-01 1 6.73 9.33 1.00
Xylene (o) A3-v0 2.68E-01 2.68E-01 2.81E-01 2.56E-01 1 6.73 9.33 1.00

Four-Stroke Internal Combustion Engines Firing Field Gas, Rich

CARB tainty, Detect
Benzene A3-v0 1.05E-02 1.05E-02 1.06E-02 1.04E-02 1 1.15 1.59 1.00
Toluene A3-v0 3.28E-03 3.28E-03 3.38E-03 3.17E-03 1 4.56 6.32 1.00
Xylene (m,p) A3-v0 5.11E-04 5.11E-04 5.36E-04 4.87E-04 1 6.73 9.33 1.00
Xylene (o) A3-v0 2.56E-04 2.56E-04 2.68E-04 2.43E-04 1 6.73 9.33 1.00
D-47

Four-Stroke Internal Combustion Engines Firing Natural Gas, Lean

CARB tainty, Detect
Acrolein A1-v1 1.63E+00 1.10E+00 5.48E+00 2.01E-01 5 92.02 48.20 1.00
Benzene A1-v1 1.21E+00 1.22E+00 2.47E+00 2.47E-01 7 47.28 24.77 1.00
Chrysene A3-v0 9.64E-05 1.11E-04 1.12E-04 6.61E-05 1 27.27 30.86 1.00
Fluorene A3-v0 4.60E-04 4.16E-04 6.28E-04 3.37E-04 1 32.63 36.92 0.00
Pyrene A3-v0 1.23E-04 1.42E-04 1.62E-04 6.61E-05 1 41.10 46.51 1.00
Toluene A1-v0 4.12E-01 3.96E-01 5.70E-01 1.65E-01 7 33.98 17.80 1.00
Xylene (m,p) A1-v0 8.63E-02 8.13E-02 1.59E-01 3.53E-02 7 46.27 24.24 1.00
Xylene (o) A1-v0 4.94E-02 4.95E-02 9.14E-02 1.06E-02 7 39.52 20.70 0.95
D-48

Four-Stroke Internal Combustion Engines Firing Natural Gas, Lean

CARB tainty, Detect
Acrolein A1-v1 1.56E-03 1.05E-03 5.22E-03 1.92E-04 5 92.02 48.20 1.00
Benzene A1-v1 1.15E-03 1.17E-03 2.35E-03 2.35E-04 7 47.28 24.77 1.00
Chrysene A3-v0 9.19E-08 1.05E-07 1.07E-07 6.29E-08 1 27.27 30.86 1.00
Fluorene A3-v0 4.38E-07 3.96E-07 5.98E-07 3.21E-07 1 32.63 36.92 0.00
Pyrene A3-v0 1.18E-07 1.35E-07 1.55E-07 6.29E-08 1 41.10 46.51 1.00
Toluene A1-v0 3.92E-04 3.77E-04 5.43E-04 1.57E-04 7 33.98 17.80 1.00
Xylene (m,p) A1-v0 8.22E-05 7.74E-05 1.51E-04 3.36E-05 7 46.27 24.24 1.00
Xylene (o) A1-v0 4.71E-05 4.71E-05 8.70E-05 1.01E-05 7 39.52 20.70 0.95
D-49

Four-Stroke Internal Combustion Engines Firing Natural Gas, Rich

CARB tainty, Detect
Acrolein A3-v2 5.40E-01 4.16E-01 1.37E+00 2.59E-03 2 105.82 84.67 1.00
Benzene A3-v0 9.87E+00 9.87E+00 1.02E+01 9.51E+00 1 5.05 7.00 1.00
Formaldehyde A3-v1 5.32E+00 4.48E+00 1.14E+01 4.21E-01 2 100.69 80.57 1.00
Toluene A3-v0 2.51E+00 2.51E+00 2.62E+00 2.41E+00 1 6.15 8.52 1.00
Xylene (m,p) A3-v0 4.41E-01 4.41E-01 4.54E-01 4.28E-01 1 4.04 5.60 1.00
Xylene (o) A3-v0 2.17E-01 2.17E-01 2.22E-01 2.12E-01 1 3.29 4.56 1.00

