Draft Regulatory Impact Analysis:
Tier 3 Motor Vehicle Emission and
Fuel Standards
&EPA
United States
Environmental Protection
Agency
-------
Draft Regulatory Impact Analysis:
Tier 3 Motor Vehicle Emission and
Fuel Standards
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
NOTICE
This technical report does not necessarily represent final EPA decisions or
positions. It is intended to present technical analysis of issues using data
that are currently available. The purpose in the release of such reports is to
facilitate the exchange of technical information and to inform the public of
technical developments.
United States
Environmental Protection
Agency
EPA-420-D-13-002
March 2013
-------
Table of Contents
EXECUTIVE SUMMARY
CHAPTER 1 VEHICLE PROGRAM TECHNOLOGICAL FEASIBILITY
1.1 Introduction 1-1
1.2 FTP NMOG+NOx Feasibility 1-1
1.3 SFTP NMOG+NOx Feasibility 1-16
1.4 Technology Description for NMOG+NOX Control 1-18
1.5 PM Feasibility 1-30
1.6 Evaporative Emissions Feasibility 1-40
CHAPTER 2 VEHICLE PROGRAM COST AND EFFECTIVENESS
2.1 Vehicle Technology Costs 2-1
2.2 Vehicle Package Costs 2-17
2.3 Vehicle Program Costs 2-22
CHAPTER 3 ESTABLISHING NEW EMISSION TEST FUEL PARAMETERS
3.1 Assessment of Current Gasoline Properties 3-1
3.2 Projected E15 Implications 3-17
3.3 Proposed Gasoline Emission Test Fuel Specifications 3-19
3.4 Changes to ASTM Test Methods 3-20
CHAPTER 4 FUEL PROGRAM FEASIBILITY
4.1 Overview of Refining Operations 4-1
4.2 Feasibility of Removing Sulfur from Gasoline 4-4
4.3 Lead Time Assessment 4-27
CHAPTER 5 FUEL PROGRAM COSTS
5.1 Methodology 5-1
5.2 Estimated Tier 3 Sulfur Control Costs 5-44
5.3 Other Cost Studies 5-51
5.4 Projected Energy Impacts and Impacts on Permitting 5-59
11
-------
CHAPTER 6 HEALTH AND ENVIRONMENTAL EFFECTS ASSOCIATED
WITH EXPOSURE TO CRITERIA AND TOXIC POLLUTANTS
6.1 Health Effects of Criteria and Toxic Pollutants 6-1
6.2 Environmental Effects of Criteria and Toxic Pollutants 6-19
CHAPTER 7 IMPACTS OF THE PROPOSED RULE ON EMISSIONS AND AIR
QUALITY
7.1 Criteria and Toxic Pollutant Emission Impacts 7-1
7.2 Criteria and Toxic Pollutant Air Quality Impacts 7-46
7.3 Greenhouse Gas Emission Impacts 7-123
CHAPTER 8 COMPARISON OF PROGRAM COSTS TO PROGRAM
EMISSION REDUCTIONS AND AIR QUALITY BENEFITS
8.1 Cost-Benefit Analysis 8-1
8.2 Cost Per Ton of Emissions Reduced 8-39
CHAPTER 9 ECONOMIC IMPACT ANALYSIS
9.1 Introduction 9-1
9.2 Impacts on Vehicle Manufacturing Sector 9-1
9.3 Impacts on Petroleum Refinery Sector 9-7
CHAPTER 10 INITIAL REGULATORY FLEXIBILITY ANALYSIS
10.1 Introduction 10-1
10.2 Overview of the Regulatory Flexibility Act 10-1
10.3 Need for the Rulemaking and Rulemaking Objectives 10-2
10.4 Definition and Description of Small Entities 10-3
10.5 Summary of Small Entities to Which the Rulemaking Will Apply 10-4
10.6 Related Federal Rules 10-5
10.7 Projected Reporting, Recordkeeping, and Other Compliance Requirements 10-6
10.8 Regulatory Alternatives 10-6
10.9 Projected Economic Effects of the Proposed Rulemaking 10-16
in
-------
List of Acronyms
A/F
AAM
ABT
ACS
AGO
AHS
AIRS
AML
ANPRM
API
ASTM
bbl
BCA
BenMAP
BTU
C-R
CAA
CAP
CARB
CASAC
CBI
DF
CG
CMAQ
CML
CO
CO2
COI
COPD
cpsi
CR
CRC
CRDM
DMC
DOE
EO
E10
E15
ECA
EGR
EHC
EIA
EISA
EPA or Agency
air/fuel ratio
Alliance of Automobile Manufacturers
averaging, banking, and trading
American Cancer Society
atmospheric gasoil
U.S. Census Bureau's American Housing Survey
Aerometric Information Retrieval System
acute myeloid leukemia
Advanced Notice of Proposed Rulemaking
American Petroleum Institute
American Society for Testing and Materials
barrel
benefit-cost analysis
Environmental Benefits Mapping and Analysis Program
British Thermal Unit
concentration response
Clean Air Act
Compliance Assurance Program (2000)
California Air Resources Board
Clean Air Science Advisory Committee
confidential business information
Deterioration Factor
conventional gasoline
Community Multiscale Air Quality model
chronic myeloid leukemia
carbon monoxide
carbon dioxide
cost of illness
chronic obstructive pulmonary disease
cells per square inch
concentration-response
Coordinating Research Council
Climatological Regional Dispersion Model
direct manufacturing costs
U.S. Department of Energy
ethanol-free gasoline
gasoline containing 10 percent ethanol by volume
gasoline containing 15 percent ethanol by volume
Emission Control Area
exhaust gas recirculation
electrically heated catalyst
Energy Information Administration
Energy Independence and Security Act of 2007
U.S. Environmental Protection Agency
IV
-------
EPAct
ERIC
ESPN
EvOH
FBP
FCC
FTP
GC/MS
GDI
GDP
GPA
GVWR
HAP
HAPEM
HC
HCUP
HDGV
HDV
HEGO
HEI
I/M
IBP
ICI
ICM
IRFA
IRIS
LCO
LOT
LDV
LEV
LM
LML
LPG
MDPV
MECA
MLE
MRAD
MSAT
MSAT2
MSCF
MTBE
MY
NAAQS
NAC
NAICS
NAPAP
Energy Policy Act of 2005
Emissions Reduction and Intercept Control (system)
EPA speciation network
ethyl vinyl alcohol
final boiling point
fluidized catalytic cracker
Federal Test Procedure
gas chromatography/mass spectrometry
gasoline direct injection
gross domestic product
Geographic Phase-in Area
gross vehicle weight rating
Hazardous Air Pollutant
Hazardous Air Pollutant Exposure Model
hydrocarbon
Healthcare Cost and Utilization Program
heavy-duty gasoline vehicle
heavy-duty vehicle
heated exhaust gas oxygen (sensor)
Health Effects Institute
inspection/maintenance
initial boiling point
independent commercial importer
indirect cost multiplier
initial regulatory flexibility analysis
Integrated Risk Information System
light cycle oil
light-duty truck
light-duty vehicle
low emission vehicle
locomotive and marine diesel fuel
lowest measured level
liquid petroleum gas
medium-duty passenger vehicle
Manufacturers of Emission Controls Association
maximum likelihood estimate
minor restricted activity days
mobile source air toxic
Regulations for Control of Hazardous Air Pollutants from Mobile
Sources, 72 FR 8428, 2/26/07
thousand standard cubic feet
methyl tertiary-butyl ether
model year
National Ambient Air Quality Standards
NOx adsorption catalyst
North American Industrial Classification System
National Acid Precipitation Assessment Program
-------
NATA
NEMA
NESHAP
NFRAQS
NGL
NLEV
NMHC
NMMAPS
NMOG
NO2
NOX
NPC
NPRA
NPRM
NRC
NSR
OAQPS
OAR
OBD
OC/EC
OMB
QMS
ORNL
OSC
OSTP
OTAG
PADD
PAN
PCM
Pd
PFI
PGM
PM
PNGV
POM
ppm
PSD
Pt
R+M/2
R&D
REL
RFA
RfC
RfD
RFG
RFS2
National-Scale Air Toxics Assessment
Northeast Mid-Atlantic
National Emissions Standards for Hazardous Air Pollutants
Northern Front Range Air Quality Study
natural gas liquids
national low emission vehicle
non-methane hydrocarbons
National Morbidity, Mortality, and Air Pollution Study
non-methane organic gases
nitrogen dioxide
oxides of nitrogen
National Petroleum Council
National Petrochemical & Refiners Association
Notice of Proposed Rulemaking
National Research Council
New Source Review
Office of Air Quality Planning and Standards
EPA's Office of Air and Radiation
on-board diagnostics
organic carbon/elemental carbon
Office of Management and Budget
Office of Mobile Sources
Oak Ridge National Laboratory
oxygen storage components
(White House) Office of Science and Technology Policy
Ozone Transport Assessment Group
Petroleum Administrative Districts for Defense
peroxy acetyl nitrate
powertrain control module
palladium
port fuel injection
platinum group metals
particulate matter
Partnership for a New Generation of Vehicles
polycyclic organic matter
part per million
Prevention of Significant Deterioration
platinum
average octane, or antiknock index
research and development
reference exposure level
Regulatory Flexibility Act
reference concentration
reference dose
reformulated gasoline
Renewable Fuel Standard Program, 75 FR 14670, 3/26/2010
VI
-------
Rh
ROI
ROTR
RPE
RRF
RVP
S-R
S&P DRI
SAB
SBA
SBARP or the Panel
SBREFA
SCR
SER
SFTP
SIC
SIGMA
SIP
SMAT
SO2
SOX
SRU
SULEV
SVM
SVOC
SwRI
T10
T50
T90
TC
TGDI
THC
TOG
TW
UAM
UCL
UEGO
ULEV
UMRA
UV
VGO
VTB
VMT
VNA
voc
VSL
rhodium
return on investment
Regional Ozone Transport Rule
retail price equivalent
relative reduction factor
Reid vapor pressure
Source-Receptor Matrix
Standard & Poor's Data Research International
Science Advisory Board
U.S. Small Business Administration
Small Business Advocacy Review Panel
Small Business Regulatory Enforcement Fairness Act
selective catalytic reduction
Small Entity Representative
Supplemental Federal Test Procedure
Standard Industrial Classification
Society of Independent Gasoline Marketers of America
State Implementation Plan
Speciated Modeled Attainment Test
sulfur dioxide
oxides of sulfur
sulfur recovery unit
super ultra low emission vehicle
small volume manufacturer (of vehicles)
semivolatile organic compound
Southwest Research Institute
average temperature at which 10 percent of gasoline is distilled
average temperature at which 50 percent of gasoline is distilled
average temperature at which 90 percent of gasoline is distilled
total technology costs
turbocharged gasoline direct injection
total hydrocarbons
total organic gases
test weight
Urban Airshed Model
upper confidence limit
universal exhaust gas oxygen (sensor)
ultra low emission vehicle
Unfunded Mandates Reform Act
ultra violet
vacuum gasoil
vacuum tower bottoms
vehicle miles traveled
Voronoi Neighbor Averaging
volatile organic compound
value of a statistical life
vn
-------
WLD work loss days
WTP willingness to pay
Vlll
-------
This Page Intentionally Left Blank
IX
-------
Executive Summary
EPA is proposing a comprehensive program to address air pollution from passenger cars
and trucks. The proposed program, known as "Tier 3," would establish more stringent vehicle
emissions standards and reduce the sulfur content of gasoline, considering the vehicle and its fuel
as an integrated system The proposed Tier 3 standards would reduce levels of multiple air
pollutants (ambient levels of ozone, particulate matter (PM), nitrogen dioxide (NO2), and mobile
source air toxics (MSATs)) across the country and help state and local agencies in their efforts to
attain and maintain health-based National Ambient Air Quality Standards (NAAQS).
This Regulatory Impact Analysis provides technical, economic, and environmental
analyses of the proposed new standards. Chapter 1 contains our technical feasibility justification
for the proposed vehicle emission standards, and Chapter 2 contains the estimated costs of the
proposed vehicle standards. In addition to the vehicle emission and gasoline standards, we are
proposing to update the specifications of the emission test fuel with which vehicles demonstrate
compliance with emissions standards; our analysis of the proposed emission test fuel parameter
changes is found in Chapter 3. Chapters 4 and 5 contain our technical feasibility and cost
analyses for the proposed gasoline sulfur standards, respectively. Chapter 6 describes the health
and welfare effects associated with the air pollutants that would be impacted by the rule.
Chapter 7 describes our analysis of the emission and air quality impacts of the Tier 3 rule. Our
estimates of the program-wide costs, the societal benefits, and the cost per ton of emissions
reduced due to the proposed Tier 3 program are presented in Chapter 8. Chapter 9 contains our
analysis of the proposed rule's economic impacts, and Chapter 10 provides the results of our
small business flexibility analysis.
Proposed Tier 3 Standards
Vehicle Emission Standards
The proposed Tier 3 standards include light- and heavy-duty vehicle tailpipe emission
standards and evaporative emission standards.
Light-Duty Vehicle, Light-Duty Truck, and Medium-Duty Passenger Vehicle Tailpipe
Emission Standards
The proposed standards in this category would apply to all light-duty vehicles (LDVs, or
passenger cars), light-duty trucks (LDTls, LDT2s, LDT3s, and LDT4s) and Medium-Duty
Passenger Vehicles, or MDPVs. We are proposing new standards for the sum of NMOG and
NOx emissions, presented as NMOG+NOx, and for PM. For these pollutants, we are proposing
standards as measured on test procedures that represent a range of vehicle operation, including
the Federal Test Procedure (or FTP, simulating typical driving) and the Supplemental Federal
Test Procedure (or SFTP, a composite test simulating higher temperatures, higher speeds, and
quicker accelerations).
ES-1
-------
The proposed FTP and SFTP NMOG+NOx standards would be fleet-average standards,
meaning that a manufacturer would calculate the weighted average emissions of the vehicles it
sells in each model year and compare that average to the applicable standard for that model year.
The proposed fleet average standards for NMOG+NOx evaluated over the FTP would begin in
MY 2017 and then decline through MY 2025, as summarized in Table ES-1. Similarly, the
proposed NMOG+NOx standards measured over the SFTP would also be fleet-average
standards, declining from MY 2017 until MY 2025, as shown in Table ES-2.
Table ES-1 Proposed LDV, LDT, and MDPV Fleet Average NMOG+NOx FTP Standards
(mg/mi)
LDV/LDTlb
LDT2,3,4 and
MDPV
Model Year
2017a
86
101
2018
79
92
2019
72
83
2020
65
74
2021
58
65
2022
51
56
2023
44
47
2024
37
38
2025
and later
30
30
" For vehicles above 6000 Ibs GVWR, the fleet average standards would apply beginning in MY 2018
6 These proposed standards would apply for a 150,000 mile useful life. Manufacturers could choose to certify
their LDVs and LDVls to a useful life of 120,000 miles. If any of these families are certified to the shorter
useful life, a proportionally lower numerical fleet-average standard would apply, calculated by multiplying the
respective 150,000 mile standard by 0.85 and rounding to the nearest mg.
Table ES-2 Proposed LDV, LDT, and MDPV Fleet-Average NMOG+NOX SFTP Fleet
Average Standards (mg/mi)
NMOG + NOX
Model Year
2017a
103
2018
97
2019
90
2020
83
2021
77
2022
70
2023
63
2024
57
2025
and later
50
a Forvehicles above 6000 Ibs GVWR, the fleet average standards would apply beginning in MY 2018.
The proposed PM standard on the FTP for certification testing is 3 mg/mi for all vehicles
and for all model years. Manufacturers could phase in their vehicle models as a percent of sales
through MY 2022. The proposed FTP PM standards would apply to each vehicle separately (i.e.,
not as a fleet average). The proposed program also includes a separate FTP PM requirement of 6
mg/mi for the testing of in-use vehicles that would apply during the percent phase-in period only.
Table ES-3 presents the FTP certification and in-use PM standards and the phase-in percentages.
ES-2
-------
Table ES-3 Phase-In for Proposed PM Standards
Phase-In
(percent of U.S. sales)
Certification Standard
(mg/mi)
In-Use Standard
(mg/mi)
2017a
20
O
6
2018
20
3
6
2019
40
O
6
2020
70
3
6
2021
100
O
6
2022 and
later
100
3
3
a Forvehicles above 6000 Ibs GVWR, the proposed FTP PM standards would apply beginning in MY 2018.
The proposed Tier 3 program also includes certification PM standards evaluated over the
SFTP (specifically the US06 component of the SFTP procedure) at a level of 10 mg/mi for
lighter vehicles and 20 mg/mi for heavier vehicles. As with the FTP PM standard, we propose
separate in-use US06 PM standards during the percent phase-in only of 15 and 25 mg/mi for cars
and trucks, respectively. The US06 PM standards would also phase in on the same schedule as
the FTP PM standards.
Heavy-Duty Tailpipe Emission Standards
We are proposing Tier 3 exhaust emissions standards for complete heavy-duty vehicles
(HDVs) between 8,501 and 14,000 Ib GVWR. Vehicles in this GVWR range are often referred
to as Class 2b (8,501-10,000 Ib) and Class 3 (10,001-14,000 Ib) vehicles, and are typically full-
size pickup trucks and work vans. The key elements of these proposed standards include a
combined NMOG+NOx declining fleet average standard, new stringent PM standards phasing in
on a separate schedule, extension of the regulatory useful life to 150,000 miles, and a new
requirement to meet standards over the SFTP that would address real-world driving modes not
well-represented by the FTP cycle alone. Table ES-4 presents the proposed HDV fleet average
NMOG+NOx standard, which becomes more stringent in successive model years from 2018 to
2022, with voluntary standards available in 2016 and 2017.
The proposed PM standards are 8 mg/mi for Class 2b vehicles and 10 mg/mi for Class 3
vehicles, to be phased in on a percent-of-sales basis at 20-40-70-100 percent in 2018-2019-2020-
2021, respectively.
Table ES-4 Proposed HDV Fleet Average NMOG+NOx Standards (mg/mi)
Model Year
Class 2b
Class 3
Voluntary
2016
333
548
2017
310
508
Required Program
2018
278
451
2019
253
400
2020
228
349
2021
203
298
2022 and later
178
247
The proposed new SFTP requirements for FtDVs include NMOG+NOx, carbon
monoxide (CO) and PM standards. Compliance would be evaluated from a weighted composite
of measured emissions from testing over the FTP cycle, the SC03 cycle, and an aggressive
ES-3
-------
driving cycle, with the latter tailored to various HDV sub-categories: the US06 cycle for most
HDVs, the highway portion of the US06 cycle for low power-to-weight Class 2b HDVs, and the
LA-92 cycle for Class 3 HDVs.
Evaporative Emission Standards
To control evaporative emissions, EPA is proposing more stringent standards that would
require covered vehicles to have essentially zero fuel vapor emissions in use, including more
stringent evaporative emissions standards, new test procedures, and a new fuel/evaporative
system leak emission standard. The Tier 3 proposal also includes refueling emission standards
for a portion of heavy-duty gasoline vehicles (HDGVs) over 10,000 Ibs GVWR. EPA is
proposing phase-in flexibilities as well as credit and allowance programs. The proposed
standards, harmonized with California's zero evaporative emissions standards, are designed to
essentially eliminate fuel vapor-related evaporative emissions.
Table ES-5 presents the proposed evaporative hot soak plus diurnal emission standards
by vehicle class. Manufacturers may comply on average within each of the four vehicle
categories but not across these categories. The proposal also includes separate high altitude
emission standards for these vehicle categories.
Table ES-5 Proposed Evaporative Emission Standards (g/test)
Vehicle Category
LDV, LDT1
LDT2
LDT3, LDT4, MDPV
HDGVs
Highest Diurnal + Hot Soak Level
(over both 2-day and 3 -day diurnal tests)
0.300
0.400
0.500
0.600
EPA is proposing a new testing requirement referred to as the bleed emission test
procedure. Under the proposal, manufacturers would be required to measure diurnal emissions
over the 2-day diurnal test procedure from just the fuel tank and the evaporative emission
canister and comply with a 0.020 g/test standard for all LDVs, LDTs, and MDPVs without
averaging. The canister bleed emission standard test would apply only for low altitude testing
conditions, but EPA expects proportional control at higher altitudes.
EPA is proposing to include these Tier 3 evaporative emission controls for HDGVs as
part of the overall scheme for LDVs and LDTs. The individual vehicle emission standard would
be 0.600 g/test for both the 2-day and 3-day evaporative emission tests, the high altitude standard
would be 1.75 g/test and the canister bleed test standard would be 0.030 g/test.
We are also proposing to add a new emission standard and test procedure related to
controlling vapor leaks from vehicle fuel and vapor control systems. The standard, which would
apply to all LDVs, LDTs, MDPVs, and Class 2b/3 HDGVs, would prohibit leaks larger than 0.02
inches of cumulative equivalent diameter in the fuel/evaporative system. The proposed Tier 3
evaporative emission standards program requirements would be phased in over a period of six
ES-4
-------
model years between MYs 2017 and 2022, with the leak test phasing in beginning in 2018 MY
as a vehicle is certified to meet Tier 3 evaporative emission requirements.
EPA is proposing new refueling emission control requirements for HDGVs equal to or
less than 14,000 Ibs GVWR (i.e., Class 2b/3 HDGVs), starting in the 2018 model year. Under
this proposal, EPA would extend current refueling emission control requirements for Class 2b
HDGVs to Class 3 HDGVs.
We are also proposing to adopt and incorporate by reference the current CARB onboard
diagnostic system (OBD) regulations effective for the 2017 MY plus two minor provisions to
enable OBD-based leak detection to be used in IUVP testing. EPA would retain the provision
that certifying with CARB's program would permit manufacturers to seek a separate EPA
certificate on that basis.
Emissions Test Fuel Requirements
We are proposing several changes to our federal gasoline emissions test fuel. Key
changes include:
• Moving away from "indolene" (EO) to a test fuel containing 15 percent ethanol by
volume (El5);
• Lowering octane to match regular-grade gasoline (except for premium-required
vehicles);
• Adjusting distillation temperatures, aromatics and olefms to better match today's in-
use fuel and to be consistent with anticipated El5 composition; and
• Lowering the existing sulfur specification and setting a benzene specification to be
consistent with proposed Tier 3 gasoline sulfur requirements and recent MSAT2
gasoline benzene requirements.
Gasoline Sulfur Standards
Under the Tier 3 fuel program, we are proposing that federal gasoline contain no more
than 10 parts per million (ppm) sulfur on an annual average basis by January 1, 2017. We are
proposing an averaging, banking, and trading (ABT) program that would allow refiners and
importers to spread out their investments through an early credit program and rely on ongoing
nationwide averaging to meet the 10-ppm sulfur standard. We are also proposing a three-year
delay for small refiners and "small volume refineries" processing less than or equal to 75,000
barrels of crude oil per day. In addition, we also proposing to either maintain the current 80-ppm
refinery gate and 95-ppm downstream per-gallon caps or lower them to 50 and 65 ppm,
respectively. A summary of the proposed Tier 3 sulfur standards is provided in Table ES-6.
ES-5
-------
Table ES-6 Proposed Tier 3 Gasoline Sulfur Standards
Proposed Tier 3 Gasoline Sulfur
Standards
Refinery annual average standard
Refinery gate per-gallon cap
Downstream per-gallon cap
Cap Option 1
Limit
10 ppm
80 ppm
95 ppm
Effective
January 1, 2017a
Already
Already
Cap Option 2
Limit
10 ppm
50 ppm
65 ppm
Effective
January 1, 2017a
January 1, 2020
March 1, 2020
"Effective January 1, 2020 for eligible small refiners and small volume refineries.
Projected Impacts
Emission and Air Quality Impacts
The proposed Tier 3 vehicle and fuel-related standards would together reduce emissions
of NOx, VOC, PM2.5, and air toxics. The gasoline sulfur standards, which would take effect in
2017, would provide large immediate reductions in emissions from existing gasoline vehicles
and engines. The emission reductions would increase over time as newer vehicles become a
larger percentage of the fleet, e.g., in 2030, when 80 percent of the light-duty fleet (and 90
percent of the vehicle miles travelled) consists of Tier 3 vehicles. Projected emission reductions
from the Tier 3 standards for 2017 and 2030 are shown in Table ES-7. We expect these
reductions to continue beyond 2030 as more of the fleet continues to turn over to Tier 3 vehicles.
Table ES-7 Estimated Emission Reductions from the Proposed Tier 3 Standards
(Annual U.S. short tons)"
NOX
VOC
CO
Direct PM2.5
Benzene
SO2
1,3-Butadiene
Formaldehyde
Acetaldehyde
Acrolein
Ethanol
2017
Tons
284,381
44,782
746,683
121
1,625
16,261
322
727
762
23
2,684
Percent of Onroad
Inventory
8%
3%
4%
0.1%
4%
51%
5%
3%
3%
1%
2%
2030
Tons
524,790
226,028
5,765,362
7,458
8,582
17,267
1,087
2,707
4,414
184
27,821
Percent of Onroad
Inventory
28%
23%
30%
10%
36%
51%
37%
12%
26%
15%
23%
a This analysis assumed emissions reductions from the Tier 3 vehicle standards would occur in all states. For
the final rule we will account for LEV III vehicle standards in states that have subsequently adopted it.
Reductions in emissions of NOx, VOC, PM2.5 and air toxics are projected to lead to
nationwide decreases in ambient concentrations of ozone, PM2.5, NO2, CO, and air toxics.
Specifically, the proposed Tier 3 standards would significantly decrease ozone concentrations
ES-6
-------
across the country, with a population-weighted average decrease of 0.47 ppb in 2017 and 1.55
ppb in 2030. The magnitude of reductions is significant enough to bring ozone levels in some
areas from above the standard to below the standard, even without any additional controls. Few
other strategies exist that would deliver the reductions needed for states to meet the current
ozone standards. The proposed Tier 3 standards would decrease ambient annual PM2.5
concentrations across the county as well, with a population-weighted average decrease of 0.06
|ig/m3 by 2030. Decreases in ambient concentrations of air toxics are also projected with the
proposed standards, including notable nationwide reductions in benzene concentrations.
Costs and Benefits
The costs that would be incurred from our proposed program fall into two categories -
costs from the Tier 3 vehicle exhaust and evaporative standards and from reductions in sulfur
content of gasoline. All costs represent the fleet-weighted average of light-duty vehicles and
trucks. All costs are represented in 2010 dollars.
Vehicle Costs
The vehicle costs include the technology costs projected to meet the proposed exhaust
and evaporative standards, as show in Table ES-8. The fleet mix of light-duty vehicles, light
duty trucks, and medium-duty trucks represents the 2016 MY fleet used in the 2012-2016 MY
light-duty GHG final rulemaking.
Table ES-8 Annual Vehicle Technology Costs, 2010$
Year
2017
2030
Vehicle Exhaust
Emission Control
Costs
(SMillion)
$634
$1,790
Vehicle Evaporative
Emission Control
Costs
($Million)
$71
$253
Facility Costs
($Million)
$4
$4
Total Vehicle
Costs
($Million)a
$709
$2,050
a These estimates include costs associated with the proposed Tier 3 vehicle standards in all
states except California.
Fuel Costs
The fuel costs consist of the additional operating costs and capital costs to the refiners to
meet the proposed sulfur average of 10 ppm. The sulfur control costs assume a cost of 0.89 cents
per gallon which includes the refinery operating and capital costs. The annual fuel costs of the
proposed program are listed in Table ES-9.
ES-7
-------
Table ES-9 Annual Fuel Costs, 2010$
Year
2017
2030
Fuel Sulfur Control Costs ($Million)a
$1,289
$1,320
a These estimates include costs associated with the
proposed Tier 3 fuel standards in all states except
California.
Total Costs
The sum of the vehicle technology costs to control exhaust and evaporative emissions, in
addition to the costs to control the sulfur level in the fuel, represent the total costs of the
proposed program, as shown in Table ES-10. The proposed fuel standards are projected to lead
to an average cost of 0.89 cents per gallon of gasoline, and the proposed vehicle standards would
have an average cost of $134 per vehicle
Table ES-10: Total Annual Vehicle and Fuel Control Costs, 2010$
Year
2017
2030
Total Vehicle and Fuel Control Costs
($Million)a
$1,999
$3,367
a These estimates include costs associated with the
proposed Tier 3 vehicle and fuel standards in all states
except California.
Cost Per Ton of Emissions Reduced
We have calculated the aggregate cost per ton of the emissions reduced due to the
proposed program using the projected costs and emission reductions. Note that, even though we
are setting new standards for PM, we believe that those standards will be met in complying with
the NMOG+NOx standards with additional care being given to proper engineering/calibration, so
there is no cost associated with the new PM standard and therefore no separate cost per ton of
emissions reduced analysis for PM.
The total program costs, NOx+VOC reductions, and results of our cost per ton of
emissions reduced analysis are provided in Table ES-11. The costs of the proposed program
would be higher immediately after it is implemented than they would be after several years, since
both vehicle manufacturers and refiners can take advantage of decreasing capital and operating
costs over time. In addition, the reductions in NOx and VOC emissions would become greater as
a larger percentage of the fleet contains the technologies required to meet the proposed
standards.
ES-8
-------
Table ES-11 Cost Per Ton of Emissions Reduced in 2017 and 2030
Year
2017
2030
Total Proposed Program Cost
(Smillion, 2010$)
$1,999
$3,367
Total NOX + VOC
Reductions (tons)
329,162
750,818
Cost Per Ton of
Emissions Reduced
($/ton)
$6,072
$4,484
Benefits
Exposure to ambient concentrations of ozone, PM2.5, and air toxics is linked to adverse
human health impacts such as premature deaths as well as other important public health and
environmental effects. The proposed Tier 3 standards would reduce these adverse impacts and
yield significant benefits, including those we can monetize and those we are unable to quantify.
The range of quantified and monetized benefits associated with this program are
estimated based on the risk of several sources of PM- and ozone-related mortality effect
estimates, along with other PM and ozone non-mortality related benefits information. Overall,
we estimate that the proposed rule would lead to a net decrease in PM2.5- and ozone-related
health and environmental impacts. The range of total monetized ozone- and PM-related health
impacts is presented in Table ES-12.
Table ES-12: Estimated 2030 Monetized PM-and Ozone-Related Health Benefits"
Description
Total Estimated Health Benefitsb'°'d'e't
3 percent discount rate
7 percent discount rate
2030
$8.0 - $23
$7.4 - $21
Notes:
a Totals are rounded to two significant digits and may not sum due to rounding.
b The benefits presented in this table have been adjusted to remove emission reductions attributed to the
Tier 3 program in California. We will account for emissions in states that have adopted California's LEV
III program in the baseline air quality modeling for the final rule.
0 Total includes ozone and PM2.5 benefits. Range was developed by adding the estimate from the Bell et
al., 2004 ozone premature mortality function to PM2 5-related premature mortality derived from the
American Cancer Society cohort study (Pope et al., 2002) for the low estimate and ozone premature
mortality derived from the Levy et al., 2005 study to PM2 5-related premature mortality derived from the
Six-Cities (Laden et al., 2006) study for the high estimate.
d Annual benefits analysis results reflect the use of a 3 percent and 7 percent discount rate in the valuation
of premature mortality and nonfatal myocardial infarctions, consistent with EPA and OMB guidelines for
preparing economic analyses.
e Valuation of premature mortality based on long-term PM exposure assumes discounting over the SAB
recommended 20-year segmented lag structure described in the Regulatory Impact Analysis for the 2006
PM National Ambient Air Quality Standards (September, 2006).
f Not all possible benefits are quantified and monetized in this analysis; the total monetized benefits
presented here may therefore be underestimated.
We estimate that by 2030, the annual emission reductions of the Tier 3 standards would
annually prevent between 670 and 1,700 PM-related premature deaths, between 160 and 710
ozone-related premature deaths, 81,000 work days lost, and approximately 1.4 million minor
ES-9
-------
restricted-activity days. The estimated annual monetized health benefits of the proposed Tier 3
standards in 2030 (2010$) would be between $8.0 and $23 billion, assuming a 3-percent discount
rate (or between $7.4 billion and $21 billion assuming a 7-percent discount rate).
Note that the air quality modeling conducted for the Tier 3 program included emission
reductions both in California (which was recently granted a waiver for the adoption of its LEV
III program) and in several other states that have adopted the LEV III program under Section 177
of the Clean Air Act. As a result, the benefits cited here have been adjusted to remove emission
reductions attributed to the Tier 3 program in California. We will account for emissions in states
that have adopted California's LEV III program in the baseline air quality modeling for the final
rule. Refer to Chapter 8 for more information about the benefits estimated for the proposal.
Comparison of Costs and Benefits
The estimated annual monetized health benefits of the proposed Tier 3 standards in 2030
(2010$) would be between $8.0 and $23 billion, assuming a 3 percent discount rate (or between
$7.4 billion and $21 billion assuming a 7 percent discount rate). The annual benefits of the Tier
3 standards outweigh the annual cost of the overall program in 2030, which would be
approximately $3.4 billion.
Economic Impact Analysis
The proposed rule will affect two sectors directly: vehicle manufacturing and petroleum
refining. The estimated increase in vehicle production cost because of the proposed rule is
expected to be small relative to the costs of the vehicle. Some or all of this production cost
increase would be expected to be passed through to consumers. This increase in price is
expected to lower the quantity of vehicles sold, though because the expected cost increase is
small, we expect the decrease in sales to be negligible. This decrease in vehicle sales is expected
to decrease employment in the vehicle manufacturing sector. However, costs related to
compliance with the rule should also increase employment in this sector. While it is unclear
which of these effects will be larger, because the increase in vehicle production costs and the
decrease in vehicle sales are minor, the impact of the rule on employment in the vehicle
manufacturing sector is expected to be small as well. The key change for refiners from the
proposed standards would be more stringent sulfur requirements. Analogous to vehicle sales,
this change to fuels is expected to increase manufacturers' costs of fuel production. Some or all
of this increase in production costs is expected to be passed through to consumers which should
lead to a decrease in fuel sales. As with the vehicle manufacturing sector, we would expect the
decrease in fuel sales to negatively affect employment in this sector, while the costs of
compliance with the rule would be expected to increase employment. It is not evident whether
the proposed rule would increase or decrease employment in the refining sector as a whole.
However, given the small anticipated increase in production costs of less than one cent per gallon
and the small likely decrease in fuel sales, we expect that the rule would not have major
employment consequences for this sector.
ES-10
-------
Chapter 1 Vehicle Program Technological Feasibility
1.1 Introduction
For the vehicles and emissions addressed in this proposed rule, EPA has comprehensively
assessed the technological phenomena related to the generation of emissions of interest, the
nature of the technological challenges facing manufacturers to produce emission reductions of
the scale proposed in the Preamble, and the technologies that we expect be available to
manufacturers to meet those challenges in the time frame of this NPRM. Our feasibility
assessment recognizes that the proposed Tier 3 program is composed of several new
requirements for all types of new vehicles, including everything from small cars to large pick-up
trucks and MDPVs with diverse applications and specific engine designs. This assessment also
recognizes the critical role of gasoline sulfur content in making it possible for us to propose
emission standards at these very stringent levels. We provide below a full assessment of our
current knowledge of the effects of gasoline sulfur on current vehicle emissions as well as our
projections of how sulfur can be expected to affect compliance with the proposed Tier 3
standards.
Since there are multiple aspects to the Tier 3 program, it is necessary to consider
technical feasibility in light of the different program requirements and their interactions with
each other. For example, the technical feasibility of the proposed FTP NMOG+NOx and the PM
standards is directly related to the specifications of the fuel, including fuel sulfur, RVP and
ethanol content. Additionally, as mentioned above, the feasibility assessment must consider that
different technologies may be needed on different types of vehicle applications (i.e., cars versus
trucks) and must consider the effectiveness of these technologies to reduce emissions for the full
useful life of the vehicle while operating on in-use fuels. Certain smaller vehicles with
correspondingly small engines may be less challenged to meet FTP standards than larger
vehicles with larger engines. Conversely, these smaller vehicles may have more difficulty
meeting the SFTP requirements than the larger and more powerful vehicles. Additionally, the
ability to meet the proposed SFTP emission requirements can also be impacted by the path taken
to meet the FTP requirements (i.e., larger volume catalysts for US06 emissions control vs.
smaller catalysts for improved FTP cold-start emissions control).
1.2 FTP NMOG+NOx Feasibility
The proposed new emission requirements include stringent NMOG+ NOx standards over
the FTP that would require new vehicle hardware and additional control of gasoline sulfur levels
in order to achieve the 30 mg/mi fleet average level in 2025. The type of new hardware that
would be required would vary depending on the specific application and emission challenges.
Smaller vehicles with corresponding smaller engines would generally need less new hardware
while larger vehicles and other vehicles with larger engines may need additional hardware and
improvements beyond what would be needed for the smaller vehicles. Additionally, the fleet-
average nature of the standards would allow more challenged vehicles to be offset by vehicles
that could outperform the required fleet averages.
1-1
-------
In order to assess the technical feasibility of a 30 mg/mi NMOG+NOx national fleet
average FTP standard and a 3 mg/mi PM vehicle standard, EPA conducted two primary analyses.
The initial analyses performed were of the current Tier 2 and LEV II fleets. This provided a
baseline for the current federal fleet emissions performance, as well as the emissions
performance of the California LEV II fleet. The second consideration was a modal analysis of
typical vehicle emissions under certain operating conditions. In this way EPA determined the
specific emissions performance challenges that vehicle manufacturers would face in meeting the
lower fleet average emission standards. Each of these considerations is described in greater
detail below.
1.2.1 Assessment of the Current Federal Fleet Emissions
The current federal fleet is certified to an average of Tier 2 Bin 5, equivalent to 160
mg/mi NMOG+NOx A For MY 2009, 92 percent of passenger cars and LDTls were certified to
Tier 2 Bin 5 and 91 percent of LDT2s through LDT4s were certified to Tier 2 Bin 5. This was
not an unexpected result as there is currently no motivation for vehicle manufacturers to produce
a federal fleet that over-complies with respect to the current Tier 2 standards. By comparison, in
the MY 2009 California fleet, where compliance with the "PZEV" program encourages
manufacturers to certify to cleaner levels, only 30 percent of the passenger cars and LDTl's are
certified to Tier 2 Bin 5 and 60 percent are certified to Tier 2 Bin 3. The situation regarding the
truck fleet in California is similarly stratified, with 37 percent of the LDT2s through LDT4s
being certified to Tier 2 Bin 5 and 55 percent being certified to Tier 2 Bin 3. In many cases
vehicles in hardware and calibration are being certified to a lower standard in California and a
higher standard federally. Although we recognize that there are relatively minor differences
between the Tier 2 and LEV II programs that might affect manufacturers' selection of which bins
to choose for federal and California certification in some cases (i.e., certification fuel
differences), EPA believes that the patterns described above indicate that much of the existing
Tier 2 fleet could be certified to a lower federal fleet average immediately, with no major
feasibility concerns.6
To support this conclusion, we analyzed the measured emissions data for NMOG and
NOx from MY 2009 certifications. The results of our analysis are shown in Table 1-1 and Table
1-2 below, listing vehicles that perform at 30 mg/mi NMOG+NOx and cleaner, and 50 mg/mi
NMOG+NOx and cleaner, respectively. The vehicles shown in Table 1-1 are mainly PZEV
vehicles that are sold in California and Section 177 states. Although the vehicles presented in
Table 1-1 represent only a small fraction of the federal fleet, they demonstrate that the Tier 3
standards we are proposing are feasible today for a range of vehicle sizes and types. In addition,
one can note that although their NOx performance is equal to the NOx performance of several of
the vehicles shown in Table 1-2, their respective NMOG performance is substantially lower.
This supports our earlier statements that manufacturers would most likely be required to focus on
A The current Tier 2 program does not combine NMOG and NOX emissions into one fleet-average standard. The
fleet-average standard in that program is for NOX emissions alone.
B Compliance with full useful life standards in California occurs at much lower in-use gasoline sulfur levels than is
the case with federally certified vehicles. For further discussion of the impact of gasoline fuel sulfur on light-vehicle
emissions feasibility and in-use compliance, please refer to Section 1.2.3.1.
1-2
-------
NMOG to meet the proposed Tier 3 FTP standards. The vehicles shown in Table 1-2
demonstrate a significant range of vehicle types and sizes that are currently performing at 50
mg/mi NMOG+NOx or cleaner. This is approximately 30 percent of the current Tier 2 fleet
standard. However, the majority of these vehicles are currently being certified to Tier 2 Bin 5.
Table 1-1: 2009 MY Engine Families with One or More Vehicle Configurations with Full
Useful Life NMOG+NOx Certification Levels at or below 30 mg/mi
Manufacturer
AUDI
BUICK
CHEVROLET
CHEVROLET
FORD
HONDA
HONDA
MERCEDES-
BENZ
MERCURY
SUBARU
VOLKSWAGEN
Models
A3
LUCERNE
COBALT
MALIBU
FOCUS FWD
CIVIC HYBRID
ACCORD 4DR
SEDAN
E350 4MATIC
(WAGON)
MARINER
HYBRID 4WD
LEGACY
WAGON AWD
JETTA
SPORTWAGEN
NMOG
Level
(mg/mi)
5.2
8
6
8
6
4
6
2.2
10
8.4
4.3
NOX
Level
(mg/mi)
11
10
0
10
0
10
10
24
20
12
0
NMOG+NOx
(mg/mi)
16.2
18
6
18
6
14
16
26.2
30
20.4
4.3
Federal
Certification
Bin 2
Bin5
Bin5
BinS
Bin 3
Bin 2
Bin 2
BinS
Bin 3
BinS
Bin 2
Table 1-2: 2009 MY Engine Families with One or More Vehicle Configurations with Full
Useful Life NMOG+NOx Certification Levels at or below 50 mg/mi
Manufacturer
ACURA
ACURA
ACURA
AUDI
BENTLEY
BUICK
CHEVROLET
CHEVROLET
CHEVROLET
CHRYSLER
FORD
FORD
Models
RDX4WD
TSX
TL4WD
S4 CABRIOLET
CONTFNENTAL
GTC
ENCLAVE AWD
MALIBU
IMP ALA
AVEO5
PTCRUISER
CONV/CABRIO
ESCAPE 4WD
FUSION FWD
NMOG Level
(mg/mi)
19
24
18
27
20
32
30
26
36
35
25
19
NOX Level
(mg/mi)
30
20
20
20
21
10
20
10
10
10
20
30
NMOG+NOx
(mg/mi)
49
44
38
47
41
42
50
36
46
45
45
49
Federal
Certification
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 4
Bin 4
Bin 5
Bin 4
Bin 5
1-3
-------
Manufacturer
FORD
HONDA
HONDA
HONDA
HONDA
HYUNDAI
MOTOR
COMPANY
HYUNDAI
MOTOR
COMPANY
HYUNDAI
MOTOR
COMPANY
HYUNDAI
MOTOR
COMPANY
HYUNDAI
MOTOR
COMPANY
HYUNDAI
MOTOR
COMPANY
HYUNDAI
MOTOR
COMPANY
ISUZU
JEEP
KIA MOTORS
CORPORATION
KIA MOTORS
CORPORATION
KIA MOTORS
CORPORATION
KIA MOTORS
CORPORATION
KIA MOTORS
CORPORATION
KIA MOTORS
CORPORATION
KIA MOTORS
CORPORATION
Models
MUSTANG
CR-V 4WD
RIDGELINE 4WD
FIT
ACCORD 4DR
SEDAN
SANTA FE 2WD
SANTA FE 4WD
TUCSON 2WD
ACCENT
ELANTRA
SONATA
GENESIS
ASCENDER 5-
PASSENGER
4WD
WRANGLER
4WD
SPORTAGE 2WD
SPORTAGE 4WD
SORENTO 2WD
RIO
SPECTRA
OPTIMA
RONDO
NMOG Level
(mg/mi)
30
20
22
21
19
30
23.6
23.7
30.7
16.1
28.9
29.9
36
27
23.7
18.9
28
23.3
19.4
26.6
33.5
NOX Level
(mg/mi)
20
30
20
10
20
14
11
20
13
27
20
14
10
10
23
23
10
17
22
21
16
NMOG+NOX
(mg/mi)
50
50
42
31
39
44
34.6
43.7
43.7
43.1
48.9
43.9
46
37
46.7
41.9
38
40.3
41.4
47.6
49.5
Federal
Certification
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 4
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
1-4
-------
Manufacturer
LAND ROVER
LTD
LINCOLN
LINCOLN-
MERCURY
MAZDA
MAZDA
MERCEDES-
BENZ
MERCEDES-
BENZ
MITSUBISHI
NISSAN
NISSAN
PORSCHE
SAAB
SATURN
Shelby
Automobiles
Incorporated
Models
LR3
NAVIGATOR
4WD FFV
TOWN CAR
MAZDAS
MAZDA 6
ML63 AMG
G55 AMG
ECLIPSE
SPYDER
MAXIMA
ALTIMA
CAYENNE
TRANSSYBERIA
9-7X AWD
ASTRA 4DR
HATCHBACK
SHELBY
MUSTANG
GT500KR
NMOG Level
(mg/mi)
12
28
29
28.2
21
9.9
12.8
29
20
30
28
36
25
41
NOX Level
(mg/mi)
30
20
10
21
20
35
36
16
20
20
20
10
20
9
NMOG+NOX
(mg/mi)
42
48
39
49.2
41
44.9
48.8
45
40
50
48
46
45
50
Federal
Certification
Bin 5
Bin 8
Bin 4
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 5
Bin 4
Bin 5
Bin 5
To support the FTP emission levels proposed for heavy duty vehicles, we analyzed the
certification emission results from the 2010 and 2011 MY vehicles. The proposed fleet average
NMOG+NOx standard in 2022 for Class 2b vehicles is 178 mg/mi while the proposed level for
Class 3 vehicles is 247 mg/mi. Shown in Table 1-3 below are the emission levels of 2010 and
2011 MY heavy-duty vehicle models operating on various fuels. It is important to note that while
we are proposing a useful life of 150,000 miles, the current heavy duty vehicle requirements and
therefore the reported emission results represent the 120,000 miles deteriorated results either
calculated using deterioration factors applied to the 4,000 mile test or actual aged vehicles and
components. It will be important for manufacturers to carefully manage emissions deterioration
throughout the useful life of the vehicle to meet useful life emission requirements, consistent
with the challenge for light-duty applications.