Four-Stroke Internal Combustion Engines Firing Natural Gas, Rich

CARB tainty, Detect
Acrolein A3-v2 5.15E-04 3.97E-04 1.31E-03 2.50E-06 2 105.65 84.54 1.00
Benzene A3-v0 9.40E-03 9.40E-03 9.73E-03 9.06E-03 1 5.05 7.00 1.00
Toluene A3-v0 2.40E-03 2.40E-03 2.50E-03 2.29E-03 1 6.15 8.52 1.00
Xylene (m,p) A3-v0 4.20E-04 4.20E-04 4.32E-04 4.08E-04 1 4.04 5.60 1.00
Xylene (o) A3-v0 2.06E-04 2.06E-04 2.11E-04 2.02E-04 1 3.29 4.56 1.00
D-50

Steam Generators Firing Natural Gas

CARB tainty, Detect
Acrolein A3-v0 1.84E-02 1.75E-02 2.22E-02 1.55E-02 1 18.73 21.19 1.00
Benzene B3-v0 3.86E-03 3.87E-03 3.88E-03 3.83E-03 2 0.57 0.46 0.00
Propylene C3-v0 1.09E-01 1.09E-01 1.09E-01 1.09E-01 1 0.36 0.41 0.00
Toluene B3-v0 1.20E-02 1.20E-02 1.20E-02 1.19E-02 2 0.57 0.46 0.00
Xylene (Total) B3-v0 2.76E-02 2.76E-02 2.77E-02 2.74E-02 2 0.57 0.46 0.00

Steam Generators Firing Natural Gas

CARB tainty, Detect
Acrolein A3-v0 1.96E-05 1.86E-05 2.37E-05 1.65E-05 1 18.73 21.19 1.00
Benzene B3-v0 3.94E-06 3.93E-06 4.10E-06 3.78E-06 2 4.28 3.42 0.00
Toluene B3-v0 1.22E-05 1.22E-05 1.27E-05 1.17E-05 2 4.28 3.42 0.00
D-51

Steam Generators Firing Natural Gas and Casing Vapor Recovery Gas

CARB tainty, Detect
Acrolein A2-v0 1.05E-02 1.07E-02 1.82E-02 5.02E-03 4 37.86 21.42 0.88
Benzene B1-v0 3.33E-03 2.68E-03 6.21E-03 2.36E-03 5 44.41 22.47 0.00
Chrysene A3-v0 1.55E-06 1.83E-06 2.16E-06 6.82E-07 2 38.63 30.91 0.33
Fluorene A3-v0 5.63E-06 2.47E-06 1.30E-05 1.77E-06 2 95.48 76.40 1.00
Hydrogen Sulfide C3-v0 1.48E-01 1.43E-01 2.28E-01 7.39E-02 1 51.99 58.83 1.00
Pyrene A3-v1 6.00E-06 2.78E-06 1.74E-05 8.19E-07 2 109.69 87.77 1.00
Toluene B1-v0 1.37E-02 9.47E-03 3.08E-02 7.33E-03 5 59.41 30.07 0.64
D-52

Steam Generators Firing Natural Gas and Casing Vapor Recovery Gas

CARB tainty, Detect
Acrolein A2-v0 1.69E-05 1.92E-05 2.62E-05 5.47E-06 4 41.86 23.68 0.88
Benzene B1-v0 4.45E-06 4.40E-06 6.12E-06 2.75E-06 5 23.61 11.95 0.00
Chrysene A3-v0 1.60E-09 1.80E-09 2.42E-09 7.12E-10 2 39.17 31.35 0.33
Fluorene A3-v0 5.66E-09 2.64E-09 1.28E-08 1.85E-09 2 92.16 73.74 1.00
Hydrogen Sulfide C3-v0 1.59E-04 1.60E-04 2.38E-04 8.02E-05 1 49.59 56.11 1.00
Pyrene A3-v1 6.02E-09 2.93E-09 1.72E-08 8.55E-10 2 106.75 85.42 1.00
Toluene B1-v0 1.73E-05 1.43E-05 3.04E-05 1.34E-05 5 33.16 16.78 0.64
D-53

Gas Turbines with No Duct Burners Firing Natural Gas

CARB tainty, Detect
Acrolein A3-v0 1.81E-02 1.81E-02 1.82E-02 1.79E-02 1 0.68 0.77 0.00
Benzene A3-v0 1.24E-02 1.23E-02 1.85E-02 6.96E-03 2 47.07 37.67 0.00
Chrysene A3-v0 5.08E-06 5.99E-06 6.34E-06 2.89E-06 2 31.73 25.39 0.31
Fluorene A3-v0 1.55E-05 1.42E-05 3.17E-05 8.05E-06 2 54.85 43.89 1.00
Pyrene A3-v0 2.31E-05 2.09E-05 4.23E-05 5.78E-06 2 63.17 50.55 1.00
Toluene A3-v1 7.17E-02 5.89E-02 1.68E-01 8.22E-03 2 88.92 71.15 0.98
D-54

Gas Turbines with No Duct Burners Firing Natural Gas

CARB tainty, Detect
Acrolein A3-v0 1.72E-05 1.73E-05 1.73E-05 1.71E-05 1 0.68 0.77 0.00
Benzene A3-v0 1.20E-05 1.19E-05 1.82E-05 6.64E-06 2 48.32 38.66 0.00
Chrysene A3-v0 4.93E-09 5.88E-09 6.05E-09 2.76E-09 2 32.43 25.95 0.31
Fluorene A3-v0 1.49E-08 1.40E-08 3.02E-08 7.90E-09 2 53.86 43.09 1.00
Pyrene A3-v0 2.25E-08 2.03E-08 4.15E-08 5.51E-09 2 64.16 51.34 1.00
Toluene A3-v1 7.01E-05 5.75E-05 1.65E-04 7.85E-06 2 89.58 71.67 0.98
D-55

Gas Turbines with No Duct Burners Firing Natural Gas Controlled with
Selective Catalytic Reduction and Carbon Monoxide Oxidation Catalyst

CARB tainty, Detect
Acrolein A3-v0 1.81E-02 1.81E-02 1.82E-02 1.79E-02 1 0.68 0.77 0.00
Benzene A3-v0 1.24E-02 1.23E-02 1.85E-02 6.96E-03 2 47.07 37.67 0.00
Chrysene A3-v0 5.08E-06 5.99E-06 6.34E-06 2.89E-06 2 31.73 25.39 0.31
Fluorene A3-v0 1.55E-05 1.42E-05 3.17E-05 8.05E-06 2 54.85 43.89 1.00
Pyrene A3-v0 2.31E-05 2.09E-05 4.23E-05 5.78E-06 2 63.17 50.55 1.00
Toluene A3-v1 7.17E-02 5.89E-02 1.68E-01 8.22E-03 2 88.92 71.15 0.98
D-56

Gas Turbines with No Duct Burners Firing Natural Gas Controlled with

CARB tainty, Detect
Acrolein A3-v0 1.72E-05 1.73E-05 1.73E-05 1.71E-05 1 0.68 0.77 0.00
Benzene A3-v0 1.20E-05 1.19E-05 1.82E-05 6.64E-06 2 48.32 38.66 0.00
Chrysene A3-v0 4.93E-09 5.88E-09 6.05E-09 2.76E-09 2 32.43 25.95 0.31
Fluorene A3-v0 1.49E-08 1.40E-08 3.02E-08 7.90E-09 2 53.86 43.09 1.00
Pyrene A3-v0 2.25E-08 2.03E-08 4.15E-08 5.51E-09 2 64.16 51.34 1.00
Toluene A3-v1 7.01E-05 5.75E-05 1.65E-04 7.85E-06 2 89.58 71.67 0.98
D-57
Table D-24a. Summary of Data for Emission Factor Development – Gas Turbines with Duct
Burners Firing Natural Gas Controlled with Selective Catalytic Reduction

CARB tainty, Detect
Formaldehyde C3-v0 6.22E+00 6.09E+00 6.87E+00 5.71E+00 1 9.57 10.83 1.00
Table D-24b. Summary of Data for Emission Factor Development – Gas Turbines with Duct
Burners Firing Natural Gas Controlled with Selective Catalytic Reduction

CARB tainty, Detect
D-58
Table D-25a. Summary of Data for Emission Factor Development – Gas Turbines with
Duct Burners Firing Natural Gas and Refinery Fuel Gas Controlled with