1-5
-------
Table 1-3: 2010/11 MY Heavy Duty Vehicle FTP 120k Certification Results
Class 2b
b
Class 3
Manufacturer
Altech
Chrysler
Chrysler
Ford
General
Motors
Baytech
Chrysler
Chrysler
Ford
Ford
Ford
Ford
Ford
General
Motors
General
Motors
Models
F250
Ram 2500
Ram 2500
F250
Silverado
2500
Silverado
3500
Ram 3500
Ram 3500
E350C
E350C
F350
F350
F350
Silverado
3500
Silverado
3500
Fuel
Type
CNG
Gasoline
Diesel
Diesel
Diesel
CNG
Gasoline
Diesel
Gasoline
E85
Diesel
Gasoline
E85
Gasoline
Diesel
NMOG
Level a
(mg/mi)
10
101
82
53
94
11
119
52
51
70
51
79
76
131
76
NOX
Level
(mg/mi)
100
60
200
180
166
100
120
400
82
65
320
130
83
150
184
NMOG
+NOX
(mg/mi)
110
161
282
233
260
111
239
452
133
135
371
209
159
281
260
CO
(g/mi)
5.9
2.4
.1
.6
1
1.3
3.6
.2
2
1.1
.6
3.2
1.8
3.4
.8
PM
(mg/
mi)
-
-
3
10
0
-
-
3
-
-
10
-
-
-
0
Notes:
"Diesel reported as NMHC
* Gasoline Class 2b models from General Motors and Ford certified using worst case Class 3 data
c Tested at LVW with MDPVs
1.2.2 NMOG and NOX Emissions on the FTP
To understand how the several currently-used technologies described below could be
used by manufacturers to reach the stringent proposed Tier 3 NMOG+NOx standards, it is
helpful to consider emissions formation in common modes of operation for gasoline engines, or
/-i
modal analysis. As previously stated during the discussion of the NMOG+NOx standard, many
gasoline engines produce very little NOx over the FTP. Thus, the challenge faced by
manufacturers for producing Tier 3 compliant gasoline powertrains would be to reduce the
NMOG portion of the emissions. Based on modal analysis of a gasoline powered vehicle being
operated on the FTP cycle, approximately 90 percent of the NMOG emissions occur during the
first 50 seconds after a cold start. In addition, about 60 percent of the NOx emissions occur in
these early seconds. Figure 1-1 below shows the second-by-second emissions for NMOG and
NOx following a cold start.
c A modal analysis provides a second-by-second view of the total amount of emissions over the entire cycle being
considered.
1-6
-------
Thus, effective control of these cold-start emissions, especially NMOG emissions, would
be the primary technological goal of manufacturers complying with the proposed Tier 3 FTP
standards. As discussed below, manufacturers are already applying several technologies capable
of significant reductions in these cold start emission to vehicles currently on the road.
Cumulative NOx + THC (Normalized)
Time (sec)
Figure 1-1 Modal Analysis of NMOG and NOX Emissions (LA92)
1.2.3 Compliance Margin
Vehicle manufacturers generally design vehicles to meet emissions targets which are 50-
70 percent of the emission standards after the catalytic converters have been thermally aged and
exposed to catalyst poisons (e.g., sulfur from fuel, phosphorous from lubricating oil, etc.) out to
the vehicle's full useful life. This difference is referred to as "compliance margin" and is a result
of manufacturers' efforts to address all the sources of variability and emissions control system
degradation that could occur during the certification or in-use testing processes and during in-use
operation. Thus, the emission design targets for Tier 3 standards at full useful life would be
approximately 15 mg/mi MOG+NOx for a Bin 30 certified vehicle. These sources of variability
include:
• Test-to-test variability (within one test site and lab-to-lab)
• Build variation and manufacturing tolerances
• Vehicle operation (for example: driving habits, ambient temperature, etc.)
• Fuel composition
o The deleterious effects of fuel sulfur on exhaust catalysts and oxygen
sensors
o Other fuel composition impacts
• Oil consumption
o The impact of direct emission of lubricating oil on semi-volatile organic
PM emissions
1-7
-------
o The impact of oil additives and other components (e.g., phosphorous) and
oil ash on exhaust catalysts and oxygen sensors
For MY 2009, the compliance margin for a Tier 2 Bin 5 vehicle averaged approximately
60 percent. In other words, actual vehicle emissions performance was on average about 40
percent of a 160 mg/mi NMOG+NOx standard, or about 64 mg/mi. By comparison, for MY
2009 California-certified vehicles, the average SULEV compliance margin was somewhat less
for the more stringent standards, approximately 50 percent. We believe that the recent California
experience is a likely indicator of compliance margins that manufacturers would design for in
order to comply with the proposed Tier 3 FTP standards. Thus, a typical Tier 2 Bin 5 vehicle,
performing at 40 percent of the current standard (i.e., at about 64 mg/mi) would need
improvements sufficient to reach about 15 mg/mi (50 percent of a 30 mg/mi standards).
1.2.3.1 Impact of Gasoline Sulfur Control on the Feasibility of the Proposed
Vehicle Emission Standards
1.2.3.1.1 Fuel Sulfur Impacts on Exhaust Catalysts
Modern three-way catalytic exhaust systems utilize platinum group metals, metal oxides
and other active materials to selectively oxidize organic compounds and carbon monoxide in the
exhaust gases. These systems simultaneously reduce nitrogen oxides when air-to-fuel ratio
control operates in a condition of relatively low amplitude/high frequency oscillation about the
stoichiometric point. Sulfur is a well-known catalyst poison. There is a large body of work
demonstrating sulfur inhibition of the emissions control performance of platinum group metals
(PGM) three-way exhaust catalyst systems.1'2'3'4'5'6'7'8'9'10 The nature of sulfur interactions with
active catalytic materials, catalyst washcoating materials, and catalyst substrate materials is
complex and varies with catalyst composition and exhaust gas composition and exhaust
temperature. The variation of these interactions with exhaust gas composition and temperature
means that the operational history of a vehicle is an important factor; continuous light-load
operation, throttle tip-in events and enrichment under high-load conditions can all impact sulfur
interactions with the catalyst.
Sulfur from gasoline is oxidized during spark-ignition engine combustion primarily to
SC>2 and, to a much lesser extent, S(V2. Sulfur oxides selectively chemically bind (chemisorb)
with, and in some cases react with, active sites and coating materials within the catalyst, thus
inhibiting the intended catalytic reactions. Sulfur oxides inhibit pollutant catalysis chiefly by
selective poisoning of active PGM, ceria sites, and alumina washcoatings (see Figure 1-2) .u The
amount of sulfur retained by the catalyst is primarily a function of its operating temperature, the
active materials and coatings used within the catalyst, the concentration of sulfur oxides in the
incoming exhaust gases, and air-to-fuel ratio feedback and control by the engine management
system.
1-8
-------
Active Catalytic Site
Catalytic Site Deactivated by Sulfur Poisoning
Alumina Washcoat
Monolith Substrate
U- 100A -»)
•
•-•» « » « »
* « • •
^ »
Figure 1-2 Functional schematic showing selective poisoning of active catalytic sites by
sulfur compounds. Adapted from Heck and Farrauto 2002.11
Selective sulfur poisoning of platinum (Pt) and rhodium (Rh) is primarily from surface-
layer chemisorption. Sulfur poisoning of palladium (Pd) and ceria appears to be via
chemisorption combined with formation of more stable metallic sulfur compounds, e.g. PdS and
€62628, present in both surface and bulk form (i.e., below the surface layer).4'5'12'13 Ceria,
zirconia and other oxygen storage components (OSC) play an important role that is crucial to
NOx reduction over Rh as the engine air-to-fuel ratio oscillates about the stoichiometric closed-
loop control point.14 Water-gas-shift reactions are important for NOx reduction over catalysts
combining Pd and ceria. This reaction can be blocked by sulfur poisoning and may be
responsible for observations of reduced NOx activity over Pd/ceria catalysts even with exposure
919
to fairly low levels of sulfur (equivalent to 15 ppm in gasoline). ' Pd is also of increased
importance for meeting Tier 3 standards due to its unique application in the closed-coupled-
catalysts location required for vehicles certifying to very stringent emission standards. Pd is
required in closed-coupled catalysts due to its resistance to high temperature thermal sintering.
Sulfur removal from Pd requires rich operation at higher temperatures than required for sulfur
removal from other PGM catalysts.2
In addition to its interaction with catalyst materials, sulfur can also react with the
washcoating itself to form alumina sulfate, which in turn can block coating pores and reduce
gaseous diffusion to active materials below the coating surface.11 This may be a significant
mechanism for the observed storage of sulfur compounds at light and moderate load operation
with subsequent, rapid release as sulfate particulate matter when high-load, high-temperature
conditions are encountered.15
1-9
-------
Operating the catalyst at a sufficiently high temperature under net reducing conditions
(e.g., air-to-fuel equivalence that is net fuel-rich of stoichiometry) can effectively release the
sulfur oxides from the catalyst components. Thus, regular operation at sufficiently high
temperatures at rich air-to-fuel ratios can minimize the effects of fuel sulfur levels on catalyst
active materials and catalyst efficiency. However, it cannot completely eliminate the effects of
sulfur poisoning. Furthermore, regular operation at these temperatures and at rich air-to-fuel
ratios is not desirable, for several reasons. The temperatures necessary to release sulfur oxides
are high enough to lead to thermal degradation of the catalyst over time via thermal sintering of
active materials. Sintering reduces the surface area available to participate in reactions.
Additionally, it is not always possible to maintain these catalyst temperatures (because of cold
weather, idle conditions, light load operation) and the rich air-to-fuel ratios necessary can result
in increased PM, NMOG and CO emissions. Thus, reducing fuel sulfur levels has been the
primary regulatory mechanism to minimize sulfur contamination of the catalyst and ensure
optimum emissions performance over the useful life of a vehicle.
The impact of gasoline sulfur has become even more important as vehicle emission
standards have become more stringent. Some studies have suggested an increase in catalyst
sensitivity to sulfur (in terms of percent conversion efficiency) when standards increase in
stringency and emissions levels decrease. Emission standards under the programs that preceded
the Tier 2 program (Tier 0, Tier 1 and National LEV, or NLEV) were high enough that the
impact of sulfur was considered negligible. The Tier 2 program recognized the importance of
sulfur and reduced the sulfur levels in the fuel from 300 ppm to 30 ppm in conjunction with the
new emission standards.16 At that time, very little work had been done to evaluate the effect of
further reductions in fuel sulfur - especially on in-use vehicles that may have some degree of
catalyst deterioration due to real-world operation.
In 2005 EPA and several automakers jointly conducted a program that examined the
effects of sulfur and other gasoline properties, benzene, and volatility on emissions from a fleet
of nine Tier 2 compliant vehicles, the "MS AT (Mobile Source Air Toxics) Study.17 Reductions
for FTP-weighted emissions for the sulfur changes in this program were 33 percent for NOx, 11
percent for THC, 17 percent for CO, and 32 percent for methane. Given the prep procedures
related to catalyst clean-out and loading, these results may represent a "best case" scenario that
magnifies what would be expected under more typical driving conditions. Nonetheless, these
data suggested the effect of sulfur loading was reversible for Tier 2 vehicles, and that there were
likely to be significant emission reductions possible with further reductions in gasoline sulfur
level. For more discussion of the impact of gasoline fuel sulfur on the current light-duty vehicle
fleet, see Chapter 7 of this draft RIA.
1.2.3.2 EPA Tier 2 In-Use Gasoline Sulfur Effects Study
Goals of this study included assessment of reversible sulfur loading in catalysts of Tier 2
compliant light duty gasoline vehicles in the in-use fleet, as well as characterization of the effects
of fuel sulfur level on emissions as a function of accumulated mileage. The study sample
consisted of 81 cars and light trucks recruited from owners in southeast Michigan, covering
model years 2007-9 with approximately 20,000-40,000 odometer miles. The makes and models
targeted for recruitment were chosen to be representative of high sales vehicles covering a range
of types and sizes. Test fuels were two non-ethanol gasolines with properties typical of
1-10
-------
certification fuel, one at a sulfur level of 5 ppm and the other at 28 ppm. A nominal
concentration of approximately 30 ppm was targeted for the high level to be representative of
retail fuel available to the public in the vehicle recruiting area. All emissions data were collected
using the FTP cycle at a nominal temperature of 75°F.
After vehicles were received at the test facility, the high-sulfur test fuel was installed and
a set of emission tests were performed to establish a baseline representative of the vehicle's as-
received state. A high-speed/load "clean-out" procedure consisting of two back-to-back US06
cycles was then performed with the intent to produce a reduction in catalyst sulfur loading. This
was followed by another set of emission tests to assess the change in emissions performance
from baseline. A statistical analysis of this data showed highly significant reductions in several
pollutants including NOx and hydrocarbons (Table 1-4), suggesting that reversible sulfur loading
exists in the in-use fleet and has a measurable effect on aftertreatment performance.
Table 1-4: Average clean-out effect on in-use emissions using 28 ppm test fuel
Bagl
Bag 2
Bag3
FTP Composite
Bag 1 - Bag 3
NOX
(p-value)
31.9%
(0.0009)
38.3%
(O.OOOl)
11.4%
(O.OOOl)
"
THC
(p-value)
16.5%
(0.0024)
21.4%
(O.OOOl)
4.1%
(0.0187)
"
CO
(p-value)
4.7%
(0.0737)
"
19.5%
(0.0011)
7.6%
(0.0008)
4.2%
(0.0714)
NMHC
(p-value)
17.8%
(0.0181)
27.8%
(O.OOOl)
3.0%
(0.0751)
"
CH4
(p-value)
15.3%
(0.0015)
12.0%
(O.OOOl)
6.9%
(0.0003)
"
PM
(p-value)
15.4%
(< 0.0001)
"
24.5%
(O.OOOl)
13.7%
(O.OOOl)
"
Note:
The clean-out effect is not significant at a = 0.10 when no reduction estimate is provided.
Next, a subset of approximately one in five vehicles (one of each make/model) was kept
for an extended test schedule consisting of additional emission test replicates alternated with
mileage accumulation. This dataset was used to assess the behavior of emissions as sulfur
reloaded toward and beyond the baseline level observed in the vehicles as-received. The fuel
was then changed to the low-sulfur test fuel and the procedure repeated, starting with a clean-out
procedure followed by alternating emission tests and mileage accumulation. This dataset was
used to assess differences in the rate at which sulfur reloading occurred as a function of fuel
sulfur level.
Comparing results of emission tests immediately following (<50 miles) the clean-out
procedures at each fuel sulfur level shows a statistically significant difference of 35 percent for
Bag 2 NOx. This suggests that the effectiveness of a high-temperature regeneration cycle in
reducing emissions is limited by fuel sulfur level. Moreover, it also suggests that the clean-out
1-11
-------
effects shown in Table 1-4 would likely be larger if the low-sulfur test fuel had been used for the
cleanout procedure and tests immediately following the as-received baseline emissions.
Mixed model analysis of all emissions data as a function of fuel sulfur level and miles
driven after cleanout found highly significant model fits for several pollutants. These results
show cold-start and hot-running NOx emissions were reduced by 11 percent and 59 percent,
respectively, comparing low vs. high-sulfur test fuels. In these particular results, the model
fitting did not find a miles-by-sulfur interaction to be significant, suggesting the relative
differences were not dependent on miles driven after cleanout, confirming the substantial
magnitude of what was found for tests immediately following the clean-out procedures, as
described above. Other results, such as FTP-composite NOx, as well as cold-start and FTP-
composite hydrocarbons, did show a significant miles-by-sulfur interaction. In this case,
determining a sulfur level effect for the in-use fleet requires estimation of a miles-equivalent
level of sulfur loading, which can be gleaned from the cleanout results obtained from the
baseline testing on the vehicles as-received. Figure 1-3 shows this graphically for FTP-
composite NOx. Other mixed model results are summarized in Table 1-5.
0.025
0.02
X 0.015
O
OJ
0.005
In-use equivalent loading
Pre-cleanout
(as-received level)
In-use sulfur level effect
•High Sulfur
•Low Sulfur
clean-out
50 100 150 200
Miles since clean-out procedure
Figure 1-3: Estimation of sulfur level effect on FTP-composite NOx in in-use fleet
1-12
-------
Table 1-5: Summary of mixed model results for emission reductions from 28 to 5 ppm
sulfur, adjusted for in-use sulfur loading where appropriate
Bagl
Bag 2
Bag3
FTP
Composite
Bag 1 - Bag 3
NOX
(p-value)
10.7%
(0.0033)
59.2%
(< 0.0001)
62.1%
(< 0.0001)
23.0%t
(0.0180)
-t
THC
(p-value)
8.5%t
(0.0382)
48.8%
(< 0.0001)
40.2%
(< 0.0001)
13.0%t
(0.0027)
5.2%
(0.0063)
CO
(p-value)
7.5%t
(0.0552)
-t
20.1%
(< 0.0001)
11.9%t
(0.0378)
4.3%
(0.0689)
NMHC
(p-value)
7.5%
(< 0.0001)
44.8%t
(0.0260)
49.9%
(< 0.0001)
10.6%t
(0.0032)
5.1%
(0.0107)
CH4
(p-value)
13.9%t
(< 0.0001)
49.9%
(< 0.0001)
29.2%
(< 0.0001)
25.8%t
(< 0.0001)
4.6%
(0.0514)
NOx+NMOG
(p-value)
N/A
N/A
N/A
17.3%
(0.0140)
N/A
PMJ
—
—
—
—
—
Notes:
t Model with significant sulfur and mileage interaction term.
* Sulfur level not significant at a = 0.10. For THC Bag 1 and CH4 Bag 1, because the effect of clean-out was not
statistically significant, the reduction estimates are based on the estimates of least squares means.
Major findings from this study include:
• Reversible sulfur loading is occurring in the in-use fleet of Tier 2 vehicles and has a
measureable effect on emissions of NOx, hydrocarbons, and other pollutants of interest
• The effectiveness of high speed/load procedures in restoring catalyst efficiency is a
function of fuel sulfur level
• Reducing fuel sulfur levels from 28 to 5 ppm is likely to achieve significant reductions in
emissions of NOx, hydrocarbons, and other pollutants of interest in the in-use fleet
Findings of this study are in general agreement with data presented in the Alliance of
Automobile Manufacturers National Clean Gasoline Study published in 2009.18 Section 3 of that
document presents analysis of the Tech 5 vehicle class portion of the California Air Resources
Board Predictive Model, as well as the MS AT test program conducted jointly by EPA and
several automakers in 2005-6. Figure 6 of the Alliance study shows changes in FTP-composite
NOx emissions as a function of sulfur as predicted by these two data sources. The results of this
study are bracketed by the Tech 5 results lower in magnitude, and the MSAT data as higher in
magnitude. This seems reasonable given the fact that Tech 5 vehicles are compliant with a lower
NOx standard than the Tier 2 vehicles used in this study, and thus sulfur loading would be
expected to have lower leverage on emission performance, and the fact that the MSAT test
program made a comparison between emissions data for the low sulfur fuel collected
immediately after a more aggressive cleanout procedure than used in this study (CRC E-60) and
data from the high sulfur fuel collected after a sulfur loading procedure consisting of > 100 miles
of low-speed cruising.
1-13
-------
Figure 5 in Section 3 of the Alliance study shows hydrocarbon emissions (as THC) from
the same data sources. The results of this study show a slightly larger emission effect than
predicted by the MSAT program results; a possible explanation for this is that the larger database
in this study allowed the model to discern a clearer effect for this pollutant (whose sulfur effect
appears to be smaller and more variable across vehicle types than for NOx). In any case, both
studies showed statistically significant results in the same direction and of similar magnitude.
1.2.3.3 Fuel Sulfur Impacts on Vehicles at the Proposed Tier 3 Levels
The Tier 3 Program would reduce fleet average NMOG+NOx emissions by over 80
percent. The feasibility of the proposed 30 mg/mi NMOG+NOx fleet average standard depends
on a degree of emissions control from exhaust catalyst systems that will require gasoline at 10
ppm sulfur or lower. The most likely control strategies will involve using exhaust catalyst
technologies and powertrain calibration to reduce NOx emissions to near-zero levels. This
would allow sufficient NMOG compliance margin to meet the combined NMOG+NOx
emissions standards for the full useful life.
Achieving the proposed Tier 3 emission standards would require very careful control of
the exhaust chemistry and exhaust temperatures to ensure high catalyst efficiency. The impact of
sulfur on OSC in the catalyst makes this a challenge even at relatively low (10 ppm) gasoline
sulfur levels.
Light-duty vehicles certified to CARB SULEV and Federal Tier 2 Bin 2 exhaust
emission standards accounted for approximately 3.5% and 1%, respectively of vehicle sales for
MY2009. Nonhybrid vehicles certified in California as SULEV are typically not certified to
Federal Tier 2 Bin 2 emissions standards even though the numeric limits for NOx and NMOG
are shared between the California LEV II and Federal Tier 2 programs for SULEV and Bin 2.
Confidential business information shared by the auto companies indicate that the primary reason
is an inability to demonstrate compliance with SULEV/Bin 2 emission standards after vehicles
have operated in-use on gasoline with greater than 10 ppm sulfur and with exposure to gasoline
up to the Tier 2 80-ppm gasoline sulfur cap. While vehicles certified to the SULEV and Tier 2
Bin 2 standards both demonstrate compliance using 15-40 ppm certification gasoline, in-use
compliance of SULEV vehicles in California occurs after operation on gasoline with an average
of 10 ppm sulfur and a 30-ppm sulfur cap while federally certified vehicles operate on gasoline
with an average of 30 ppm sulfur and a maximum cap of 80 ppm sulfur. Although the SULEV
and Tier 2 Bin 2 standards are numerically equivalent, the increased sulfur exposure of in-use
vehicles certified under the Federal Tier 2 program results in certification of California SULEV
vehicles to emissions standards under the Federal Tier 2 program that are typically 1-2
certification bins higher (e.g., SULEV certified as Tier 2 Bin 3 or Bin 4) in order to ensure in-use
compliance with emissions standards out to the full useful life of the vehicle when operating on
higher-sulfur gasoline.
Emissions of vehicles certified to the SULEV standard of the California LEV II program,
or the equivalent Tier 2 Bin 2 standards, can provide some insight into the impact of fuel sulfur
on vehicles at the very low proposed Tier 3 emissions levels. Vehicle testing by Toyota of LEV
I, LEV IIULEV and prototype SULEV vehicles showed larger percentage increases in NOx and
HC emissions for SULEV vehicles as gasoline sulfur increased from 8 ppm to 30 ppm, as
1-14
-------
compared to other LEV vehicles they tested. Testing of a SULEV-certified PZEV vehicle by
Umicore showed a pronounced, progressive trend of increasing NOx emissions (referred to as
"NOx creep") when switching from a 3 ppm sulfur gasoline to repeated, back-to-back FTP tests
using 33 ppm sulfur gasoline.10 The PZEV Chevrolet Malibu, after being aged to an equivalent
of 150,000 miles, demonstrated emissions at a level equivalent to the compliance margin for the
Tier 3 Bin 30 NMOG+NOx standard when operated on 3 ppm sulfur fuel and for at least one
FTP test after switching to 33 ppm certification fuel. Following operation over 2 FTP cycles on
33 ppm sulfur fuel, NOx emissions alone were more than double the proposed Tier 330 mg/mi
NMOG+NOx standard.19 This represents a NOx percentage increase that is approximately 2-3
times of what has been reported for similar changes in fuel sulfur level for Tier 2 and older
90 91
vehicles over a similar difference in fuel sulfur. '
Although there are no vehicles larger than LDT2 and no non-HEV vehicles above the
LDV D-segment size, we expect that additional catalyst technologies, for example increasing
catalyst surface area (volume or substrate cell density) and/or increased PGM loading, would
need to be applied to larger vehicles in order to achieve the catalyst efficiencies necessary to
comply with the proposed Tier 3 standards. Any sulfur impact on catalyst efficiency would have
a larger impact on vehicles and trucks that rely more on very high catalyst efficiencies in order to
achieve very low emissions.
The negative impact of gasoline sulfur on NOx, NMOG and air toxic emissions occurs
across all vehicle categories. However, the impact of gasoline sulfur on NOx emissions control
of catalysts in the fully-warmed-up condition is particularly of concern for larger vehicles (E-
and F-segment LDVs, LDT3, LDT4 trucks, and MDPVs). Manufacturers face the most
significant challenges in reducing cold-start NMOG emissions for these vehicles. Because of the
need to reach near-zero NOx levels, any significant degradation in NOx emissions control over
the useful life of the vehicle would likely prevent some if not most larger vehicles from reaching
a combined NMOG+NOx low enough to comply with the 30 mg/mi fleet-average standard .
These vehicles represent a sufficiently large segment of light-duty vehicle sales now and in the
foreseeable future that their emissions could not be offset (and thus the fleet-average standard
achieved) by certifying vehicles to bins below the fleet average. Any degradation in catalyst
performance due to gasoline sulfur would reduce or eliminate the margin necessary to ensure in-
use compliance with the proposed Tier 3 emissions standards. Certifying to a useful life of
150,000 miles vs. the current 120,000 miles would further add to manufacturers' compliance
challenge for Tier 3 large light trucks (see Figure 1-6). (See Section IV.7.b in the preamble for
more on the proposed useful life requirements.)
1.2.3.4 Gasoline Sulfur Control Required to Meet Tier 3 Emissions Standards
The impact of gasoline sulfur poisoning on exhaust catalyst performance and the relative
stringency of the Tier 3 standards, particularly for larger vehicles and trucks, when considered
together make a compelling argument for the virtual elimination of sulfur from gasoline. The
10-ppm standard for sulfur in gasoline represents the lowest practical limit from a standpoint of
fuel handling and transport. A gasoline sulfur standard of 10 ppm also represents the highest
level of gasoline fuel sulfur that will allow compliance with a national fleet average of 30 mg/mi
NMOG+NOx.
1-15
-------
1.3 SFTP NMOG+NOx Feasibility
The proposed new emission requirements include stringent NMOG+ NOx composite
standards over the SFTP that would generally only require additional focus on fuel control of the
engines and diligent implementation of new technologies like gasoline direct injection (GDI) and
turbocharged engines. Additionally, the fleet-average nature of the standards would allow more
challenged vehicles to be offset by vehicles that could outperform the required fleet averages.
In order to assess the technical feasibility of a 50 mg/mi NMOG+NOx national fleet
average SFTP composite standard EPA conducted an analysis of data from the in-use
verification program (IUVP). The IUVP vehicles are tested by manufacturers at various ages and
mileages and the results are reported to EPA. The analysis was performed on Tier 2 and LEV II
vehicles. This provided a baseline for the current federal fleet emissions performance, as well as
the emissions performance of the California LEV II fleet.
1.3.1 Assessment of the Current Federal Fleet Emissions
To investigate feasibility, we acquired and analyzed certification results for model years
2010 and 2011. These data included FTP composites, as well as results for the US06, and SC03
cycles. We focused on results for hydrocarbons (HC) and NOx. For the FTP results HC
represents non-methane organic gases (NMOG). The US06 and SC03 results represent
NMHC+NOx.
As a first step, we averaged the results by model year and test group (engine family).
After compiling results on all three cycles for each test group, we calculated SFTP composite
estimates for each engine family as
SFTP - 0.35 • (¥TPNOx + FTIVOG) + 0.28 • US06 + 0.37 • SC03
As a second step, we then averaged the SFTP composite results by standard level and
vehicle class, focusing on results in Bins 2, 3 and 5, as well as vehicles certified to LEV-II LEV
and SULEV standards. In averaging, we treated Bin 2 and LEV-II/SULEV standards as
equivalent, and accordingly, pooled their results. Table 1-6 shows the numbers of test groups in
each combination of standard level and vehicle class.
Table 1-6: Numbers of Test Groups Certified to Selected Tier-2 and LEV-II Standards in
Model Years 2010 and 2011
Standard Level
Bin 2 + LEV-II/SULEV
Bin3
Bin5
LEV-II/LEV
Vehicle Class
LDV-LDTl
88
26
331
124
LDT2
3
1
37
17
LDT3
13
4
LDT4
1
14
4
Figure 1-4 shows results for Bin-5 and LEV-II/LEV vehicles. It is clear that vehicles in
all four vehicle classes, from LDV to LDT4, are certified to these standards. The means show a
1-16
-------
modest, but not striking increase with vehicle class, from approximately 30 mg/mi for LDV to
approximately 50 mg/mi for LDT4. However, an equivalent trend among the maxima is not
evident. The results also show that assuming equivalence between these two standards is
reasonable. On average it is clear that test groups certified at the Bin-5 level are capable of
meeting the target level of 50 mg/mi, although with small compliance margins. However,
relatively small numbers of families exceed this level, ranging to over 100 mg/mi.
Additionally, Figure 1-5 shows results for test groups certified to Bin-2 and Bin-3
standards. For these test groups, a trend with vehicle class is not evident, although very small
numbers of test groups are certified as trucks. In contrast to the Bin-5 vehicles, most families
certified at the Bin-2 and Bin-3 levels are well below the 50-mg/mi level, and maxima are no
higher than 7 percent below this level.
LDV-Tl
LDT2
LDT3
LDT4
Figure 1-4: Mean and Maximum Composite SFTP Results for HC+NOX for Test Groups
certified to Bin-5 and LEV-II/LEV Standards (bars and error-bars represent means and
maxima for sets of test groups, respectively)
1-17
-------
Bin 2/LEVII-SULEV
Bin3
LDV-T1
LDT2
LDT3
LDT4
Figure 1-5: Mean and Maximum Composite SFTP Results for HC+NOX for Test Groups
certified to Bin-2 and Bin-3 Standards (bars and error-bars represent means and maxima
for sets of test groups, respectively)
1.4 Technology Description for NMOG+NOX Control
A range of technology options exist to reduce NMOG and NOx emissions from both
gasoline fueled spark ignition and diesel engines below the current Tier 2 standards. Available
options include modifications to the engine calibration, engine design, exhaust system and after
treatment systems. The different available options have unique benefits and limitations. This
section describes the technical challenge to reducing emissions from current levels, describes
available technologies for reducing emissions, estimates the potential emissions reduction of the
different technologies, describes if there are other ancillary benefits to engine and vehicle
performance with the technology, and reviews the limits of each technology. Except where
noted, these technologies are applicable to all gasoline vehicles covered by this proposed rule.
Unique diesel technologies are addressed in Section 1.4.2.
1.4.1 Summary of the Technical Challenge for NMOG+NOx control
The proposed Tier 3 rule would require vehicle manufacturers to reduce the level of both
NMOG and NOx emissions from the existing Tier 2 fleet by approximately 80 percent over the
FTP by 2025. The FTP measures emissions during cold start, hot start, and warmed-up vehicle
city driving. The majority of NMOG and NOx emissions from gasoline fueled vehicles
measured during the FTP test occur during the cold start phase. Figure 1-1, above, graphically
demonstrates when NMOG and NOx emissions are produced during a cold start. As shown in
the figure, approximately 90 percent of the NMOG emissions occur during the first 50 seconds
after the cold start. In addition, about 60 percent of the NOx emissions occur during this same
50 second period. Unlike NMOG which is mostly controlled after the first 50 seconds, NOx
emissions tend to be released throughout the remainder of the FTP test. Achieving the proposed
Tier 3 NMOG+NOx FTP emissions standard may require manufacturers to reduce both cold start
NMOG and NOx emissions and further reduce NOx emissions when the vehicle is warmed up.
1-18
-------
The proposed Tier 3 rule would also require manufacturers to maintain their current high
vehicle load NMOG+NOx emission performance as measured during the US06 operation of the
Supplemental Federal Test Procedure (SFTP). The US06 component of the SFTP is designed to
simulate higher speeds and acceleration rates during warmed up vehicle operation. Significant
quantities of NMOG and NOx emissions are produced during the US06 portion of the SFTP if
enrichment events occur to reduce exhaust temperatures during high-load operation. Most
vehicles are now avoiding these enrichment events during the US06 and achieve relatively low
NMOG+NOx emissions.
It is anticipated that manufacturers will change the design of their exhaust and catalyst
systems to reduce catalyst light-off times to achieve the proposed Tier 330 mg/mi FTP
NMOG+NOx standard. Design changes to reduce catalyst light-off time can also result in higher
catalyst temperatures during high-load operation as seen during the US06 test. To achieve the
NMOG+NOx Tier 3 SFTP standard manufacturers will need to develop and implement
technologies to manage catalyst temperatures during high-load operation without using fuel
enrichment.
In addition, it is anticipated that the technologies manufacturers will use for reducing
warmed up NOx emissions during the FTP will also reduce NOx emissions during warmed up
operation on the US06.
For the catalyst to effectively reduce NMOG+NOx emissions it must reach the light-off
temperature of approximately 250 °C. Emissions during the catalyst warm up period can be
reduced by reducing the emissions produced by the engine during the catalyst warm up phase.
Emissions can also be reduced by shortening the time period required for the catalyst to reach the
light-off temperature. Reducing warmed-up NOx emissions requires improving the efficiency of
the catalyst system.
To achieve the proposed NMOG+NOx FTP emissions standards it is anticipated that
vehicle manufacturers will focus on three areas to reduce emissions:
• reducing the emissions produced by the engine before the catalyst reaches the
light-off temperature;
reducing the time required for the catalyst to reach the light-off temperature; and,
improving the NOx efficiency of the catalyst during warmed-up operation.
It is anticipated that improvements in all three areas will be required for heavier
passenger cars, light-duty trucks in classes LDT3 and LDT4, and MDPVs. The NOx efficiency
during warmed-up operation of vehicles certified to the Tier 2 Bin 4 emission level are such that
it is anticipated that reductions in cold start emissions are all that will be required to achieve the
proposed Tier 3 NMOG+NOX standard.
Heavy-duty trucks (8,501 up to 14,000 Ibs) will have a similar challenge to meet their
proposed Tier 3 standards along with the new SFTP requirements for this vehicle class. In
addition to the new test requirements and tighter standards, these vehicles useful life is being
1-19
-------
extended from 120,000 to 150,000 miles. Unlike lighter weight vehicles, heavy-duty trucks tend
to operate at higher loads for greater periods of time and therefore have different constraints to
meet the new requirements and more stringent proposed standards.
For spark-ignition engines, the higher operating load of these vehicles limits the ability to
move the catalyst close to the cylinder head due to durability concerns from higher thermal
loading. This limit will constrain the ability of these trucks to quickly light-off the catalyst, it
will, however, allow them to stay out of fuel-enriched operation to maintain catalyst
temperatures when the vehicle is being operated under high load. The emissions produced
during fuel-enrichment events, which occur at high loads would be expected to be significantly
greater than the reductions achievable during the cold start and idle phase. Fuel enrichment
events cause criteria pollutant and CC>2 emission rates to increase and also reduces the vehicle's
fuel economy. To achieve the proposed NMOG+NOx FTP emissions standards while also
meeting the new SFTP requirements it is anticipated that heavy-duty vehicle manufacturers will
focus on four areas for spark ignition engines:
• reducing the emissions produced by the engine before the catalyst reaches light-
off temperature;
• reducing the time required for the catalyst to reach the light-off temperature;
• improving the NOx efficiency of the catalyst during warmed up operation; and,
• minimizing the time spent in fuel enrichment to reduce the operating temperature
of the catalyst.
Compression ignition or diesel engines also have limitations with thermal goals and
location of the emission control system on the vehicle. With the similar goal of providing engine
exhaust heat to the catalysts, SCR and DPF, these emission control systems may compete with
each other for thermal energy. Additionally, the SCR system and the DPF generally require
sufficient capacity or size to handle the emissions from the engine which may limit the ability to
locate them in the optimal location.
To achieve the proposed NMOG+NOx FTP emissions standards while also meeting the
new SFTP requirements it is anticipated that heavy-duty vehicle manufacturers will focus on
three areas for compression ignition:
• reducing the emissions produced by the engine while the catalysts and SCR
system are being brought to proper operating temperature;
• reducing the time required for the catalysts and SCR system to reach the proper
operating temperature;
• improving the NOx efficiency of the SCR during warmed up operation through
refinement in engine out emission controls and SCR strategies.
1-20
-------
1.4.1.1 Reducing Engine Emissions Produced Before Catalyst Light-Off
During the first 50 seconds of the cold start phase of the FTP the engine is operating
either at idle or low speed and load in non-hybrid vehicles. The engine temperature is between
20 and 30 °C (68 and 86 °F). At these temperatures and under these low loads the cold engine
produces lower concentrations of NOx than NMOG. As the engine warms up and as the load
increases the concentration of NOx produced by the engine increases and the concentration of
NMOG decreases.
The design of the air induction system, combustion chamber, spark plug, and fuel
injection system determines the quantity of fuel required for stable combustion to occur in the
cold engine. Optimizing the performance of these components can provide reductions in the
amount of fuel required to produce stable combustion during these cold operating conditions.
Reductions in the amount of fuel required leads to reductions in cold start NMOG emissions.
The design considerations to minimize cold start emissions are also dependent on the fuel
injection method. Port fuel injected (PFI) engines have different design constraints than gasoline
direct injection (GDI) spark ignition engines. For both PFI and GDI engines, however, attention
to the details affecting the in cylinder air/fuel mixture can reduce cold start NMOG emissions.
It has been shown that cold start NMOG emissions in PFI engines can be reduced by
reducing the size of the fuel spray droplets and optimizing the spray targeting. Fuel impinging
on cold engine surfaces in the cylinder does not readily vaporize and does not combust.
Improving injector targeting to reduce the amount of fuel reaching the cylinder walls reduces the
amount of fuel needed to create a combustible air fuel mixture. Reducing the size of the spray
droplets improves the vaporization of the fuel and creation of a combustible mixture. 22
Droplet size can be reduced by modifying the injector orifice plate and also by increasing
the fuel pressure. Reducing droplet size and improving fuel vaporization during cold start has
been shown to reduce cold transient emissions by up to 40 percent during the cold start phase of
the FTP emission test.23 This and other PFI injector technology improvements have been used to
optimize the cold start performance of today's vehicles certified to the CA LEV II SULEV
standards.
The mixture formation process in a DISI engine is different than a PFI engine. In a PFI
engine the fuel is injected during the intake stroke of the engine in the intake runner. The fuel
has time to evaporate during the intake stroke as the fuel and air are drawn into the cylinder. In
addition, as the engine warms up the fuel can be injected into the intake runner and engine heat
can assist in evaporating the fuel prior to the intake valve opening.
The DISI engine injects fuel at higher fuel pressures than PFI engines directly into the
combustion chamber. In a DISI engine the fuel droplets need to evaporate and mix with the air
in the cylinder in order to form a flammable mixture.
Injecting directly into the cylinder reduces the time available for the fuel to evaporate and
mix with the intake air in a DISI engine compared to a PFI engine. An advantage of the DISI
1-21
-------
design is that the fuel spray does not impinge on the walls of the intake manifold or other
surfaces in the cylinder.
DISI systems have the ability to split the injection timing event. At least one study has
indicated that significant reductions in hydrocarbon emissions can be achieved by splitting the
injections during the cold start of a DISI engine. An initial injection occurs during the intake
stroke and a second injection is timed to occur during the compression stroke. This injection
method reduced unburned hydrocarbon emissions 30 percent compared to a compression stroke
only injection method.24
These are two examples of specific engine design characteristics, fuel injector design and
fuel system pressure on PFI engines and injection timing on GDI engines which can be used to
reduce cold start NMOG emissions significantly during the engine warm up prior to the catalyst
reaching the light-off temperature.
Optimizing the fuel injection system design and calibration is anticipated to be used in all
vehicle classes, including heavy-duty vehicles. It is anticipated that these described
improvements, along with improvements to other engine design characteristics, will be used to
reduce cold start emissions for passenger cars, LDTs, MDPVs, and HDTs.
Because the engine is relatively cold and the operating loads are low during the first 50
seconds of the FTP the engines typically do not produce significant quantities of NOx emissions
during this phase. In addition manufacturers tend to retard the combustion timing during the
catalyst warm up phase. Retarding combustion timing has been shown also to reduce the
concentration of NMOG in the exhaust. This calibration method further reduces peak
combustion temperatures while increasing the exhaust gas temperature compared to optimized
combustion timing. The increased exhaust gas temperature leads to improved heating of the
catalyst and reduced catalyst light-off times. Retarding combustion and other technologies for
reducing catalyst light-off time are discussed in the following section.
1.4.1.2 Reducing Catalyst Light-Off Time
The effectiveness of current vehicle emissions control systems depends in large part on
the time it takes for the catalyst to light-off, which is typically defined as the catalyst reaching a
temperature of 250°C. In order to reduce catalyst light-off time, it is expected manufacturers will
use technologies that will improve heat transfer to the catalyst during the cold start phase and
improve catalyst efficiency at lower temperatures. Technologies to reduce catalyst light-off time
include calibration changes, thermal management, close-coupled catalysts, catalyst PGM
loading, and secondary air injection. It is anticipated that in some cases where the catalyst light-
off time and efficiency are not sufficient to reduce cold start NMOG emissions, hydrocarbon
adsorbers may be utilized. The adsorbers trap hydrocarbons until such time that the catalyst is
fully warmed up and the emissions can be oxidized by the catalyst. Note that with the exception
of hydrocarbon adsorbers each of these technologies addresses NMOG and NOx performance.
The technologies are described in greater detail below.
1-22
-------
1.4.1.2.1 Calibration Changes to Reduce Catalyst Light-Off Time
These include calibration changes to increase the temperature and mass flow of the
exhaust prior to the catalyst reaching the light-off temperature. By reducing the time required for
the catalyst to light-off engine calibration changes can affect NMOG and NOx emissions.