CARB tainty, Detect
Benzene C3-v0 1.66E-01 1.68E-01 1.81E-01 1.46E-01 2 9.66 7.73 0.00
Cadmium C3-v0 2.08E-03 1.06E-03 4.16E-03 1.03E-03 1 86.47 97.85 0.67
Chrysene C3-v0 1.16E-04 6.85E-05 2.43E-04 3.47E-05 1 96.95 109.70 1.00
Copper A3-v0 1.89E-03 1.98E-03 2.64E-03 1.04E-03 1 42.72 48.34 1.00
Fluorene C3-v1 1.90E-04 8.02E-05 4.46E-04 4.26E-05 1 117.53 132.99 1.00
Manganese A3-v0 3.49E-03 2.36E-03 6.79E-03 1.32E-03 1 83.26 94.21 1.00
Mercury A3-v0 4.63E-03 4.01E-03 8.27E-03 2.55E-03 2 51.31 41.05 1.00
Nickel C3-v0 8.18E-03 6.79E-03 1.43E-02 3.40E-03 1 68.51 77.52 1.00
Phenol C3-v0 1.53E-02 1.58E-02 1.59E-02 1.43E-02 1 5.81 6.57 0.00
Pyrene C3-v0 1.28E-04 1.04E-04 2.29E-04 5.21E-05 1 71.00 80.34 1.00
Toluene E3-v0 1.79E-01 1.78E-01 1.87E-01 1.72E-01 1 4.37 4.94 0.00
Xylene (Total) E3-v0 4.13E-01 4.14E-01 4.15E-01 4.11E-01 1 0.50 0.56 0.00
Zinc A3-v0 1.65E-02 1.56E-02 1.93E-02 1.45E-02 1 15.37 17.39 1.00
D-59
Table D-25b. Summary of Data for Emission Factor Development – Gas Turbines with
Duct Burners Firing Natural Gas and Refinery Fuel Gas Controlled with

CARB tainty, Detect
Benzene C3-v0 1.52E-04 1.54E-04 1.69E-04 1.32E-04 2 11.10 8.88 0.00
Cadmium C3-v0 1.94E-06 9.84E-07 3.87E-06 9.55E-07 1 86.47 97.85 0.67
Chrysene C3-v0 1.07E-07 6.37E-08 2.26E-07 3.23E-08 1 96.95 109.70 1.00
Copper A3-v0 1.75E-06 1.84E-06 2.46E-06 9.65E-07 1 42.72 48.34 1.00
Fluorene C3-v1 1.76E-07 7.46E-08 4.15E-07 3.96E-08 1 117.53 132.99 1.00
Manganese A3-v0 3.25E-06 2.19E-06 6.32E-06 1.23E-06 1 83.26 94.21 1.00
Mercury A3-v0 4.35E-06 3.73E-06 7.82E-06 2.37E-06 2 51.85 41.49 1.00
Nickel C3-v0 7.60E-06 6.32E-06 1.33E-05 3.16E-06 1 68.51 77.52 1.00
Phenol C3-v0 1.43E-05 1.47E-05 1.48E-05 1.33E-05 1 5.81 6.57 0.00
Pyrene C3-v0 1.19E-07 9.63E-08 2.13E-07 4.84E-08 1 71.00 80.34 1.00
Toluene E3-v0 1.62E-04 1.61E-04 1.69E-04 1.55E-04 1 4.30 4.87 0.00
Zinc A3-v0 1.53E-05 1.45E-05 1.80E-05 1.35E-05 1 15.37 17.39 1.00
D-60
Table D-26a. Summary of Data for Emission Factor Development – Gas Turbines with
Duct Burners Firing Natural Gas, Refinery Fuel Gas, and Liquefied Petroleum Gas
Controlled with Selective Catalytic Reduction and Carbon Monoxide Oxidation Catalyst

CARB tainty, Detect
Arsenic B3-v0 1.70E-02 1.71E-02 1.79E-02 1.61E-02 1 5.27 5.96 0.00
Cadmium B3-v0 7.41E-03 6.85E-03 8.94E-03 6.44E-03 1 18.06 20.43 0.40
Copper B3-v0 4.08E-02 1.88E-02 8.74E-02 1.61E-02 1 99.07 112.11 0.87
Lead B3-v0 6.82E-02 6.85E-02 7.16E-02 6.44E-02 1 5.25 5.94 0.00
Nickel B3-v0 2.78E-01 9.70E-02 6.60E-01 7.79E-02 1 118.86 134.50 1.00
Phenol C3-v1 5.80E-02 2.31E-02 1.45E-01 5.55E-03 1 131.32 148.60 1.00
Selenium B3-v0 1.70E-02 1.71E-02 1.79E-02 1.61E-02 1 5.27 5.96 0.00
Zinc B3-v0 4.12E-01 4.08E-01 6.80E-01 1.47E-01 1 64.69 73.20 1.00
Table D-26b. Summary of Data for Emission Factor Development – Gas Turbines with
Duct Burners Firing Natural Gas, Refinery Fuel Gas, and Liquefied Petroleum Gas
Controlled with Selective Catalytic Reduction and Carbon Monoxide Oxidation Catalyst