Retarding combustion in a cold engine by retarding the spark advance is a well known
method for reducing the concentration of NMOG emissions in the exhaust and increasing the
exhaust gas temperature.25'26 The reduction in NMOG concentrations is due to a large fraction of
the unburned fuel within the cylinder combusting before the flame is extinguished on the
cylinder wall. Reductions of total hydrocarbon mass of up to 40 percent have been reported
from studies evaluating the effect of spark retard on exhaust emissions.
In addition to reducing the NMOG exhaust concentrations retarding the spark advance
reduces the torque produced by the engine. In order to produce the same torque and maintain the
engine speed and load at the desired level when retarding the spark advance, the air flow into the
engine is increased causing the manifold pressure to increase which also improves combustion
stability. Retarding the combustion process also results in an increase in the exhaust gas
temperature. The retarded ignition timing during the cold start phase in addition to reducing the
NMOG emissions therefore increases the exhaust mass flow and exhaust temperature. These
changes lead to a reduction in the time required to heat the catalyst.
The torque produced by the engine will begin to vary as the spark retard increases. As
the torque variations increase, the combustion process is deteriorating and the engine
performance degrades due to the partial burning. It is the level of this variability which defines
the absolute maximum reduction in spark advance that can be utilized to reduce NMOG
emissions and reduce the catalyst light-off time.
Retarding combustion during cold start can be applied to spark-ignition engines in all
vehicle classes. The exhaust temperatures and NMOG emission reductions will vary based on
engine design. This calibration methodology is anticipated to be used to improve catalyst warm-
up times and reduce cold start NMOG emissions for all vehicle classes, passenger cars, LDTs,
MDPVs, and HDTs.
With the penetration of variable valve timing technology increasing in gasoline-fueled
engines additional work is being performed to characterize the impact of valve timing on cold
start emissions. The potential exists that calibration changes to the valve timing during the cold
start phase will lead to additional reductions in cold start NMOG emissions. 2?
1.4.1.2.2 Exhaust System Thermal Management to Reduce Catalyst Light-Off Time
This category of technologies includes all design attributes meant to conduct combustion
heat into the catalyst with minimal cooling. This includes insulating the exhaust piping between
the engine and the catalyst, reducing the wetted area of the exhaust path, reducing the thermal
mass of the exhaust system, and/or using close-coupled catalysts (i.e., the catalysts are packaged
as close to the engine cylinder head as possible to mitigate the cooling effects of longer exhaust
1-23
-------
piping). By reducing the time required to light-off the catalyst, thermal management
technologies reduce NMOG and NOx emissions.
Moving the catalyst closer to the cylinder head is a means manufacturers have been using
to reduce both thermal losses and the catalyst light-off time. Many vehicles today use close-
coupled catalysts, a catalyst which is physically located as close as possible to the cylinder head.
Moving the catalyst from an underbody location to within inches of the cylinder head reduces the
light-off time significantly.
Another means for reducing heat losses are to replace cast exhaust manifolds with thin-
wall stamped manifolds. Reducing the mass of the exhaust system reduces the heat losses of the
system. In addition an insulating air gap can be added to the exhaust system which further
reduces the heat losses from the exhaust system, insulating air gap manifolds are also known as
dual-wall manifolds.
With thin- and dual-wall exhaust manifolds, close-coupled catalyst housings can be
welded to the manifold. This reduces the needed for manifold to catalyst flanges which further
reduces the thermal inertia of the exhaust system. Close coupling of the catalyst and reducing
the thermal mass of the exhaust system significantly reduces the light-off time of the catalyst
compared to an underbody catalyst with flanges and pipes connected to a cast exhaust manifold.
Using close-coupled catalysts reduces the heat losses between the cylinder head and
catalyst. While reducing the time required to light-off the catalyst the close-coupled catalyst can
be subject to higher temperatures than underbody catalysts during high load operating
conditions. To ensure the catalyst does not degrade manufacturers currently use fuel enrichment
to maintain the exhaust temperatures below the levels which would damage the catalyst. It is
anticipated that to meet the proposed Tier 3 SFTP standards, manufacturers will need to ensure
that fuel enrichment is not required on the US06 portion of the FTP. Calibration measures, other
than fuel enrichment, may be required to ensure the catalyst temperature does not exceed the
maximum limits.
Another technology beginning to be used for both reducing heat loss in the exhaust and
limiting exhaust gas temperatures under high load conditions is integrating the exhaust manifold
into the cylinder head. Honda utilized this technology on the Insight's 1.0 L VTEC-E engine.
The advantage of this technology is that it minimizes exhaust system heat loss during warm-up.
In addition with the exhaust manifold integrated in the cylinder head the cooling system can be
used to reduce the exhaust temperatures during high load operation. It is anticipated that
manufacturers will further develop this technology as means to both quickly light-off the catalyst
and reduce high-load exhaust temperatures.
To achieve the proposed Tier 3 NMOG and NOx emissions standards it is expected that
manufacturers will optimize the thermal inertia of the exhaust system to minimize the time
needed for the catalyst to achieve the light-off temperature. In addition, the manufacturers will
need to ensure the high load performance does not cause thermal degradation of the catalyst
system. It is expected that methods and technologies will be developed to reduce the need to use
fuel enrichment to reduce high load exhaust temperatures.
1-24
-------
Optimizing the catalyst location and reducing the thermal inertia of the exhaust system
are design options manufacturers can apply to all vehicle classes (PCs, LDTs, MDPVs, and
HDTs) for improving vehicle cold start emission performance.
It is not anticipated HDTs with spark-ignition engines will utilize catalysts close-coupled
to the exhaust manifold. The higher operating loads of these engines results in durability
concerns due to high thermal loading. It is expected that manufacturers will work to optimize
the thermal mass of the exhaust systems to reduce losses along with optimizing the underbody
location of the catalyst. These changes are expected to improve the light-off time while not
subjecting the catalysts to the higher thermal loadings from a close coupled location.
1.4.1.2.3 Catalyst Design Changes
A number of different catalyst design changes can be implemented to reduce the time for
the catalyst to light-off. Changes include modifying the substrate design, replacing a large
volume catalyst with a cascade of two or more catalysts, and optimizing the loading and
composition of the platinum group metals (PGM).
Progress continues to be made in the development of the catalyst substrates which
provide the physical support for the catalyst components which typically include a high surface
area alumina carrier, ceria used for storing oxygen, PGM catalysts, and other components. A
key design parameter for substrates is the cell density. Today catalyst substrates can be
fabricated with cell densities up to 900 and 1,200 cells per square inch (cpsi) with wall
thicknesses approaching 0.05 mm.
Increasing the surface area of the catalyst improves the performance of the catalyst.
Higher substrate cell densities increases the surface area for a given catalyst volume. Higher
surface areas improve the catalyst efficiency and durability reducing NMOG and NOx emissions.
The limitation of the higher cell density substrates include increased exhaust system
pressures at high load conditions. The cell density and substrate frontal area are significant
factors that need to be considered to optimize the catalyst performance while limiting flow loss
at high load operation.
During the cold start phase of the FTP the engine speeds and load are low during the first
50 seconds of the test. One method for reducing the catalyst light-off time is to replace a larger
volume catalyst with two catalysts which total the same volume as the single catalyst. The
reduced volume close-coupled catalyst reduces the heat needed for this front catalyst to reach the
light-off temperature. The front catalyst of the two catalyst system will reach operating
temperature before the larger volume single catalyst, reducing the light-off time of the system.
All other parameters held constant, increasing the PGM loading of the catalyst also
improves the efficiency of the catalyst. The ratio of PGM metals is important as platinum,
palladium, and rhodium have different levels of effectiveness promoting oxidation and reduction
reactions. Therefore, as the loading levels and composition of the PGM changes the light-off
performance for both NMOG and NOx need to be evaluated. Based on confidential
1-25
-------
conversations with manufacturers it appears that there is an upper limit to the PGM loading,
beyond which further increases do not improve light-off or catalyst efficiency.
To achieve the proposed Tier 3 NMOG and NOx emission standards it is anticipated that
9R
manufacturers will make changes to catalyst substrates and PGM loadings. To achieve the
emission levels required to meet the proposed Tier 3 NMOG+NOx standard of 30 mg/mi with a
compliance margin will require very low sulfur levels in the fuel. As described in Section
1.2.3.3 even low levels (greater than 10 ppm) of sulfur in gasoline inhibit the ability of PGM
catalysts to achieve the low levels NOx emission levels of the proposed Tier 3 standard. For the
Tier 3 FTP emission standards to be achieved and maintained, particularly in use, it is required
that the sulfur content of the fuel be reduced to 10 ppm or lower.
Manufacturers will optimize the design of their aftertreatment systems for their different
vehicles. Primary considerations include cost, light-off performance, warmed-up conversion
efficiency and the exhaust temperatures encountered by the vehicle during high load operation.
Vehicles having low power to weight ratios will tend to have higher exhaust gas temperatures
and exhaust gas flow which will result in a different design when compared to vehicles having
higher power to weight ratios.
Manufacturers and catalyst suppliers perform detailed studies evaluating the cost and
emission performance of aftertreatment systems. It is anticipated that manufacturers will
optimize their aftertreatment systems to achieve the Tier 3 standards and meet the durability
criteria for all vehicle classes (PCs, LDTs, MDPVs, and HDTs).
1.4.1.2.4 Secondary Air Injection
By injecting air directly into the exhaust stream, close to the exhaust valve, combustion
can be maintained within the exhaust, creating additional heat thereby further increasing the
catalyst temperature. The air/fuel mixture must be adjusted to provide a richer exhaust gas for
the secondary air to be effective.
Secondary air injection has been used by a variety of passenger vehicle manufacturers to
assist with achieving the emission levels required of the CA LEV II SULEV standard.
Secondary air injection systems are used after the engine has started and once exhaust port
temperatures are sufficiently high to sustain combustion in the exhaust port. When the
secondary air pump is turned on the engine control module increases the amount of fuel being
injected into the engine. Sufficient fuel is added so that the air/fuel ratio in the cylinder is rich of
stoichiometry. The exhaust contains significant quantities of CO and hydrocarbons. The rich
exhaust gas mixes with the secondary air in the exhaust port and the combustion process
continues increasing the temperature of the exhaust and rapidly heating the manifold and close-
coupled catalyst.29'30
Engines which do not use secondary air injection cannot operate rich of stoichiometry as
the added enrichment would cause increased NMOG emissions. The richer cold start calibration
used with vehicles that have a secondary air injection system provides a benefit as combustion
stability is improved. In addition, the richer calibration is not as sensitive to changes in fuel
volatility. Less volatile fuels found in the market may result in poor start and idle performance
1-26
-------
on engines calibrated to run lean during the cold operation. Engines which use secondary air and
have a richer warm up calibration would have a greater combustion stability margin.
Manufacturers may perceive this to be a benefit for the operation of their vehicles during the cold
start and warm up phase.
Installing a secondary air injection system combined with calibration changes can be used
by manufacturers to reduce the cold start emissions and improve the catalyst light-off on existing
engine designs. It is anticipated that manufacturers will utilize this technology to improve the
cold start performance on heavier vehicles and vehicles with low power to weight ratios.
Secondary air injection has been used on vehicles to achieve the CA LEVII SULEV emission
standards. This technology can be used on engines in all vehicle classes.
It is anticipated that secondary air injection will be used primarily in combination with
close coupled catalysts. Therefore, it is not anticipated that this technology will be used with
HDTs as it is not expected that the catalyst in these vehicles will be moved to a location
sufficiently close to the exhaust manifold to provide any improvement in catalyst light-off.
HDTs tend to operate at higher loads and catalyst durability is a concern due to the
increased thermal loading as the catalyst is moved closer to the cylinder head. Moving the
catalyst closer to the exhaust manifold would result in increasing the time spent in fuel
enrichment modes to ensure the temperatures are maintained below the threshold which would
reduce the durability of the catalyst. Using fuel enrichment to control catalyst temperature
causes significant increases in criteria pollutant emissions, COz emissions and reductions in fuel
economy.
1.4.1.2.5 Hydrocarbon Adsorbers
Hydrocarbon adsorbers trap hydrocarbons emitted by the engine when the adsorber is at
low temperatures. As the temperature of the hydrocarbon adsorber increases the trapped
hydrocarbons are released. Passive adsorbers use an additional washcoat on an existing three-
way catalyst. The adsorber is a zeolite-based material which absorbs hydrocarbons at low
exhaust temperatures and desorbs hydrocarbons as the temperature increases. A significant
technical challenge to using a passive adsorber is to design the system such that the three-way
catalyst has reached the light-off temperature prior to the adsorber coating releasing the adsorbed
hydrocarbons.
Active adsorbers use a substrate with an adsorber washcoat over which the exhaust is
directed when the exhaust temperature is below the desorption temperature of the material. Once
the exhaust temperature reaches the desorption temperature the exhaust is routed such that it no
longer passes over the adsorber. As the adsorber continues to heat in the exhaust the captured
hydrocarbons are released and oxidized by the warmed-up catalyst system.
Adsorbers have been used to reduce cold start NMOG emissions on CA LEV II SULEV
vehicles. Additional work is being performed to further improve the performance of
hydrocarbon adsorbers.
1-27
-------
It is anticipated that if manufacturers have difficulty achieving the proposed Tier 3
NMOG+NOx emission standards that manufacturers will evaluate including a hydrocarbon trap
to further reduce the NMOG emissions during the cold start. One manufacturer used this
approach to achieve the CA LEV II SULEV standard on a large displacement V-8 engine with
the application of an active hydrocarbon adsorber.31
Hydrocarbon adsorbers can be used on all spark-ignition engines and all classes of
vehicles. It is anticipated that these technologies may be required for engines with larger
displacement and in some of the larger vehicles. It is anticipated that HDTs will be able to
achieve the emissions levels required without the use of hydrocarbon adsorbers to meet the
proposed standard.
1.4.1.3 Improving catalyst NOx efficiency during warmed-up operation
Significant quantities of NOx emissions are produced by vehicles during warmed-up
vehicle operation on the FTP for Tier 2 Bin 5 certified vehicles. The stabilized NOx emission
levels will need to be reduced to achieve the proposed Tier 3 NMOG+NOx emission standard.
Improving the NOx performance of the vehicle can be achieved by improving the catalyst
efficiency during warmed-up operation. As previously described the performance of the catalyst
can be improved by modifications to the catalyst substrate, increasing cell density, increasing
PGM loadings and reducing the sulfur level of gasoline. Three-way catalyst efficiency is also
affected by frequency and amplitude of the air/fuel ratio. For some vehicles warmed-up catalyst
NOx efficiency can be improved by optimizing the air/fuel ratio control and limiting the
amplitude of the air fuel ratio excursions. It is anticipated that a combination of changes will be
made by manufacturers including further improvements to air/fuel ratio calibration and catalyst
changes including cell density and PGM loadings.
A requirement to ensure that the NOx emission performance of the vehicles is maintained
at or below the 30 mg/mi NMOG+NOx emission standard is reduced fuel sulfur concentrations.
As described in detail in Section 1.2.3.3 further reductions in fuel sulfur concentration are
required to ensure the catalyst performance is not degraded which causes increases in NOx
emissions beyond the proposed Tier 3 standard.
It is anticipated that manufacturers will use these catalyst and calibration technologies to
improve the warmed up NOx emissions performance of vehicles in all classes, passenger cars,
LDTs, MDPVs, and HDTs.
1.4.1.4 EPA Estimates of Technology Improvements Required for Large Light-
Duty Trucks
Discussions between EPA, CARB, vehicle manufacturers and major component suppliers
indicated that large light-duty trucks (e.g., pickups and full-size SUVs in the LDT3 and LDT4
categories) will be the most challenging light-duty vehicles to bring into compliance with the
proposed Tier 3 NMOG+NOx standards at the 30 mg/mi corporate average emissions level. A
similar challenge was addressed when large light-duty trucks were brought into compliance with
the Tier 2 standards in the previous decade. Figure 1-6 provides a graphical representation of the
effectiveness of Tier 3 technologies when combined with gasoline sulfur control for large light-
1-28
-------
duty truck applications. The Tier 3 technologies shown are those that can be utilized on existing
vehicles and do not require engine design changes. A compliance margin is shown in both cases.
Note that the graphical representation of the effectiveness of catalyst technologies on NOx and
NMOG when going from Tier 2 to Tier 3 also includes a reduction in gasoline sulfur levels from
30 ppmto lOppm.
180
E
o
§
BO
VI
C
'o
i/i
.2
UJ
X
O
|
Tier 2 Bin 5 NMOG+NOx
NMOG+NOx Emissions of Tier 2 Bin 5 LDT3/4
NMOG+NOx Emissons of Tier 3 LD3/T4
rocarbon Adsorber L_
Reduced Thermal Mass I
Calibration Changes I
ncreased Catalyst
Active Materials+
Calibration Changes
77% Emissions
Reduction
Secondary
Air Injection
+ Calibration Changes
Tier 3 NMOG+NOx Standard
50% Compliance Margin
Tier 2 Bin 5 LDT3/4
w/30 ppm S Gasoline
Tier 3 LDT3/4
w/10 ppm S Gasoline
Figure 1-6: Contribution of the expected Tier 3 technologies to large light-duty truck
compliance with the Tier 3 standards with a comparison to Tier 2 Bin 5. The technologies
and levels of control are based on a combination of confidential business information
submitted by auto manufacturers and suppliers, public data and EPA staff engineering
judgment.
1.4.2 Diesel Technologies for Achieving Tier 3 NMOG and NOx Emission Requirements
Compared to spark-ignition engines, diesel engines typically produce very low NMOG
emissions. However, diesel engines do not operate at stoichiometry preventing them from using
emission control approaches similar to spark-ignition engines to control NOx emissions. The
technical challenge for diesel engines to achieve the proposed Tier 3 NMOG+NOx emission
levels will be to obtain significant NOx emission reductions. It is anticipated that improvements
in NOx emissions performance of diesel exhaust catalysts during the cold start phase will be a
major technical challenge. Depending on the performance of the exhaust catalyst system,
1-29
-------
additional reductions in warmed-up NOx emissions may also be required to achieve the proposed
Tier 3 emission levels.
It is not anticipated that diesel engines will have difficulty achieving the proposed Tier 3
SFTP emission standards. The exhaust catalyst system is fully warmed up and operational on the
high load portion of the SFTP, the US06. It is anticipated that manufacturers may need to
optimize the calibration of the selective catalytic reduction (SCR) system or the NOx adsorption
catalyst (NAC) system to ensure the systems achieve the required performance.
The technical task for achieving the Tier3 standards on all diesel engines in all vehicle
classes will be to have the exhaust catalysts reach operating temperatures early in the cold-start
phase of the FTP. To achieve these improvements it is anticipated that diesel manufacturers will
focus on means to reduce NOx emissions during the engine warm-up phase after the cold start
and reducing the time required for the SCR or NAC system to begin reducing (SCR) or capturing
and reducing (NAC) the NOx.
By controlling the timing of the fuel injection event, the number of fuel injection events
and the timing of intake and exhaust valve events, the temperature of the exhaust can be
increased. Diesel engine manufacturers will optimize the injection and valvetrain calibration to
increase the exhaust temperature after the engine is started and before it has reached operating
temperature.
As with gasoline engines, locating the exhaust catalyst system closer to the cylinder head
and air-gap insulating the exhaust system or reducing the mass of the exhaust components
upstream of the catalysts can be used to build and maintain heat in the exhaust system. A variety
of technologies are available to conduct combustion heat into the exhaust catalyst system with
minimal cooling. This includes uses of dual-wall, air-gapped exhaust piping between the engine
and the catalyst or trap; reducing the wetted area of the exhaust path; and reducing the thermal
mass of the exhaust system through use of thinner wall materials. By reducing the time required
to light-off the catalysts, thermal management technologies can reduce NOx emissions from
diesel engines. Once light-off has been achieved, NOx emissions reduction for modern, base-
metal zeolite SCR systems approaches that of modern three-way catalyst systems used for
stoichiometric gasoline spark-ignition applications.32
1.5 PM Feasibility
Particulate matter emitted from internal combustion engines is a multi-component
mixture composed of elemental carbon (or soot), semi-volatile organic compounds, sulfate
compounds (primarily sulfuric acid) with associated water, nitrate compounds and trace
quantities of metallic ash. At temperatures above 1,300K, fuel hydrocarbons without access to
oxidants can pyrolize to form particles of elemental carbon. Fuel pyrolysis can occur as the result
of operation at richer than stoichiometric air-to-fuel ratio (primarily PFI gasoline GDI engines),
direct fuel impingement onto surfaces exposed to combustion (primarily GDI and diesel engines)
and non-homogeneity of the air-fuel mixture during combustion (primarily diesel engines).
Elemental carbon particles that are formed can be oxidized during later stages of combustion via
in-cylinder charge motion and reaction with oxidants. Semi-volatile organic compounds (SVOC)
are composed primarily of organic compounds from lubricant and partial combustion products
1-30
-------
from fuel. PM emissions from SVOC are typically gas phase when exhausted from the engine
and contribute to PM emissions via particle adsorption and nucleation after mixing with air and
cooling. Essentially, PM-associated SVOC represent the condensable fraction of NMOG
emissions. Sulfur and nitrogen compounds are emitted primarily as gaseous species (SO2, NO
and NO2). Sulfate compounds can be a significant contributor to PM emissions from stratified
lean-burn gasoline engines and diesel engines, particularly under conditions where PGM-
containing exhaust catalysts used for control of gaseous and PM emissions oxidize a large
fraction of the SO2 emissions to sulfate (primarily sulfuric acid). Sulfate compounds do not
significantly contribute to PM emissions from spark-ignition engines operated at near
stoichiometric air-fuel ratios due to insufficient availability of oxygen in the exhaust for
oxidation of SO2 over PGM catalysts.
Elemental carbon PM emissions can be controlled by:
• Reducing fuel impingement on piston and cylinder surfaces
• Inducing charge motion and air-fuel mixing via charge motion (e.g., tumble and
swirl) or via multiple injection (e.g., GDI and diesel/common rail applications)
• Reducing or eliminating operation at net-fuel-rich air-to-fuel ratios (PFI gasoline
and GDI applications)
• Use of wall-flow or partial-wall-flow exhaust filters (diesel applications)
SVOC PM emissions can be controlled by:
• Reducing lubricating oil consumption
• Improvements in exhaust catalyst systems used to control gaseous NMOG
emissions (e.g., increased PGM surface area in the catalyst, improvements in
achieving catalyst light-off following cold-starts, etc.)
Sulfate PM emission can be controlled by:
• Reducing or eliminating sulfur from fuels
1.5.1 PM Emissions from Light-duty Tier 2 Vehicles
In order to establish the feasibility of the proposed PM emission standards for the Tier 3
rule, EPA conducted a test program to measure PM emissions from Tier 2 light-duty vehicles.
The test program was designed to measure PM emissions from late model year vehicles that
represented a significant volume of annual light duty-sales and included vehicles that ranged
from small cars through trucks. In addition, GDI vehicles were included in the program as were
vehicles with known high oil consumption.
The Agency investigated PM emissions from Tier 2 light-duty vehicles. Seventeen model
year 2005-2010 Tier 2 Bin 4, 5, and 8 vehicles were tested at the U.S. EPA National Vehicle and
1-31
-------
Fuel Emissions Laboratory (NVFEL) facility. A summary of their characteristics are provided in
Table 1-7. They included ten cars and seven trucks. Fifteen of these vehicles had accumulated
102,000-124,000 miles prior to the launch of the test program. One vehicle had accumulated
75,000 miles and another accumulated 21,000. Three cars and one truck were equipped with
GDI engines. Twelve of the fifteen test vehicles had previously been used in the DOE V4
Program. The remaining five vehicles were recruited in southeastern Michigan. One vehicle
(vehicle K) was suspected of having atypically high oil consumption and had only 37,000 miles
of mileage accumulation. Vehicle K was a duplicate of Vehicle C and was determined to
consume two and one half times the average oil consumption of vehicle C and three vehicles of
the same make, model and model year when tested within the DOE V4 program.0
The twelve vehicles acquired from the DOE V4 Program were selected to represent a
broad cross section of some of the highest sales vehicles in the U.S. market for model years
2005-2009. These vehicles had originally been purchased by DOE with odometer readings
ranging from 10,000-60,000 miles, placed in a mileage accumulation program and operated over
the EPA Standard Road Cycle on a test track or on mileage accumulation dynamometers to
110,000-120,000 miles.33 Immediately prior to inclusion in the EPA PM Test Program, the test
vehicles were serviced per the manufacturer's published service schedule and maintenance
procedures and underwent engine oil aging over a distance of 1,000 miles accumulated over the
EPA Standard Road Cycle to stabilize engine oil contribution to PM emissions34.
Three recruited test vehicles were selected because they used GDI technology. An
additional GDI equipped vehicle was obtained from the DOE V4 Program An attempt was made
to only recruit vehicles approaching the 120,000 mile useful life level. Testing was completed
for two of the four vehicles in time for the proposed rule. All of the recruited test vehicles were
thoroughly inspected, but otherwise tested as received.
All vehicles were tested on an El5 fuel with RVP, aromatic content, sulfur content, T50
and T90 of 9.1 psi, 23.8 vol%, 7 ppm, 160F and 31 IF, respectively. The properties of this fuel
approximated those of a projected El 5 market fuel.
The test program included three cold start and three hot start UDDS tests and three US06
tests conducted on each vehicle. FTP results were calculated for gaseous and PM emissions by
applying the cold-start and hot-start weighting factors to the complete cold and hot UDDS
results, respectively. This eliminated separate analysis of the typically very low concentration
FTP phase-2 gaseous and PM emissions samples and represented one method proposed within
40 CFR 1066 for increasing sample integration of measured gaseous and PM mass. During these
tests, triplicate PM samples were collected in parallel on PTFE membranes and single
(composite) PM samples were collected on primary and secondary quartz filters for TOT/TOR
OC/EC PM speciation analysis. Additional quartz filters were collected to determine the
contribution of gas-phase artifact to the OC collected on the quartz filter samples. The
compositing of quartz filters over three repeats of each test was done to enhance the precision of
D Vehicle K consumed approximately 1 quart per 3,000 miles vs. an average of approximately 1 quart per 8,000
miles for the other four vehicles of this make, model and year tested within the DOE V4 program.
1-32
-------
subsequent OC/EC thermogravimetric measurements. Single background (dilution air) PM
samples were also taken during each emissions test. Weekly tunnel blank and field blank PM
samples were also collected.
The following parameters were measured: NOx, NMHC, NMOG, alcohols, carbonyls,
CO, CH4, CC>2 and fuel consumption and PM mass as per the 40 CFR 1065 and the proposed 40
CFR part 1066 test procedures. Limited exhaust HC speciation was also performed.
PM composition was determined from filter samples taken on both quartz filters and
PTFE membranes. PM compositional analyses include determination of the contribution of
elemental and organic carbon to PM mass,35 elemental analysis via EDXRF, sulfate analysis via
ion chromatography and determination of the contribution of unresolved complex organic PM
compounds by GC/MS.
Note that during the compositional analysis of the PM, EPA discovered a significant
amount of silicon deposited on some of the filters. The source of the silicon was determined to
be a silicone elastomer transfer tube used to connect vehicles to the emissions measurement
equipment. The data below reflect test results that are not subject to silicone contamination. For
additional information, refer to our memo to the docketE which describes the original analysis
and corrective actions in greater detail.
Table 1-7: Vehicles Tested as part of EPA's Light-Duty Vehicle PM Test Program
Vehicle Make, Model
and Designation
Honda Civic
Toyota Corolla
Honda Accord
Dodge Caliber
Chevrolet Impala
Ford Taurus
Toyota Tundra
Chrysler Caravan
Jeep Liberty
Ford Explorer
Honda Accord
Ford F 150
Chevrolet
Silverado
VW Passat
A
B
C
D
E
F
G
H
I
J
K
L
P
M
Model
Year
2009
2009
2007
2007
2006
2008
2005
2007
2009
2009
2007
2005
2006
2006
Certified to
Emissions
Standard
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 4
Tier 2/Bin 5
Tier 2/Bin 8
Tier 2/Bin 8
Tier 2/Bin 5
Odometer at
Start of
Program, miles
121,329
120,929
123,695
114,706
114,284
115,444
121,243
116,742
121,590
121,901
36,958
111,962
110,898
102,886
Fuel
Delivery51
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
TGDI
Used in
DOEV4
Program?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
E Sobotowski, R. (February, 2013). Test Program to Establish LDV Full Useful Life PM Performance.
Memorandum to the docket.
1-33
-------
Manufacturer' s
Development
Vehicle13
Saturn Outlook
Cadillac STS4
N
O
Q
PC
2009
2010
None (Tier
2/LEV II
Prototype)*
Tier 2/Bin 5
Tier 2/Bin 5
120,011
123,337
21,266
TGDI
GDI
GDI
No
Yes
No
Notes:
a PFI is Naturally aspirated, port fuel injected; GDI is Naturally aspirated, gasoline direct
injection; TGDI is Turbocharged, gasoline direct injection
b Manufacturer's developmental vehicle. Vehicle used a spray-guided GDI fuel system
with a centrally-mounted injector. Emissions were targeted at Tier 2 Bin 5 or better.
1.5.1.1 PM Emissions Test Results
The results of exhaust emissions tests conducted in this program are summarized in Table
1-8 and Table 1-9 for the FTP and US06 test cycles, respectively. FTP NMOG +NOX and PM
results are also plotted in Figure 1-7 and Figure 1-8, respectively. The US06 NMOG+NOx and
PM results are shown in Figure 1-9 and Figure 1-10, respectively. In all figures, the vehicles are
divided into two groups: PFI and GDI. Within each group they are listed in the sequence of
increasing CO2 emissions on the FTP test cycle. The bars shown in the figures represent the
means of triplicate measurements. The individual data points are indicated in all figures together
with the corresponding standard deviations. Vehicle Q only had one valid PM test on the FTP
test cycle and no error bars are plotted.
Table 1-8: FTP Composite Emissions Results
Vehicle
A
B
C
D
E
F
G
H
I
J
K
CO2
g/mile
284.6
286.3
324.4
364.8
410.8
419.2
447.2
462.9
495.7
554.8
332.5
CO
g/mile
0.358
0.434
0.382
6.740
0.571
0.271
0.626
1.617
0.719
1.072
0.202
NOX
g/mile
0.0239
0.0461
0.0231
0.1432
0.0600
0.0151
0.0424
0.0507
0.0317
0.0281
0.0165
NMOG
g/mile
0.0316
0.0408
0.0299
0.0663
0.0359
0.0206
0.0439
0.0493
0.0429
0.0525
0.0171
NOx+NMOG
g/mile
0.056
0.087
0.053
0.210
0.096
0.036
0.086
0.100
0.075
0.081
0.034
PM
mg/mile
0.27
0.22
0.18
0.45
0.14
0.11
0.36
0.40
1.36
0.10
0.93
1-34
-------
L
M
N
O
P
Q
568.8
365.2
411.0
505.2
547.0
465.0
2.264
0.346
0.735
0.599
0.649
0.475
0.1024
0.0342
0.0279
0.0173
0.3578
0.0279
0.0822
0.0261
0.0258
0.0399
0.0429
0.0221
0.185
0.060
0.054
0.057
0.401
0.050
0.39
_
2.55
4.72
0.18
7.15
Table 1-9: US06 Emissions Results
Vehicle
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
CO2
g/mile
289.0
312.8
318.2
413.7
393.3
422.8
490.9
467.0
516.0
555.9
320.4
595.6
352.8
401.7
547.4
529.1
436.6
CO
g/mile
7.092
9.315
1.293
9.077
0.660
1.237
3.462
1.128
0.833
3.015
1.800
5.519
9.225
0.330
9.862
2.728
2.595
NOX
g/mile
0.0212
0.0530
0.0257
0.1324
0.1019
0.0274
0.0369
0.0910
0.1852
0.1121
0.0247
0.0036
0.0481
0.1614
0.0377
0.1427
0.0265
NMOG
g/mile
0.0162
0.0248
0.0105
0.0127
0.0163
0.0124
0.0172
0.0134
0.0037
0.0159
0.0079
0.0125
0.0297
0.0048
0.0282
0.0116
0.0204
NOx+NMOG
g/mile
0.0374
0.0779
0.0362
0.1451
0.1183
0.0398
0.0540
0.1044
0.1889
0.1280
0.0326
0.0160
0.0779
0.1662
0.0659
0.1543
0.0470
PM
mg/mile
0.76
2.05
1.05
_
0.46
1.61
_
2.04
3.31
0.27
2.84
2.13
_
2.37
_
1.83
_
1-35
-------
As shown in Figure 1-7, with the exception of one PFI passenger car (vehicle D), the FTP
NMOG+NOx emissions of all tested vehicles remained below their respective proposed fleet
average 2017 standards, but none performed below the proposed 2025 standard.
FTP PM emissions increased with CC>2 emissions for the PFI vehicles and remained well
below the proposed 3 mg/mile standard, confirming that most current light duty vehicles are
already capable of meeting the proposed Tier 3 PM standard (Figure 1-8). Two GDI vehicles
demonstrated FTP PM emissions above the proposed standard, indicating that additional
combustion system development would be necessary in some vehicles to achieve compliance.
As shown in Figure 1-9, with the exception of two LDTs (vehicles I and J), all vehicles
met their respective proposed fleet average 2017 (for vehicles below 6,000pounds GVWR) or
2018 (for vehicles above 6,000 pounds GVWR) US06 NMOG+NOX standards. Five vehicles,
four passenger cars (vehicles A, B, F and L) and one LDT (vehicle L), produced US06
NMOG+NOx emissions lower than the proposed 2025 standard.
As in the case of FTP results, US06 PM emissions increased with the increase in CO2
emissions in PFI vehicles (Figure 1-10). All PFI passenger cars remained well below the
proposed 10 mg/mile standard. One GDI passenger car (vehicle N) performed well below its
respective US06 PM standard and achieved PM emissions over the US06 comparable to its
performance over the FTP. In summary, all of the vehicles tested performed below the proposed
US06 PM standards.
The suspected high oil consumption vehicle (vehicle K) emitted 3 and 2.3 times more PM
in this program than a comparable vehicle with average oil consumption (vehicle C) in the FTP
and US06 tests, respectively.
1-36
-------
PFI (PC)
PFI (LPT2)
PFI (LDT3/4)
GDI
0.5 -
PC
LDT2
LDT3/4
PC PC PC LDT3
£0.4
S
CD
O
0.3 -
r
~z.
Q.0.1
I—
LJ_
0.0
•I.
•I
I I I I I I III
BCKDEF GHI
P J L
i i i i
M N Q O
Figure 1-7: Composite FTP NMOG+NOX Emissions Results
PFI (PC)
PFI (LPT2)
PFI (LDT3/4)
GDI
E
a
^'
Q_
Q.
2 -
0 -
PC
LDT2
LDT3/4
PC PC LDT3
i i i i i i i
A B C K D E F
GHI PJL NQO
Figure 1-8: Composite FTP PM Emission Results
1-37
-------
PFI (PC)
PFI (LPT2)
PFI (LDT3/4)
GDI
0.5 -
Efl.4 -
CD
O0.3 -
X0.2 -\
o
§0.1 -\
o.o -
PC
LDT2
LDT3/4
PC PC PC LDT3
^..IL
i i i i i i iii
BCKDEF GHI
PJL
i i i i
MNQO
Figure 1-9 US06 NMOG+NOX Emissions Results
PFI (PC)
PFI (LPT2)
PFI (LDT3/4)
GDI
O)
.4 -
Q_
CD
O
C/)2
Z)
o -
PC
LDT2
LDT3/4
i
PC
B C K
H I
Figure 1-10: US06 PM Emissions Results
1-38
-------
1.5.2 FTP PM Feasibility
With regard to the feasibility of the light-duty fleet to meet the proposed PM standards
for testing on the FTP and US06, we based our conclusions on the PM performance of the
existing fleet. Data on both low and high mileage light-duty vehicles demonstrate that the
majority of vehicles are currently achieving levels in the range of the proposed Tier 3 FTP
standards. A small number of vehicles are at or just over the proposed standard at low mileage
and would require calibration changes, catalyst changes and/or further combustion system
improvements to meet the new standards. It is our expectation that the same calibration and
catalyst changes required to address NMOG would also provide some additional PM control.
Vehicles that are currently demonstrating higher PM emissions over the FTP at higher mileages
would likely be required to control oil consumption and combustion chamber deposits.
1.5.3 SFTPPM
Also, US06 test data shows that many vehicles are already at or below the proposed
standards for US06. Vehicles that are demonstrating high PM on the US06 would need to
control enrichment and oil consumption. The oil consumption strategies are much like that
described above for controlling oil consumption on the FTP. However, given the higher engine
RPMs experienced on the US06 and the commensurate increase in oil consumption,
manufacturers will most likely focus on oil sources stemming from the piston to cylinder
interface and positive crankcase ventilation (PCV). With respect to enrichment, changing the
fuel/air mixture by increasing the fuel fraction is no longer the only tool that manufacturers have
available to them to protect engine and exhaust system components from over-temperature
conditions. With application of electronic throttle controls on nearly every light-duty vehicle,
the manufacturer has the option to richen the air/fuel mixture by maintaining the amount of fuel
being delivered and closing the throttle plate. Previously, on manual throttle control vehicles,
the throttle plate position was established by the driver and the engine controls had no capability
to change the amount of air in the intake. While this solution may result in a small reduction in
vehicle performance we believe that it is an effective way to reduce PM emissions over the US06
cycle.
1.5.4 Full Useful Life: Durability and Oil Consumption
Manufacturers have informed us that they have already or are planning to reduce oil
consumption by improved sealing of the paths of oil into the combustion chamber, including
improved piston-to-cylinder interfaces. They are taking or considering these actions to address
issues of customer satisfaction, cost of ownership and improved emission control system
performance as vehicles age.
Over the past decade, many manufacturers have extended oil change intervals from the
historically required 3,000 miles interval to a now typical 10,000 mile interval or more in some
vehicle models. In order to allow for these longer intervals, improvements were made to limit
pathways for oil to enter the combustion chamber, resulting in significantly reduced oil
consumption. While customer satisfaction and longer oil change intervals, particularly for leased
vehicles where owners may be less inclined to perform frequent oil changes, were a motivation
for reducing oil consumption, improvements in the performance of the emission control system
1-39
-------
are a secondary benefit of reduced oil consumption. Oil consumption can damage catalytic
converters by coating the areas of the catalyst that convert and oxidize the pollutants. Over time,
this can cause permanent inactivity of those areas, resulting in reduced catalytic conversion
efficiency. Reductions in oil consumption can extend the life of the catalytic converter and help
manufacturers meet longer useful life requirements. This is particularly important on vehicles
meeting the most stringent emission standards, because they will need to maintain high catalyst
efficiencies in order to meet the stringent emission standards at higher mileage.
1.6 Evaporative Emissions Feasibility
The basic technology for controlling evaporative emissions was first introduced in the
1970s. Manufacturers routed fuel tank and carburetor vapors to a canister filled with activated
carbon, where vapors were stored until engine operation allowed for purge air to be drawn
through the canister to extract the vapors for delivery to the engine intake. Over the past 30
years, evaporative emission standards have changed several times, most notably in the mid-
1990s when enhanced evaporative controls were required to address 2- and 3- day diurnal
emissions and running losses. Refueling emission controls were added with phase-in beginning
in the 1998 MY. Almost universally manufacturers elected to integrate evaporative and
refueling emission control systems. In the mid-2000s more stringent evaporative emission
standards with E10 durability gasoline led to the development and adoption of technology to
identify and eliminate permeation of fuel through fuel tanks, fuel lines, and other fuel-system
components.
The proposed evaporative emission requirements include more stringent hot-soak plus
diurnal standards that are expected to require new vehicle hardware and improved fuel system
designs. The type of new hardware that would be required will vary depending on the specific
application and emission challenges and are described in the following section. Additionally, the
fleet-average nature of the proposed standards would allow more challenged vehicles to be offset
by vehicles that could outperform the required fleet averages.
In order to assess the technical feasibility of the proposed evaporative emission standards,
EPA conducted two analyses. The first analysis performed was based on the certification results
for the current EPA-certified evaporative families. This provided a baseline for the current
federal fleet emissions performance. The second analysis looked at the list of PZEV-certified
vehicles in the California LEV II fleet. The proposed Tier 3 evaporative emission standards are
similar to the current evaporative requirements for PZEVs in California. Both of these analyses
are described in greater detail below.
1.6.1 Assessment of the Current EPA Certification Emissions
EPA's current evaporative emission standards vary by vehicle category. Table 1-10
shows the currently applicable hot-soak plus diurnal emission standards.
1-40
-------
Table 1-10 Current EPA Evaporative Emission Standards
Vehicle Category
LDV
LDT1/LDT2
LDT3/LDT4
MDPV
Complete HDV
< 14,000 pounds GVWR
Hot-soak plus Diurnal (2-day)
0.65 g/test
0.85 g/test
1.1 5 g/test
1.25 g/test
1.75 g/test
Hot-soak plus Diurnal (3-day)
0.50 g/test
0.65 g/test
0.90 g/test
1.00 g/test
1.4 g/test
Based on MY2010 certification data, EPA analyzed the certification hot-soak plus diurnal
emission levels for all certified vehicle categories that will be subject to the proposed Tier 3
standards. The following figure shows the hot-soak plus diurnal certification levels (based on
the 2-day diurnal test) for each vehicle category ordered from the lowest to the highest emission
levels. (While not presented in this analysis, the data based on the 3-day diurnal tests shows a
similar trend.) Figure 1-11 also shows the proposed Tier 3 evaporative emission standards.
MY2010 EPA Certification Levels
{Hot-soak + 2-day Diurnal)
Proposed Tier 3 Standards
LDT2 f U LDT4
LDT1 LDT3
MDV
Numberof Families
Figure 1-11: MY2010 Hot-soak Plus Diurnal (2-Day) Emission Certification Levels
1-41
-------
It should be noted that the current evaporative emission data is based on a different
certification test fuel than what is being proposed for the Tier 3 program. While both the current
and proposed certification fuels have a Reid vapor pressure of nominally 9.0 psi, EPA's current
certification test fuel contains no ethanol, whereas the proposed certification fuel contains 15
percent ethanol. While the discussion below is based on the current certification fuel without
ethanol, EPA believes it is still useful in gauging the level of effort needed by manufacturers to
comply with the proposed Tier 3 standards. It is generally understood that ethanol can impact
permeation emissions from the fuel tank and fuel lines to some degree, but the bulk of
evaporative emissions are from diurnal emissions which are primarily a function of the Reid
vapor pressure of the fuel which will be maintained at 9.0 psi and therefore should not be
impacted by the presence of ethanol in the proposed certification fuel.
As can be seen from the figure, there are many families certified below the proposed Tier
3 hot-soak plus diurnal standards. Of the nearly 200 evaporative families included in the
analysis, 40 percent had certification levels below the proposed Tier 3 standards. Some of these
families are certified to the more stringent PZEV standards, upon which the proposed Tier 3
evaporative emission standards are based, but many of the families are not. However, the
proposed Tier 3 evaporative emission standards include a new canister bleed test that is not
required under the current EPA regulations. (The families certified to the PZEV requirements
are subject to a similar requirement and would likely meet that new canister bleed test
requirement without further modification.) Therefore, even though many families are certified
below the proposed Tier 3 evaporative emission standards, manufacturers would still need to
make additional changes with many of the evaporative control systems to ensure compliance
with the proposed standards.
1.6.2 Assessment of California-certified PZEVs
Based on the California Air Resources Board's MY2011 certification list, EPA identified
the vehicles certified by manufacturers to the PZEV requirements. As noted earlier, the
proposed Tier 3 evaporative emission standards are very similar to the PZEV evaporative
emission requirements and, as allowed with one of the proposed options for MY2017,
manufacturers could sell their evaporative emission compliant PZEV vehicles nationwide in
MY2017. Manufacturers have certified over 50 models of passenger cars and light-duty trucks
to the PZEV requirements. EPA believes that manufacturer's experience with PZEV
technologies will assist them as they work to apply similar technologies across their fleets to
comply with the proposed Tier 3 evaporative emission standards. As described in more detail
below, EPA expects manufacturers will employ a number of technologies to meet the proposed
Tier 3 standards. The anticipated control technologies to comply with the proposed emission
standards have already been included on many of the PZEVs and include improved carbon
canister designs to even better capture vapor emissions from the canister, air intake designs to
prevent the escape of unburned fuel from the engine's crankcase, various upgrades to further
limit potential micro-sized leaks, and further steps to reduce permeation rates. Table 1-11 shows
the manufacturers and models certified to the PZEV standards in MY2011.
1-42
-------
Table 1-11 List of MY2011 Models Certified to CARB's PZEV Requirements
Manufacturer
AUDI/VOLKSWAGEN
BMW
CHRYSLER
FORD
GENERAL MOTORS
HONDA
HYUNDAI
KIA
MAZDA
MERCEDES
MITSUBISHI
NISSAN
SUBARU
TOYOTA
VOLVO
Models
Audi A3, Volkswagen GTI, Volkswagen Jetta, Volkswagen Golf,
Volkswagen Jetta Sportwagen, Volkswagen EOS, Volkswagen CC
BMW 1281, BMW 3281, BMW 328C1
Chrysler 200, Dodge Avenger
Ford Escape Hybrid, Mercury Mariner Hybrid,
Mazda Tribute Hybrid, Ford Focus, Ford Fusion Hybrid,
Mercury Milan Hybrid, Lincoln MKZ Hybrid
Chevy Malibu, Buick LaCrosse, Buick Regal
Honda Civic GX (CNG-fueled), Honda Civic Hybrid,
Honda CR-Z Hybrid, Honda Insight Hybrid, Honda Accord
Hyundai Tucson, Hyundai Sonata
Kia Spoilage, Kia Forte, Kia Forte ECO, Kia Forte KOUP,
Kia Optima, Kia Optima Hybrid
Mazda 3, Mazda 6
Mercedes S400 Hybrid, Mercedes C300,
Mercedes C350/E350/GLK350
Mitsubishi Outlander Sport, Mitsubishi RVR, Mitsubishi Lancer,
Mitsubishi Lancer Sportback, Mitsubishi Galant
Nissan Sentra, Nissan Altima, Nissan Altima Hybrid
Subaru Legacy AWD, Subaru Outback Wagon AWD,
Subaru Forester AWD
Toyota Camry, Toyota Camry Hybrid, Toyota Prius Hybrid
Volvo S80, Volvo XC60, Volvo XC70
1.6.3 Hot Soak Plus Diurnal
The current baseline technology for LDVs, LDTs, and MDPVs is a properly designed
and assembled fuel/evaporative system for controlling emissions over the 2-and 3-day test
sequences to meet the current standard of 0.650 grams/test. This involves activated carbon
canisters which capture gasoline vapors, with engine calibrations designed to maximize canister
purge over the test sequence. Fuel systems generally include widespread use of various grades
of permeation-resistant materials.
The anticipated control technologies to comply with the proposed hot soak plus diurnal
evaporative emission standards include an improved carbon canister designs to even better
capture vapor emissions from the canister, air intake designs to prevent the escape of unburned
fuel from the engine's crankcase, various upgrades to further limit potential micro-sized leaks,
and further steps to reduce permeation rates. Applying these new or improved technologies will
allow manufacturers to meet the proposed 300 mg standard for LDVs/LDTls. The proposed
evaporative emission standards are slightly higher for larger vehicles to account for potentially
higher background emissions and in some cases larger surface area components, but the baseline
and anticipated control technologies follow a very similar path. These baseline and control
technologies are described further in the rest of this section.
1-43
-------
Current evaporative canisters use high working-capacity activated carbon, usually with
multiple compartments, to optimize vapor loading and purging behavior. These canisters
sometimes employ carbons of different working capacities within each chamber. Testing
indicates that the total canister adsorption capacity in grams of gasoline vapor is generally
dictated by the requirements of the refueling emission test and standard rather than the
evaporative emission test (either the 2 or 3-day sequence).
Manufacturers have identified the engine's intake system as another source of
evaporative emissions. These result from crankcase vapors and from unburned fuel injectors, or
sometimes from an injection event that occurred shortly before engine shutdown. We estimate a
typical emission rate of about 40 mg associated with each engine shutdown event; however,
since the actual emission rates depend on timing of individual injection events and cylinder
position at shutdown, baseline emission rates can vary significantly. These vapors must follow a
contorted path before reaching the ambient air, which would generally cause these emissions to
show up during the first day of the diurnal test rather than the hot soak test. One way to prevent
these emissions is to add activated carbon to the air intake downstream of the air filter, typically
in the form of reticulated foam coated with activated carbon. This device would have only a few
grams of working capacity and would be designed to purge easily to ensure that the vapor
storage is available at engine shutdown. This carbon insert would almost completely eliminate
any vapor emissions from the air intake system.
Manufacturers wanting to avoid adding any specialized emission control component to
control evaporative emissions from the air intake could pursue alternative approaches. First, it is
possible to allow the engine to continue rotating for 2-3 revolutions after engine shutdown to
sweep any hydrocarbon vapors from the intake system into or through the cylinder. These
vapors could be burned in the cylinder, oxidized at the catalyst, or stored until the engine starts
again. This may still allow for a small amount of residual vapor release, but this should be a
very small quantity. Vehicle owners would be unlikely to notice this amount of engine operation
after shutdown. Second, to the extent that manufacturers use direct injection, there should be no
fuel vapor coming from the intake system. Any unburned fuel coming from the injectors would
be preserved in the cylinder or released to the exhaust system and the catalyst. A small amount
of crankcase vapor might remain, but this would likely not be enough to justify adding carbon to
the intake system.
Fuel tanks are designed to limit permeation emissions. Fuel tanks are typically made of
high-density polyethylene with an embedded barrier layer of ethyl vinyl alcohol (EvOH)
representing about 1.8 percent of the average wall thickness. The EvOH layer is effective for
reducing permeation emissions. Recent developments in production processes have led to
improved barrier coverage around the ends of the tank where the molded plastic is pinch-welded
to form a closed vessel. We are expecting manufacturers to increase the EvOH barrier thickness
to about 3 percent of the average wall thickness to provide a more uniform barrier layer, to
provide better protection with ethanol-based fuels, and to improve permeation resistance
generally. These changes are expected to decrease emission rates over the diurnal test from
about 40 mg per day to 15 mg per day from the fuel tank assembly.
Fuel lines are also already designed for low permeation rates. The biggest portion of fuel
and vapor lines are made of metal, but that may still leave several feet of nonmetal fuel line.
1-44
-------
There may be development of new materials to further reduce permeation rates, but it is more
likely that manufacturers will adjust the mix of existing types of plastic fuel lines to achieve the
desired performance at the lowest possible price.
The bigger area of expected development with respect to fuel lines is to re-engineer fuel
systems to further reduce the number of connections between fuel-system components and other
fuel-line segments. Today these systems may involve more than the optimum number of
connections and segments due to assembly and production considerations or other factors.
Designing the fuel system more carefully to minimize connection points will limit possible paths
for fuel vapors to escape. This would reduce emission rates and it should also improve system
durability by eliminating potential failure points. A broader approach to addressing this source
of emissions is to integrate designs and to move fuel-system components inside the fuel tank,
which eliminates the concern for vapor emissions and permeation from those components and
connections.
A remaining area of potential evaporative emissions is the connection between the fill
neck and the fuel tank. Manufacturers can reduce emissions by perhaps 10 mg per day by
making this connection permeation-resistant. The challenge is to design a low-cost solution that
is easily assembled and works for the demanding performance needs related to
stiffness/flexibility. The best approach is likely either to use mating parts made from low-
permeation materials, or to use conventional materials but cover this joint with material that acts
as a barrier layer.
Purge rates are currently designed to flow relatively large volumes of outside air through
the canister when the purge solenoid is activated. This involves using available manifold
vacuum to create purge flow, with limits in place to avoid drawing too much unmetered fuel
vapor from the canister. Tightening the evaporative emission standard would lead manufacturers
to address remaining emission sources from micro-size leak points, permeation, and diffusion, as
noted above. Since the amount of additional vapor being captured by the carbon canister is small
and the test procedure is not changing, we do not expect the change in standards to drive changes
in purge strategy, rates, or canister capacity. Nonetheless, vehicle system and engine changes to
improve fuel economy could impact future purge strategies. Thus, as part of this approach,
manufacturers may incorporate designs to reduce vapor volume/mass directed to the canister and
thus potentially reduce the purge air volume requirements. In addition, canister designs can be
optimized to increase the effectiveness of a given volume of purge air. This could involve
selecting different combinations of carbon characteristics and canister architecture types and by
adding features to add heat (or preserve heat) in the canister during a purge event.
The technology discussed above is in use to varying degrees on many of the CARB
PZEV zero evap vehicles mentioned above. Taken together, we believe these technologies
provide manufacturers with effective tools for reducing emissions sufficiently to meet the
proposed evaporative emission standards.
1.6.4 Canister Bleed Emission Standard
More stringent evaporative emission standards have led to more careful measurements,
which led manufacturers to discover that 80 mg or more of fuel vapor would diffuse from the
1-45
-------
canister vent as a result of the normal redistribution of vapors within the activated carbon while
the vehicle is at rest. The emission rate depends on the tank volume, its fill quantity, and the size
and architecture of the canister and the characteristics of the carbon itself. While the biggest
effect of this vapor distribution is a uniform concentration within the canister, it can also cause
vapors to escape through the canister vent even without continued canister loading that would
result from fuel tank heating. These are referred to as canister bleed emissions. These emissions
occur to some degree during the 2- and 3-day evaporative emissions test, but a separate standard
is needed if the goal of near zero fuel vapor emissions is to be achieved.
The design to address this concern is a supplemental "scrubber" canister (or canister
compartment) with a very low working capacity carbon. Adding 100 or 200 ml of this type of
carbon near the canister vent provides a margin of "reserve capacity" to capture diffusion losses
from the canister. Since this extra carbon has low working capacity and it purges readily, it is
typically cleared of hydrocarbon vapors and ready to perform its function after any amount of
engine operation or even with natural back purge which occurs when the fuel in the tank cools.
This scrubber element is expected to eliminate all but 5-10 mg of emissions from the evaporative
canister over the measurement procedure.
1.6.5 Leak Emission Standard
Vapor leaks in the vehicle fuel/evaporative system can arise from micro-cracks or other
flaws in various fuel/evaporative system component structures or welds, problems with
component installations, and more generally from connections between components and fuel
lines and vapor lines. Because these emissions from these areas would occur in the 2-3 day
evaporative emissions test if the problems were present, manufacturers have taken steps to
address these potential problem areas as part of their overall evaporative emissions control
strategy. Since the 2-3-day hot soak plus diurnal standards are proposed to become even more
stringent and leak emissions occur during the evaporative emissions test, we expect
manufacturers to take the measures described above in Section 1.6.1. These include reducing
connections, improving the quality of fuel and vapor line connections, use of improved
component materials and revised installation practices. Manufacturers could also review their
OBD leak warranty data and related information from OBD queries to help inform their
strategies. One of the key reasons for focusing on a leak emission standard is to increase focus
on designs which will yield improved in-use emissions performance. EPA believes this focus on
in-use durability is important because a vehicle with even a small leak, say the 0.020-0.040 inch
orifice diameter monitored by OBD systems would likely emit above the proposed hot soak plus
diurnal evaporative emission standard in use. See Appendix 1 A: Technical Report on Leak Test
Procedure.
1-46
-------
References
1 Beck, D.D., Sommers, J. W., DiMaggio, C.L. "Impact of sulfur on model palladium-only catalysts under simulated
three-way operation." Applied Catalysis B: Environmental 3 (1994) 205-227.
2 Beck, D.D., Sommers, J. W. "Impact of sulfur on the performance of vehicle aged palladium monoliths." Applied
Catalysis B: Environmental 6 (1995) 185-200.
3 Beck, D.D., Sommers, J.W., DiMaggio, C.I. "Axial characterization of oxygen storage capacity in close coupled
lightoff and underfloor catalytic converters and impact of sulfur." Applied Catalysis B: Environmental 11 (1997)
273-290.
4 Waqif, M, Bazin, P., Saur, O. Lavalley, J.C., Blanchard, G., Touret, O. "Study of ceria sulfation." Applied
CatalysisB: Environmental 11 (1997) 193-205.
5 Bazin, P., Saur, O. Lavalley, J.C., Blanchard, G., Visciglio, V., Touret, O. "Influence of platinum on ceria
sulfation." Applied Catalysis B: Environmental 13 (1997) 265-274.
6 Takei, Y., Kungasa, Y., Okada, M., Tanaka, T. Fujimoto, Y. "Fuel Property Requirement for Advanced
Technology Engines." SAE Technical Paper 2000-01-2019.
7 Takei, Y., Kungasa, Y., Okada, M., Tanaka, T. Fujimoto, Y. "Fuel properties for advanced engines." Automotive
Engineering International 109 (2001) 12 117-120.
8 Kubsh, J.E., Anthony, J.W. "The Potential for Achieving Low Hydrocarbon and NOx Exhaust Emissions from
Large Light-Duty Gasoline Vehicles." SAE Technical Paper 2007-01-1261.
9 Shen, Y., Shuai, S., Wang, J. Xiao, J. "Effects of Gasoline Fuel Properties on Engine Performance." SAE
Technical Paper 2008-01-0628.
10 Ball, D., Clark, D., Moser, D. "Effects of Fuel Sulfur on FTP NOx Emissions from a PZEV 4 Cylinder
Application." SAE Technical Paper 2011-01-0300.
11 Heck, R.M., Farrauto, RJ. "Chapter 5: Catalyst Deactivation" in "Catalytic Air Pollution Control, 2nd Edition."
John Wiley and Sons, Inc., Publ., 2002.
12 Luo, T. Gorte, RJ. "A Mechanistic Study of Sulfur Poisoning of the Water-Gas-Shift Reaction Over Pd/Ceria."
Catalysis Letters, 85 (2003) Issues 3-4, pg. 139-146.
13 Li-Dun, A., Quan, D.Y. "Mechanism of sulfur poisoning of supported Pd(Pt)/Al2O3 catalysts for H2-O2 reaction."
Applied Catalysis 61 (1990) Issue 1, pg. 219-234.
14 Heck, R.M., Farrauto, RJ. "Chapter 6: Automotive Catalyst" in "Catalytic Air Pollution Control, 2nd Edition."
John Wiley and Sons, Inc., Publ., 2002.
15 Marcq, M. M. Chace, R.E., Xu, N., Podsiadlik, D.H. "The Effects of the Catalytic Converter and Fuel Sulfur
Level on Motor Vehicle Particulate Matter Emissions: Gasoline Vehicles." Environmental Science and Technology,
36 (2002) No. 2 pg. 276-282.
16 Tier 2 Regulatory Impact Analysis (EPA 420-R-99-023, December 22, 1999)
17 This test program is described in Chapter 6 of the RIA of the MSAT2 final rulemaking, available at
http://www.epa.gov/otaq/regs/toxics/420r07002chp6.pdf.
1-47
-------
18 Shapiro, E. "National Clean Gasoline - An Investigation of Costs and Benefits" Published by the Alliance of
Automobile Manufacturers, June 2009. Available at
http://www.autoalliance.org/files/NationalCleanGasolineJune09.pdf.
19 Ball, D., Clark, D., Moser, D. "Effects of Fuel Sulfur on FTP NOx Emissions from a PZEV 4 Cylinder
Application." SAE Technial Paper No. 2011-01-0300.
20 Butler, A. and Choi, D. "The Effects of Fuel Sulfur Level on Emissions In-Use Tier 2 Vehicles [expected
publication early 2012].
21 Shapiro, E. "National Clean Gasoline - An Investigation of Costs and Benefits" Published by the Alliance of
Automobile Manufacturers, June 2009. Available at
http://www.autoalliance.org/files/NationalCleanGasolineJune09.pdf.
22 Samenfink, W., Albrodt, H., Frank, M., Gesk, M., Melsheimer, A., Thurso, J., Matt, M. "Strategies to Reduce
HC-Emissions During the Cold Starting of a Port Fuel Injected Engine." SAE Technical Paper 2003-01-0627.
23 Samenfink, W., Albrodt, H., Frank, M., Gesk, M., Melsheimer, A., Thurso, J., Matt, M. "Strategies to Reduce
HC-Emissions During the Cold Starting of a Port Fuel Injected Engine." SAE Technical Paper 2003-01-0627.
24 Yi, J., Wooldridge, S., Coulson, G., Hilditch, J., Iver, C., Moilanen, P., Papaioannou, G., Reiche, D., Shelby, M.,
VanDerWege, B., Weaver, C., Xu, Z., Davis, G., Hinds, B., Schamel, A. "Development and Optimization of the
Ford 3.5L V6 EcoBoost Combustion System." SAE Technical Paper 2009-01-1494.
25 Choi, M., Sun, H., Lee, C., Myung, C., Kim, W., Choi, J. "The Study of HC Emission Characteristics and
Combustion Stability with Spark Timing Retard at Cold Start in Gasoline Engine Vehicle." SAE Technical Paper
2000-01-0182.
26 Eng, James A. "The Effect of Spark Retard on Engine-out Hydrocarbon Emissions." SAE Technical Paper 2005-
01-3867.
27 Hattori, M., Inoue, T., Mashiki, Z., Takenaka, A., Urushihata, H., Morino, S., Inohara, T. "Development of
Variable Valve Timing System Controlled by Electric Motor." SAE Technical Paper 2008-01-1358.
28 Ball, D., Zammit, M., Wuttke, J., Buitrago, C. "Investigation of LEV-HI Aftertreatment Designs." SAE Technical
Paper 2011-01-0301.
29 Serrano, D., Lavy, J., Kleeman, A., Zinola, S., Dumas, J., Le Mirronet, S., Heitz, D. "Post Oxidation Study During
Secondary Exhaust Air Injection for Fast Catalyst Light-Off." SAE Technical Paper 2009-01-2706.
30 Lee, D., Heywood, J. "Effects of Secondary Air Injection During Cold Start of SI Engines." SAE Technical Paper
2010-01-2124.
31 Sano, K., Kawai, T., Yoshizaki, S., Iwamoto, Y. "HC Adsorber System for SULEVs of Large Volume
Displacement." SAE Technical Paper 2007-01-0929.
32 McDonald, J.F., Schenk, C., Sanchez, L. J., Nelson, B.J. "Testing of Catalytic Exhaust Emission Control Systems
Under Simulated Locomotive Exhaust Conditions." SAE Technical Paper No. 2011-01-1313.
33 U.S. Code of Federal Regulations, Title 40, § 86.1823-08 Durability demonstration procedures for exhaust
emissions.
34 Christiansen, Michael G. "Impact of Lubricating Oil Condition on Exhaust Paniculate Matter Emissions from
Light Duty Vehicles" SAE Technical Paper No. 2010-01-1560
1-48
-------
35 NIOSH Reference Method 5040 - Elemental Carbon (Diesel Paniculate Matter). NIOSH Manual of Analytical
Methods (NMAM), Fourth Edition, 2003.
1-49
-------
I.A. Appendix to Chapter 1: Technical Report on Leak Test
Procedure
Technical Report
Draft Recommended Test Procedure
and Supporting Testing Data for the Proposed
Evaporative Emissions Leak Test
By
Peter Smith
July 2012
NOTICE
Technical Reports do not necessarily represent final EPA decisions or positions. They are
intended to present technical analysis of issues using data or testing results which are currently
available. The purpose in the release of such reports is to facilitate the exchange of technical
information and to inform the public of technical developments which may form the basis for a
final EPA decision, position or regulatory action.
Assessment and Standards Division
Office of Air and Radiation
U.S. Environmental Protection Agency
1A-1
-------
*** E.O. 12866 Review-Revised Version-Do Not Cite, Quote, or Release During Review **
I. Introduction
Evaporative emission control systems have been present on new light-duty vehicles since the
1971 model year. These systems have gone through several generations of upgrades over the years and
have become more sophisticated and complex as fuel properties have varied, test procedures have
changed, emission standards have increased in stringency, and new requirements have been added to meet
greater demands for improved air quality. These onboard systems are currently responsible for capturing
evaporative emissions produced from gasoline during diurnal, hot soak and running conditions as well as
in vehicle refueling situations. Toluene and benzene, which currently account for about five percent of
the tank headspace vapor composition, have been linked to possible adverse health effects.1'2 In addition
these compounds and others commonly found in gasoline vapor such as butane and pentane are also
responsible for the creation of photochemical smog. Due to these concerns, EPA has sought to increase
the capability of these systems to strictly limit the release of these compounds to the atmosphere.
Over the past several months, the Agency has developed an evaporative emissions leak test
procedure which could be used in both laboratory and field conditions to quickly test the integrity of a test
vehicle's evaporative system. Such a test is independent of On-Board Diagnostics (OBD) leak detection
requirements and as such would not be bound by the limitations and restrictions commonly present in
such a monitoring system. The leak test as presently envisioned would be a stand-alone test designed to
find any vapor leaks in the fuel/evaporative system. This test is not designed to indentify leaks outside of
the fuel and vapor containment portion of the vehicle fuel system.
Past and current work at the Agency has found that not all evaporative emissions leaks are being
identified by current OBD systems. This is especially important since these emissions are a potentially
significant portion of the evaporative emissions inventory. This test procedure and accompanying leak
standard could serve as a means to reduce vapor from system leaks in use. Concurrent with Tier 3
emissions levels it is being proposed that the leak test emission standard be proposed at a value of 0.02
inches cumulative diameter orifice.
This report details the evaporative emissions leak test procedure as presently proposed as well as
the vehicle preconditioning necessary prior to undergoing the leak test. It also presents the various testing
that has been performed at NVFEL both in the creation of the test procedure as well as supporting data to
ensure that the leak test provides accurate results on as many different vehicle configurations as possible.
Examples of calculations of relative leak orifice sizes are also presented.
II. History of Evaporative Emissions Controls and Standards
Even though present day fuel evaporative emission controls are relatively mature, the
development of the present emission standards and related requirements and the control technology
applied by the manufacturers has evolved over about 40 years. The activated carbon canister and the
basics of evaporative system as it is known today for controlling hot soak and diurnal emissions were
required by the for 1971 model year light -duty vehicles even before the creation of EPA. The purpose
of this system is to capture fuel vapors created by the presence or operation of the motor vehicle and store
these hydrocarbon vapors in an onboard activated charcoal media known as the charcoal canister. Until
recently, most manufacturers placed this canister under the hood where it was responsible for capturing
the fuel vapors generated when the fuel tank was heated via atmospheric conditions (diurnal cycle) as
well as the vapors generated from the carburetor float bowls (hot soak) and fuel tank when the vehicle
was parked. In uncontrolled vehicles fuel tank and carburetor evaporative emissions were vented to the
atmosphere. The fuel vapors generated during vehicle operation, known as "running loss" were basically
not controlled in early designs.
1A-2
-------
*** E.O. 12866 Review-Revised Version-Do Not Cite, Quote, or Release During Review **
These early systems did not focus on full vehicle evaporative emissions thus early carbon trap test
procedures were deficient in gaining control. Early evaporative systems did not have any type of
microprocessor interface for controlling vapor loads and purge and were controlled via a series of vacuum
check valves, mechanical linkages, and/or solenoids. The canister had a finite storage volume, much like
present day systems, and was purged of stored vapors during vehicle operation. Unfortunately, many of
these early systems were disconnected or otherwise modified by unknowing motorists thinking these
systems had a detrimental effect on the performance or running operation of their vehicle.
Over the past 40 years there have been about five major changes in evaporative emission control
requirements. These have involved test procedure changes, more stringent numerical standards, control
of evaporative emissions from other sources on the vehicle, or control of emissions from various other
operating modes. Following the initial requirements for the 1971 model year, noteworthy changes in the
regulatory requirements as indicated in the Code of Federal Regulations included: (1) the change from the
carbon trap to SHED test method in 1978, (2) the numerically more stringent 1981 model year hot soak
plus diurnal evaporative emission standards, (3) enhanced evaporative requirements which phased-in for
the 1996—1999 model years and incorporated multi-day hot soak plus test requirements and emission
standards, fuel spit back standards, and running loss controls,(4) new onboard refueling vapor recovery
requirements which phased in from model years 1998-2006, and (5) the Tier 2/MSAT requirements for
2004-2010 model years which further tightened the hot soak plus standards, addressed resting losses, and
brought in to play consideration of ethanol-blend fuels. New provisions for "zero evaporative" vehicles
are now in place in California as part of the PZEV allowance in the ZEV program and have recently been
included in the evaporative emission control standards for California's LEV III program. EPA expects to
propose "zero evaporative" emission standards as parts of its Tier 3 rule to include a more stringent multi-
day hot soak plus diurnal emission standard and a canister bleed emission standard and test procedure.
Computerized emission control system monitoring became necessary when the California Air
Resources Board (ARB) mandated that all vehicles sold in that state in 1988 have an onboard diagnostic
system to detect problems in the performance of emission control systems. These early systems were
known as On Board Diagnostics (OBD) monitored the oxygen sensor, the fuel delivery system, engine
control module, and the exhaust gas recirculation (EGR) system. If a problem is detected with a vehicle
or engine system that is monitored by OBD, the OBD monitor illuminates a Check Engine or Malfunction
Indicator Light (MIL) and stores a diagnostic trouble code in memory which can be later retrieved. While
some of these codes and problems may be noticed through vehicle drivability changes or other discernible
problems with the vehicle performance, some are not. Thus, the MIL provides the driver notice of a need
to check for a problem in the operation or performance of a given vehicle/engine system based on the
diagnostic trouble code (DTC) set in the OBD system. Vehicle-based computers also created the
opportunity for more sophisticated and precise control of canister purge and management of the effects of
canister purge on exhaust emission rates.
These controls were further refined with the advent of OBDII which was required beginning in
1996 with full compliance by 1999. OBDII requirements enhanced the requirements of the original OBD
systems and added specific requirements related to detecting problems in the fuel /evaporative control
system. Current OBD systems are required to monitor for vapor leaks of various diameters within the
fuel /evaporative control system and to check for proper operation of the purge system, and other general
malfunctions related to items such as pressure sensors. It is clear that OBD and the use of the onboard
computer control system has greatly increased the ability of vehicle evaporative and refueling control
systems to function more precisely and effectively and to monitor for problems in performance.
1A-3
-------
*** E.O. 12866 Review-Revised Version-Do Not Cite, Quote, or Release During Review **
III. Reasons for Creation of an Evaporative Emissions Leak Test
At certification current evaporative and refueling emission standards provide reductions in excess
of 95 percent over uncontrolled levels when evaluated over the specified test procedures. While programs
such as OBD, extended useful life requirements, and in-use verification programs (IUVP) testing have
improved the overall in-use performance of these systems, full life durability and performance in off cycle
situations remain important objectives.
Even with the upgrades in technology and test procedures over the past years, it was possible for
a vehicle with a small leak in the fuel/evaporative control system to meet the hot soak plus diurnal
emission standards. This is because leaks are not expected in certification vehicles and more generally
because the test procedures and standards were not as rigorous and did not require control under all of the
various in-use conditions which affect evaporative emissions . Some leak emissions could be
accommodated and the vehicle would still pass the standard. The leak detection requirements of the
OBDII regulations have helped to focus more attention on preventing leaks over the vehicle life.
However, from an in-use performance perspective OBD alone is not sufficient for at several reasons.
First, OBD systems do not operate under all conditions and do not require that the owner/operator seek a
repair if a MIL is indicated. While repairs are far more likely areas with inspection /maintenance (I/M)
programs, repairs are not always made and if so they may well wait until the I/M check is due. Second, is
the issue of OBD system readiness. There are operating circumstances when an OBD system is not
required to monitor the evaporative system. These are commonly termed global disables and deficiencies.
For example, global disables include specifications on outside air temperature, elevation (i.e., barometric
pressure), and tank fuel volume. Deficiencies include manufacturer specific exceptions provided on a
case-by-case basis when the system cannot meet the OBD monitoring requirements either due to an
element of fuel/evaporative system design or if there is some operating condition in which the OBD leak
detection methodology does not operates reliably in the sense of not giving false fails or false passes. If
the vehicle is operating in one of these "disable" or "deficiency" conditions the OBD system will not run
and not detect a leak. Third, OBD systems are not required to detect leaks of less than 0.020". While the
prevalence of these very small leaks is not known, it is reasonable to project that they are at least as
prevalent as larger leaks. Limited data indicates that the leak emission rate for vehicles with leaks less
than a 0.020" diameter orifice is not significantly less than at 0.020". Thus, they are potentially
significant.
Tier 3 evaporative emission standards are intended to allow for essentially zero fuel vapor
emissions over the vehicle useful life. Full life Tier 3 evaporative emission standards coupled with
upgraded OBD will help to encourage focus on designs and technology which will improve in-use
durability. However, a vehicle with a vapor leak in the fuel evaporative emission system will emit VOCs
at a greater rate than anticipated from the "zero evaporative" technology on which the proposed Tier 3
evaporative emission standard is based. To help insure that the reductions envisioned by proposed Tier 3
evaporative emission standards are achieved in-use, EPA is considering an evaporative emissions leak test
procedure and standard. The leak emission standard would help to ensure added focus on efforts to
improve in-use durability by reducing the diameter and prevalence of leaks.
The remainder of this memorandum covers the following areas. Section IV, which follows next,
discusses current leak measurement methods and work done by EPA. Following this is a presentation on
the basics of the recommended test procedure, (section V), a validation of the recommended test
procedure by testing on some more unusual fuel/evaporative system configurations (section VI), and
finally an assessment of the leak test procedure results to several key pre-conditioning variables (section
VII).
1A-4
-------
*** E.O. 12866 Review-Revised Version-Do Not Cite, Quote, or Release During Review **
IV. Measurement Methods
At present, at least two different methods exist for determining evaporative system leaks. The
first and more commonly used method is via positive pressure and flow measurement of the system. In
this instance, the evaporative system is pressurized with a gas (normally inert nitrogen) to a set and
measured pressure. In-line with the test apparatus is a flow meter designed to measure the flow of the test
gas being introduced to the evaporative system. Once the system is pressurized to a certain level, the
resulting gas flow is measured and compared to that at known conditions and leak sizes. In this manner,
any resultant leaks on the vehicle's evaporative system are determined.
The second manner in which leaks are measured and quantified is through a process of measured
vacuum decay. In this method, a certain amount of pressure or vacuum is placed on the evaporative
system and then monitored as to how long the evaporative system takes to release any such pressure or
vacuum. By comparing this rate to that of known decay rates for an identical system, any such leak
present in the test system can be quantified.
a. Pressure and Flow Measurement Systems
The most common devices for finding evaporative leaks are based on the pressure and flow
measurement basis. There are many versions on the market today with some being used in individual
states' I/M programs. Most of these devices will use an external source of inert test gas so as to mitigate
the combustion hazard of mixing oxygen with flammable hydrocarbon vapors.
i. Current Production Versions
Two examples of current production leak testers are shown in Figures 6 and 7. They are the
EELD (Evaporative Emissions Leak Detector) 500 tester produced by Snap On and the Delphi 2010
produced by Delphi. Both devices operate on the same principle of flow and pressure measurement into
the test vehicle.
Figure 1 - Snap On EELD 500 Tester Figure 2 - Delphi 2010 Evaporative Tester
The Snap On unit has the capability to also introduce smoke into the test vehicle's evaporative
system in order to determine not only the resultant leak size but also its location. The Delphi unit also has
an optional smoke generation feature. Both machines are similar in that they also provide for individual
calibration of the units. The EELD performs a calibration check each time the unit is switched on, while
the Delphi unit comes with a test tank and calibrated orifices to perform a field calibration of the unit.
1A-5
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
Both devices are portable with onboard rechargeable batteries or 110V power and both devices pressurize
the evaporative system with up to 18" of water pressure.
The EELD unit was extensively tested by NVFEL staff over the course of several months as well
as compared to the EPA Test Console created using laboratory grade equipment also designed to measure
leaks. While we used this tool in our evaluation program and in developing the test procedure,
manufacturers can use any tool or other method which meets the requirements of the regulations.
Several tests were conducted so as to ascertain the EELD's performance. These test units were
purchased by NVFEL for testing through normal distribution channels. All measurements were
obtained using a series of stainless or brass orifices obtained from O'Keefe Controls. The orifices tested
ranged in sizes under 0.010 to over 0.050. Figure 8 shows the relative performance between two different
EELD units, serial #'s SA 3890 and SA 3888, respectively. In addition # SA 3890 was also tested in the
same manner on a different day to determine any day to day variability with the units. Upon analysis of
the results, it was noticed that both SA 3890 and SA 3888 units underestimated the size of the resultant
evaporative leak on the order of 0.003-0.004" as compared to the orifices obtained from O'Keefe
Controls. Discussion with the parent company Star Envirotech, which produces the unit for Snap On,
resulted in the procurement on a new prototype test device which did not include the smoke feature. The
performance of this new version is indicated by the "New Board" data on the graph in Figure 3. Its
performance was greatly improved versus the production versions. Upon further testing it was realized
that the flowing of smoke through the test orifice displaced a certain volume of measured test gas which
caused the units to indicate a leak size less than actual.
Snap On Evap Tester Performance
•SA 3890 -10/3
•SA3B90
SA3SS8
•Target curve
Board
0-010 0-020
D.030 0.040
Orifice Size
Figure 3 - Production EELD Samples with Smoke vs Non Smoke Version
As seen in Figure 4, the EELD without the smoke feature was also compared directly to the
results obtained via the NVFEL Test Console.
1A-6
-------
' K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
Snap On Tester v2 vs EPA Test Console
"Target Orifice j
-Evap Console
• Snap On
Testtf
Figure 4 - Comparison of EELD Non Smoke Board vs Test Console
In all instances, the EELD version without the smoke compared favorably to the Test Console.
Some variation was seen at the extreme ends of the orifice sizes, primarily under 0.010" and over 0.045".
Since leaks at these ranges were not of concern for the ongoing test program, the EELD was used for
validation of the test procedure due to its ease of use and fast reporting of test results. The raw data from
the testing is presented below in Figure 5. It is important to note that in the range of interest 0.010 -
0.030, that both the Test Console and the EELD displayed similar errors in the calculation of the test
orifice size. This could possibly be due to slightly inconsistent sizing in the test orifices themselves
leading to even better unit performance once these sizing errors are taken into account.
Sraa OnTtitEr Neiv Boa^ and Ev^p Teit Conso* Resuls
Test Console Snap On
O-ifcc Sie iin|i Output Value (ir I New Eo; -d (ir |
UM
0.008
0.012
0-014
OJHB
O.OZO
0022
0.023
O.O24
0025
ooze
0.027
ana
0.031
0.035
OU050
0060
MM
Q.OOS1
o.oiia
0.0112
0.0192
0.0198
0.0219
0.0219
0.024
0.0262
0.0258
0.0273
0.0293
0.0311
0.0332
OJM9
0.056
0.001
0.006
o.cai
0.013
0.019
0.02
0.021
0.024
0.023
0.025
0.024
0.026
0.02S
0.03
0.034
0056
0.072
Test Con&de
Error
•5JH
12M
-L7K
MS
6.7*
-UK
-0.5V
tK
OIK
4.8K
-O.R
11*
LBX
0.3V
-5.W
-2.0K
-«.7JS
Srap On
Error
-75.0S
-25.014
-a.m
-7.1U
s.eu
O.Oli
•4.5*4
4.IU
-4>2«
o.ov
-7.71*
-3.7K
-l.«
•3.2V
-2.9VI
12.01i
2am
Figure 5 - Data Obtained from Comparison of Test Console and EELD
1A-7
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
ii. NVFEL Constructed Device
The EPA constructed test device is also based on the principle of flow and pressure measurement.
The test device (Test Console) was built using a Honeywell FP200 pressure transducer calibrated in the
range of 0-2.5 psi as well as Teledyne Hastings HFM-300 flow meter s calibrated in ranges of 0-2 slpm
and 0-10 slpm. Figure 6 depicts the flow layout and construction of the test console.
Supply Gas
AA
3-50 psi check valve
AA
„
AA
AA
M
3 LPM ' 10 LPM
1/3 psi
check valve
Vent valve
To tank
Figure 6 - NVFEL Test Console Flow Diagram
The equation presented in the test procedure (see below) was developed through NVFEL testing.
It allows the calculation of an effective orifice leak diameter in inches based on the type of test gas being
used, the temperature of the test gas, as well as the measured pressure and flow of the test gas through the
Test Console. Figure 5 shows the correlation between the test console equation constants and empirical
data obtained from a wide range of O'Keefe orifices. In all measurements, error was less than 3 percent
from actual to measured values.
The overall equation as presented in the test procedure was then implemented in the control
software for the Test Console. Flow and pressure values were recorded by flowing N2 gas through the
test console and test orifice. A variety of orifice sizes were tested ranging from 0.008" to 0.031". The
pressure was varied from 0.25 to 1 psi and the corresponding flow recorded at each pressure point. The
orifice diameter was calculated via software using the supplied equation and plotted in Figure 7.
1A-8
-------
* * * K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
0.045
0.04
0.035
0.03
1/1
01
« 0.02
6
0.015
0.01
0.005
= 0.2147x°5053
R2 = 0.9993
'Empirical Data
iTheoretical Data
Power (Empirical Data)
0 0.01 0.02 0.03 0.04 0.05
Cv
0.06
Figure 7 - Validation of Test Console Equation Constants
1A-9
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
The actual orifice size is modeled as a straight line and is indicated as the orifice size - TGT on
the plot legend from Figure 8. For all orifice sizes tested, the equation deviated from the actual value by
less than 0.001" in actual reading. Good correlation was seen throughout the pressure ranges tested.
Evaporative Test Equation Results
0.035
0.03
£ 0.025
cu
S 0.02
E
5 0.015
01
o 0.01
0.005
0
-.008_TGT
0.012
-.012 TGT
-.014 TGT
•0.02
-.020 TGT
0.025
0.5 1
Test Pressure (psi)
1.5
0.031
-.031 TGT
.008 2
Figure 8 - Validation of Supplied Test Procedure Equation
b. Vacuum or Pressure Decay Measurement Systems
Most commonly available evaporative test devices operate on the principle of pressure and flow
measurement to find evaporative system leaks. At the time of the report, initial conversations were
underway with Mahle Test Systems regarding their new tester for evaporative systems that operated on
the principle of drawing a fixed vacuum on the evaporative system and measuring the decay over a short
period of time. Currently such a system is being used in several OEM assembly plants to find leaks prior
to the vehicles being fueled. According to Mahle, proprietary algorithms allow leak detection down to
0.010" and smaller can be found on less than 0.5 psig of vacuum. Typical test times are on the order of
10-30 seconds. No vacuum or pressure decay systems were available for evaluation.
1A-10
-------
*** E.O. 12866 Review-Revised Version-Do Not Cite, Quote, or Release During Review **
V. Recommended Test Procedure
Based on our developmental work with the EPA NVFEL Console and the Snap-On EELD 500
(no smoke) meter, EPA developed a draft recommended test procedure. The proposed test procedure for
fuel/evaporative system leak testing provided in Code of Federal Regulation format is as follows. It does
not require the use of any specific instrument.
§86.167-18 Leak test procedure.
(a) Scope. Verify that there are no significant leaks in your fuel storage system using the leak test
described in this section. Perform this check as required in the standard-setting part.
(b) Measurement principles. A leak may be detected by measuring pressure, temperature, and flow to
calculate an equivalent orifice diameter for the system. Use good engineering judgment to develop and
implement leak test equipment. Your leak test equipment must meet the following requirements:
(1) Pressure, temperature, and flow sensors must be calibrated with NIST-traceable standards.
(2) Correct flow measurements to standard temperature and pressure of 20 °C and 101.3 kPa.
(3) Leak test equipment must have the ability to pressurize fuel storage systems to at least 4.1 kPa and
have an internal leak rate of less than 0.20 slpm.
(4) You must be able to attach the test equipment to the vehicle without permanent alteration of the fuel
storage or evaporative emission control systems.
(5) The point of attachment to the fuel storage system must allow pressurization to test system integrity of
the fuel tank and fuel and vapor lines reaching up to and including the gas cap and the evaporative
canister. An example of an effective attachment point is the evaporative emission system test port
available on some vehicles.
(c) Leak test procedure. Test a vehicle's fuel storage system for leaks as follows:
(1) Fill the vehicle's fuel tank to 40% capacity.
(2) Soak the vehicle for 6 to 24 hours at a temperature of 20 to 30 °C and maintain this temperature
throughout the leak test.
(3) Before performing the test, purge the fuel storage system of any residual pressure, bringing the system
into equilibration with the ambient.
(4) Seal the evaporative canister's vent to atmosphere and ensure that the system purge valve is closed.
(5) Attach the leak test equipment to the vehicle.
(6) Pressurize the fuel storage system with nitrogen or another inert gas to at least 2.4 kPa. Use good
engineering judgment to avoid over pressurizing the system.
(7) Maintain gas flow through the system for at least three minutes, ensuring that the flow reading is
stable for an effective leak diameter of ± 0.002 inches.
1A-11
-------
* * * K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
(8) Use the following equation, or a different equation you develop based on good engineering judgment,
to calculate the effective leak diameter, Z)eff.
x0.5057
Deff =0.2153-
gas
960-
Pl-P2 • Pl+P2
G- T + 459.67
Where:
Z)eff = Effective leak diameter, rounded to the nearest 0.01 inch.
Pgas = Volumetric flow of gas (scfh).
PI = Inlet pressure to orifice (psia).
P2 = Atmospheric pressure (psia).
G = Specific gravity of gas at 14.7 psia and 60°F.
T= Temperature of flowing medium (°F).
(9) You may perform any number of replicate tests; however, you must perform the same number of tests
at the same attachment points for every vehicle from a given model year. The average value of replicate
tests is the official result for a given vehicle.
(10) You may use special or alternative test procedures as described in 40 CFR 1065.10(c). For example,
a manufacturer may use a vacuum-based instrument or approach if it can demonstrate correlation and
meet the requirements of the applicable regulations.
(d) Equipment calibration. Use good engineering judgment to calibrate the leak check device.
For example, you may compare measured and calculated values to a calibration orifice such as that
defined as an O'Keefe Controls Co, Type B precision-machined brass or stainless steel orifice having a
gas flow path depicted in the following figure:
1A-12
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
Figure 1 of §86.167
Orifice 'hameTei 'i21"m laraer
Orifice diameter ,020" or smaller
Validation of Test Procedure
There are many minor design and architecture variations, but essentially all vehicle evaporative
control systems involve the use of a carbon canister to capture vapors from the gasoline tank and a
vent and purge solenoid system for refreshing the storage capacity. Leak detection approaches as part
of this system are based on either natural vacuum or an integrated pump and orifice approach. The
leak detection instruments described above are commercially available and in use in the auto industry
today; there is a high degree of confidence in their capability to detect leaks when applied properly
for the intended purpose. This was confirmed during the testing we conducted on a number of
different production vehicles in developing this recommended practice. For further validation
purposes, we looked at several unique design and test conditions to confirm the utility of the test
equipment and test procedure in these applications. These included testing on a Ford F150 with the
"No Cap" style fuel fill, a Chrysler 300 equipped with both an ESIM valve and a "No Cap" style fuel
fill, a Ford E350 cargo van with a fuel tank larger than 20 gallons and a Honda Accord. For all the
vehicles tested, the following EELD tester procedure was used.
Evaporative Emissions EELD Tester Attachment Procedure
Connect EELD Tester to Test Vehicle
a. Locate purge solenoid (typically located under hood nearby intake manifold or throttle
body)
b. Remove vapor line on fuel tank side of purge solenoid (various clips can be used)
c. Insert hose adapter and modify as necessary to ensure a proper seal to the vapor line
Attach Scan Tool to OBD port
a. Verify engine size and details from under hood emissions label
b. Follow options and screens necessary to allow tool to control evaporative vent solenoid
c. Arrive at vent solenoid command screen (typically under special tests or emissions test
sub-menus)
Power up EELD Tester
a. Attach N2 gas to tester (set to 85-100 psi inlet)
b. Select menu for proper gas while tester is booting (must do this at the right time or tester
will default to AIR for test medium) Menu selection is done by pressing display select on
the side of the tester until N2 is displayed. Tester can accommodate N2, CO2 and air for
test gas.