CARB tainty, Detect
Arsenic B3-v0 8.95E-06 9.00E-06 9.40E-06 8.46E-06 1 5.25 5.94 0.00
Cadmium B3-v0 3.89E-06 3.60E-06 4.70E-06 3.39E-06 1 18.07 20.45 0.40
Copper B3-v0 2.14E-05 9.89E-06 4.59E-05 8.46E-06 1 99.09 112.13 0.87
Lead B3-v0 3.58E-05 3.60E-05 3.76E-05 3.39E-05 1 5.23 5.92 0.00
Nickel B3-v0 1.46E-04 5.09E-05 3.47E-04 4.09E-05 1 118.88 134.52 1.00
Phenol C3-v1 3.05E-05 1.21E-05 7.65E-05 2.91E-06 1 131.46 148.76 1.00
Selenium B3-v0 8.95E-06 9.00E-06 9.40E-06 8.46E-06 1 5.25 5.94 0.00
Zinc B3-v0 2.16E-04 2.14E-04 3.58E-04 7.75E-05 1 64.70 73.22 1.00
D-61
Table D-27a. Summary of Data for Emission Factor Development – Gas Turbines with No Duct
Burners Firing Refinery Fuel Gas Controlled with Carbon Monoxide Oxidation Catalyst

CARB tainty, Detect
Arsenic B3-v0 4.09E-03 4.15E-03 4.18E-03 3.95E-03 1 3.09 3.49 0.00
Benzene C3-v0 1.49E-01 1.51E-01 1.53E-01 1.44E-01 1 3.34 3.78 0.00
Cadmium B3-v0 7.41E-03 5.21E-03 1.29E-02 4.15E-03 1 64.24 72.70 1.00
Copper B3-v0 5.78E-02 5.49E-02 9.72E-02 2.12E-02 1 65.90 74.57 1.00
Lead B3-v0 3.99E-02 4.15E-02 4.18E-02 3.64E-02 1 7.62 8.62 0.30
Mercury A3-v0 2.15E-02 1.55E-02 3.63E-02 1.27E-02 1 59.94 67.83 1.00
Nickel B3-v0 2.33E-01 2.64E-01 2.86E-01 1.48E-01 1 31.93 36.13 1.00
Phenol C3-v0 9.41E-03 5.99E-03 1.72E-02 5.07E-03 1 71.54 80.95 1.00
Selenium B3-v0 5.42E-03 4.18E-03 7.93E-03 4.15E-03 1 40.11 45.39 0.00
Toluene E3-v1 1.09E+00 4.34E-01 2.58E+00 2.56E-01 1 118.63 134.23 1.00
Xylene (Total) E3-v1 3.14E+00 3.39E-01 8.85E+00 2.33E-01 1 157.44 178.16 1.00
Zinc B3-v0 6.99E+00 8.23E+00 9.17E+00 3.56E+00 1 42.95 48.60 1.00
D-62
Table D-27b. Summary of Data for Emission Factor Development – Gas Turbines with No Duct
Burners Firing Refinery Fuel Gas Controlled with Carbon Monoxide Oxidation Catalyst