Conduct Test
a. Command tester to flow nitrogen to test vehicle.
b. Close vent solenoid on test vehicle using the scan tool
1A-13
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
c. Monitor flow and pressure in tank and read resultant orifice leak size on tester screen.
(Test value will start off high and stabilize to a final reading as fuel tank is pressurized.
Test runs for 5 minutes and then tester will turn off. To continue flowing nitrogen simply
command the EELD to test again)
a. Ford "No Cap" System
The Ford vehicle tested was a 2011 Ford F150 pickup VIN# 1FTFX1CT5BFA93342. The
vehicle had 252 miles on it at the time of testing. The fuel tank was drained and filled to 40 percent with
certification fuel. The vent solenoid was commanded by a Genisys OTC scan tool. The system was
pressurized by disconnecting the evaporative line from the purge solenoid and pressurizing the system via
this interface point. All testing was done with the updated version of the EELD500 which did not include
the provision for smoke. The inlet pressure for nitrogen test gas for the EELD was set to 90 psi.
Readings were taken approximately four minutes after the start of each test. Seven test repeats were run
with the tester disconnected and the system allowed to vent for several minutes between tests.
This vehicle was chosen in order to asses any differences in testing necessary to accommodate
Ford Motor Company's "No Cap" style of fuel tank system. In all cases, a well sealed system was
indicated. The indicated leak value on the EELD was at .000 at four minutes duration for each of the
seven tests. It took approximately 3.5 minutes for the tester to pressurize the tank and indicate a .000
reading. The final pressure indicated by the EELD was relatively consistent as well, with readings
ranging from 17.3 - 17.6 inches water. No changes to the test procedure were necessary for this
application.
b. Chrysler " No Cap" ESIM System
The next configuration tested was a 2012 Chrysler 300 VIN# 2C3CCACG2CH266160 equipped
with Chrysler's version of the "No Cap" fuel filler system as well as the ESIM style of vent valve. The
vehicle had 1942 miles on it at time of testing. The updated EELD unit (no smoke) was used to test this
vehicle. The fuel tank was drained and filled identically as the F150 previously tested. Because of the
year of the vehicle, available scan tool equipment was unable to communicate with the vehicle's ECM.
In all testing, the vent line was disconnected from the ESIM valve. The evaporative system was
pressurized via the line going to the under hood purge solenoid. The ESIM valve and charcoal canister
were located behind the right hand rear wheel well liner.
This vehicle was chosen in order to asses any differences in testing necessary to accommodate
Chrysler's "No Cap" style of fuel tank system as well as any special procedures needed in testing a
system which did not have a traditional solenoid controlled vent valve. In all cases, a well sealed system
was indicated. The indicated leak value on the EELD was at .000 at four minutes duration for each of the
seven tests. It took approximately 2.0 minutes for the tester to pressurize the tank and indicate a .000
reading. The final pressure indicated by the EELD was relatively consistent as well with readings ranging
from 17.4 to 17.5 inches water. No changes to the test procedure were necessary for the application with
the exception of the understanding that sealing the vent line from the ESIM valve acted in a similar
fashion as commanding the vent valve closed on a traditional system. Figure 9 below shows the "No
Cap" style of filler system. The Ford system is similar with the exception that it does not have the sealing
cover for the filler inlet pipe.
1A-14
-------
' K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
Figure 9 - Chrysler 300 "No Cap" Fuel Filler Inlet
d. Fuel Tank Systems Greater Than 20 Gallons
The vehicle tested was a 2012 Ford E350 Passenger Van VIN# 1FBSS3BL3CDA75451. The
vehicle had 3452 miles on it at the time of testing. The fuel tank was drained and filled to 40 percent
certification test fuel. The capacity of the fuel tank was 40 gallons as indicated in the owner's manual and
via measurement of the external tank dimensions. The vent solenoid was commanded by a Genisys OTC
scan tool. The system was pressurized by disconnecting the evaporative line from the purge solenoid and
pressurizing the system via this interface point. Figure 10 shows the EELD attachment point. All testing
was done with the updated version of the EELD500 which did not include the provision for smoke. The
inlet pressure for nitrogen test gas for the EELD was set to 90 psi. Readings were taken approximately
four and a half minutes after the start of each test. Seven test repeats were run with the tester
disconnected and the system allowed to vent for several minutes between tests.
Figure 10 - Ford E350 EELD Attachment Point
This vehicle was chosen in order to asses any differences in testing necessary to accommodate
tank sizes larger than 20 gallons. The indicated leak value on the EELD ranged from 0.001 to 0.005 for
each of the seven test repeats for this vehicle. There was indication that the default five minute test time
on the EELD was not sufficient for this vehicle to fully pressurize such a large tank volume leading to the
variation in "found" leak values. In addition, it was seen that the variation in indicated leak values was
1A-15
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
affected by the depressurize time between tests. In prior testing, two minutes was allocated between tests
to allow the evaporative system and fuel tank to equalize to atmospheric pressure with an open vent valve
and fuel cap. With the larger tank system of 40 gallons two minutes was seen as not enough
depressurization time. If the subsequent test was started within two minutes, the ramp time down to the
indicated leak value was faster than that measured between tests with more than a two minute wait time.
Figure 11 shows this effect. Based on this testing , it is recommended, when testing large tank sizes, to
wait at least 5 minutes between test intervals if repeated testing is necessary as well as to test longer than
5 minutes (in the case of the EELD) in order to give larger systems the appropriate pressurization time
needed to achieve representative results. It should be noted that even though the larger system took
longer to fill and achieve a final reading, indicated leak values were within .004" of each other after three
minutes of testing.
Indicated Leak Value Indicated Leak Value
Test Time (min)
w/ 2 Min Wait
0.014
0.006
0.002
0.001
w/ 5 Min Wait
0.021
0.01
0.006
0.004
0.002
Figure 11 - EELD Indicated Values on 40 Gallon Test Tank
e. Honda Produced Fuel Systems
The Honda vehicle tested was a 2007 Accord sedan VIN# 1HGCM56727A141399. The vehicle
had 123821 miles on it at the time of testing. The fuel tank was drained and filled to 40 percent with
certification test fuel. The vent solenoid was commanded by a Vectronix Mastertech scan tool. The
system was pressurized by disconnecting the evaporative line from the purge solenoid and pressurizing
the system via this interface point. All testing was done with the updated version of the EELD500 which
did not include the provision for smoke. The inlet pressure for nitrogen test gas for the EELD was set to
95 psi. Readings were taken approximately four minutes after the start of each test. Seven test repeats
were run with the tester disconnected and the system allowed to vent for several minutes between tests.
This vehicle was chosen based on concerns initially expressed by Honda regarding the interface
of the leak detection test methodology with their OBD leak detection scheme. We assessed a Honda
1A-16
-------
*** E.O. 12866 Review-Revised Version-Do Not Cite, Quote, or Release During Review **
vehicle equipped with a Honda evaporative control system using a solenoid controlled vent and purge
valve. In all cases, a well sealed system was indicated. The indicated leak value on the EELD was at
.000 at four minutes duration for each of the seven tests. It took approximately 3 minutes for the tester to
pressurize the tank and indicate a .000 reading. The final pressure indicated by the EELD was relatively
consistent as well with readings ranging from 17.5 - 17.7 inches water. No changes to the test procedure
were necessary for this application.
VI. Tested Vehicle Variables for Preconditioning
Current preconditioning required prior to conducting the evaporative test is as follows:
1. 6-24 our soak at 20-30 °C prior and during testing
2. 40 percent fuel fill capacity using Tier III certification fuel (9 RVP El5)
3. Purge of any residual fuel tank pressure prior to testing
The cold soak and fuel tank fill level is consistent with the preconditioning needed for the FTP test so as
to minimize any additional testing burden associated with running the leak test. Prior to setting these
preconditioning requirements, several factors that were thought to possibly influence the leak test results
were investigated. These factors were the fuel fill level, the fuel tank temperature, test vehicle inclination
and the leak location relative to the insertion point of the leak tester. These factors were investigated by
using a first generation Snap On EELD500 tester. The EELD had the provision to introduce smoke into
the evaporative system but was not filled with the smoke solution for this testing. The vehicle used for all
of the preconditioning investigation was a 2006 Chevrolet Silverado, VIN# 1GCEK19B6GZ154114.
This vehicle had approximately 110,056 miles on it and the vent valve was controlled via a GM Tech 2
service tool. An evaporative test port was located under the hood and was used as an attachment point for
the EELD tester.
a. Fuel Fill Level
To investigate the influence of fuel fill level on the leak test, the test vehicle was drained and
filled to various levels of certification test fuel. The levels investigated were 10, 40, 80, and 100 percent
tank capacity. Each test was performed at a temperature of 25°C. At each fuel level, seven repeat tests
were performed with the EELD tester. After each reading the tester was removed and the system was
allowed to depressurize for several minutes. A simulated leak was created by tying in a 0.020 stainless
O'Keefe orifice at the purge valve. The results are presented below in Figure 12. The error bars
represent the inherent repeatability of the unit. It was felt that the EELD was accurate to a +/- .001"
reading while operating under the testing conditions. While the 10 percent fill levels had the highest
indicated leak reading, on average these were .001" or less than what was indicated at the 40 percent fill
levels. This difference was also within the accuracy of the EELD unit. So based on the collected data, it
was determined that the fuel fill level had a minimal, if any, effect on the test reading.
1A-17
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
Fuel
On?
Omo -
0)
N f\ fli Ji
>
0) I
LJ n m 7 - —
£ 0.017 j ,
~ T T
On mfi -
T3 T 1
0)
*j n m i Sk
•*; u.uu
^ n nid -
Om ^ -
Om 7
012
Tank Fill Effects (25 C, 0.020")
0.020" Implanted Leak
T T T !
1 i * I I
I T I
X * i i •
4 40% Fill
• 10% Fill
A 80% FIN
X 100% Fill
345678
Test Point
Figure 12 - EELD Readings for Several Fuel Fill Levels
b. Fuel Temperature
To investigate any fuel temperature effects on the leak test, the same test vehicle was used and
filled to a 40 percent fill level with certification test fuel. The temperature of the vehicle was varied by
subjecting it and the fuel to an overnight soak at the prescribed temperature in a VT SHED. The three
temperatures tested were 25, 30, and 35 °C. Data obtained from the testing is presented in Figure 13. As
can be seen, most values are within the accuracy of the EELD, however those obtained at 35 °C show a
slight trend for higher readings. Since the existing FTP preconditioning was from 20-30 °C, it was
determined to stay within that range. In this range, fuel tank temperature did not affect the EELD
readings.
1A-18
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
On?
Omo -
0)
N o 018 -
In
0)
Un m7 -
^
"s_
O 0 016 -
•a
0)
4_> n m ^ -
TO °-0l;>
_U
"O n m A -
Om ^ -
Om 7 -
fuel Temperature Effects (25-35°C)
0.020" Implanted Leak
I
I
1 I I T
t T T I 1
T T T * t I
Most measurements within accuracy of unit
0 1
425C
• 30C
A35C
2345678
Test Point
Figure 13 - EELD readings for Various Fuel Temperatures
c. Vehicle and Tank Inclination
Another factor investigated was the inclination of the test vehicle. Again, the same vehicle was
used and filled to a 40 percent level with certification test fuel. The test vehicle was then placed on a two
post hoist and the appropriate lift arms positioned to cant the vehicle in various directions. The
inclinations tested were 5° left and right (roll), as well as 5° front and back (pitch). All readings for the
EELD were within the accuracy of the unit for the various vehicle inclinations tested. These results can
be seen in Figure 14.
1A-19
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
Vehicle Inclination Effects (5°cant)
0.020" Implanted Leak
On?
Omo -
0)
•— n niR -
^ u.uio
0)
(j n ni7 -
j^
's_
On DTK -
T3
2J no/I1; -
(0
u
1_ r> m/i -
_C
Om 3 -
Om 7 -
1 * * x T 1
IDl >O< N> _ MX \
• A /N /
T I 1 i
I *
f V
k ;
L
; ;
+ 5 deg Front
• 5 deg Left
A 5 deg Rear
X 5 deg Right
012345678
Test Point
Figure 14 - EELD Readings for Various Vehicle Inclinations
d. Leak Location
The last possible influencing factor tested was the location of the leak relative to the insertion or
test point of the EELD. In these tests, the identical test vehicle was used and filled to a 40 percent fuel
level with certification test fuel. In all cases, the EELD attachment point remained consistent at the under
hood evaporative test port. The location of the leak was varied between the purge valve (near to the test
attachment point) and the top of the fuel cap (far from the attachment point). The results are presented in
Figure 15. Within these results a definitive trend could be observed. The EELD indicated higher leak
values when the orifice was placed near the test attachment point versus those where the leak was placed
far away from the attachment point. On average the near leak values were higher than the far leak values
by 0.002-0.003". This was outside of the margin of error of the EELD was considered a definitive test
influence.
A plausible explanation for this difference was due to the way the EELD and similar devices
compute the leak values. For the EELD and similar test instruments, the effective leak values are
calculated via onboard pressure and flow sensors. Normally they are calibrated to provide an accurate
reading at the instrument or attachment hose. When the instrument is placed on a test vehicle some
pressure drop is always present as long as there is flow through the evaporative system and attaching
lines. It is conceivable that the pressure drop present in the evaporative system is greater when the leak
location is farther away from the test insertion point. In the data obtained in NVFEL testing, leak
locations farther away from the test point were reported as 0.002-0.003" smaller than those that were
implanted closer to the test point. Based on this testing it may be advisable to consider multiple test
points as part of the requirement to show compliance with the requirements
1A-20
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
Leak Location Effects
0.020" Implanted Leak
.020 Purge
.020 Cap
345
Test Point
Figure 15 - EELD readings for Various Leak Locations
VII. Conclusions
This report is the summary of several months of work of NVFEL testing and test procedure
development. Pressure leak test equipment was both constructed in-house as well as purchased from
outside vendors for testing. Both types of equipment produced good results and correlation when tested
against known orifice sizes. In addition the in-house construction of evaporative test equipment gave
particularly well placed insight on the governing physical properties and equations necessary to
accurately predict leak sizes from flow and pressure measurements of an evaporative emissions system.
Several influencing factors affecting the outcome of the leak test were investigated. Within tested
limits, the test temperature, vehicle inclination, and fill level all displayed minor influences on the overall
test readings obtained. The only influence that was seen among the factors investigated was the leak
location, which cannot be controlled in a typical testing fashion. In this case, existing instruments can
under predict the actual leak by a few thousandths of an inch.
The test procedure was validated against known vehicle types which represented a difference
from the "norm" in evaporative emission system control or construction. In all instances, the test
procedure was effective either as is or with minor well understood modifications.
Based on the work so far, the test procedure is seen as robust and applicable to many different
types of vehicles and evaporative emission systems. However, alternative test procedures are allowed as
per 40 CFR 1065.10(c) for those systems which cannot be tested under the given procedure or if new fuel
evaporative control system or OBD leak detection methods are developed which are not compatible with
this methodology.
1A-21
-------
: K O. 12866 Review - Revised Version - Do Not Cite, Quote, or Release During Review
VIII. References
"SAE Paper 860086, Composition of Vapor Emitted from a Vehicle Gasoline Tank During
Refueling, Robert L. Furey and Bernard F. Nagel, 1986.
VOC composition of current motor vehicle fuels and vapors, and collinearity analyses for
receptor modeling", J. Chin, Chemosphere (Oxford). 2012;86:951-958.
"Automotive Evaporative Emissions Systems", M. Schager, University of Southern Colorado.
2011,pp.l-5.
"Chrysler Evaporative Leak Detection Systems", Dixon Matthew, 2008. Presentations. Paper 7.
http://opensiuc.lib.siu.edu/auto_pres/7.
1A-22
-------
Chapter 2 Vehicle Program Costs
2.1 Vehicle Technology Costs
Although the proposed increase in standard stringency is substantial for the vehicles
affected by this proposed rule, we do not expect that the associated vehicle costs will be high.
Our analysis shows that the federal fleet is already demonstrating actual emissions performance
that is much cleaner than the level to which it is currently being certified. Although the vehicles
that make up the federal light-duty fleet are capable of meeting lower standards there is no
impetus for vehicle manufacturers to certify their respective fleets to anything lower than the
current requirements. In addition, we anticipate that not every technology will be required on all
vehicles to meet the proposed standards. While catalyst loading and engine calibration changes
will most likely be applied on all vehicles, only the most difficult powertrain applications will
require very expensive emissions control solutions such as active hydrocarbon adsorbers. We
expect that manufacturers will implement emission control solutions as a function of increasing
cost and will avoid implementing very expensive designs whenever possible.
To determine the cost for vehicles, we first determined which technologies were most
likely to be applied by vehicle manufacturers to meet the proposed standards. These
technologies were then combined into technology packages which reflected vehicle design
attributes that directly contribute to a vehicle's emissions performance. The attributes
considered included vehicle type: car or truck, number of cylinders, engine displacement and the
type of fuel used, either gas or diesel. We also created separate packages for light-duty and
heavy-duty trucks and vans.
2.1.1 Direct Manufacturing Costs
In making our estimates for both direct manufacturing cost (DMC) and application of
technology, we have relied on our own technology assessments. These assessments include
publicly available information, such as that developed by the California Air Resources Board, as
well as confidential information supplied by individual manufacturers and suppliers.1 We have
also considered the results of our own in-house testing.2 The technology packages that we
developed represent what we consider to be the most likely average emissions control solution
for each vehicle type.
In general, we expect that the majority of vehicles will be able to comply with the Tier 3
standards which we are proposing through refinements of current emissions control components
and systems. Some vehicles, for example large trucks with large displacement engines, in
particular LDT3s and LDT4s, may require additional emission controls. Overall, smaller lighter-
weight vehicles will require less extensive improvements than larger vehicles and trucks.
Specifically, we anticipate a combination of technology upgrades for reducing exhaust emissions
including:
Catalyst Platinum Group Metal (PGM) Loading: Increased application of precious
metals in the catalyst is expected to be one of the primary means of mitigating NMOG and NOx
to meet the proposed Tier 3 standards. Increasing the catalyst PGM loading results in greater
2-1
-------
catalyst efficiency. Vehicle manufacturers and suppliers have supplied CBI that estimates the
cost of increasing the PGM loading and modifications to increase the surface area within the
catalyst. These costs ranged from $80 to $260 and were estimated as being incremental to an
existing Tier 2 Bin 5 compliant vehicle. As our feasibility study in Chapter 1 points out many
vehicles are performing much cleaner than the Tier 2 Bin 5 fleet average they are required to
meet. As a result we concluded that the incremental costs for PGM loading would be less than
the estimates we received. We estimate the costs to be $60, $80, and $100 for an 14, V6 and V8
respectively. All DMC are in 2009 dollars.
Optimized Close-coupled catalyst: Close-coupled catalyst technologies include
improvements to the catalyst system design, structure, and packaging to reduce light-off time.
As catalysts are moved closer to the engine the temperature of the exhaust gases to which
catalysts are exposed under high load operation goes up substantially. As a result some of the
materials used in the catalyst construction, as well as the precious metals used in close-coupled
applications, must be improved to survive in the higher operating temperatures. Cost estimates
for close-coupled catalyst designs received from vehicle manufacturers ranged from $25 to $50,
however, they did not include all of the considerations identified above. As a result we have
estimated the cost for an 14 gasoline engine to be $20, a V6 at $40, and a V8 at $60. All DMC
are in 2009 dollars.
Optimized Thermal Management: Overall thermal management of the emissions control
system to shorten the time it takes for the catalyst to light-off will most likely be a primary
technology for mitigating NMOG on gasoline vehicles and NOx on diesel vehicles. This
technology includes dual wall exhaust manifolds and pipe that will help maintain exhaust gas
temperatures from the exhaust port of the engine to the close-coupled catalyst or, in the case of
diesel engines, the Selective Catalyst Reduction (SCR) system. In some cases the packaging of
the exhaust system will be modified to reduce the wetted area of the exhaust path. This will in
turn reduce the decrease in exhaust gas temperatures associated with a longer exhaust path.
Based on CBI submitted by exhaust system suppliers and vehicle manufacturers we estimate that
the cost of implementing dual wall exhaust designs are approximately $30 for all engine
applications. All DMC are in 2009 dollars.
Secondary Air Injection: Secondary Air Injection is a technology that provides a source
of combustion air such that a portion of the exhaust gases are burned in the exhaust manifold.
This technology provides increased heat in the exhaust system that provides for faster catalyst
light-off It is used only during cold start and requires that the air/fuel mixture is rich such that a
small amount of fuel is available for combustion outside of the combustion chamber. We expect
that some gasoline V6's and V8's will require the application of secondary air injection to reduce
NMOG emissions. The secondary air injection system consists of an air pump (normally
electrically powered), plumbing from the pump to the exhaust manifold, an electrically
controlled valve, control circuitry in the powertrain control module, wiring and calibration. CBI
estimates received from vehicle manufacturers and suppliers ranged from $50 to $310. We have
estimated that the final direct manufacturing cost for secondary air is $100 for all applications.
All DMC are in 2009 dollars.
Engine Calibration: Product changes considered for Engine Calibration include engine
control and calibration modifications to improve air and fuel mixtures, particularly at cold start
2-2
-------
and/or to control secondary air and hydrocarbon adsorbers. While typically there are no direct
manufacturing costs associated with the calibration itself, we recognize that some additional
engineering efforts will be required to implement the changes described above. We have
estimated that 2 additional engineers would be required per vehicle at an annual salary and
benefits of $300,000. Assuming they are supporting an annual powertrain volume of 150,000
units, the per vehicle cost is $2. All DMC are in 2010 dollars.
Hydrocarbon Adsorber: Hydrocarbon Adsorbers trap hydrocarbons during cold start and
release the hydrocarbons after the catalyst lights off. Hydrocarbon adsorbers can be applied in
two different manners: The first is a passive device which traps hydrocarbons at cold start and
releases them as the temperature of the device increases. The catalyst may or may not have lit
off at the time of desorption. The second is an active hydrocarbon adsorber. This device
controls the adsorber exposure to exhaust gases based on temperature and is able to trap the
hydrocarbons until the catalyst has lit off. The effectiveness of the active hydrocarbon system is
much greater than the passive system. We anticipate manufacturers will apply only active
systems due limited ability of passive systems to mitigate NMOG. Estimates for active
hydrocarbon adsorber systems ranged from $150 to $450. For our cost estimates we concluded
that the lower estimate is appropriate. In addition, we expect hydrocarbon adsorbers to be
applied in only large displacement V8 powered vehicles. All DMC are in 2009 dollars.
This rulemaking also addresses control of evaporative emissions. We expect also for
evaporative emissions that most vehicles will be able to comply with the proposed Tier 3
standards through refinements of current emission control components and systems. Many of
these technologies have already seen commercial use, while others are variations on established
technologies. For example, manufacturers have designed vehicles to comply with CARS's
PZEV evaporative emission standards, which are very similar to the new standards proposed in
this rulemaking. We anticipate that manufacturers will use some combination of the following
technology upgrades:
Evaporative canisters: Current evaporative canisters use high working-capacity
activated carbon, usually with multiple compartments, to optimize vapor loading and purging
behavior. These canisters sometimes employ carbons of different working capacities within each
chamber. Manufacturers may adjust the shapes and sizes of internal compartments, including
design variations to include different grades of carbon in different areas to best manage rapid
purge following engine starting, back purge during overnight parking, vapor loading at different
loading rates, and vapor redistribution and migration while the vehicle is not operating. The
biggest expected change to evaporative emission canisters is the addition of a secondary canister
element, either attached to the canister body, or integral to it, in which a carbon with very low
working capacity is available to capture diffusion emissions (also known as bleed emissions).
This is commonly referred to as a canister scrubber. While this carbon element can hold only a
few grams of hydrocarbon, it back purges easily and purges readily with a short amount of
driving, so it is always ready to capture the small amount of hydrocarbon that escapes the body
of the evaporative canister as a result of diffusion from vapor migration within the carbon bed.
For purposes of this analysis, we expect that all vehicles covered by the proposal will need a
canister scrubber. The scrubbers will vary in size, but a typical unit would cost about $10. We
expect that in most cases these will be built as an integral part of the current canister to avoid
2-3
-------
extra packaging costs. In some cases, dual tank HDGVs may employ two evaporative emission
canisters. All DMC are in 2009 dollars.
Air intake scrubbers: Manufacturers have identified the engine's intake system as
another source of evaporative emissions. These result from crankcase vapors and from unburned
fuel from injectors, or sometimes from an injection event that occurred shortly before engine
shutdown. One way to prevent these emissions is to add a device containing activated carbon to
the air intake downstream of the air filter, typically in the form of reticulated foam coated with
activated carbon. This device would have only a few grams of working capacity and would be
designed to purge easily to ensure that the vapor storage is available any time the engine shuts
down. This carbon insert would almost completely eliminate vapor emissions from the air intake
system. Manufacturers wanting to avoid adding specialized emission control components to
control evaporative emissions from the air intake could pursue other approaches. First, it is
possible to allow the engine to continue rotating for 2-3 revolutions after engine shutdown to
sweep any hydrocarbon vapors from the intake system into or through the combustion chamber.
These vapors could be burned in the cylinder, oxidized at the catalyst, or stored until the engine
starts again. This may still allow for some residual vapor release, but this should be a very small
quantity. Vehicle owners would be unlikely to notice this amount of engine operation after
shutdown. Second, to the extent that manufacturers use direct injection, there should be no fuel
vapor coming from the intake system. Any unburned fuel coming from the injectors would be
preserved in the cylinder or released to the exhaust system and the catalyst. A small amount of
crankcase vapor might remain, but this would likely not be enough to justify adding carbon to
the intake system. These alternative approaches would generally not involve any incremental
costs. We estimate that 25 percent of vehicles will use the air intake scrubber to address this
source of emissions. The intake scrubber would cost approximately $7.50 per vehicle. All DMC
are in 2009 dollars.
Fuel tank permeation: Fuel tanks are already designed to limit permeation emissions.
Fuel tanks are typically made of high-density polyethylene with an embedded barrier layer of
ethyl vinyl alcohol (EvOH) representing about 1.8 percent of the average wall thickness. The
EvOH layer is effective for reducing permeation emissions. Recent developments in production
processes have led to improved barrier coverage around the ends of the tank where the molded
plastic is pinch-welded to form a closed vessel, which is an important step in eliminating a
permeation path through the wall of the fuel tank. We are expecting manufacturers to increase
the EvOH barrier thickness to about 3 percent of the average wall thickness to provide a more
uniform barrier layer, to provide better protection with ethanol-based fuels, and to improve
permeation resistance generally. The incremental material cost for this thicker layer of EvOH
comes to about $3.50, which we would anticipate for about 30 percent of the fleet. Heavier
HDGVs (HHDGVs) usually have one large volume (>25 gallon) fuel tank per vehicle mounted
on the frame rail or underbody although in the past there have been dual tank offerings. They are
usually metal, but the use of limited offerings of plastics now or in the future cannot be
eliminated. If they are metal there is no permeation cost, if they are plastic they will incur
incremental permeation emission control costs as discussed above for each tank. We assume the
market is 90 percent metal and 90 percent single tank. All DMC are in 2009 dollars.
Fuel line permeation: Fuel lines in use today also are designed for low permeation rates.
The biggest portion of fuel and vapor lines are made of metal, but that may still leave several feet
2-4
-------
of nonmetal fuel line on a vehicle. There may be development of new materials to further reduce
permeation rates, but it is more likely that manufacturers will adjust the mix of existing types of
plastic fuel lines, and perhaps use more metal fuel lines, to achieve the desired performance at
the lowest possible price. This would likely vary significantly among vehicle models. As an
industry average figure, we estimate that 40 percent of vehicles will make upgrades, each
involving $1.60 of additional cost for materials with greater permeation resistance. All DMC are
in 2009 dollars.
Fuel-system architecture: The bigger area of expected development with respect to fuel
lines is to re-engineer whole systems to reduce the number of connections between fuel-system
components and other fuel-line segments. While manufacturers have already made some
changes in this direction, these systems may still involve more than the optimum number of
connections and segments due to assembly and production considerations or other factors.
Designing the fuel system more carefully to minimize connection points will limit possible paths
for fuel vapors to escape. This would reduce emission rates and it should also improve system
durability by eliminating potential failure points. A broader approach to addressing this source
of emissions is to integrate designs and to move fuel-system components inside the fuel tank,
which eliminates the concern for vapor emissions and permeation from those components and
connections. Most of the costs associated with these upgrades lie in development and tooling.
There may be some additional part costs, but the overall trend should ultimately allow for
reduced costs from reducing the number of components and reducing assembly time. To the
extent that fuel-system components are moved inside the fuel tank, there may be further cost
savings since those components would no longer need to be made from low-permeation
materials. Overall, we estimate that this initiative will involve no net change in costs.
Filler neck: A remaining area of potential evaporative emissions is the connection
between the fill neck and the fuel tank. The challenge is to design a low-cost solution that is
easily assembled and works for the demanding performance needs related to stiffness and
flexibility. The best approach is likely either to use mating parts made from low-permeation
materials, or to use conventional materials but cover this joint with material that acts as a barrier
layer. Final designs to address this might vary widely. However, we estimate that a technology
upgrade costing $5 will be applied to 30 percent of vehicles. An alternative scenario would
involve a simpler $2 upgrade to be applied to 75 percent of vehicles. Heavy HDGVs with metal
tanks mounted on the outside of the frame rail do not have filler necks and thus would not incur
this cost. All DMC are in 2009 dollars.
Purge/vapor generation strategies: Recent and projected engine design changes are
increasing the challenge to maintain manifold vacuum for drawing purge air over the evaporative
canister. Several different technology options would help to address this increasing challenge.
Different grades of carbon and canister configurations can lead to a more effective canister purge
for a given volume of air flowing over the canister. Adding a heating element to the canister has
been shown to decrease the desorption energy for drawing hydrocarbon vapors away from
activated carbon. Purge pumps could replace or supplement manifold vacuum as the driving
force for drawing air through the evaporative canister. All of these approaches have merit.
Another effective solution to this new and increasing challenge may be to install a vapor
blocking valve that would allow for pressurized fuel tanks. Before refueling emission controls, it
was common for standard, high-density polyethylene fuel tanks to be pressurized up to about 2
2-5
-------
psi. Vapor blocking valves could be designed to relieve pressure at any appropriate set point.
Pressurized fuel tanks would contain vapors in the fuel tank instead of routing them to the
evaporative canister. The hydrocarbon load and the corresponding purge demand would
therefore be reduced. The design could be targeted to reduce diurnal vapor loading, which
would then reduce purge demands between refueling events. Vapor blocking valves and the
corresponding control systems could have widely varying complexity and cost. We believe a
relatively simple design would be sufficient to achieve the objective safely and effectively. We
estimate that 15 percent of vehicles would employ such canister purge or vapor valve approaches
at a cost of about $6 per vehicle.
Vapor leaks: As part of the Tier 3 evaporative emission requirements EPA is proposing a
vapor leak emission standard. EPA expects that many of the technologies and approaches for
reducing evaporative emissions described above will assist in addressing potential vapor leak
problems and that in most cases no specific additional measures would be needed. Nevertheless,
there might be two additional cost areas. First would be certification testing. However, EPA is
allowing certification requirements for the vapor leak emission standard to be met by written
attestation rather than by testing since the certification vehicle would fail the hot soak plus
diurnal evaporative emissions standard if it had a 0.02 inch leak. Manufacturers agree this is
appropriate. Second, EPA is proposing to include assessment of the vapor leak emission standard
within the in-use verification testing program (IUVP). However, we have structured the program
to minimize additional costs. Testing will be required on all vehicles otherwise procured for
exhaust emissions. All vehicles tested for exhaust emissions must also be tested for the leak
emission standard. Thus, we generally expect multiple leak test results per group but in no case
may there be fewer than one test group representative for each evaporative/refueling/leak family.
Unless there are performance problems, no additional vehicle procurement costs are expected.
Also, we are proposing to permit the manufacturer to use its current evaporative system leak
monitoring OBD hardware to screen vehicles from IUVP testing for leaks and/or to use as an
option to the proposed EPA test procedure if testing is needed. The additional costs for leak
emission testing for IUVP (approximately $0.25 (2009$) per vehicle) are included in the indirect
costs discussed below.
Taken together, these technologies applied to the fleet to the degree described in the
paragraphs above result in an estimated incremental cost of $17 per vehicle in 2009 dollars.
Onboard Refueling Vapor Recovery (ORVR): Current EPA standards require vehicle-
based control of refueling emissions for all LDVs and LDTs up to 10,000 Ibs GVWR. We are
proposing to extend EPA's refueling emission standard to heavy-duty gasoline vehicles
(HDGVs) up to 14,000 Ibs GVWR starting with the 2018 model year. Today these HDGVs are
produced by only two OEMs. Their chassis and fuel system configurations are very similar to
their slightly lighter GVWR LDT counterparts, which are now covered by the refueling emission
standard. Because annual sales of these 10,001-14,000 Ib GVWR HDGVs is small relative to
their similar lighter GVWR LDT counterparts, for uniformity of production and other cost
savings reasons, manufacturers have installed ORVR on these vehicles since about 2006.
However, they have not been certified since there were no emission control requirements to
certify them against. We are including refueling emission control requirements for these vehicles
but expect no additional costs beyond current practice.
2-6
-------
Onboard Diagnostics (OBD): EPA and CARB have similar but not identical OBD
requirements for LDVs, LDTs, MDPVs, and HDGVs up to 14,000 Ibs GVWR. Within the past
five years CARB has revised their implementation scheme and upgraded requirements to
improve the effectiveness of their systems in addressing potential exhaust and evaporative
system performance issues in use. EPA regulations permit manufacturers to meet CARB's most
recent requirements and to seek a Federal certificate based on meeting CARB's requirements.
Certification based on meeting CARB's requirements and application of those OBD systems
nationwide is common practice in the industry with only a few exceptions. EPA is proposing to
adopt current CARB OBD certification, verification, and monitoring requirements. As part of
our proposal, we are also seeking to include two new elements; (1) certification that the OBD
evaporative system leak monitor is able to find a 0.020 inch leak and (2) a requirement that the
OBD computer store information on when the full OBD leak monitoring protocol was last run
successfully and the result of that assessment. Since current CARB OBD requirements are being
met by manufacturers, additional costs are attributable to certification to the 0.020 inch leak
detection requirement and software modification to retain information on the last successful run
of the OBD evaporative system leak monitor. EPA estimates these two items to cost on average
approximately $0.50 (2009$) per vehicle or less. These are reflected in evaporative emission and
calibration costs
2.1.2 Indirect Costs
We are using an approach to estimating indirect costs that is consistent with that used in
our 2012-2016 Greenhouse Gas (GHG) final rule and our recent 2017-2025 GHG final rule.3
Rather than a traditional retail price equivalent markup (RPE), as described below we are
marking up DMCs using an indirect cost multiplier (ICM). Furthermore, we are applying the
ICMs in a manner that differs from the traditional RPE approach in which the DMC would be
multiplied by the RPE factor in any given year. As such, as the DMC decreased with learning,
the product of the RPE factor and the DMC decreased along with it. However, we have more
recently decided that learning impacts (discussed below) should be applied only to the DMC and
not to the indirect costs. Our approach with ICMs, consistent with the recent 2017-2025 GHG
final rule, is to determine the indirect costs based on the initial value of direct costs and then hold
that constant until the long-term ICM is applied. This is done for all ICM factors except
warranties, which are influenced by the learned value of direct costs.
2.1.2.1 Cost markups to account for indirect costs
To produce a unit of output, auto manufacturers incur direct and indirect costs. Direct
costs include the cost of materials and labor costs. Indirect costs may be related to production
(such as research and development [R&D]), corporate operations (such as salaries, pensions, and
health care costs for corporate staff), or selling (such as transportation, dealer support, and
marketing). Indirect costs are generally recovered by allocating a share of the costs to each unit
of goods sold. Although it is possible to account for direct costs allocated to each unit of goods
sold, it is more challenging to account for indirect costs allocated to a unit of goods sold. To
make a cost analysis process more feasible, markup factors, which relate total indirect costs to
total direct costs, have been developed. These factors are often referred to as retail price
equivalent (RPE) multipliers.
2-7
-------
Cost analysts and regulatory agencies including EPA have frequently used these
multipliers to estimate the resultant impact on costs associated with manufacturers' responses to
regulatory requirements. The best approach to determining the impact of changes in direct
manufacturing costs on a manufacturer's indirect costs would be to actually estimate the cost
impact on each indirect cost element. However, doing this within the constraints of an agency's
time or budget is not always feasible, and the technical, financial, and accounting information to
carry out such an analysis may simply be unavailable.
RPE multipliers provide, at an aggregate level, the relative shares of revenues (Revenue =
Direct Costs + Indirect Costs + Net Income) to direct manufacturing costs. Using RPE
multipliers implicitly assumes that incremental changes in direct manufacturing costs produce
common incremental changes in all indirect cost contributors as well as net income. A concern
in using the RPE multiplier in cost analysis for new technologies added in response to regulatory
requirements is that the indirect costs of vehicle modifications are not likely to be the same for
different technologies. For example, less complex technologies could require fewer R&D efforts
or less warranty coverage than more complex technologies. In addition, some simple
technological adjustments may, for example, have no effect on the number of corporate
personnel and the indirect costs attributable to those personnel. The use of RPEs, with their
assumption that all technologies have the same proportion of indirect costs, is likely to
overestimate the costs of less complex technologies and underestimate the costs of more
complex technologies.
To address this concern, the agency has developed modified multipliers. These
multipliers are referred to as indirect cost multipliers (ICMs). In contrast to RPE multipliers,
ICMs assign unique incremental changes to each indirect cost contributor
ICM = (direct cost + adjusted indirect cost + profit)/(direct cost)
Developing the ICMs from the RPE multipliers requires developing adjustment factors
based on the complexity of the technology and the time frame under consideration. This
methodology was used in the cost estimation for the MYs 2012-2016 GHG final rule. The ICMs
were developed in a peer-reviewed report from RTI International and were subsequently
discussed in a peer-reviewed journal article.4 Note that the cost of capital (reflected in profit) is
included because of the assumption implicit in ICMs (and RPEs) that capital costs are
proportional to direct costs, and businesses need to be able to earn returns on their investments.
The capital costs are those associated with the incremental costs of the new technologies.
As noted above, for the analysis supporting this proposed rulemaking, EPA is using the
ICM approach but we have made some changes to both the ICM factors and to the method of
applying those factors to arrive at a final cost estimate since publishing the 2012-2016 GHG final
rule. Both of these changes make the ICMs used in this analysis consistent with those used in
the 2017-2025 GHG final rule. The first of these changes was done in response to continued
thinking about how past ICMs have been developed and what are the most appropriate data
sources to rely upon in determining the appropriate ICMs. We have a detailed discussion of this
change in Chapter 3 of the joint TSD supporting the 2017-2025 GHG rule.5 Because that
discussion is meant to present changes made in the time between the 2012-2016 GHG final rule
2-8
-------
and the 2017-2025 GHG final rule, the full text is not really relevant in the context of this Tier 3
proposal. The second change has been done both due to staff concerns and public feedback
suggesting that the agency was inappropriately applying learning effects to indirect costs via the
multiplicative approach to applying the ICMs. This change is detailed below because it is
critical to understanding how indirect costs are calculated in the context of this Tier 3 proposal.
Table 2-1 shows the ICMs used in this analysis. As noted, these ICMs are consistent
with those used in our recent 2017-2025 GHG final rule. Despite the fact that these ICMs were
developed with GHG technologies in mind, we are using them here to estimate indirect costs
associated with criteria emission control technology. We believe the ICMs are applicable here
because, as with the GHG requirements, the technologies considered in this Tier 3 proposal are
provided to the auto maker by suppliers and their integration into the end vehicle involves the
same sorts of methods and demands as integrating GHG improving technologies.
Table 2-1 Indirect Cost Multipliers Used in this Analysis
Complexity
Low
Medium
Highl
High2
Near term
1.24
1.39
1.56
1.77
Long term
1.19
1.29
1.35
1.50
The second change noted above made to the ICMs has to do with the way in which they
are applied. In the 2012-2016 GHG final rule, we applied the ICMs, as done in any analysis that
relied on RPEs, as a pure multiplicative factor. This way, a direct manufacturing cost of, say,
$100 would be multiplied by an ICM of 1.24 to arrive at a marked up technology cost of $124.
However, as learning effects (discussed below) are applied to the direct manufacturing cost, the
indirect costs are also reduced accordingly. Therefore, in year two the $100 direct
manufacturing cost might reduce to $97 and the marked up cost would become $120 ($97 x
1.24). As a result, indirect costs would be reduced from $24 to $20. Given that indirect costs
cover many things such as facility-related costs, electricity, etc., it is perhaps not appropriate to
apply the ICM to the learned direct costs, at least not for those indirect cost elements unlikely to
change with learning. EPA believes that it is appropriate to allow only warranty costs to
decrease with learning, since warranty costs are tied to direct manufacturing costs (since
warranty typically involves replacement of actual parts which should be less costly with
learning). The remaining elements of the indirect costs should remain constant year-over-year,
at least until some of those indirect costs are no longer attributable to the rulemaking effort that
imposed them (such as R&D).
As a result, the ICM calculation has become more complex. We must first establish the
year in which the direct manufacturing costs are considered "valid." For example, a cost
estimate might be considered valid today, or perhaps not until high volume production is
reached—which will not occur until MY 2015 or later. That year is known as the base year for
the estimated cost. That cost is the cost used to determine the "non-warranty" portion of the
indirect costs. For example, the non-warranty portion of the medium complexity ICM in the
short-term is 0.343 (the warranty versus non-warranty portions of the ICMs are shown in Table
2-2). For example, consider a technology with an estimated direct manufacturing cost of $70 in
2-9
-------
MY 2015. For this technology, the non-warranty portion of the indirect costs would be $24.01
($70 x 0.343). This value would be added to the learned direct manufacturing cost for each year
through 2018, the hypothetical last year of short term indirect costs for this technology.
Beginning in 2019, when long-term indirect costs begin, the additive factor would become
$18.13 ($70 x 0.259). Additionally, the $70 cost in 2015 would become $67.90 in MY 2016 due
to learning ($70 x (1-3%)). So, while the warranty portion of the indirect costs would be $3.15
($70 x 0.045) in 2015, the warranty portion would decrease to $3.06 ($67.90 x 0.045) in 2016 as
warranty costs decrease with learning. The resultant indirect costs of the example technology
would be $27.16 ($24.01+$3.15) in MY 2015 and $27.07 ($24.01+$3.06) in MY2016, and so on
for subsequent years.
Table 2-2 Warranty and Non-Warranty Portions of ICMs
Complexity
Low
Medium
Highl
High2
Near term
Warranty
0.012
0.045
0.065
0.074
Non-warranty
0.230
0.343
0.499
0.696
Long term
Warranty
0.005
0.031
0.032
0.049
Non-warranty
0.187
0.259
0.314
0.448
2.1.3 Cost reduction through manufacturer learning
For this proposal, we have not changed our estimates of learning and how learning will
impact costs going forward from what was employed in the analysis for the MYs 2012-2016
light-duty vehicle rule. However, we have updated our terminology in an effort to clarify that we
consider there to be one learning effect—learning by doing—which results in cost reductions
occurring with every doubling of production.A This updated terminology is entirely consistent
with our approach in the HD GHG rule and the recent 2017-2025 GHG final rule.6 In the past,
we have referred to volume-based and time-based learning. Our terms were meant only to
denote where on the volume learning curve a certain technology was—"volume-based learning"
meant the steep portion of the curve where learning effects are greatest, while "time-based
learning" meant the flatter portion of the curve where learning effects are less pronounced.
Unfortunately, our terminology led some to believe that we were implementing two completely
different types of learning—one based on volume of production and the other based on time in
production. Our new terminology—steep portion of the curve and flat portion of curve—is
simply meant to make more clear that there is one learning curve and some technologies can be
considered to be on the steep portion while others are well into the flatter portion of the curve.
These two portions of the volume learning curve are shown in Figure 2-1.
A Note that this new terminology was described in the recent heavy-duty GHG final rule (see 76 FR 57320). The
learning approach used in this analysis is entirely consistent with that used and described for that analysis.
2-10
-------
Volume Leaning Curve - Steep & Flat Portions
120%
100%
Steep portion of volume learning curve
Flat portion of volume learning curve
20%
0%
Cumulative Production
Figure 2-1 Steep and Flat Portions of the Volume Learning Curve
For some of the technologies considered in this analysis, manufacturer learning effects
would be expected to play a role in the actual end costs. The "learning curve" or "experience
curve" describes the reduction in unit production costs as a function of accumulated production
volume. In theory, the cost behavior it describes applies to cumulative production volume
measured at the level of an individual manufacturer, although it is often assumed—as both
agencies have done in past regulatory analyses—to apply at the industry-wide level, particularly
in industries like the light duty vehicle production industry that utilize many common
technologies and component supply sources. We believe there are indeed many factors that
cause costs to decrease over time. Research in the costs of manufacturing has consistently
shown that, as manufacturers gain experience in production, they are able to apply innovations to
simplify machining and assembly operations, use lower cost materials, and reduce the number or
complexity of component parts. All of these factors allow manufacturers to lower the per-unit
cost of production. We refer to this phenomenon as the manufacturing learning curve.
EPA included a detailed description of the learning effect in the MYs 2012-2016 light-
duty GHG rule and the more recent heavy-duty GHG rule.7 Most studies of the effect of
experience or learning on production costs appear to assume that cost reductions begin only after
some initial volume threshold has been reached, but not all of these studies specify this threshold
volume. The rate at which costs decline beyond the initial threshold is usually expressed as the
percent reduction in average unit cost that results from each successive doubling of cumulative
production volume, sometimes referred to as the learning rate. Many estimates of experience
curves do not specify a cumulative production volume beyond which cost reductions would no
2-11
-------
longer occur, instead depending on the asymptotic behavior of the effect for learning rates below
100 percent to establish a floor on costs.
In past rulemaking analyses, as noted above, EPA has used a learning curve algorithm
that applied a learning factor of 20 percent for each doubling of production volume. EPA has
simplified the approach by using an "every two years" based learning progression rather than a
pure production volume progression (i.e., after two years of production it was assumed that
production volumes would have doubled and, therefore, costs would be reduced by 20 percent).6
In the MYs 2012-2016 light-duty GHG rule and the recent heavy-duty GHG final rule,
the agencies employed an additional learning algorithm to reflect the volume-based learning cost
reductions that occur further along on the learning curve. This additional learning algorithm was
termed "time-based" learning in the 2012-2016 rule simply as a means of distinguishing this
algorithm from the volume-based algorithm mentioned above, although both of the algorithms
reflect the volume-based learning curve supported in the literature. As described above, we are
now referring to this learning algorithm as the "flat portion" of the learning curve. This way, we
maintain the clarity that all learning is, in fact, volume-based learning, and that the level of cost
reductions depend only on where on the learning curve a technology's learning progression is.
We distinguish the flat portion of the curve from the steep portion of the curve to indicate the
level of learning taking place in the years following implementation of the technology (see Table
2-3). We have applied learning effects on the steep portion of the learning curve for those
technologies considered to be newer technologies likely to experience rapid cost reductions
through manufacturer learning, and learning effects on the flat portion learning curve for those
technologies considered to be more mature technologies likely to experience only minor cost
reductions through manufacturer learning. As noted above, the steep portion learning algorithm
results in 20 percent lower costs after two full years of implementation (i.e., the MY 2016 costs
would be 20 percent lower than the MYs 2014 and 2015 costs). Once two steep portion learning
steps have occurred, flat portion learning at 3 percent per year becomes effective for 5 years.
Beyond 5 years of learning at 3 percent per year, 5 years of learning at 2 percent per year, then 5
at 1 percent per year become effective.
For this analysis, learning effects are applied to all technologies because, while most are
already widely used, the technologies would undergo changes relative to their Tier 2 level
design, and we believe auto makers will find ways to reduce costs in the years following
introduction. The steep portion learning algorithm has not been applied to any technologies in
this analysis because we believe that the technologies considered in this analysis have already
experienced the large cost reductions due to learning in the early years of use. The learning
B To clarify, EPA has simplified the steep portion of the volume learning curve by assuming that production
volumes of a given technology will have doubled within two years time. This has been done largely to allow for a
presentation of estimated costs during the years of implementation, without the need to conduct a feedback loop that
ensures that production volumes have indeed doubled. The assumption that volumes have doubled after two years is
based solely on the assumption that year two sales are of equal or greater number than year one sales and, therefore,
have resulted in a doubling of production. This could be done on a daily basis, a monthly basis, or, as we have done,
a yearly basis.
2-12
-------
algorithm applied to each technology and the applicable timeframes are summarized in Table
2-3.
Table 2-3 Learning Effect Algorithms Applied to Technologies Used in this Analysis
Technology
Catalyst Loading
Optimized Close-coupled Catalyst
Optimized Thermal Management
Secondary Air Injection
Engine Calibration
Hydrocarbon Adsorber
Evaporative Emissions Controls/OBD
Selective Catalytic Reduction Optimization
Steep learning
Flat learning
2015-2025
2015-2025
2015-2025
2015-2025
2015-2025
2015-2025
2015-2025
2015-2025
No learning
2.1.4 Costs Updated to 2010 Dollars
For this analysis, we are estimating all costs in terms of 2010 dollars. We have updated
any non-2010 dollar values to 2010 dollars using GDP price deflator as reported by the Bureau
of Economic Analysis on June 27, 2011. The factors used to update costs from 2009 dollars are
shown below in Table 2-4.
Table 2-4 Factors used to Convert 2009 dollars to 2010 dollars
Price Index for Gross Domestic Product
Factor applied to convert to 2010 dollars
2009
109.6
1.010
2010
110.7
1.000
Source: Bureau of Economic Analysis, Table 1.1.4. Price Indexes for Gross Domestic Product, downloaded
1/27/2011, last revised 12/22/2010.
2.1.5 Technology Costs
The total costs (TC) of a given technology are the direct manufacturing costs (DMC) plus
the indirect costs (1C). These costs change over time due to learning effects and different levels
of indirect costs as discussed above. Here we summarize our actual technology cost estimates by
year for each technology. Below, we present our approach to developing package costs—a
package being a group of individual technologies added to a given vehicle—and then our
approach to moving from package costs to program costs.
2-13
-------
Each of the technology costs we have estimated in this analysis are considered to be
applicable to the 2015 MY and, as shown in Table 2-3, we consider each to be on the flat portion
of the learning curve going forward. For all but one technology, we have applied a low
complexity ICM of 1.24 through 2022 then 1.19 thereafter. For the hydrocarbon adsorber
technology, we have applied a medium complexity ICM of 1.39 through 2022 then 1.29
thereafter. Table 2-5 presents the 2015 MY DMC we have estimated for each technology along
with the engine type to which that DMC applies.
The evaporative emissions standards and OBD system upgrades that we are proposing for
LDVs, LDTs, MDPVs, and complete HDGVs under 14,000 Ibs GVWR are feasible with
relatively small cost impacts. We estimate the DMC of system improvements to be about $17
per vehicle, for all car classes. This incremental cost reflects the cost of moving to low
permeability materials, reduced number of connections, longer contiguous lengths of plumbing,
and low-loss connectors for evaporative control and minor costs for OBD upgrades. We have
applied the same learning and ICMs to evaporative emission controls as to exhaust controls
discussed above.
Table 2-5 Technology Direct Manufacturing Costs for the 2015MY (2010$)
Technology
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions Controls/OBD
Engine Calibration
Optimized Thermal Management
SCR Optimization
Gasoline
14
$61
$20
$17
$2
$30
V6
$81
$40
$101
$17
$2
$30
V8
$101
$61
$101
$152
$17
$2
$30
HDV8
$51
$61
$101
$152
$17
$2
$30
Diesel
14
$2
$30
$51
V6
$2
$30
$51
V8
$2
$30
$51
HDV8
$2
$30
$51
Note: Empty cells reflect the fact that the technology is not considered as an enabler for compliance with
the proposed standards.
The following tables present our estimated DMC over time. These changing DMC by
year reflect the effects of learning as described above. The tables also show the indirect costs
(1C) by year. These changing 1C by year reflect the effects of learning on warranty costs and the
effects of the long term ICM (seen in 2023 for each technology in this analysis). The tables also
show the total costs for each technology. Note that these tables do not reflect penetration rates of
technologies or phase-in rates of standards. These impacts are reflected in our package level
costs discussed below. The tables that follow present costs for, in order: passenger car and
light-truck 14 gasoline; passenger car and light-truck V6 gasoline; passenger car and light-truck
V8 gasoline; heavy-duty gasoline; passenger car, light-truck and heavy-duty diesel.
Table 2-6 Technology Costs by Year for 14 Gasoline (2010$)
Technology
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Cost
type
DMC
DMC
DMC
2017
$57
$19
2018
$55
$18
2019
$54
$18
2020
$52
$17
2021
$51
$17
2022
$50
$17
2023
$49
$16
2024
$48
$16
2025
$47
$16
2-14
-------
Technology
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal
Management
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal
Management
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal
Management
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
2017
$16
$2
$29
$15
$5
$4
$0
$7
$72
$24
$20
$2
$36
2018
$16
$2
$28
$15
$5
$4
$0
$7
$70
$23
$20
$2
$35
2019
$15
$2
$27
$15
$5
$4
$0
$7
$69
$23
$19
$2
$34
2020
$15
$2
$26
$15
$5
$4
$0
$7
$67
$22
$19
$2
$33
2021
$14
$2
$25
$15
$5
$4
$0
$7
$66
$22
$18
$2
$32
2022
$14
$2
$25
$15
$5
$4
$0
$7
$65
$22
$18
$2
$32
2023
$14
$2
$24
$12
$4
$3
$0
$6
$61
$20
$17
$2
$30
2024
$14
$2
$24
$12
$4
$3
$0
$6
$60
$20
$17
$2
$30
2025
$13
$2
$24
$12
$4
$3
$0
$6
$59
$20
$16
$2
$30
Note: Empty cells reflect the fact that the technology is not considered as an enabler for compliance with the
proposed standards; $0 values reflect rounding for presentation but are non-zero.
Table 2-7 Technology Costs by Year for V6 Gasoline (2010$)
Technology
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal
Management
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
2017
$76
$38
$95
$16
$2
$29
$19
$10
$24
$4
$0
$7
2018
$74
$37
$92
$16
$2
$28
$19
$10
$24
$4
$0
$7
2019
$72
$36
$89
$15
$2
$27
$19
$10
$24
$4
$0
$7
2020
$69
$35
$87
$15
$2
$26
$19
$10
$24
$4
$0
$7
2021
$68
$34
$85
$14
$2
$25
$19
$10
$24
$4
$0
$7
2022
$67
$33
$83
$14
$2
$25
$19
$10
$24
$4
$0
$7
2023
$65
$33
$82
$14
$2
$24
$15
$8
$19
$3
$0
$6
2024
$64
$32
$80
$14
$2
$24
$15
$8
$19
$3
$0
$6
2025
$63
$31
$78
$13
$2
$24
$15
$8
$19
$3
$0
$6
2-15
-------
Technology
Management
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal
Management
Cost
type
TC
TC
TC
TC
TC
TC
TC
2017
$95
$48
$119
$20
$2
$36
2018
$93
$47
$116
$20
$2
$35
2019
$91
$46
$113
$19
$2
$34
2020
$88
$45
$111
$19
$2
$33
2021
$87
$44
$109
$18
$2
$32
2022
$86
$43
$107
$18
$2
$32
2023
$80
$41
$101
$17
$2
$30
2024
$79
$40
$99
$17
$2
$30
2025
$78
$39
$97
$16
$2
$30
Note: Empty cells reflect the fact that the technology is not considered as an enabler for compliance with the
proposed standards; $0 values reflect rounding for presentation but are non-zero.
Table 2-8 Technology Costs by Year for V8 Gasoline (2010$)
Technology
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
2017
$95
$57
$95
$143
$16
$2
$29
$24
$15
$24
$58
$4
$0
$7
$119
$72
$119
$201
$20
$2
$36
2018
$92
$55
$92
$138
$16
$2
$28
$24
$15
$24
$58
$4
$0
$7
$116
$70
$116
$196
$20
$2
$35
2019
$89
$54
$89
$134
$15
$2
$27
$24
$15
$24
$58
$4
$0
$7
$113
$69
$113
$192
$19
$2
$34
2020
$87
$52
$87
$130
$15
$2
$26
$24
$15
$24
$58
$4
$0
$7
$111
$67
$111
$188
$19
$2
$33
2021
$85
$51
$85
$127
$14
$2
$25
$24
$15
$24
$58
$4
$0
$7
$109
$66
$109
$185
$18
$2
$32
2022
$83
$50
$83
$125
$14
$2
$25
$24
$15
$24
$58
$4
$0
$7
$107
$65
$107
$183
$18
$2
$32
2023
$82
$49
$82
$122
$14
$2
$24
$19
$12
$19
$43
$3
$0
$6
$101
$61
$101
$165
$17
$2
$30
2024
$80
$48
$80
$120
$14
$2
$24
$19
$12
$19
$43
$3
$0
$6
$99
$60
$99
$163
$17
$2
$30
2025
$78
$47
$78
$118
$13
$2
$24
$19
$12
$19
$43
$3
$0
$6
$97
$59
$97
$161
$16
$2
$30
Note: $0 values reflect rounding for presentation but are non-zero. There is at present one V-10 gasoline engine
product offering for HHDGVs
Table 2-9 Technology Costs by Year for HD Gasoline (2010$)
Technology
Cost
type
2017
2018
2019
2020
2021
2022
2023
2024
2025
2-16
-------
Technology
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
2017
$48
$57
$95
$143
$16
$2
$29
$12
$15
$24
$58
$4
$0
$7
$60
$72
$119
$201
$20
$2
$36
2018
$46
$55
$92
$138
$16
$2
$28
$12
$15
$24
$58
$4
$0
$7
$58
$70
$116
$196
$20
$2
$35
2019
$45
$54
$89
$134
$15
$2
$27
$12
$15
$24
$58
$4
$0
$7
$57
$69
$113
$192
$19
$2
$34
2020
$43
$52
$87
$130
$15
$2
$26
$12
$15
$24
$58
$4
$0
$7
$55
$67
$111
$188
$19
$2
$33
2021
$42
$51
$85
$127
$14
$2
$25
$12
$15
$24
$58
$4
$0
$7
$54
$66
$109
$185
$18
$2
$32
2022
$42
$50
$83
$125
$14
$2
$25
$12
$15
$24
$58
$4
$0
$7
$54
$65
$107
$183
$18
$2
$32
2023
$41
$49
$82
$122
$14
$2
$24
$10
$12
$19
$43
$3
$0
$6
$51
$61
$101
$165
$17
$2
$30
2024
$40
$48
$80
$120
$14
$2
$24
$10
$12
$19
$43
$3
$0
$6
$50
$60
$99
$163
$17
$2
$30
2025
$39
$47
$78
$118
$13
$2
$24
$10
$12
$19
$43
$3
$0
$6
$49
$59
$97
$161
$16
$2
$30
Notes: $0 values reflect rounding for presentation but are non-zero.
Table 2-10 Technology Costs by Year for 14, V6, V8 & HD Diesel (2010$)
Technology
Engine Calibration
Optimized Thermal Management
SCR Optimization
Engine Calibration
Optimized Thermal Management
SCR Optimization
Engine Calibration
Optimized Thermal Management
SCR Optimization
Cost type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
2017
$2
$29
$48
$0
$7
$12
$2
$36
$60
2018
$2
$28
$46
$0
$7
$12
$2
$35
$58
2019
$2
$27
$45
$0
$7
$12
$2
$34
$57
2020
$2
$26
$43
$0
$7
$12
$2
$33
$55
2021
$2
$25
$42
$0
$7
$12
$2
$32
$54
2022
$2
$25
$42
$0
$7
$12
$2
$32
$54
2023
$2
$24
$41
$0
$6
$10
$2
$30
$51
2024
$2
$24
$40
$0
$6
$10
$2
$30
$50
2025
$2
$24
$39
$0
$6
$10
$2
$30
$49
Note: $0 values reflect rounding for presentation but are non-zero.
2.2 Vehicle Package Costs
As stated above, we have developed our costs with respect to a given vehicle type and the
type of engine with which it is equipped. Although the cost of achieving the proposed Tier 3
standards will increase with both the size of the vehicle and the displacement of the engine we
have concluded that the cost by engine type is consistent. The final cost per vehicle is the result
of not only the cost per technology but also the application rate of that technology for each
vehicle type. For example, while the cost of Secondary Air Injection is the same, $119, for both
a V6 and V8 application we anticipate that only 25 percent of the V6 applications will require
the technology while 75 percent of the V8 applications will require the technology. This
2-17
-------
technology penetration rate, or application rate, is the first step in developing our vehicle
package costs.
Table 2-11 presents our estimates of application rates of each enabling technology by
engine type to meet the proposed standards.
Table 2-11 Technology Application Rates
Technology
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD
Engine Calibration
Optimized Thermal
Management
SCR Optimization
Gasoline
14
100%
50%
0%
0%
100%
100%
25%
0%
V6
100%
60%
25%
0%
100%
100%
25%
0%
V8
100%
75%
75%
15%
100%
100%
25%
0%
HD
100%
0%
0%
0%
100%
100%
25%
0%
Diesel
All
0%
0%
0%
0%
0%
100%
25%
100%
Note: 0% entries reflect the fact that the technology is not considered to be an enabler for
compliance with the proposed standards.
MDPVs were included in the light-duty fleet as part of Tier 2. Given their current
certification requirements for criteria pollutants, we have included the costs for MDPVs to meet
the Tier 3 standards with the LDT4 cost estimates. We do not expect that the technologies
required to meet the Tier 3 standards for MDPVs will be very different from those applied to
LDT4s as in many cases identical powertrains and chassis exist between the LDT4 and MDPV
platforms.
The next step in developing vehicle package costs is to consider the phase-in rate of the
proposed standards. For example, the proposed standards do not reach maximum stringency
until the 2025 MY, ramping down from a presumed Tier 2 Bin 5 level in the 2016 MY to the
final levels in 2025. Manufacturers would be required to start the phase-in of Tier 3 standards on
both LDVs and LDTs in MY 2017. Based on the declining fleet averages for cars and trucks, we
have apportioned our estimates for full compliance across of the phase-in years as a percentage
of the final standard. Manufacturers would be required to move from a Tier 2 Bin 5 fleet
average in 2016 MY (for vehicles <6,000 Ibs GVW) to the proposed standards. This results in a
significant step in stringency in 2017. As a result, a large portion of the costs are expected to be
incurred in the initial model years. It is also important to note that while we are aligned with
CARB on the individual cost of each technology and their application rates, our costs are
different from California's estimated LEV III costs due to the fact that the California fleet is
currently slightly cleaner than the federal fleet. Finally, manufacturers will have the opportunity
in 2015 and 2016 MY to earn Tier 3 credits by producing a fleet that is cleaner than the current
Tier 2 requirements. While we expect that most manufacturers will earn credits, either by selling
2-18
-------
California vehicles as 50 state vehicles or by certifying existing vehicles to lower Tier 2 bins, we
have not reflected these credits in our cost analysis.
The ramp down in standards can also be expressed as an increasing percentage of the
fleet meeting the proposed standards, moving from 0 percent compliance in the 2016 MY to 100
percent compliance in the 2025 MY (see Section IV of the preamble, which presents the
proposed standards and how they change by MY). This changing percentage of vehicles
complying is treated as being equal in this analysis to the percentage of costs being incurred.
Table 2-12 shows the percentage of vehicles complying with the new standards and, therefore,
the percentage of costs being incurred by manufacturers.
Table 2-12 Percentage of Vehicles Complying with the Proposed Standards
MY
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
Exhaust Standards - Gasoline & Diesel
Passenger
Car
0%
57%
62%
68%
73%
78%
84%
89%
95%
100%
Light
truck
0%
0%
52%
59%
66%
73%
80%
87%
94%
100%
Class 2b
0%
0%
54%
65%
77%
88%
100%
100%
100%
100%
Class 3
0%
0%
47%
60%
73%
87%
100%
100%
100%
100%
Evaporative Standards - Gasoline only
Passenger
car
0%
40%
60%
60%
80%
80%
100%
100%
100%
100%
Light
truck
0%
0%
60%
60%
80%
80%
100%
100%
100%
100%
HDGV
0%
0%
60%
60%
80%
80%
100%
100%
100%
100%
We then use the application rates shown in Table 2-11 along with the compliance
percentages shown in Table 2-12, along with the technology costs shown in Table 2-6 through
Table 2-10 to generate the total technology costs for use in our vehicle package cost estimates.
Table 2-13 Total Technology Costs
Standard Phase-ins -
after Applying Penetration Rates and Proposed
Gasoline Passenger Cars (2010$)
Technology
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Engine
14
14
14
14
14
14
14
V6
V6
V6
V6
V6
V6
V6
2017
$41
$7
$0
$0
$8
$1
$5
$54
$16
$17
$0
$8
$1
$5
2018
$44
$7
$0
$0
$12
$1
$5
$58
$18
$18
$0
$12
$1
$5
2019
$47
$8
$0
$0
$11
$1
$6
$62
$19
$19
$0
$11
$1
$6
2020
$49
$8
$0
$0
$15
$1
$6
$64
$20
$20
$0
$15
$1
$6
2021
$52
$9
$0
$0
$14
$2
$6
$68
$21
$21
$0
$14
$2
$6
2022
$55
$9
$0
$0
$18
$2
$7
$72
$22
$22
$0
$18
$2
$7
2023
$54
$9
$0
$0
$17
$2
$7
$71
$22
$23
$0
$17
$2
$7
2024
$57
$9
$0
$0
$17
$2
$7
$75
$23
$23
$0
$17
$2
$7
2025
$59
$10
$0
$0
$16
$2
$8
$78
$23
$24
$0
$16
$2
$8
2-19
-------
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
V8
V8
V8
V8
V8
V8
V8
$68
$31
$51
$17
$8
$1
$5
$72
$33
$54
$18
$12
$1
$5
$76
$35
$57
$19
$11
$1
$6
$81
$37
$61
$21
$15
$1
$6
$86
$39
$64
$22
$14
$2
$6
$90
$41
$67
$23
$18
$2
$7
$90
$41
$68
$22
$17
$2
$7
$94
$43
$70
$23
$17
$2
$7
$97
$44
$73
$24
$16
$2
$8
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-11) and/or compliance rate (see
Table 2-12).
Table 2-14 Total Technology Costs after Applying Penetration Rates and Proposed
Standard Phase-ins - Gasoline Light-duty Trucks (2010$)
Technology
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Engine
14
14
14
14
14
14
14
V6
V6
V6
V6
V6
V6
V6
V8
V8
V8
V8
V8
V8
V8
2017
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2018
$37
$6
$0
$0
$12
$1
$5
$49
$15
$15
$0
$12
$1
$5
$61
$27
$46
$15
$12
$1
$5
2019
$41
$7
$0
$0
$11
$1
$5
$54
$16
$17
$0
$11
$1
$5
$67
$31
$50
$17
$11
$1
$5
2020
$44
$7
$0
$0
$15
$1
$5
$58
$18
$18
$0
$15
$1
$5
$73
$33
$55
$19
$15
$1
$5
2021
$48
$8
$0
$0
$14
$1
$6
$64
$19
$20
$0
$14
$1
$6
$80
$36
$60
$20
$14
$1
$6
2022
$52
$9
$0
$0
$18
$2
$6
$69
$21
$21
$0
$18
$2
$6
$86
$39
$64
$22
$18
$2
$6
2023
$53
$9
$0
$0
$17
$2
$7
$70
$21
$22
$0
$17
$2
$7
$88
$40
$66
$22
$17
$2
$7
2024
$56
$9
$0
$0
$17
$2
$7
$74
$23
$23
$0
$17
$2
$7
$93
$42
$70
$23
$17
$2
$7
2025
$59
$10
$0
$0
$16
$2
$8
$78
$23
$24
$0
$16
$2
$8
$97
$44
$73
$24
$16
$2
$8
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-11) and/or compliance rate (see
Table 2-12).
Table 2-15 Total Technology Costs after Applying Penetration Rates and Proposed
Standard Phase-ins - Gasoline Heavy-duty Class 2b Trucks (2010$)
Technology
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Engine
All
All
All
All
All
All
2017
$0
$0
$0
$0
$0
$0
2018
$31
$0
$0
$0
$12
$1
2019
$37
$0
$0
$0
$11
$1
2020
$42
$0
$0
$0
$15
$2
2021
$48
$0
$0
$0
$14
$2
2022
$54
$0
$0
$0
$18
$2
2023
$51
$0
$0
$0
$17
$2
2024
$50
$0
$0
$0
$17
$2
2025
$49
$0
$0
$0
$16
$2
2-20
-------
Optimized Thermal Management
All
$0
$5l$6
$6
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-11) and/or compliance rate (see
Table 2-12).
Table 2-16 Total Technology Costs after Applying Penetration Rates and Proposed
Standard Phase-ins - Gasoline Heavy-duty Class 3 Trucks (2010$)
Technology
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls/OBD Upgrades
Engine Calibration
Optimized Thermal Management
Engine
All
All
All
All
All
All
All
2017
$0
$0
$0
$0
$0
$0
$0
2018
$27
$0
$0
$0
$12
$1
$4
2019
$34
$0
$0
$0
$11
$1
$5
2020
$40
$0
$0
$0
$15
$1
$6
2021
$47
$0
$0
$0
$14
$2
$7
2022
$54
$0
$0
$0
$18
$2
$8
2023
$51
$0
$0
$0
$17
$2
$8
2024
$50
$0
$0
$0
$17
$2
$8
2025
$49
$0
$0
$0
$16
$2
$8
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-11) and/or compliance rate (see
Table 2-12).
Table 2-17 Total Technology Costs after Applying Penetration Rates and Proposed
Standard Phase-ins - Diesel Passenger Cars (2010$)
Technology
Engine Calibration
Optimized Thermal Management
SCR Optimization
Engine
All
All
All
2017
$1
$5
$34
2018
$1
$5
$36
2019
$1
$6
$39
2020
$1
$6
$40
2021
$2
$6
$42
2022
$2
$7
$45
2023
$2
$7
$46
2024
$2
$7
$47
2025
$2
$8
$49
Table 2-18 Total Technology Costs after Applying Penetration Rates and Proposed
Standard Phase-ins - Diesel Light-duty Trucks (2010$)
Technology
Engine Calibration
Optimized Thermal Management
SCR Optimization
Engine
All
All
All
2017
$0
$0
$0
2018
$1
$5
$30
2019
$1
$5
$34
2020
$1
$5
$36
2021
$1
$6
$39
2022
$2
$6
$43
2023
$2
$7
$44
2024
$2
$7
$47
2025
$2
$8
$49
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-11) and/or compliance rate (see
Table 2-12).
Table 2-19 Total Technology Costs after Applying Penetration Rates and Proposed
Standard Phase-ins - Diesel Heavy-duty Class 2b Trucks (2010$)
Technology
Engine Calibration
Optimized Thermal Management
SCR Optimization
Engine
All
All
All
2017
$0
$0
$0
2018
$1
$5
$31
2019
$1
$6
$37
2020
$2
$6
$42
2021
$2
$7
$48
2022
$2
$8
$54
2023
$2
$8
$51
2024
$2
$8
$50
2025
$2
$8
$49
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-1 1) and/or compliance rate (see
Table 2-12).
2-21
-------
Table 2-20 Total Technology Costs after Applying Penetration Rates and Proposed
Standard Phase-ins - Diesel Heavy-duty Class 3 Trucks (2010$)
Technology
Engine Calibration
Optimized Thermal Management
SCR Optimization
Engine
All
All
All
2017
$0
$0
$0
2018
$1
$4
$27
2019
$1
$5
$34
2020
$1
$6
$40
2021
$2
$7
$47
2022
$2
$8
$54
2023
$2
$8
$51
2024
$2
$8
$50
2025
$2
$8
$49
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-11) and/or compliance rate (see
Table 2-12).
The final package costs are simply the sum of the costs shown in each of Table 2-13
through Table 2-20. These results are shown in Table 2-21 for gasoline vehicles and Table 2-22
for diesel vehicles. Evaporative system costs associated with heavy heavy-duty gasoline
vehicles (HHDGVs, those with GVWR over 14,000 pounds) are not included in these tables.
EPA estimates that there are only about 50,000 HHDGVs sold each year, and the evaporative
system costs are less than $20 per vehicle (see Table 2-16).
Table 2-21 Vehicle Package Costs by Year for All Gasoline Vehicles (2010$)
Vehicle category
Passenger car
Passenger car
Passenger car
Light-duty truck
Light-duty truck
Light-duty truck
Class 2b
Class 3
Engine
14
V6
V8
14
V6
V8
V8
V8
2017
$62
$102
$181
$0
$0
$0
$0
$0
2018
$69
$112
$196
$60
$96
$167
$49
$44
2019
$73
$118
$207
$65
$105
$182
$56
$52
2020
$80
$127
$222
$74
$116
$202
$65
$63
2021
$83
$133
$233
$78
$124
$218
$71
$70
2022
$90
$143
$247
$87
$137
$237
$82
$82
2023
$89
$141
$246
$87
$138
$240
$78
$78
2024
$92
$147
$256
$92
$146
$254
$77
$77
2025
$95
$151
$264
$95
$151
$264
$75
$75
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-11) and/or compliance rate (see
Table 2-12).
Table 2-22 Vehicle Package Costs by Year for All Diesel Vehicles (2010$)
Vehicle category
Passenger car
Passenger car
Passenger car
Light-duty truck
Light-duty truck
Light-duty truck
Class 2b
Class 3
Engine
14
V6
V8
14
V6
V8
V8
V8
2017
$40
$40
$40
$0
$0
$0
$0
$0
2018
$43
$43
$43
$36
$36
$36
$37
$32
2019
$46
$46
$46
$40
$40
$40
$44
$41
2020
$48
$48
$48
$43
$43
$43
$50
$48
2021
$50
$50
$50
$47
$47
$47
$57
$55
2022
$54
$54
$54
$51
$51
$51
$64
$64
2023
$54
$54
$54
$53
$53
$53
$61
$61
2024
$56
$56
$56
$56
$56
$56
$60
$60
2025
$59
$59
$59
$59
$59
$59
$59
$59
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-11) and/or compliance rate (see
Table 2-12).
2.3 Vehicle Program Costs
With the package costs presented in Table 2-21 and Table 2-22, we can begin to develop
vehicle program costs associated with the proposal. The program costs multiply package costs
by appropriate vehicle sales per year to estimate the annual costs of the proposed program. The
first step to this is determining the sales of each type of vehicle, or package, as presented in
Table 2-21 and Table 2-22. To do this, we have started with the baseline database developed in
2-22
-------
support of the 2012-2016 GHG final rule. That baseline database provides vehicle sales in the
years 2017-2025 for each of the vehicle category/engine/fuel combinations listed in Table 2-21
and Table 2-22. However, that baseline database is not reflective of the 2012-2016 GHG final
rule which is expected to have an impact on the sales mix of the vehicle category/engine/fuel
combinations largely due to an expectation that engines will be turbocharged and downsized to
achieve better GHG performance while maintaining vehicle performance. This downsizing is
expected to provide downward effects on overall Tier 3 costs since vehicles with smaller engines
are expected to incur lower costs than vehicles with larger engines. Therefore, using the baseline
database and the technology penetration rates expected from the 2012-2016 GHG final rule, we
have developed a Tier 3 reference case fleet. This reference fleet is the fleet we have used in
developing Tier 3 vehicle program costs. Table 2-23 shows the baseline fleet mix—representing
the best estimates of the future fleet absent any GHG rules—and Table 2-24 shows the Tier 3
reference fleet—representing the future fleet in the presence of the 2012-2016 GHG final rule.
Table 2-26 shows projected sales of light-duty and heavy-duty vehicles excluding sales in the
State of California which we have estimated to be 10 percent of nationwide sales.
Table 2-23 Baseline Light-Duty Fleet Mix and Sales
Vehicle
category
Passenger
Car
Light-duty
truck
Electric
Vehicle
All Light-
duty
Engine
14
14
V6
V6
V8
V8
All
14
14
V6
V6
V8
V8
All
All
Fuel
Gas
Dies
Gas
Dies
Gas
Dies
All
Gas
Dies
Gas
Dies
Gas
Dies
All
All
2016
38%
0.0%
21%
0.0%
3.2%
0.0%
62%
4.4%
0.0%
22%
0.1%
11%
0.0%
37%
0.2%
100%
2017
38%
0.0%
22%
0.0%
3.1%
0.0%
63%
4.0%
0.0%
23%
0.1%
9.3%
0.0%
37%
0.2%
100%
2018
38%
0.0%
22%
0.0%
3.2%
0.0%
63%
3.9%
0.0%
23%
0.1%
9.0%
0.0%
36%
0.2%
100%
2019
39%
0.0%
22%
0.0%
3.2%
0.0%
64%
3.7%
0.0%
23%
0.1%
8.6%
0.0%
36%
0.2%
100%
2020
39%
0.0%
22%
0.0%
3.2%
0.0%
65%
3.6%
0.0%
23%
0.1%
8.5%
0.0%
35%
0.2%
100%
2021
39%
0.0%
22%
0.0%
3.2%
0.0%
65%
3.6%
0.0%
23%
0.1%
8.2%
0.0%
35%
0.2%
100%
2022
40%
0.0%
22%
0.0%
3.1%
0.0%
65%
3.6%
0.0%
23%
0.1%
8.0%
0.0%
35%
0.2%
100%
2023
40%
0.0%
23%
0.0%
3.1%
0.0%
66%
3.5%
0.0%
23%
0.1%
7.5%
0.0%
34%
0.2%
100%
2024
40%
0.0%
23%
0.0%
3.2%
0.0%
66%
3.5%
0.0%
23%
0.1%
7.2%
0.0%
34%
0.2%
100%
2025
41%
0.0%
23%
0.0%
3.2%
0.0%
67%
3.5%
0.0%
22%
0.1%
7.0%
0.0%
33%
0.2%
100%
Table 2-24 Tier 3 Reference Light-Duty Fleet Mix
Vehicle
category
Passenger
Car
Light-duty
Engine
14
14
V6
V6
V8
V8
All
14
Fuel
Gas
Dies
Gas
Dies
Gas
Dies
All
Gas
2016
41%
0.2%
18%
0.1%
2.6%
0.0%
62%
7.8%
2017
41%
0.2%
19%
0.1%
2.7%
0.0%
63%
7.7%
2018
42%
0.2%
19%
0.1%
2.7%
0.0%
63%
7.6%
2019
42%
0.2%
19%
0.1%
2.7%
0.0%
64%
7.5%
2020
43%
0.2%
19%
0.1%
2.7%
0.0%
65%
7.3%
2021
43%
0.2%
19%
0.1%
2.7%
0.0%
65%
7.3%
2022
43%
0.2%
19%
0.1%
2.8%
0.0%
65%
7.2%
2023
43%
0.2%
19%
0.1%
2.8%
0.0%
66%
7.1%
2024
44%
0.2%
20%
0.1%
2.8%
0.0%
66%
7.0%
2025
44%
0.2%
20%
0.1%
2.8%
0.0%
67%
6.9%
2-23
-------
Vehicle
category
track
Electric
Vehicle
All Light-
duty
Engine
14
V6
V6
V8
V8
All
All
Fuel
Dies
Gas
Dies
Gas
Dies
All
All
2016
0.2%
20%
0.2%
9.0%
0.0%
37%
0.2%
100%
2017
0.2%
20%
0.2%
8.9%
0.0%
37%
0.2%
100%
2018
0.2%
20%
0.2%
8.8%
0.0%
36%
0.2%
100%
2019
0.2%
19%
0.2%
8.6%
0.0%
36%
0.2%
100%
2020
0.2%
19%
0.2%
8.5%
0.0%
35%
0.2%
100%
2021
0.2%
19%
0.2%
8.5%
0.0%
35%
0.2%
100%
2022
0.2%
19%
0.2%
8.4%
0.0%
35%
0.2%
100%
2023
0.2%
18%
0.2%
8.2%
0.0%
34%
0.2%
100%
2024
0.2%
18%
0.2%
8.1%
0.0%
34%
0.2%
100%
2025
0.2%
18%
0.2%
8.0%
0.0%
33%
0.2%
100%
For heavy-duty Class 2b and 3, we expect no downsizing of engines or other changes to
engines that might influence Tier 3 costs as a result of the Heavy-duty GHG rule. Therefore, we
are using the baseline fleet mix as the reference fleet mix for this analysis.9 This fleet mix is
shown in Table 2-25.
Table 2-25 Tier 3 Reference Heavy-Duty Fleet Mix
Vehicle
category
Class 2b
Class 3
A11HD
Engine
V8
V8
All
Fuel
Gas
Dies
Gas
Dies
All
2016
37%
38%
13%
13%
100%
2017
37%
38%
13%
13%
100%
2018
37%
38%
13%
13%
100%
2019
37%
38%
13%
13%
100%
2020
37%
38%
13%
13%
100%
2021
37%
38%
13%
13%
100%
2022
37%
38%
13%
13%
100%
2023
37%
38%
13%
13%
100%
2024
37%
38%
13%
13%
100%
2025
37%
38%
13%
13%
100%
Table 2-26 Projected Non-California Sales by Year"
Year
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
Light-duty
14,575,658
14,225,690
14,018,769
14,020,792
14,306,345
14,570,160
14,795,795
14,989,940
15,240,678
15,525,414
Heavy-duty
641,095
637,249
644,894
661,595
689,149
709,140
726,608
739,953
753,308
767,900
"Based on AEO 2011.
Using the reference fleet mix and the projected sales, we can calculate the annual costs of
the proposed vehicle program for each vehicle category/engine/fuel combination. We can then
add the passenger car and light-duty truck results to get the costs for light-duty and add the Class
2b and 3 costs to get the costs for heavy-duty. We have done this separately for the proposed
exhaust and evaporative standards and then the combined standards. The results are shown in
Table 2-28.
In addition to considering the costs associated with improving the emission control
systems on vehicles, we also expect that manufacturers will be required to improve their
2-24
-------
capability to measure Particulate Matter (PM) at the levels we are proposing. For additional
information on the test procedure changes we are proposing, see Section IV.F of the preamble.
To determine the appropriate costs for upgrading test facilities for PM measurement we
have used two sources of information: The first was the cost that the EPA incurred in upgrading
its own PM measurement equipment, and the second was information provided vehicle
manufacturers reflecting estimates for upgrading their internal facilities. The cost estimates
ranged from $250,000 to $500,000 per PM test site. We recognize that the number of sites that a
manufacturer will require is dependent on the number of vehicle models it expects to develop
and certify in a given model year. As stated in Section IV. A, we have limited the number of
certifications required per model year to 25 percent of the represented durability groups, thereby
potentially reducing the number of test sites that require upgrade. In addition, costs will vary by
manufacturer depending on the state of their current test facilities.
Our estimated costs for each manufacturer are show in Table 2-27. With a certification
responsibility of 25 percent of its given model year durability groups we believe that
manufacturers with annual sales of 1 million units or less will require 2 facility upgrades at an
average cost of $375,000. For manufacturers with greater than 1 million units per year annual
sales we believe that 4 facility upgrades may be required to meet the Tier 3 requirements.
Table 2-27: PM Facility Costs
Annual
Sales
Volume
=!
million
> 1 million
Fleet
# of PM Sites
to be
upgraded
2
4
Cost per site
$375,000
Weigh
Room
Costs
Facility
Cost/Manufacturer
$750,000
$1,500,000
#of
Manufacturers
20
5
25
Total Costs
$15,000,000
$7,500,000
$22,500,000
We also anticipate that each manufacturer would hire a new full time employee to cover
additional PM measurement-related work. We have estimated this employee to cost each
manufacturer $150,000 per year. With 25 manufacturers, the total cost would be $3,750,000 per
year every year going forward. In contrast, the PM facility costs shown in Table 2-27 represent
one-time costs we expect to be incurred in the year prior to implementation of the proposed
standards. These costs are shown in Table 2-28.
Table 2-28 Undiscounted Annual Costs and Costs of the Proposed Program Discounted
back to 2012 at 3 and 7 Percent Discount Rates (Millions of 2010 dollars)
2016
2017
2018
2019
2020
2021
2022
Exhaust
Light-duty
$0
$634
$1,150
$1,240
$1,350
$1,470
$1,580
Heavy-duty
$0
$0
$23.1
$28.6
$34.2
$40.0
$46.5
All
$0
$634
$1,170
$1,270
$1,380
$1,510
$1,630
Evaporative
Light-duty
$0
$71.4
$167
$159
$216
$208
$264
Heavy-duty
$0
$0
$3.86
$3.76
$5.22
$5.09
$6.52
All
$0
$71.4
$171
$162
$221
$213
$271
PM Facility
& staff
$22.5
$3.75
$3.75
$3.75
$3.75
$3.75
$3.75
Total
$22.5
$709
$1,340
$1,440
$1,600
$1,730
$1,900
2-25
-------
2023
2024
2025
2030
2040
2050
PVat3%
PVat7%
Exhaust
Light-duty
$1,610
$1,720
$1,830
$1,750
$1,750
$1,750
$29,900
$14,700
Heavy-duty
$44.8
$44.8
$44.9
$46.5
$51.9
$58.6
$814
$384
All
$1,660
$1,770
$1,870
$1,790
$1,800
$1,810
$30,700
$15,100
Evaporative
Light-duty
$253
$257
$246
$246
$246
$246
$4,250
$2,090
Heavy-duty
$6.27
$6.38
$6.12
$6.74
$7.52
$8.49
$116
$55.0
All
$259
$263
$253
$253
$254
$255
$4,360
$2,150
PM Facility
& staff
$3.75
$3.75
$3.75
$3.75
$3.75
$3.75
$89.1
$52.2
Total
$1,920
$2,040
$2,130
$2,050
$2,060
$2,070
$35,100
$17,300
Note: Costs shown include costs for the proposed Tier 3 standards on vehicles sold in all states except California.
By then sales weighting the exhaust and evaporative/OBD results by sales in each of the
vehicle category/engine/fuel combinations, we can calculate the annual costs for passenger cars,
light-duty trucks and heavy-duty trucks. We show these cost per vehicle results for the proposed
exhaust standards in Table 2-29, for the proposed evaporative standards in Table 2-30 and for the
combined exhaust and evaporative standards in Table 2-31. The costs shown in these three
tables include all direct and indirect costs for new vehicle hardware. They also include the
effects of learning, and the expected penetration rates and phase-ins of the proposed standards.
Table 2-29 Cost per Vehicle for the Proposed Exhaust Emission Standards (2010$)
Passenger car
Light-duty truck
All light-duty
Class 2b
Class 3
All heavy-duty
All LD and HD
2016
$0
$0
$0
$0
$0
$0
$0
2017
$71
$0
$45
$0
$0
$0
$43
2018
$75
$93
$82
$37
$32
$36
$80
2019
$80
$103
$88
$44
$41
$43
$86
2020
$84
$112
$94
$50
$48
$50
$92
2021
$89
$122
$101
$57
$55
$56
$99
2022
$94
$132
$107
$64
$64
$64
$105
2023
$94
$134
$108
$61
$61
$61
$105
2024
$98
$143
$113
$60
$60
$60
$110
2025
$102
$150
$118
$59
$59
$59
$115
Note: Costs shown include costs for the proposed Tier 3 standards on vehicles sold in all states except
California.