CARB tainty, Detect
Arsenic B3-v0 2.92E-06 2.96E-06 2.98E-06 2.81E-06 1 3.09 3.49 0.00
Benzene C3-v0 1.06E-04 1.07E-04 1.09E-04 1.02E-04 1 3.34 3.78 0.00
Cadmium B3-v0 5.28E-06 3.71E-06 9.18E-06 2.96E-06 1 64.24 72.70 1.00
Copper B3-v0 4.12E-05 3.91E-05 6.93E-05 1.51E-05 1 65.90 74.57 1.00
Lead B3-v0 2.84E-05 2.96E-05 2.98E-05 2.59E-05 1 7.62 8.62 0.30
Mercury A3-v0 1.53E-05 1.11E-05 2.59E-05 9.05E-06 1 59.94 67.83 1.00
Nickel B3-v0 1.66E-04 1.88E-04 2.04E-04 1.05E-04 1 31.93 36.13 1.00
Phenol C3-v0 6.71E-06 4.27E-06 1.22E-05 3.62E-06 1 71.54 80.95 1.00
Selenium B3-v0 3.86E-06 2.98E-06 5.65E-06 2.96E-06 1 40.11 45.39 0.00
Toluene E3-v1 7.77E-04 3.10E-04 1.84E-03 1.83E-04 1 118.63 134.23 1.00
Zinc B3-v0 4.98E-03 5.87E-03 6.54E-03 2.54E-03 1 42.95 48.60 1.00
D-63
D-64

Protocol Report 2015

Uploaded by

Copyright:

Available Formats

Protocol Report 2015

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Protocol Report 2015

Uploaded by

Copyright:

Available Formats

April 2015

Emissions Estimation Protocol

Office of Air Quality Planning and Standards

10. Fugitive Dust Sources ....................................................................................................................... 10-1

[This page intentionally left blank.]

List of Acronyms and Abbreviations

µg/L micrograms per liter

g/mol grams per mole

NMOC nonmethane organic compounds

TDS total dissolved solids

PM Emission Inventory Nomenclature

Table 1-1. Summary of Pollutants and Emission Sources Inclusion

Fugitive Dust Sources

630-08-0 Carbon monoxide ● ● ● ● ● ● ○ ○

Particulate matter (PM)

PM25-FIL Filterable PM ≤ 2.5 µm ● ● ● ● ● ● ● ● ● ○ ○

10102-44-0 Nitrogen dioxide ● ● ● ● ● ● ○ ○

NOX Nitrogen oxides ● ● ● ● ● ● ○ ○

7446-09-5 Sulfur dioxide ● ● ● ● ● ● ● ●

Table 1-1. Summary of Pollutants and Emission Sources Inclusion

Fugitive Dust Sources

142-82-5 Heptane (and isomers) ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○

111-65-9 Octane (and isomers) ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○

78-93-3 Methyl ethyl ketone ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○

Hazardous Air Pollutants (HAPs)

Volatile Organic HAPs

Table 1-1. Summary of Pollutants and Emission Sources Inclusion

Fugitive Dust Sources

108-10-1 Methyl isobutyl ketone ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

1634-04-4 Methyl tert-butyl ether ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○

593-60-2 Vinyl bromide ○ ○ ○ ○ ○

Table 1-1. Summary of Pollutants and Emission Sources Inclusion

Fugitive Dust Sources

75-01-4 Vinyl chloride ○ ○ ● ○ ○

1330-20-7 Xylenes (total) ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

83-32-9 Acenaphthene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●

208-96-8 Acenaphthylene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●

120-12-7 Anthracene POM

56-55-3 Benzo(a)anthracene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●

50-32-8 Benzo(a)pyrene POM

192-97-2 Benzo(e)pyrene POM

207-08-9 Benzo(k)fluoranthene POM

92-52-4 Biphenyl POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●

1319-77-3 Cresols (total) ● ● ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ● ● ●

218-01-9 Chrysene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●

84-74-2 di-n-Butyl phthalate ○ ● ● ○

Table 1-1. Summary of Pollutants and Emission Sources Inclusion

Fugitive Dust Sources

206-44-0 Fluoranthene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●

86-73-7 Fluorene POM

91-57-6 2-Methylnaphthalene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●

218-01-9 Chrysene POM

91-20-3 Naphthalene POM ● ● ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ● ○ ● ●

198-55-0 Perylene POM

85-01-8 Phenanthrene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●

129-00-0 Pyrene POM ● ○ ● ● ● ● ● ○ ○ ● ● ○ ● ● ● ● ○ ○ ● ●

1746-01-6 Dioxin: 4D 2378j ● ● ● ● ● ○ ○ ○

40321-76-4 Dioxin: 5D 12378 j

39227-28-6 Dioxin: 6D 123478j ○ ○ ○ ○ ○ ○ ○ ○

57653-85-7 Dioxin: 6D 123678 j