Table 2-30 Cost per Vehicle for the Proposed Evaporative Emission Standards/OBD
Upgrades (2010$)
Passenger car
Light-duty truck
All light-duty
Class 2b
Class 3
All heavy-duty
All LD and HD
2016
$0
$0
$0
$0
$0
$0
$0
2017
$8
$0
$5
$0
$0
$0
$5
2018
$12
$12
$12
$6
$6
$6
$12
2019
$11
$11
$11
$6
$6
$6
$11
2020
$15
$15
$15
$8
$8
$8
$15
2021
$14
$14
$14
$7
$7
$7
$14
2022
$18
$18
$18
$9
$9
$9
$17
2023
$17
$17
$17
$8
cc
q>o
$8
$16
2024
$17
$17
$17
$8
cc
q>o
$8
$16
2025
$16
$16
$16
$8
$8
$8
$15
Note: Costs shown include costs for the proposed Tier 3 standards on vehicles sold in all states except
California.
Table 2-31 Cost per Vehicle for the Proposed Exhaust, Evaporative Emission Standards
and OBD Upgrades (2010$)
II 2016 I 2017 I 2018 I 2019 I 2020 I 2021 I 2022 I 2023 I 2024 I 2025 I
2-26
-------
Passenger car
Light-duty truck
All light-duty
Class 2b
Class 3
All heavy-duty
All LD and HD
$0
$0
$0
$0
$0
$0
$0
$78
$0
$50
$0
$0
$0
$47
$87
$105
$94
$43
$38
$42
$91
$92
$114
$100
$50
$46
$49
$97
$99
$127
$109
$58
$55
$57
$107
$103
$136
$115
$64
$63
$64
$113
$112
$150
$125
$73
$73
$73
$122
$111
$151
$124
$69
$69
$69
$122
$115
$159
$130
$68
$68
$68
$127
$118
$165
$134
$66
$66
$66
$130
Note: Costs shown include costs for the proposed Tier 3 standards on vehicles sold in all states except
California.
2-27
-------
References
1 California Air Resources Board Initial Statement of Reasons, Public Hearing to Consider LEV III, December 7,
2011, Workshop Document (http://www.arb.ca.gov/regact/2012/leviiighg2012/leviiighg2012.htm).
2 "The Effects of Fuel Sulfur Level onEmissions from Tier 2 Vehicles in the In-Use Fleet," EPA-420-D-13-003.
Available in docket number EPA-HQ-OAR-2011-0135.
3 The 2012-2016 GHG final rule can be found at 75 FR 25374; the 2017-2025 GHG final rule can be found at
http://epa.gov/otaq/climate/regs-light-duty.htm.
4 RTI International. Automobile Industry Retail Price Equivalent and Indirect Cost Multipliers. February 2009.
http://www.epa.gov/otaq/ld-hwy/420r09003.pdf; Rogozhin, A.,et al., "Using indirect cost multipliers to estimate the
total cost of adding new technology in the automobile industry," International Journal of Production Economics
(2009), doi: 10.1016/j.ijpe.2009.11.031. The peer review for the RTI report is at http://www.epa.gov/otaq/ld-
hwy/420r09004.pdf.
5 "Joint Technical Support Document: Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards," EPA-420-R-12-901, August 2012.
6 76 FR 57106 and http://epa.gov/otaq/climate/regs-light-duty.htm.
7 76 FR 57106.
8 See Chapter 1 of "Final Rulemaking to Establish Light-Duty Vehicle Greenhouse Gas Emission Standards and
Corporate Average Fuel Economy Standards: Regulatory Impact Analysis," EPA-420-R-10-009, April 2010.
9 See Chapter 5 of "Final Rulemaking to Establish Greenhouse Gas Emissions Standards and Fuel Efficiency
Standards for Medium- and Heavy-Duty Engines and Vehicles: Regulatory Impact Analysis," EPA-420-R-11-901,
August 2011.
2-28
-------
Chapter 3 Establishing New Emission Test Fuel Parameters
3.1 Assessment of Current Gasoline Properties
In-use gasoline has changed considerably since EPA's emission test fuel specifications
were first set and last revised. Gasoline sulfur and benzene have been reduced and, perhaps most
importantly, gasoline containing 10 percent ethanol by volume (E10) has replaced clear gasoline
(EO) across the country. According to the Energy Information Administration (EIA), ethanol is
now blended into almost every gallon of U.S. gasoline, bringing the average gasoline ethanol
content to 9.3 percent by volume (vol%) as shown in Figure 3-1.
14
12 -
03
n
v>
~5
c
<0
ffi
R -
CO
=>
2 -
10%
• U.S. Fuel Ethanol Usage
-Avg. % Ethanol in U.S. Gasoline
9.3%
3.9°y
2.1%
2.9%/
m
n 3%
1.0%
Year
Figure 3-1 Average Gasoline Ethanol Content Over Time1
The increase in fuel ethanol has resulted in fuel property changes according to the
Alliance of Automobile Manufacturers (AAM) North American Fuel Survey. Each summer and
winter, the AAM takes over 300 gasoline samples from 29 major metropolitan areas in 23 states
plus the District of Columbia. Areas currently sampled include: Albuquerque, NM; Atlanta, GA;
Billings, MT; Boston, MA; Chicago, IL; Cleveland, OH; Dallas, Houston, and San Antonio, TX;
Denver, CO; Detroit, MI; Fairbanks, AK; Kansas City and St. Louis, MO; Las Vegas, NV; Los
Angeles and San Francisco, CA; Memphis, TN; Miami, FL; Minneapolis/St. Paul, MN; New
Orleans, LA; New York, NY; Philadelphia and Pittsburgh, PA; Phoenix, AZ; Salt Lake City,
UT; Seattle, WA; Washington, D.C.; and Watertown, SD. Although the AAM North American
Fuel Survey does not represent all U.S. gasoline, we believe it represents a sizeable percentage of
3-1
-------
the domestic pool and encompasses the major gasoline formulations (conventional and
reformulated) and grades (regular and premium) available to consumers. As such, we relied
heavily on the AAM North American Fuel Survey trends (as well as refinery compliance data
and information provided by CARB) to help inform our proposed vehicle test fuel changes.A
3.1.1 Octane
Finished U.S. gasoline has not experienced an increase in average octane (also known as
antiknock index or (R+M)/2) due to increased ethanol blending. Denatured fuel ethanol has an
average octane rating of 115 (R+M)/2 which is greater than that of conventional regular-grade
gasoline, 87 (R+M)/2. With the shift to E10 nationwide, many refiners have backed off on
octane production at the refinery in the form of aromatics and olefins. Refiners are currently
producing sub-octane blendstocks to avoid giving away expensive octane. We estimate that
refiners are currently producing 84 (R+M)/2 blendstocks for 87 octane regular-grade gasoline
and 88 (R+M)/2 blendstocks for 91 octane premium-grade gasoline to maintain (but not exceed)
minimum octane ratings for E10.
According to AAM summer fuel surveys, the average octane of finished gasoline has
remained constant around 89-90 (R+M)/2 over the past decade (refer to Figure 3-2). The
reported octane is higher than expected because AAM takes roughly an equal number of regular
and premium gasoline samples as part of their North American Fuel Survey. In reality, the
majority of consumers fill up on 87 octane regular-grade gasoline based on price and other
factors. According to EIA's Petroleum Marketing Annual, regular-grade gasoline represents
over 85 percent of U.S. sales.2 Accordingly, we believe our proposed 87-88.4 R+M/2
specification for test fuel is representative of in-use gasoline. However, the AAM survey does
tell us one important thing about octane - refiners are doing their best not to give it away. We
anticipate that this trend will continue into the future as E15 replaces E10.
A Since Alaska gasoline varies in fuel properties due to crude oil supply, the manner in which it's refined, and
applicable EPA regulations, we have opted to exclude AAM fuel samples taken from Fairbanks, AK in the average
gasoline properties presented in this chapter of the draft RIA.
3-2
-------
100
5
-Avi. Octane (AAM) —«—Avj.EtOH (EIA)I
2000 2001 2002
2003 2004 2005
2006 2007 2008 2009 2010
Figure 3-2 Average Gasoline Octane and Ethanol Levels Over Time
3.1.2 Aromatics and Olefms
The increase in fuel ethanol content has resulted in reduced aromatics and olefin levels.
According to AAM summer fuel surveys, average aromatics levels have been reduced by 15
percent and olefins levels have been reduced by 19 percent in relative terms over the past decade.
In 2010, the average aromatics content of gasoline was measured by AAM at 22.6 vol% and the
average olefins content was measured at 7.3 vol%. The average aromatics, olefin, and ethanol
levels by year are shown in Figure 3-3.
3-3
-------
15
c
o
o
11
c
10
S.9
-Avg. Corrected Aromatics (AAM)
- Avg. Correcte d Olefins (AAM)
Avg. EtOH (EIA)
2000
2001
2002
2003
2004
2005
2006
2007
200S
2009
2010
Figure 3-3 Average Gasoline Aromatics, Olefin and Ethanol Levels Over Time
Although aromatics and olefm levels have been reduced over the past decade, there
continues to be variation on a batch-by-batch and geographic basis. Our 2009 refinery batch data
suggests that gasoline aromatics levels ranged anywhere from 0 to 67 vol% with an average
concentration of 23 vol% and olefm levels ranged from 0 to 43 vol% with an average
concentration of 11 vol%. The 2009 certification data shown in Figure 3-4 and Figure 3-5 only
reflects ethanol that is match blended into reformulated gasoline. It does not account for ethanol
that may be blended into conventional gasoline since our existing gasoline regulations do not
allow refiners to take advantage of fuel ethanol properties in most compliance calculations. As a
result, AAM and other gasoline surveys may show lower aromatics and olefms than what is
suggested by EPA certification data.
3-4
-------
5000
10000
15000
Number of Batches
20000
25000
30000
Figure 3-4 Gasoline Aromatics based on 2009 Refinery Batch Data
3-5
-------
5000
10000 15000 20000
Number of Batches
25000
Figure 3-5 Gasoline Olefins based on 2009 Refinery Batch Data
3-6
-------
In the summer of 2010, according to the AAM North American Fuel Survey, measured
in-use aromatics levels ranged from 3 to 47 vol% (refer to Figure 3-6) while olefm levels ranged
from 0.6 to 17 vol% (refer to Figure 3-7). California tends to have lower, tighter in-use gasoline
properties based on their existing CaRFG3 regulations. As shown below, gasoline samples taken
from Los Angeles and San Francisco, CA had aromatics levels ranging from 10 to 30 vol% and
olefm levels ranging from 1 to 8 vol%. Accordingly, the Federal test fuel parameters we're
proposing for aromatics and olefms are slightly broader than California's, yet still inclusive of
CARB's planned LEV III specs.
so
45
IS
O
O
U
15
10
Cities Sampled, Summer 2010
Figure 3-6 Measured Aromatics by AAM City Surveyed, Summer 2010
3-7
-------
18
16
c 10
1
O
"S R
O
u
Cities Sampled, Summer 2010
Figure 3-7 Measured Olefins by AAM City Surveyed, Summer 2010
Finished gasoline samples obtained by EPA suggest that gasoline currently contains a
wide range of aromatics from basic benzene (C6), and toluene (C7), to larger more complicated
aromatics of C10+ hydrocarbons. From August 2007 to September 2011, EPA analyzed 52 fuel
samples from various locations throughout the country. Approximately 60 percent of the
samples were RFG oversight samples (as part of the RFG rule, every refiner is required to send
EPA a sample of every 33rd batch of gasoline) and the remainder were audit samples (collected,
mostly from retail, as part of the City Surveys program performed as part of the agreement with
API during the regulatory negotiation process that lead to the RFG rule). Total aromatics ranged
anywhere from 6 to 39 percent by volume, but the C6, C7, C8, C9 and C10+ contributions were
relatively consistent (refer to Figure 3-8). As shown below in Table 3-1, gasoline contained 3.2
percent benzene, 22.5 percent C7 toluene, 28.3 percent C8 aromatic hydrocarbons (xylene and
ethyl benzene), 25.5 percent C9 hydrocarbons (trimethylbenzene and other compounds), and
20.6 percent C10+ aromatic hydrocarbons.
3-8
-------
"o
W1
•M
ro
1
1
S-,
C
u c
1
- h
L
-
3
L -
<
1
3.
"
1
> 2
?z
y
5 'D
5
^
1
^ -
S"L
l(
1
.
C
3-
- h
>.
D
1
'-c
f
; *
^
J
5
1
L C
f
2,
10
i
o
3 u
1
i =
-"
3-,
= -
n
3
1
= p:
s'
C
1
^- _
5 r
L
_
U
L
i i i
.. ^ 5 H
J oj" " ••
1 = 3":
1 1 1
1^3 =
f
1
; ;
^
5'-
|
1 '
West Bend, W
1
h
f
s'
> ^
1
!f
>^
v .
5
>
1
f
ll
c
1
= F
-"
u
5
1
f
i
^c
1
zf
^ C
1
if
£-
T3
L C
5
U
3 C
1
- c
5
L 4
5 .
5L
1
if
|:
^)
a c
1
^"
= :
u -
V
:>
1
^ <
f
D
= c
J 4
L
1
ic
U
i-
^ 1
1
L -
— ""
2
10 C
= 4
2L
1
il
V
ft
D
n
1
jf
1^
L
1
- c
^
!J
J L
1
i;
--
2-,
3P.
= -
/)
/)
LJ
1
^ ;
u
D ~
J '_
•
1
= P
s;
±
-i
c
C6
1
j
s
u i.
3
1
f
J
5
D
C7
1
t{
5 '-
c -
1
^ ;
D r
•
L
C
1
^ ;
=
r>
1 I
1
^ ;
IC9
1
> ^~
> 2
S|
3
• Cl
_if c"
t is
D+
1
j
Figure 3-8 Measured Aromatics by EPA Sample Location, 2007-2011
3-9
-------
Table 3-1 In-Use Gasoline Aromatics Composition
C6
C7
C8
C9
CIO
Cll
C12
Compound
Benzene
Toluene
m/p-Xylene
o-Xylene
Ethylbenzene
1,2,4-Tri methyl benzene
3-ethyltoluene
1,3,5-Tri methyl benzene
1,2,3-Tri methyl benzene
l-Methyl-2-ethyl benzene
n-Propylbenzene
Indan
4-ethyltoluene
Isopropylbenzene
alkyl indans
CIO Benzenes
n-butylbenzene
1,4-diethyl benzene
1,2,3,5-tetramethyl benzene
Naphthalene
1,2, 4,5-tetramethyl benzene
1,2-diethyl benzene
Cll Benzenes
2-Methyl naphthalene
1-Methyl naphthalene
Pentamethyl benzene
C12 Benzenes
Total Aromatics
%of
Aromatics
3.2%
22.5%
16.8%
6.4%
5.0%
9.0%
5.6%
2.6%
2.2%
1.9%
1.6%
1.2%
1.0%
0.5%
7.1%
3.6%
1.5%
1.5%
1.3%
0.9%
0.9%
0.1%
2.0%
0.8%
0.3%
0.1%
0.5%
100.0%
Total
3.2%
22.5%
r
28.3%
r
25.5%
r
20.6%
100.0%
Research by the Japan Petroleum Energy Center (JPEC), Honda, and others have found
that larger more complicated aromatic compounds result in greater vehicle particulate matter
(PM) emissions. The JPEC study found that PM mass emissions from a light-duty gasoline
vehicle increased with increasing carbon number of aromatics in the gasoline.3 Honda has
devised a "PM Index" that correlates PM emissions to the double bond equivalent (DBE) and
3-10
-------
vapor pressure (V.P) of the fuel components. Their research was published in an SAE paper on
October 25, 2010. 4 According to Honda, the PM index of the fuel is a function of all the
gasoline components (i) and their respective weight fractions (Wt;) as shown in Equation 1 .
DBE is essentially an indication of the number of double bonds and rings present in the
molecule. For example, benzene (CeHe) would have a DBE of four (three double bonds plus one
ring) while naphthalene (CioHg), would have a DBE of seven (five double bonds plus two rings).
As expected, gasoline containing a large fraction of heavier aromatics compounds with high
DBE values result in greater vehicle PM emissions.
Equation 1 Particulate Matter Index
DBE, + 1
Zr
LI/.
. P (44 3 A*),
Since aromatics do not appear to be created equally in terms of the potential impact on
vehicle PM emissions, we believe that it is prudent that our new Tier 3 gasoline emissions test
fuel contain both a representative amount and distribution of aromatics. In Section 3.2 we
explain how we arrived at the estimated aromatics concentration of our proposed El 5 test fuel.
We also detail our proposal for ensuring that the contributions of C6, C7, C8, C9, and C10+
aromatic hydrocarbons are representative of in-use gasoline.
3.1.3 Distillation Temperatures
As shown below in Figure 3-9, the transition from EO to E10 has had little, if any, effect
on gasoline distillation temperatures. However, we are proposing modest changes in our test fuel
to make distillation temperatures more representative of today's in-use fuel. According to AAM
summer fuel surveys, the average temperature at which 10 percent of gasoline is distilled (T10)
is at/around 130°F; the average temperature at which 50 percent of gasoline is distilled (T50) is
currently at/around 200°F; the average temperature at which 90 percent of gasoline is distilled
(T90) is at/around 325°F; and the average temperature at which all gasoline is distilled (final
boiling point or FBP) is at/around 400°F. The temperatures reported in Figure 3-9 represent an
average of the gasoline formulations (from 9- and 10-psi conventional to 7-psi reformulated) and
grades (87 (R+M)/2 regular to 91-93 (R+M)/2 premium) available to consumers and sampled by
AAM.
3-11
-------
500
400
— 300
ts
b
-------
300
250
200
fT 1r p nffnl
150
t:
o
a
100
50
5000
10000
15000 20000
Number of Batches
Figure 3-10 Gasoline T50 based on 2009 Refinery Batch Data
3-13
-------
450
400
150
100
50
15000
Number of Batches
25'juu
Figure 3-11 Gasoline T90 based on 2009 Refinery Batch Data
As part of our test fuel changes, we are proposing to eliminate our initial boiling point
(IBP) specification. The IBP of gasoline is directly correlated to the vapor pressure of gasoline.
Since we already limit the vapor pressure of our general testing fuel to 8.7-9.2 psi (and are not
planning on changing this under Tier 3), we believe it would be redundant and unnecessary to
maintain an IBP specification for emission test fuel. This is consistent with CARB's current
LEV II standards and planned LEV III approach.
3.1.4 Sulfur and Benzene
Gasoline sulfur levels have declined significantly over the past decade under the Tier 2
gasoline program. According to AAM summer fuel surveys, average gasoline sulfur has gone
from over 150 ppm in 2000 to less than 30 ppm in 2010 (refer to Figure 3-12). However, it's
worth noting that The Alliance may not have sampled all the gasoline markets that had higher
sulfur during the Tier 2 phase-in as a result of extra lead time given to GPA/Rocky Mountain
refineries and small refiners. According to refinery batch reports (refer to Figure 3-13), average
gasoline sulfur was still around 37 ppm in 2009. However, refinery certification data does not
include ethanol blended into conventional gasoline (CG), so the average downstream sulfur
could have been a few ppm lower in 2009. Regardless, since this time, all U.S. refiners
(including small refiners) have come into compliance with the Tier 2 30-ppm sulfur standard.
3-14
-------
Under today's proposed Tier 3 program, average gasoline sulfur levels are expected to be
reduced even further to around 10 ppm by 2017. Accordingly, our proposed Tier 3 test fuel
sulfur specification of 8-11 ppm is expected to be representative of future in-use gasoline.
Gasoline benzene levels have also been reduced in recent years due to the MSAT2
program. According to AAM summer fuel surveys, average gasoline benzene content has
declined from almost 1 vol% in 2005 to less than 0.8 vol% in 2010 (refer to Figure 3-12).
However, The Alliance may not have sampled all the gasoline markets that had higher benzene
during the MSAT2 phase-in as a result of extra lead time given to small refiners and early credit
provisions. According to refinery batch reports (refer to Figure 3-14), average gasoline benzene
was still around 1 vol% in 2009. However, due to unaccounted for CG ethanol blending,
average downstream benzene was likely a little lower in 2009. Nevertheless, effective January 1,
2011, all gasoline refiners and importers must be in compliance with the 0.62 vol% annual
average benzene standard. So today's gasoline benzene levels are even lower than those
reflected in the AAM fuel survey and 2009 compliance data. Accordingly, our proposed Tier 3
test fuel benzene specification of 0.6-0.8 vol% is expected to be representative of future in-use
gasoline. These proposed benzene (and sulfur) specifications are consistent with CARS's
planned LEV III specifications.
Avg. Sulfur (AAM) —^~Avg. Benzene (AAM)
(3
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Figure 3-12 Average Gasoline Sulfur and Benzene Levels Over Time
3-15
-------
500
5000
10000 15000 20000
Number of Batches
25000
30000
Figure 3-13 Gasoline Sulfur based on 2009 Refinery Batch Data
3-16
-------
m
TO
o
0.
5000
10000
15000
Number of Batches
25000
30000
Figure 3-14 Gasoline Benzene based on 2009 Refinery Batch Data
3.2 Projected E15 Implications
In-use gasoline is projected to continue to change with the implementation of the RFS2
program and the further expansion of El 5 into the marketplace. As explained in Section IV.D of
the preamble, we are proposing to update our Federal emissions test fuel not only to better match
today's in-use fuel (average gasoline properties detailed in Section 3.1) but also to be forward
looking with respect to future ethanol content. We are proposing to add a 15 vol% ethanol
specification to test fuel to be forward-looking with respect to the maximum gasoline ethanol
concentration Tier 3 vehicles could expect to encounter. The additional 5 vol% ethanol will have
second-order impacts on T50, aromatics, and olefms.
Shifting to El5 test fuel will result in a T50 range that's about 15 percent lower than
today's EO test fuel and about 10 percent lower than today's in-use E10. The proposed 170-
190°F T50 range for E15 test fuel was determined by interpolating between the T50 of current
E10 market fuel which averages around 195°F (according to AAMB) and the estimated T50 of
B Based on our assessment of 60 U.S. fuel samples containing 9.5-10.5 vol% ethanol and 8.5-9.5 psi RVP based on
AAM's summer 2010 North American Fuel Survey.
3-17
-------
E20 which spans a narrow range of plus or minus several degrees Fahrenheit centered around
165°F (according to our EPAct test program5). We believe the proposed T50 range adequately
characterizes future 9-psi El5 market fuel.
As explained above in Section 3.1, the commercialization of E10 has already resulted in a
15 percent reduction in aromatics and a 19 percent reduction in olefins in relative terms.
Refiners are backing off on expensive aromatics and olefin production in anticipation for
downstream ethanol blending. We expect that refiners will continue to back off on octane
production once the transition from E10 to E15 has occurred.0 With that, we can extrapolate to
arrive at future aromatics levels in the neighborhood of 21 vol% and olefin levels in the
neighborhood of 6.5 vol% for E15. Our Tier 3 sulfur standards should also result in slightly
lower olefin production at the refinery. Accordingly, we are proposing an aromatics
specification of 19.5-24.5 vol% and an olefins specification of 4-10 vol% (reported as 4.5-11.5
mass% per ASTM D6550) for our proposed El5 test fuel as shown in Table 3-3. We believe
that these ranges encompass levels that might be typical for E15 (as well as E10) while still
providing flexibility for specialty fuel manufacturers producing Tier 3 test fuel. The proposed
aromatics and olefin specifications, although slightly broader to represent the wider range of
Federal gasoline properties (detailed in Section 3.1), encompass CARB's planned LEV III
emission test fuel specifications.
As explained earlier, in-use gasoline contains a range of aromatic compounds whose
contribution to PM emissions seems to increase with carbon number. Accordingly, we are
planning on building off the proposed 19.5-24.5 vol% total aromatics spec for E15 and the
proposed 0.6-0.8 vol% benzene spec to set test fuel specs for the remaining unspecified aromatic
hydrocarbons. For the NPRM, have relied on EPA's 2007-2011 in-use aromatics data
(presented in Section 3.1.2) to arrive at the proposed C7, C8, C9 and C10+ specifications
presented in Table 3-2. We seek comment on the appropriateness of the proposed aromatics
specifications and welcome any additional gasoline aromatics data that others may want to
provide us with through notice and comment. We also seek comment on other alternative
approaches that would result in a distribution of gasoline aromatics in emissions testing fuel.
One such approach might be to set regulations that require equal contributions of C7, C8, C9 and
C10+ aromatics in the range of 5 ± 1 vol%.
Table 3-2 Proposed Aromatics Contributions in Tier 3 Cert Fuel
C6 (benzene)
C7 (toluene)
C8
C9
C10+
Total Aromatics
Aromatics (vol%)
0.6-0.8
4.4-5.5
5.5-6.9
5.0-6.2
4.0-5.0
19.5-24.5
c In the interim, while E10 and E15 co-exist, we anticipate that refiners will make E15 using a blendstock similar to
today's E10 blendstock. While this may result in a small, temporary octane giveaway, the refinery burden will
likely be less than that of producing a separate sub-octane blendstock for E15.
3-18
-------
While revisiting emission test fuel parameters, we are taking this opportunity to propose
specifications for distillation residue, total content of oxygenates other than ethanol, copper
corrosion, solvent-washed gum content, and oxidation stability. The proposed parameters,
summarized below in Table 3-3, are consistent with ASTM's D-4814 gasoline specifications and
CARS's planned LEV III test fuel requirements.
3.3 Proposed Gasoline Emission Test Fuel Specifications
As explained in Section IV.D of the preamble, we are proposing a consolidation of all
gasoline exhaust and evaporative emission test fuels into a single general test fuel. This would
be used for all on-highway vehicle testing with the exception of cold CO vehicle testing (which
would use higher volatility test fuel) and high-altitude testing (which would be permitted to use
lower volatility fuel). Commercial gasoline or "street fuel" would continue to be used for service
accumulation (durability fuel). This is consistent with CARB's LEV III approach and should
help limit the total number of test fuels that automakers need to manage. The proposed Tier 3
test fuel specifications are summarized in Table 3-3. For more on how we arrived at the
proposed ASTM test procedures, refer to Section 3.4.
Table 3-3 Proposed Gasoline Emission Test Fuel Specifications
Property
Antiknock Index
(R+M)/2
Sensitivity (R-M)
Dry Vapor Pressure
Equivalent (DVPE)C
Distillation
10% evaporated
50% evaporated
90% evaporated
Evaporated final boiling
point
Residue
Total Aromatic
Hydrocarbons
C6 Aromatics
(benzene)
C7 Aromatics
(toluene)
C8 Aromatics
C9 Aromatics
C 10+ Aromatics
Olefmsd
Unit
-
-
kPa (psi)
°C (°F)
°C (°F)
°C (°F)
°C (°F)
milliliter
vol. %
vol. %
vol. %
vol. %
vol. %
vol. %
mass %
SPECIFICATION
General Low-
Testing Temperature
Testing
87.0 - 88.4b
High Altitude
Testing
87.0 Minimum
7.5 Minimum
60.0-63.4 77.2-81.4
(8.7-9.2) (11.2-11.8)
49-60 43-54
(120-140) (110-130)
52.4-55.2
(7.6-8.0)
49-60
(120-140)
77-88(170-190)
154-166(310-330)
193-216(380-420)
2.0 Maximum
19.5-24.5
0.6-0.8
4.4-5.5
5.5-6.9
5.0-6.2
4.0-5.0
4.5-11.5
ASTM Reference Procedure8
D2699-llandD2700-ll
D5191-10b
D86-10a
D5769-10
D6550-10
3-19
-------
Property
Ethanol6
Total Content of
Oxygenates Other than
Ethanol6
Sulfur
Lead
Phosphorus
Copper Corrosion
Solvent- Washed Gum
Content
Oxidation Stability
Unit
vol. %
vol. %
mg/kg
g/liter
g/liter
-
mg/lOOml
minute
SPECIFICATION
14.6-15.0
0.1 Maximum
8.0-11.0
0.0026 Maximum
0.0013 Maximum
No. 1 Maximum
3.0 Maximum
1 ,000 Minimum
ASTM Reference Procedure8
D5599-00 (Reapproved2010)
D2622-10,D5453-09or
D7039-07
D3237-06
D3231-11
D130-10
D381-09
D525-05
Notes:
"ASTM procedures are incorporated by reference in §1065.1010. See §1065.701(d) for other allowed procedures.
b Octane specifications apply only for testing related to exhaust emissions. For engines or vehicles that require the
use of premium fuel, as described in §1065.710(d), the adjusted specification for antiknock index is a minimum
value of 91.0; no maximum value applies. All other specifications apply for this high-octane fuel.
0 Calculate dry vapor pressure equivalent, DVPE, based on the measured total vapor pressure, pi, in kPa using the
following equation: DVPE = 0.956-/>T-2.39. DVPE is intended to be equivalent to Reid Vapor Pressure using a
different test method.
d The specified olefm concentration range equates to approximately 4-10 volume percent when measured
according to ASTM D1319.
e The reference procedure prescribes measurement of ethanol concentration in mass %. Convert results to volume %
as specified in ASTM D4815-09.
3.4 Changes to ASTM Test Methods
Many of the test methods specified in 40 CFR 86.113 for gasoline used in exhaust and
evaporative emission testing of petroleum-fueled Otto-cycle vehicles have been retained in the
proposed 40 CFR 1065.710 test fuel specification for ethanol-blended gasoline. However, some
test methods have been replaced with methods deemed more appropriate, easier to use, or more
precise. The following paragraphs highlight the new reference methods.
ASTM D323 "Standard Test Method for Vapor Pressure of Petroleum Products (Reid
Method)" is not applicable to ethanol-blended gasoline. We are planning to replace this with an
automated ASTM D5191 "Standard Test Method for Vapor Pressure of Petroleum Products
(Mini Method)," which is appropriate for ethanol-blended gasoline.
ASTM D1319 "Standard Test Method for Hydrocarbon Types in Liquid Petroleum
Products by Fluorescent Indicator Adsorption" is required by 40 CFR 86.113 for use in the
measurement of aromatics and olefms. We are planning to replace it with ASTM D5769
"Standard Test Method for Determination of Benzene, Toluene, and Total Aromatics in Finished
Gasolines by Gas Chromatography/Mass Spectrometry" and ASTM D6550 "Standard Test
Method for Determination of Olefm Content of Gasolines by Supercritical-Fluid
Chromatography." Method D5769 enables simultaneous determination of the total aromatic
hydrocarbon content, carbon number-specific content, and benzene content and is currently
3-20
-------
being used in reformulated gasoline applications. ASTM D1319 does not measure carbon
number-specific aromatic hydrocarbon content, which are now specified for the ethanol-blended
test fuel in 40 CFR 1065.710. In addition, ASTM D5769 and D6550 are more precise and less
labor-intensive than ASTM D1319.
For sulfur measurements, we are planning to replace ASTM D1266 "Standard Test
Method for Sulfur in Petroleum Products (Lamp Method)" with three automated methods:
ASTM D2622 "Standard Test Method for Sulfur in Petroleum Products by Wavelength
Dispersive X-ray Fluorescence Spectrometry", ASTM D5453 "Standard Test Method for
Determination of Total Sulfur in Light Hydrocarbons, Spark Ignition Engine Fuel, Diesel Engine
Fuel, and Engine Oil by Ultraviolet Fluorescence" and ASTM D7039 "Standard Test Method for
Sulfur in Gasoline and Diesel Fuel by Monochromatic Wavelength Dispersive X-ray
Fluorescence Spectrometry." These three new methods are significantly less labor-intensive than
ASTM D1266 and are widely used in the measurement of sulfur content in petroleum products.
We request comments on the use of three different test methods for the measurement of sulfur
content.
3-21
-------
References
1 EIA, January 2012 Monthly Energy Review, Table 3.7a, Petroleum Consumption: Residential and Commercial
Sectors; Table 3.7b, Petroleum Consumption: Industrial Sector; Table 3.7c, Petroleum Consumption: Transportation
Sector, and Table 10.3, Fuel Ethanol Overview.
2 EIA, Petroleum Marketing Annual 2009, Table 45, Prime Supplier Sales Volume of Motor Gasoline by Grade,
Formulation, PAD District and State.
3 lizuka, Masashi, Advanced Technology and Research Institute (ATRI) and Japan Petroleum Energy Center
(JPEC), Effect of Fuel Properties on Emissions from Direct Injection Gasoline Vehicle.
4 Jetter, Jeff, Honda R&D America Inc., Development of a Predictive Model for Gasoline Vehicle Particulate Matter
Emissions, SAE 2010-01-2115.
5 U.S. Environmental Protection Agency (2012). Assessing the Effect of Five Gasoline Properties on Exhaust
Emissions from Light-Duty Vehicles Certified to Tier-2 Standards (EPAct/V2/E-89: Phase 3), Part I: Program
Design and Data Collection. EPA Report Number XX.
3-22
-------
Chapter 4 Fuel Program Feasibility
4.1 Overview of Refining Operations
Figure 4-1 shows a process flow diagram for a typical complex refinery, capable of
making a wide product slate (shown on the right side of the figure) from crude oil (input on the
left). Following the figure is a brief description of key units and streams focusing more on the
Natural
Gas
Hydrogen Plant
Fuel Gas
LPG
Gasoline
Aromatics
Kerosene
Jet Fuel
Low Sulfur Diesel
Off-road Diesel
Heating Oil
Resid
Coke
Figure 4-1 Process Flow Diagram for a Typical Complex Refinery
Crude Tower
The purpose of the crude tower is to perform a distillation separation of crude oil into
different streams for additional processing in the refinery and for the production of specific
products. Crude oil is shipped to the refinery via pipeline, ship, barge, rail, or truck, whereupon
it is sampled, tested, and approved for processing. The crude oil is heated to between 650 °F and
700 °F and fed to crude distillation tower. Crude components vaporize and flow upward through
the tower. Draw trays are installed at specific locations up the tower from which desired side
cuts or fractions are withdrawn. The first side-cut above the flash zone is usually atmospheric
gasoil (AGO), then diesel and kerosene/jet fuel are the next side-cuts, in that order. The lightest
4-1
-------
components, referred to here as straight run naphtha, remain in the vapor phase until they exit the
tower overhead, following which they are condensed and cooled and sent to the naphtha splitter.
Naphtha Splitter
The purpose of the naphtha splitter is to perform a distillation separation of straight run
naphtha into light straight run naphtha and heavy straight run naphtha. The feed can be split
between the C5's and C6's in order to assure the C6's and heavier are fed to the reformer.
Naphtha Hydrotreater
The purpose of the naphtha hydrotreater is to reduce the sulfur of light and heavy straight
run streams before those streams are refined further by the isomerization and reformer units.
Isomerization Unit
The purpose for the isomerization unit is to convert the light naphtha from straight chain
hydrocarbons to branched chain hydrocarbons, increasing the octane of this stream. The
isomerate is sent to gasoline blending.
Reformer
The purpose of the reformer unit is to convert heavy straight run (C6 to C8 or C9
hydrocarbons) into aromatic and other higher octane compounds (benzene is one of the aromatic
compounds produced), typically necessary to produce gasoline with sufficient octane. To protect
the very expensive, precious metal catalyst used in reformers, heavy straight run naphtha must be
hydrotreated first before it is fed to the reformer. As the reformer converts the feed
hydrocarbons to aromatics, hydrogen and light gases are produced as byproducts. The liquid
product, known as reformate, is sent directly to gasoline blending, or to aromatics extraction.
Aromatics Extraction Unit
The purpose of aromatics extraction is to separate the aromatic compounds from the rest
of the hydrocarbons in reformate using chemical extraction with a solvent to concentrate the
individual aromatic compounds, (mainly xylene and benzene) for sale to the chemicals market.
Vacuum Tower
The purpose of the vacuum distillation tower unit is to enable a refinery to produce more
gasoline and diesel fuel out of a barrel of crude oil. It separates the vacuum gasoil (VGO), which
is fed to the FCC unit, from the vacuum tower bottoms (VTB) which is sent to the coker, or in
other refineries is made into asphalt. Because most sulfur contained in crude oil is contained in
the heaviest part of crude oil, the VGO and VTB are very high in sulfur.
Fluidized Catalytic Cracker
The purpose of the fluidized catalytic cracker is to convert heavy hydrocarbons, which
have very low value, to higher value lighter hydrocarbons. AGO and VGO are the usual feeds to
4-2
-------
a fluid catalytic cracker (FCC). The full boiling range cracked product leaves the reactor and is
sent to a fractionator. The overhead includes propane, propylene, butane, butylene, fuel gas and
FCC naphtha, which contains a significant amount of sulfur. There are two heavy streams; light
cycle oil (LCO), which can be hydrotreated and blended into diesel fuel or hydrocracked into
gasoline; and heavy cycle oil, sometimes called slurry oil, which can be used for refinery fuel.
Very simple refineries do not have FCC units, and therefore, their gasoline is very low in sulfur.
FCC Feed Hydrotreater or Mild Hydrocracker "A"
FCC feed hydrotreaters and mild hydrocrackers hydrotreat or mildly hydrocrack the feed
to the FCC unit which provides two distinct benefits. First, by increasing the amount of
hydrogen in the feed to the FCC unit, the FCC unit increases the conversion of the feed to high
value light products, particularly FCC naphtha which increases the gasoline yield. Second,
hydrotreating the feed removes some contaminants in the feed such as nitrogen and sulfur.
Nitrogen in the feed negatively affects the FCC catalyst. Removing the sulfur in the feed helps
in two ways. Some of the sulfur in the feed is released by the cracking process and results in
high SOx emissions that would otherwise have to be controlled by scrubbers - the FCC feed
hydrotreaters may prevent the need to add a scrubber. Also, FCC feed hydrotreaters remove
sulfur which can allow a refinery to comply with gasoline sulfur standards.
FCC Postreat Hydtrotreater "B"
Postreat hydrotreaters solely hydrotreat the naphtha that is produced by the FCC unit to
reduce its sulfur level which enables compliance with gasoline sulfur standards. The FCC
naphtha is high in olefms which can be saturated by postreat hydrotreaters resulting in lower
octane of the FCC naphtha. Vendor companies have developed postreat hydrotreating
technologies which minimize this octane loss.
Distillate Hydrotreater
The purpose of the distillate hydrotreater is to reduce the sulfur of distillate, which is also
called diesel fuel.
Gas Plant
The purpose of the gas plant is to use a series of distillation towers to separate various
light hydrocarbons for further processing in the alkylation or polymerization units or for sale.
Alkylation Unit
The purpose of the alkylation unit is to chemically react light hydrocarbons together to
produce a high quality, heavy gasoline product. Alkylation uses sulfuric or hydrofluoric acid as
catalysts to react butylene or propylene together with isobutane. Following the main reaction
and product separation, the finished alkylate is sent to gasoline blending. Alkylate is low in
RVP.
Polymerization Unit
4-3
-------
The purpose of the polymerization unit is to react light hydrocarbons together to form a
gasoline blendstock. A polymerization unit, often referred to as a "cat poly" is somewhat similar
to an alkylation unit, in that both use light olefins to produce gasoline blendstocks. The feed is
generally propylene and/or butylene from the gas plant. The product, called polygas is sent to
gasoline blending.
Coker Unit
The purpose of the coker unit is to process vacuum tower bottoms (VTB) to coke and to
crack a portion to various lighter hydrocarbons. The hydrocarbons produced by the coker
include cracked gases, coker naphtha, coker distillate and gas oil. The gas is fed to the gas plant,
the naphtha to the naphtha hydrotreater after which the heavy coker naphtha is typically fed to
the reformer, and the distillate either to distillate hydrotreating or to the hydrocracker.
Hydrocracker
The purpose of the hydrocracker is to crack and "upgrade" the feedstock into higher
value products. The feedstock to the hydrocracker is usually light cycle oil (LCO) and coker
distillate, poor quality distillate blendstocks, which are upgraded to diesel fuel, or cracked to
gasoline. Heavier hydrocarbons such as AGO and HVGO can be feedstocks as well.
A more complete description for naphtha hydrotreating is contained in Section 4.2.
4.2 Feasibility of Removing Sulfur from Gasoline
The case can be made in two ways that it is feasible to comply with the proposed 10-ppm
gasoline sulfur standard. First, feasibility can be demonstrated by understanding the
technologies currently available which can achieve the necessary reductions in sulfur, and that
these technologies are currently being used to achieve significant reductions in gasoline sulfur.
The second way to make the case that it is feasible to comply with the proposed 10-ppm gasoline
sulfur standard is to highlight that refiners in certain countries or other regions are currently
complying with a 10-ppm gasoline sulfur cap standard. These two cases will be made below, but
first we will review the source of sulfur in gasoline to understand how sulfur levels can be
further reduced.
4.2.1 Source of Gasoline Sulfur
Sulfur is in gasoline because it naturally occurs in crude oil. Crude oil contains anywhere
from fractions of a percent of sulfur, such as less than 500 ppm (0.05 weight percent) to as much
as 30,000 ppm (3 percent). The average amount of sulfur in crude oil refined in the U.S. is about
14,000 ppm. Most of the sulfur in crude oil is in the heaviest part, or in the heaviest petroleum
compounds, of the crude oil (outside of the gasoline boiling range). In the process of refining
crude oil into finished products, such as gasoline, some of the heavy compounds are broken up,
or cracked, into smaller compounds and the embedded sulfur can end up in gasoline. Thus, the
refinery units which convert the heavy parts of crude oil into gasoline are the units most
responsible for putting sulfur into gasoline.
4-4
-------
The fluidized catalytic cracker (FCC) unit is a refinery processing unit that creates a high
sulfur content gasoline blendstock. FCC naphtha contains from hundreds to several thousand
parts per million of sulfur. The FCC unit cracks large carbon molecules into smaller ones and
produces anywhere from 25 to 50 percent of the gasoline in those refineries with FCC units.
Because the FCC unit makes a gasoline blendstock out of the heavier, higher sulfur-containing
compounds, more than 95 percent of sulfur in gasoline blendstocks comes from streams
produced in that unit. When complying with the 30-ppm Tier 2 gasoline sulfur standard, refiners
reduced the sulfur content of the FCC naphtha. The impact of this action is described below in
subsection 4.2.2.
Straight run naphtha is a gasoline blendstock which contains a moderate amount of
sulfur. Straight run naphtha is the part of crude oil which after distillation in the atmospheric
crude oil tower falls in the gasoline boiling range. The heaviest portion of straight run, which
would have more sulfur, is normally desulfurized and reformed in the reformer (to improve its
octane), so its contribution to the gasoline pool is virtually nil.A The light straight run which
contains the five-carbon hydrocarbons contains on the order of 100 ppm sulfur and if this
material is not hydrotreated and processed in an isomerizaition unit, it is blended directly into
gasoline.
Another refinery unit which produces naphtha with a significant amount of sulfur is the
coker unit. These units produce coke from the heaviest part of the crude oil. In the process of
producing coke, a naphtha is produced that contains more than 3,000 ppm sulfur and many very
unstable olefms. Because this stream is highly olefmic and unstable, refiners tend to hydrotreat
coker naphtha. Coker naphtha is normally split into two different streams. The six- to nine-
carbon hydrocarbons are hydrotreated along with the rest of the heavy naphtha and fed to the
reformer. The five-carbon hydrocarbon part of coker naphtha is called light coker naphtha and
usually contains on the order of several hundred parts per million sulfur. Light coker naphtha is
usually hydrotreated along with the light straight run, and refined further in an isomerization unit
if the refinery has one.
Other gasoline blendstocks contain little or no sulfur. Alkylate, which is produced from
isobutene and butylenes that contain a small amount of sulfur, can end up with a small amount of
sulfur. Most refineries have less than 15 ppm sulfur in this pool, however, some refineries which
feed coker naphtha to the alkylate plant can have much more. On average, alkylate probably has
about 10 ppm sulfur. One more gasoline blendstock with either very low or no sulfur is
hydrocrackate, which is the naphtha produced by hydrocrackers. It is low in sulfur because the
hydrocracking process removes the sulfur. Ethanol which is eventually blended into gasoline
usually has very little or no sulfur. However, the hydrocarbon used as a denaturant and blended
with ethanol at 2 percent is usually natural gasoline, a C5 to C7 naphtha from natural gas
processing, and it contains anywhere from a few parts per million to a couple hundred parts per
million sulfur. After the denaturant is blended in, the denatured ethanol contains somewhere
A Sulfur interferes in the function of the precious metal catalyst used in the reforming process. As a result, refiners
historically have desulfurized the heavy straight run naphtha feed to the reformer from several hundred ppm sulfur
down to less than 1 ppm.
4-5
-------
between 0 and 10 ppm sulfur. To meet current pipeline and California specifications, denatured
ethanol must be less than 10 ppm sulfur.
4.2.2 Complying with the Current Tier 2 Gasoline Sulfur Standard
It is important to understand the steps that refiners took to comply with the 30-ppm Tier 2
gasoline sulfur standard because those capital investments and operational changes will play a
major role in determining the steps that refiners take to comply with a more stringent gasoline
sulfur standard.
The Tier 2 sulfur standard was promulgated February 10, 2000.2 The sulfur standard
requires that refiners reduce their annual average gasoline sulfur levels down to 30 ppm and each
gallon of gasoline cannot exceed a per-gallon standard of 80 ppm. The sulfur standards were
phased in from 2004 to 2006. The only exceptions were western refiners (GPA) and small
refiners who were given until 2008. Some small refiners had their gasoline sulfur deadlines
extended through 2010 in the highway diesel fuel sulfur rule in exchange for on-time compliance
there. As of January 1, 2011, all refineries are complying with the Tier 2 30-ppm sulfur
standard.
A refinery's previous average gasoline sulfur level is an important factor which
determined whether a refiner would need to make a substantial capital investment to meet the
Tier 2 gasoline sulfur standards. We believe that those refiners with low gasoline sulfur levels to
begin with (i.e., gasoline sulfur levels lower than, perhaps, 50 ppm) probably did not invest in
expensive capital. These refineries have very low sulfur levels due to one or more of a number
of possible reasons. For example, some of these refiners may not have certain refining units,
such as a fluidized catalytic cracker (FCC) unit, or a coker, which convert heavy boiling stocks
to gasoline. As stated above, these units push more sulfur into gasoline and their absence means
much less sulfur in gasoline. Alternatively, refiners may use a very low sulfur (sweet) crude oil
which can result in a low sulfur gasoline. Or, these refiners may have already installed an FCC
feed hydrotreater to improve the operations of their refinery which uses a heavier, higher sulfur
(more sour) crude oil. As described above, this unit removes much of the sulfur from the
heaviest portion of the heavy gas oil before it is converted into gasoline.
Of the refiners in this first category, the refineries with average sulfur levels below 30
ppm may not have had to do anything to meet the Tier 2 standards. On the other hand, those
refineries which had sulfur levels above 30 ppm but below some level, such as 50 ppm, probably
are meeting the 30-ppm sulfur standard by employing operational changes only and avoided
making capital investments. Most of the refineries with gasoline sulfur levels below 50 ppm
prior to the Tier 2 investments either do not have a FCC unit, or if they do, they probably have an
FCC feed hydrotreating unit.
The vast majority of gasoline which was being produced was by refineries with higher
sulfur levels, and these refiners had to either adapt some existing hydrotreating capital or install
new capital equipment in these refineries to meet the Tier 2 gasoline sulfur standards. As stated
above, the FCC unit is responsible for most of the sulfur in gasoline. Thus, investments for
desulfurizing gasoline involved the FCC unit to maximize the sulfur reduction, and to minimize
the cost. This desulfurization capital investment can be installed to treat the gas oil feed to the
4-6
-------
FCC unit, or treat the gasoline blendstock which is produced by the FCC unit. Each method has
advantages and disadvantages.
4.2.2.1 Using FCC Feed Pretreat Hydrotreating to Comply with Tier 2
Some refiners installed FCC feed hydrotreaters (also known as pretreaters) at their
refineries to comply with the Tier 2 gasoline sulfur standards. FCC pretreaters treat the vacuum
gas oil, heavy coker gas oil and, in some cases, atmospheric residual feed to the FCC unit using a
hydrotreater or a mild hydrocracker. These units are designed to operate at high pressures and
temperatures to treat a number of contaminants in the feed. Besides sulfur, FCC pretreaters also
reduce nitrogen and certain metals such as vanadium and nickel. These nonsulfur contaminants
adversely affect the FCC catalyst, so the addition of this unit would improve the functioning of
the unit. Also, the hydrotreating which occurs in the FCC pretreater reacts hydrogen in the
feedstock which increases the yield of the FCC unit by increasing the highest profit-making
products produced by refineries, such as gasoline and light olefms.3 While FCC pretreaters
provide yield benefits that offset the capital costs of adding this type of desulfurization, the costs
are still high enough that many refiners would have a hard time justifying the installation of this
sort of unit. For a medium to large refinery (i.e., 150,000-200,000 BPCD), the capital costs may
exceed $250 million. Because of the higher temperatures and pressures involved, utility costs
are expensive relative to postreat hydrotreating as explained below. Another justification for this
approach is that it allows refiners to switch to a heavier, more sour crude oil. These crude oils
are less expensive per-barrel and can offset the increased utility cost of the FCC pretreater,
providing that the combination of reduced crude oil costs and higher product revenues justify the
switch. Another benefit for using FCC pretreaters is that the portion of the distillate pool which
comes from the FCC unit would be partially hydrotreated as well. This distillate blendstock,
termed light cycle oil, comprises a relatively small portion of the total distillate produced in the
refinery (about 20 percent of on-road diesel comes from light cycle oil), and like FCC naphtha,
light cycle oil contributes a larger portion of the total sulfur which ends up in distillate. Thus,
FCC pretreaters would also help a refiner meet the 15-ppm highway and nonroad diesel fuel
standards.
In terms of desulfurization capability, FCC preateaters have different abilities to remove
sulfur from the gas oil feed depending on the unit pressure. FCC pretreaters can be subdivided
into high pressure units (1400 psi and above), medium pressure units (900 to 1400 psi), and low
pressure units (under 900 psi). High pressure FCC pretreaters are capable of removing about 90
percent of the sulfur contained in the gas oil feedstock to the FCC unit, while low and medium
pressure units are capable of removing 65 to 80 percent of the feed sulfur.4 Because there is no
postreating at many of the refineries with FCC pretreaters, control of the feed to these units is a
critical determining factor for how well the FCC pretreater will function as desulfurizers. If the
feed becomes too heavy, the concentration of contaminants increases and the catalyst may lose
its effectiveness.5
FCC pretreaters improve desulfurization indirectly by improving the desulfurization
performance of the FCC unit itself. When FCC units crack the vacuum gas oil into naphtha,
about 90 percent of the sulfur is typically cracked out of the hydrocarbons converted to FCC
naphtha (or the FCC naphtha contains only about 10 percent of the sulfur present in the feed) and
is removed as hydrogen sulfide. When FCC units are preceded by FCC pretreaters, the amount
4-7
-------
of sulfur in the feed which ends up in the FCC naphtha is only about 5 percent, which means that
about 95 percent of the sulfur in the feed is removed from the FCC feed when it is cracked into
FCC naphtha. This effect is caused by the additional hydrogen which reacts with the feed
hydrocarbons. With more hydrogen molecules available in the feedstock after hydrotreatment,
the FCC cracking reactions can react more hydrogen with the sulfur contained in the feed to
produce more hydrogen sulfide.
For complying with Tier 2, refiners which already had or which installed high pressure
FCC pretreaters were able to comply with the 30-ppm sulfur standard without the need to install
a FCC naphtha hydrotreater. However, if a refinery had a low pressure, or perhaps even a
medium pressure FCC feed hydrotreater, they were generally less able to comply with the 30-
ppm gasoline sulfur standard with the FCC hydrotreater by itself, and these refineries were more
likely to also install an FCC postreater.
4.2.2.2 Using FCC Naphtha Postreat Hydrotreating to Comply with Tier 2
A less capital intensive alternative for reducing FCC naphtha sulfur levels to comply with
Tier 2 is FCC naphtha hydrotreating (also known as postreaters). FCC postreaters only treat the
gasoline blendstock produced by the FCC unit. Understandably, this unit is much smaller
because only about 50 to 60 percent of the feed to the FCC unit ends up as FCC naphtha, a
gasoline blendstock. The unit is sometimes smaller than that as some refiners which choose to
use a fixed bed hydrotreater may only treat the heavier, higher sulfur portion of that stream with
hydrotreating, and then treat the lighter fraction with another lower desulfurization cost
technology. FCC postreaters operate at lower temperatures and pressures than FCC pretreaters,
which further reduces the capital and operating costs associated with this type of desulfurization
equipment. Furthermore, the feed to the FCC unit has corrosive properties which require that
FCC pretreaters use more corrosion-expensive metallurgy, which is not needed for postreaters.6
For a medium to large-sized refinery, the capital costs are on the order of $70 million for a
conventional FCC pretreaters - less than half that of an FCC pretreater.
One drawback of this desulfurization methodology is that the octane value and/or some
of the gasoline yield may be lost depending on the process used for desulfurization. Octane loss
occurs by the saturation of high octane olefins which are produced by the FCC unit. Most of the
olefins are contained in the lighter fraction of FCC naphtha.7 Increased olefm saturation usually
means higher hydrogen consumption. There can also be a loss in the gasoline yield caused by
mild cracking which breaks some of the gasoline components into smaller fractions which are
too light for blending into gasoline. If there is octane loss, two of the ways that the octane loss
can be made up is by blending in more ethanol, or by increasing the feed to or the severity of the
reformer, the aromatics production unit of the refinery. Sometimes vendors of FCC pretreater
technologies design octane increasing capability into their designs, which is discussed below in
the section about the individual postreater technologies.
The loss of octane and gasoline yield caused by FCC postreating is lower with
technologies which were developed prior to the implementation of the Tier 2 program.8 These
processes are termed selective because they achieve the lower sulfur while preserving much of
the octane and gasoline yield (they were designed specifically for treating FCC naphtha). Octane
is preserved because the hydrotreating units and their catalysts are specially designed to avoid
4-8
-------
saturating olefins. These selective processes, or parts of these processes, usually operate at less
severe conditions which result in less cracking preserving yield compared to conventional
hydrotreating processes. The less severe conditions also lower the capital and operating costs for
this process. The lower operating costs arise out of the reduced utility requirements (e.g., lower
pressure). For example, because these processes are less severe, there is less saturation of
olefins, which means that there is less hydrogen used. Less olefin saturation also translates into
less octane loss which would otherwise have to be made up by octane boosting processing units
in the refinery. The lower capital and operating costs of these newer FCC postreaters are
important incentives for refiners to choose this desulfurization methodology over FCC
pretreaters. For this reason, refiners chose to use the more recently developed FCC postreaters
technologies for meeting the 30-ppm Tier 2 gasoline sulfur standard.
Not saturating the olefins to preserve octane and limit hydrogen consumption provides a
different challenge. When the hydrogen sulfide is formed and there is a significant concentration
of olefins present, the hydrogen sulfide compounds tend to react with the olefmic hydrocarbon
compounds forming mercaptan sulfur compounds. This reaction is called "recombination"
because the removed sulfur recombines with the olefmic hydrocarbons contained in the naphtha.9
This is particularly a problem if the light cat naphtha is present in the hydrotreater because the
highest concentration of olefins is in the light cat naphtha. The recombination reactions occur
more readily if the hydrotreater is operated more severely (at a higher temperature) to increase
the sulfur removal, and the feed to the hydrotreater is high in sulfur. However, while operating
this type of hydrotreater more severely can result in the further removal of the original sulfur
present in the hydrocarbons, it also can result in the formation of more recombination
mercaptans that results in a "floor" reached for the amount of sulfur that can be removed from
the hydrocarbons. This cycle of increased sulfur removal and simultaneous increase in
recombination results in the saturation of more olefins and increases the consumption of
hydrogen. There are a number of different vendor-specific technologies that each vendor may
use to avoid or address recombination reactions as discussed below. It is important to note that
the technologies employed to reduce recombination may require the addition of some capital
costs which offsets some or perhaps all the capital cost savings due to the milder operating
conditions of these selective hydrotreater technologies compared to nonselective hydrotreating.
One means to achieve high levels of desulfurization while avoiding much of the problem
with recombination reactions is by using a two-stage hydrodesulfurization methodology. A two-
stage unit has two desulfurization reactors, but instead of just adding additional reactor volume,
the hydrocarbons exiting the first reactor are stripped of gaseous compounds (most importantly,
the hydrogen sulfide is removed), injected with fresh hydrogen, and then hydrodesulfurized
again in the second stage. Both reactors undergo modest desulfurization and hydrogen sulfide
concentrations remain sufficiently low to avoid recombination reactions. The downside of this
approach is that the second stage incurs greater capital costs compared to single-stage
configurations. Because Tier 2 was not too constraining, we believe that refiners installed few, if
any, two-stage desulfurization units to comply with those gasoline sulfur standards.
Whatever strategy chosen by the refiner to comply with Tier 2, a critical criterion is that
the postreater be capable of cycle lengths that match that of the FCC unit, which typically is 5
years. If the postreater were to require a catalyst changeout before the FCC unit requires a
shutdown, either the refiner would have to shutdown the FCC unit early to mirror that of the
4-9
-------
postreater, or need to store up the high sulfur FCC naphtha (this stream would be too high in
sulfur to blend directly to gasoline under the Tier 2 80-ppm cap standard) until the postreater was
started up again and is able to hydrotreat the stored up high sulfur FCC naphtha.
We know of six FCC postreater technologies that refiners used to comply with the Tier 2
gasoline sulfur standards. These are Axens (was IFF) Prime G and Prime G+, Exxon Scanfming,
CDTech's CDHydro and HDS, Sinopec's (was Phillips)S-Zorb and UOP's ISAL and
Selectfining.
Of the list of FCC postreaters, Axens Prime G+, Exxon Scanfming and UOP's ISAL and
Selectfining are fixed bed desulfurization technologies. These processes are called fixed bed
because the catalyst resides in a fixed bed reactor.10 The high sulfur gasoline blendstock is
heated to a high temperature (on the order of 600 degrees Fahrenheit) and pumped to a high
pressure to maintain the stream as a liquid. It is then combined with hydrogen before it enters
the reactor. The reactions occur within the bed of the catalyst. While the petroleum is in contact
with the catalyst in the reaction vessel, the sulfur reacts with hydrogen and is converted to
hydrogen sulfide. Also, depending on the process, some of the olefin compounds which are
present in the cracked stream are saturated which increases the amount of octane lost and
hydrogen consumed. After the reactor, the gaseous compounds, which include unreacted
hydrogen, hydrogen sulfide, and any light end petroleum compounds which may have been
produced in the reactor by cracking reactions, are separated from the liquid compounds by a
gas/liquid separator. The hydrogen sulfide must be stripped out from the other compounds and
then converted to elemental sulfur in a separate sulfur recovery unit. The recovered sulfur is
then sold. If enough hydrogen is present, and it is economical to recover, it is separated from the
remaining hydrocarbon stream and recycled. Otherwise, it is burned with light hydrocarbons as
fuel gas.
Despite the similarities, each of these desulfurization technologies has its differences.
Axens Prime G+ desulfurization process largely preserves olefins as its strategy for diminishing
octane loss.11'1213 The Axens process employs a selective hydrogenation unit (SHU) as a first
step. The role of this unit is to saturate the unstable diolefin hydrocarbons in a hydrogen rich
environment, and react the light mercaptan and sulfide hydrocarbons together. The SRU also
converts exterior olefins to interior olefins which results in a small increase in octane. The mild
operating conditions of the SHU tend to avoid the saturation of monoolefms. After exiting the
SRU, the FCC naphtha is sent to a distillation column which separates the light FCC naphtha
(typically comprising about one fourth of the total cat naphtha) from the heavy naphtha. Because
the light sulfur compounds were reacted together and those compounds no longer fall within the
light cat naphtha boiling range, the light cat naphtha is low in sulfur and can be blended directly
into gasoline. The heavy cat naphtha which is naturally high in sulfur and which also contains
the self-reacted light mercaptans and sulfides from the SHU, is sent to a fixed bed hydrotreater.
The fixed bed hydrotreater contains both cobalt-molybdenum and nickel-molybdenum catalyst.
An important way that Axens avoids recombination reactions is by separating the light sulfur
compounds from the light naphtha and keeping the light naphtha out of the fixed bed
hydrotreater. The desulfurized heavy cat naphtha is blended into the gasoline pool.
If the feed to the Axens Prime G unit is very low in sulfur, a low capital investment
option was available to the refiner by feeding the entire FCC naphtha stream to the hydrotreating
4-10
-------
reactor avoiding the SHU and splitter. This option trades lower capital cost with somewhat
higher octane loss and hydrogen consumption. Because of the low severity of the hydrotreating
reactor (low severity is possible because the lower amount of desulfurization that is occurring),
the amount of octane loss and hydrogen consumption is modest. There are more than 180 Prime
G+ units operating worldwide.
The first step in Exxon's fixed bed Scanfming process is to mildly heat the full FCC
naphtha and pass it through a small reaction vessel which reacts the diolefins to monoolefins.14 15
161? is 19 rpjie ^j p££ nap|lt|ia js t|ien heatec[ further, injected with hydrogen gas and sent to the
fixed bed hydrotreating reactor which is packed with a catalyst developed jointly between Exxon
and Akzo Nobel (now Albermele). If the degree of desulfurization is relatively modest, the
amount of recombination is low and the FCC naphtha is sent to gasoline blending. If, however,
the degree of desulfurization is higher (due to FCC naphtha with a higher sulfur content), then
there likely would be an excessive number of recombination reactions. In this case, Exxon
recommends either one of two different technologies to address the recombination reactions.
One technology is Zeromer. Zeromer is a fixed bed reactor vessel installed after the main fixed
bed hydrotreater reactor that specifically designed to hydrodesulfurize the mercaptan sulfur from
the FCC naphtha without saturating olefms.20 Another technology Exxon developed, in
conjunction with Merichem, is an extractive mercaptan removal technology named Exomer. The
Exomer technology differs from other sulfur extraction technologies in that it is capable of
extracting mercaptans from the entire FCC naphtha pool.21 Like Zeromer, the Exomer
technology would be an add-on technology installed after the Scanfming fixed bed reactor.
There are 16 Scanfming units operating in the U.S.
UOP has licensed two FCC naphtha hydrotreating technologies. When Tier 2 was being
phased-in, UOP was licensing a technology named ISAL developed by INTEVEP S.A.22 23 The
ISAL process is different than the other FCC naphtha hydrotreaters because instead of avoiding
the saturation of olefms as sulfur is being hydrotreated out of FCC naphtha, the ISAL process
completely saturates the olefms. To avoid a large octane loss, the ISAL process separates the
olefin-rich, light cat naphtha from the heavy cat naphtha. The light cat naphtha is treated by an
extractive desulfurization technology such as Merox which does not saturate olefms. Only the
heavy cat naphtha is sent to the ISAL reactor. To offset the octane loss caused by the saturation
of the olefms in the heavy cat naphtha as it is being desulfurized, the ISAL catalyst isomerizes
and conducts some mild cracking and reforming of the heavy cat naphtha. One downside of the
ISAL process is that, due to the complete saturation of olefms, the hydrogen consumption is
higher relative to the selective hydrodesulfurization technologies that avoid saturating olefms.
UOP has since developed and licenses its own FCC naphtha desulfurization technology
named SelectFining.24 SelectFining is a selective hydrodesulfurization technology that seeks to
minimize olefin saturation to minimize both octane loss and hydrogen consumption.
SelectFining treats the full FCC naphtha. The full range FCC naphtha is first sent to a diolefin
saturating reactor before being sent to the SelectFining reactor. SelectFining relies on its catalyst
design to selectively remove sulfur and prevent recombination reactions. UOP recommends a
two-stage reactor setup for high levels of desulfurization.
4-11
-------
The next two FCC naphtha desulfurization technologies, CDTech and S-Zorb do not use
fixed bed reactors, but very different technologies which are also very different from each other.
Each will be discussed separately.
Although the CDTech process is significantly different from the fixed bed hydrotreating
technologies, it still uses the same type of catalyst. The CDTech process utilizes catalytic
distillation.25'26'27 Catalytic distillation is a technology which has been applied for a number of
different purposes. CDTech is currently licensing the technology to produce MTBE and
selective hydrogenation processes, including FCC naphtha desulfurization and benzene
saturation. As the name implies, distillation and desulfurization, via catalyst, take place in the
same vessel. This design feature saves the need to add a separate distillation column sometimes
used with fixed bed hydrotreating. All refineries have a distillation column after the FCC unit
(called the main fractionation column) which separates the FCC naphtha from the most volatile
components (such as liquid petroleum gases), the distillate or diesel (light cycle oil), and the
heavy ends or residual oil. However, if a refiner only wishes to treat a portion of the FCC
naphtha, then a second distillation column would need to be added after the main FCC
fractionation column to separate off the portion of the FCC naphtha which he wishes not to treat.
With the CDTech process, the refiner can choose to treat the entire pool or a portion of the pool,
but choosing to treat a part of the pool can be an option in how the CDTech hardware is applied,
thus negating any need for an additional distillation column.
The most important portion of the CDTech desulfurization process is a set of two
distillation columns loaded with desulfurization catalyst in a packed structure. The first vessel,
called CDHydro, treats the lighter compounds of FCC gasoline and separates the heavier portion
of the FCC naphtha for treatment in the second column. The second column, called CDHDS,
removes the sulfur from the heavier compounds of FCC naphtha. All of the FCC naphtha is fed
to the CDHydro column. The five- and six-carbon petroleum compounds boil off and head up
through the catalyst mounted in the column, along with hydrogen which is also injected in the
bottom of the column. The reactions in this column are unique in that the sulfur in the column is
not hydrotreated to hydrogen sulfide, but they instead are reacted with dienes in the feed to form
thioethers. Their higher boiling temperature causes the thioethers to fall to the bottom of the
column. They join the heavier petroleum compounds at the bottom of the column and are sent to
the CDHDS column. Because the pressure and temperature of the first column is much lower
than conventional hydrotreating, saturation of olefins is reduced to very low levels. The olefm
saturation which does occur is necessary to eliminate diolefins. Thus, little excess hydrogen is
consumed. CDTech offers an option to refiners to put in an additional catalyst section in the
CDHydro column to increase octane. This octane enhancing catalyst isomerizes some of the
olefins, which increases the octane of this stream by about three octane numbers, and few of the
olefins are saturated to degrade this octane gain. The seven-carbon and heavier petroleum
compounds leave the bottom of the CDHydro unit and are fed into the CDHDS column. There,
the heavier compounds head down the column and the lighter compounds head up. Both
sections of the CDHDS column have catalyst loaded into them which serve as hydrotreating
reaction zones. Similar to how hydrogen is fed to the CDHydro column, hydrogen is fed to the
bottom of the CDHDS column.
The temperature and pressure of the CDTech process columns are lower than fixed bed
hydrotreating processes, particularly in the upper section of the distillation column, which is
4-12
-------
where most of the olefins end up. These operating conditions minimize yield and octane loss.
While the CDTech process is very different from conventional hydrotreating, the catalyst used
for removing the sulfur compounds is the same. One important difference between the CDTech
process and conventional hydrotreating is that CDTech mounts its catalyst in a unique support
system, while conventional catalyst is usually dumped into the fixed bed reactor. CDTech has
13 CDHydro/CDHDS desulfurization units in operation in the U.S.
Phillips Petroleum Co. commercialized and licensed an adsorption desulfurization
technology called S-Zorb.28 29 In 2007, Phillips sold the S-Zorb process to SINOPEC. S-Zorb
uses a chemical adsorption process, instead of hydrotreating, as the principal methodology for
the removal of sulfur from FCC naphtha. Adsorption has the benefit of operating at much lower
pressure and temperatures, which lowers operating costs. S-Zorb, uses two separate columns and
is constantly moving an adsorption catalyst from the reactor vessel to the regeneration column,
and back again.30 The untreated FCC naphtha and hydrogen are fed to the reaction vessel where
the catalyst catalytically removes the sulfur from the petroleum compound facilitated by the
hydrogen present in the reactor. The catalyst, which begins to accumulate the removed sulfur, is
transferred over to the regeneration column on a continual basis where the sulfur is removed
from the catalyst using hydrogen as the scavenging compound. Then the hydrogen disulfude is
converted to sulfur dioxide and sent to the sulfur recovery unit. Because the process still relies
upon catalytic processing in the presence of hydrogen, there is some saturation of olefins, with a
commensurate reduction in octane. Through a literature search, we believe that 7 S-Zorb
desulfurization units were originally licensed for Tier 2, but our information sources have
communicated that only 4 units are actually operating today.
We conducted a literature search or in some cases asked vendors to name which
refineries installed their FCC naphtha desulfurization technology to enable compliance with Tier
2. A summary of the total number of units by vendor and technology type is summarized in
Table 4-1.
Table 4-1 Estimated Number of FCC Desulfurization Technologies Installed to comply
with Tier 2 by Vendor Company or Technology
Axens
Prime G
40
Exxon
Scanfming
16
CDTech
15
Sinopec S-
Zorb
4
UOP ISAL
UOP Selectfming
1
FCC Feed
HT
17
No FCC
Unit
14
4.2.3 Meeting a 10-ppm Gasoline Sulfur Standard
To meet a 10-ppm average gasoline sulfur standard, we believe that the primary strategy
that refiners would adopt would be to further reduce the sulfur level of FCC naphtha. There are
three primary reasons why we settled on this as the primary strategy we chose for analyzing the
compliance costs for Tier 3. The first reason is that, even after refiners used hydrotreating to
reduce the sulfur in the FCC naphtha to comply with Tier 2, FCC naphtha is by far the largest
contributor of sulfur to the gasoline pool, by virtue of both its volume and sulfur content. Table
4-2 below summarizes the estimated average volumes and average sulfur levels for the primary
blendstocks typically blended into gasoline for the current Tier 2 situation. By using the
refinery-by-refinery model to model today's situation for the typical refinery, we estimate that
4-13
-------
the FCC naphtha contains about 70 ppm for the typical refinery complying with the 30-ppm Tier
2 sulfur standard and that gasoline blendstock typically contributes to about 40 percent of a
refiner's gasoline pool. Table 4-2 also summarizes the changes in gasoline blendstock sulfur
levels we believe would occur when complying with the proposed 10-ppm gasoline sulfur
standard. Using the refinery-by-refinery model, we project that a 10-ppm gasoline sulfur
standard can be met by a typical refinery by reducing the sulfur level of FCC naphtha from about
70 ppm to 20 ppm. In fact, for virtually all refineries that have an FCC unit, refiners would not
be able to comply with the proposed 10-ppm gasoline sulfur standard without further
desulfurizing the FCC naphtha. The second reason why we believe that refiners would address
the sulfur in the FCC naphtha is because both vendors and refiners have told us that this is the
gasoline blendstock stream that they intend to address. Both vendors and refiners have explained
to us that, for most refineries, FCC naphtha hydrotreaters will already be in place that can be
retrofitted with only a modest capital cost to realize the sulfur reduction needed. Third, further
reducing the sulfur of the FCC naphtha as the means to comply with Tier 3 is supported by other
cost studies. When these studies assessed the costs for further reducing the sulfur levels of
gasoline, they also focused further reducing the sulfur levels of the FCC naphtha. See the
subsection at the end of Chapter 5 discussing these other cost studies.
Table 4-2 Estimated Typical Gasoline Blendstock Volumes and Sulfur Levels after Tier 2
and Complying with a 10-ppm Sulfur Standard
Gasoline
Blendstock
FCC Naphtha
Reformate
Alkylate
Isomerate
Butane
Light Straight Run
Naphtha and
Natural Gas Liquids
Hydrocrackate
Ethanol
Coker Naphtha
Other Gasoline
Blendstocks
Total/Sulfur
Average
30-ppm Tier 2 Gasoline Sulfur
Standard
Volume
(Percent)
37
23
13
3
4
5
3
10
2
1
100
Sulfur (ppm)
72
0.5
5
0.5
10
34
8
5
1
10
30
10 -ppm Gasoline Sulfur Standard
Volume
(Percent)
36
22
13
3
4
5
3
12.5
2
1
100
Sulfur (ppm)
22
0.5
5
0.5
10
1
8
5
1
1
10
Reducing FCC naphtha from 70 ppm to 20 ppm would likely be accomplished in different ways
depending on the desulfurizing technology and configuration used for Tier 2, and whether the
current capital employed for lowering gasoline sulfur is severely taxed or not severely taxed. For
purposes of this discussion, we will discuss the likely steps taken to comply with Tier 3 based on
whether a refiner solely used an FCC pretreater or FCC postreater to comply with Tier 2. While
we provided an example for a typical refinery needing to reduce its FCC naphtha from 70 ppm to
4-14
-------
20 ppm to enable compliance with Tier 2, there are many refineries which are not typical and so
their starting and ending sulfur levels would be different from this example. Despite these
differences, we believe that every refinery could technically comply with a 10-ppm gasoline
sulfur standard. This is because gasoline sulfur is easy to remove - the challenge is to comply
while minimizing the cost of doing so. This challenge is further discussed in Section 4.2.3.5
below which discusses the value of the proposed averaging, banking and trading program.
The one exception is the case where a refinery does not have an FCC unit. Refineries in
this situation would likely already be producing gasoline which is 10 ppm or below. If the
refinery's gasoline is above 10 ppm, then the refiner would need to address one or more of
several different gasoline blendstocks, including light straight run, butane and natural gas liquids.
This is discussed at the end of this section about other gasoline streams.
4.2.3.1 Meeting 10 ppm if Refiners Used an FCC Feed Pretreater to Comply with
Tier 2
If a refiner relied on an FCC pretreater to comply with Tier 2 at a refinery, the refiner
would likely only be able to achieve 10-ppm sulfur gasoline if their FCC pretreater is a high
pressure unit.3132 This is because most refineries which have FCC pretreaters process sour crude
oils and if the unit is a mid or low-pressure unit, the unit pressure would likely be too low to
sufficiently desulfurize the FCC feed. This is likely true even if the refiner added reactor volume
to its existing low or medium pressure FCC pretreater which does cause additional
desulfurization. The problem with the mid and low pressure FCC pretreaters is that they just
cannot remove enough of the sulfur in the gas oil feed to the FCC unit to achieve adequately low
sulfur levels in the FCC naphtha. If a refinery processes moderate to low sulfur crude oil and has
a low to mid-pressure FCC pretreater, however, it may be able to achieve an adequate degree of
desulfurization in the FCC naphtha to enable the refiner to reduce its gasoline sulfur down to 10
ppm. Thus, if a refinery cannot achieve a sufficient level of desulfurization with its current or
revamped FCC pretreater to comply with a 10-ppm gasoline sulfur standard, then the refiner will
have to install a grassroots FCC postreater. Alternatively, refiners in this situation would be in
the best situation to take advantage of the averaging aspect of the averaging, banking and trading
program (ABT). Using the ABT provisions to its advantage, the refiner would achieve the most
desulfurization that it can with its existing FCC pretreater (perhaps 20 ppm sulfur gasoline), and
then would need to purchase credits to demonstrate the remainder of its compliance with the 10-
ppm gasoline sulfur standard. This scenario would avoid the need for a refiner to install an
expensive grassroots FCC postreater.
While they are expensive to install, FCC pretreaters provide important operating cost
advantages over postreaters. An important advantage of FCC pretreating is that it occurs
upstream of the FCC unit and therefore does not jeopardize the octane value of the olefms
produced in the FCC unit. Another advantage of the FCC pretreater is that it tends to increase
the yield of naphtha from the FCC unit which improves operating margins for the refinery with
such a unit. Thus, refiners which are able to use FCC pretreaters to comply with the Tier 3 sulfur
standard would likely yield a further return on any investment made, and offset some or all of the
increased operating costs incurred. Perhaps only 5 refineries have high pressure FCC pretreaters
in the U.S.
4-15
-------
4.2.3.2 Meeting 10 ppm if Refiners Used an FCC Postreater to comply with Tier 2
If a refiner installed an FCC postreater to comply with the Tier 2 gasoline sulfur standard,
there are several considerations about the current configuration of the postreater which would
affect how a refiner would use this unit to comply with a 10-ppm gasoline sulfur standard. The
first issue is what is the degree of desulfurization the postreater is currently facing? It makes
sense to work through several examples to understand the types of revamps and associated
investments that might occur.
For the first example, if the refinery is refining a very sour (high sulfur) crude oil and the
sulfur of the FCC naphtha exiting the FCC unit is 2,400 ppm, the postreater is currently
removing almost 97 percent of the feed sulfur assuming that the sulfur level of the FCC naphtha
exiting postreater is 70 ppm, which is a very high level of desulfurization. When attempting to
achieve further sulfur reduction in the FCC naphtha, the refiner must be concerned about the
increased occurrence of recombination reactions and the potential for much more octane loss and
hydrogen consumption. This refiner would strongly consider adding a second stage, which may
actually reduce the level of recombination reactions and the octane loss currently experienced by
the postreater. Most all the vendors offer a second stage option. In the case of CDTech, they
call the second reactor, added as part of its second stage, a polishing reactor. We contacted the
desulfurization engineer at Sinopec who explained that these units could just be turned up and
that no additional capital investments would be needed. A Conoco-Phillips hydrotreating
specialist we spoke to confirmed that this would be the strategy for their S-Zorb units. Yet one
more option, if the refiner is interested in improving its operating margins such as increased
gasoline production, and has ample capital dollars to spend, the refiner could add an FCC feed
hydrotreater to increase its yield of FCC naphtha, or a mild hydrocracker to increase its
production of low sulfur distillate.
In contrast, if a refiner is processing a very sweet (low sulfur) crude oil, the sulfur level
exiting the FCC unit may be as low as 300 ppm, and under Tier 2, the level of desulfurization
necessary to bring that stream down to 70 ppm is about 81 percent which is a very modest level
of desulfurization. Similarly, a refinery processing a moderately sour crude oil with a medium
pressure FCC feed hydrotreater could be in a similar situation. The refineries in this situation
could have a lot more capacity in their existing postreaters to achieve lower sulfur without
additional capital cost investments. However, many refiners in this situation which invested in
an FCC postreater for Tier 2 may have minimized their capital investments. For example, a
refiner may have avoided the capital and operating cost of a splitter with its postreater by
hydrotreating the full range FCC naphtha. Therefore, the increased severity of the postreater
needed to achieve 20 ppm in the FCC naphtha to meet a 10-ppm gasoline sulfur standard might
create a larger octane loss and higher hydrogen consumption than what the refinery could easily
provide without a significant additional capital investment. In this case, the refiner can invest
some capital in the postreater to minimize the increase in octane loss and hydrogen consumption.
For example a refiner with an Axens unit in this situation could add the SHU and a splitter. A
refiner with a Scanfming unit in this situation wishing to minimize the octane loss and hydrogen
consumption could add a Zeromer or an Exomer unit. Alternatively, if the refiner is processing a
moderately sour crude oil and has a moderate pressure FCC feed hydrotreater, the refinery may
choose instead to revamp the FCC feed hydrotreater for its operational benefits rather than
revamp the postreater.
4-16
-------
The last example of a postreater deserving some discussion is the case where the sulfur
level exiting the FCC unit is 800 ppm. This is probably most typical of a refinery refining a
crude oil containing an average amount of sulfur, or, perhaps a refinery refining a very sour
crude oil but treating the vacuum gas oil with a low pressure FCC feed hydrotreater. The current
FCC naphtha hydrotreater would be achieving about 90 percent desulfurization when producing
FCC naphtha with 80 ppm sulfur. In looking to reduce the FCC naphtha down to 20 ppm to
comply with a 10-ppm sulfur standard, a refiner in this position would not likely consider adding
a second stage. This is because avoiding the increased octane loss and increased hydrogen
consumption for the additional increment of sulfur reduction would probably not justify the
capital costs associated with a second stage. Instead of a second stage, a refiner could revamp
the existing FCC postreater with additional reactor volume, or add capital for addressing
recombination reactions, both likely to be a lot less capital intensive than a second stage. A no
investment option is possible for refiners in this situation, although the increase in octane loss
and hydrogen consumption is likely to be significant.
Perhaps the most important part of an FCC hydrotreater is the catalyst used in the unit.
Due to continuing research, catalysts are constantly being developed which are more active, thus
achieving greater desulfurization at a lower temperature, and minimize octane loss and hydrogen
consumption due to lower olefm saturation. When the Tier 2 naphtha desulfurizers were being
put into service the most recent catalysts were likely used in those units. These catalysts can be
changed out when the postreater is being taken down for regular maintenance, and new and
improved catalysts can be used to improve the desulfurization capacity of the unit. If refiners
indeed need to comply with a 10-ppm gasoline sulfur standard, they would be expected to
upgrade to the most recent catalyst to minimize their costs. Using the most active catalyst
available would reduce the capital cost that would need to be incurred and reduce the hydrogen
consumption and octane loss that would otherwise occur. We are aware of newer lines of
catalysts being marketed by the various vendors. We can confirm that Axens and UOP have
introduced more active catalysts since the catalysts were loaded into the FCC postreaters to
comply with Tier 2, although it is likely that all the vendors now offer improved hydrotreating
catalysts.
4.2.3.3 Desulfurizing Other Blendstocks
A more stringent gasoline sulfur standard could require refiners to have to address other
gasoline streams that are high enough in sulfur to be a concern to the refiners. This is because
without addressing these gasoline streams, the refiner would have to reduce their FCC naphtha
even lower in sulfur resulting in high per gallon costs at the lower sulfur levels. The gasoline
streams that we have identified that could require additional desulfurization include light straight
run naphtha, natural gas liquids and butane.
Light straight run naphtha (LSR) is naturally occurring in the crude oil and is
desulfurized at many refineries before it is sent to an isomerization unit. However, a number of
refineries don't have isomerization units and therefore some or perhaps many of these refineries
may not be treating this stream today. Natural gas liquids (also termed pentanes plus) are
naphtha streams sourced from natural gas wells which are purchased by refiners and blended into
the gasoline pool. Depending on the source of the specific naphtha stream being purchased,
these streams could vary widely in gasoline sulfur, ranging from a few ppm sulfur up to several
4-17
-------
hundred ppm sulfur. Butane is natural occurring in crude oil and butane is also produced by the
FCC unit, and to a lesser degree, hydrocrackers. Refiners separate the butane from these various
streams and then blend it back into their gasoline pool depending on the RVP requirements of the
gasoline market that the refiner is selling their gasoline into.
Refiners have multiple options for addressing the sulfur levels of these various streams.
The LSR and natural gas liquids can be hydrotreated in either the FCC postreaters or the naphtha
hydrotreaters. Because these naphtha streams do not have any olefms, there is essentially no
octane loss and, therefore, hydrogen consumption is lower compared to hydrotreating FCC
naphtha. Another way of treating these streams would be to use caustic extraction to extract the
mercaptan sulfur from these streams. Since only the mercaptans are removed with the extraction
technology, the final sulfur level won't be as low compared to desulfurization using
hydrotreating. Finally, the refiner could choose to simply not purchase the natural gas liquids
and sell the LSR on the open market as opposed to treating these streams. If a refiner decides to
not treat the LSR or natural gas liquids, other refiners with excess capacity in their FCC
postreaters or naphtha hydrotreaters could treat these streams.
Butane is normally treated using mercaptan extraction technologies. We are aware that
some refiners have installed at least a few of these units at their refineries to address butane
sulfur. Because butane is usually relatively low in sulfur to begin with, refiners are likely to only
pursue desulfurizing this gasoline blendstock if their butane is higher than average in sulfur, or if
they are considering producing a very low sulfur gasoline, such as 5 ppm.
In summary, to comply with a 10-ppm gasoline sulfur standard, refiners have a range of
options available to them, most of which involve reducing the sulfur in the FCC naphtha. If a
refinery has a high pressure FCC pretreater, the refiner may be able to just turn up the
hydrotreating severity of that unit. If a refinery has a low or medium pressure FCC pretreater
and no postreater, the refinery would likely need to install a grassroots FCC postreater to comply
with a 10-ppm gasoline sulfur standard, or achieve the most that it can with its current capital
and rely on the ABT program. Refiners with FCC postreaters have multiple options. If a
refinery is short on octane and hydrogen, the refiner is likely to invest in capital to avoid as much
octane loss and hydrogen consumption as possible. However, if the refiner has a lot of excess
octane and hydrogen, the refiner may choose to avoid any capital cost investments or only make
small capital investments and tolerate the higher octane loss and hydrogen consumption by
simply turning up the severity of its current FCC postreater. Refineries with postreaters could
always invest in an FCC pretreater (hydrotreater or mild hydrocracker) to improve its refinery's
margins or to produce more low sulfur diesel fuel. Finally, in blending up their gasoline, some
refiners may still be blending in some produced or purchased gasoline blendstocks that continue
to have high enough sulfur levels which would be a concern when faced with a more stringent
gasoline sulfur standard, and several options exist for addressing the sulfur in these gasoline
blendstocks.
4.2.3.4 Demonstrated Compliance with a 10-ppm Gasoline Sulfur Standard
There are multiple cases today where refiners are complying with 10-ppm or lower
gasoline sulfur programs. The State of California required gasoline sold in the State to meet a
15-ppm gasoline sulfur standard on average and a 20-ppm cap (California gasoline's per-gallon
4-18
-------
sulfur cap dropped to 20 ppm on January 1, 2012). Furthermore, refiners can produce gasoline
which varies in composition, provided that the California Predictive Emissions Model (which,
like EPA's Complex Model, estimates vehicle emissions from fuels of varying composition)
confirms that the proposed fuel formulation meets or exceeds the emissions reduction that would
occur based on the default fuel requirements. California refineries are using the flexibility
provided by the Predictive Model to surpass the prescriptive standards for gasoline sulfur and are
producing gasoline which contains around 10 ppm sulfur on average. They are making this very
low sulfur gasoline despite using Californian and Alaskan crude oils which are poorer quality
than most other crude oils being used in the U.S. today. Thus, the experience in California
demonstrates that commercial technologies already exist to permit refiners to produce very low
sulfur gasoline.
Japan currently has a 10-ppm gasoline sulfur cap that took effect January 2008. Europe
also has a 10-ppm sulfur cap that has been adopted by the 30 Member States that comprise the
European Union (EU) and the European Free Trade Association (EFTA) as well as Albania and
Bosnia-Herzegovina. Under a 10-ppm cap standard, the gasoline sulfur level likely averages
about 5 ppm. Although gasoline in Japan and Europe is made from different crude oil sources
and much of the heavier ends are cut into diesel fuel, these international fuel programs (along
with California) provide evidence that advanced gasoline desulfurization technologies have been
deployed and are readily available enable compliance with the proposed Tier 3 fuel program.
4.2.3.5 Improved Feasibility with the Proposed ABT Provisions
The averaging, banking and trading (ABT) and small refiner and small volume refinery
aspects of the proposed Tier 3 gasoline sulfur program would ease the feasibility of compliance
with the program. To make the point, it is useful to first understand compliance if the ABT and
small refiner and small volume refinery provisions did not exist. Without these provisions, all
refineries would have to comply with the 10-ppm gasoline sulfur standard by January 1, 2017.
In the approximate 4 years after finalizing this rulemaking, most refiners would have to make
capital investments in their refineries to enable compliance with the 10-ppm gasoline sulfur
standard. These investments include revamped FCC pretreaters and postreaters, and the
installation of grassroots FCC postreaters. As described above, reaching 10 ppm sulfur in the
gasoline pool is attainable by each refinery. However, refiners assess the economic feasibility of
their refineries differently depending on past and expected future economic performance. They
therefore have different tolerances for making capital investments and absorbing increased
operating costs. This is particularly true during a period of time in which gasoline demand is
projected to be flat and renewable fuel blending is expected to increase. Refiners who own small
refineries are concerned about the higher per-barrel costs for the capital installed at those small
refineries.
The small refiner and small volume refinery provisions will delay compliance for these
entities until January 1, 2020. Small refiners need more time because they have smaller
engineering staffs that they can dedicate to oversee the necessary refinery changes, thus they are
more likely to complete the necessary changes to their refineries later than large refiners. Also,
the delay allows the small refiners to experience improved margins for a couple of years when
other larger refineries are complying with the gasoline sulfur standards.
4-19
-------
The banking provisions of the ABT program effectively phase in the sulfur standard over
six years starting in 2014 through the end of 2019. The phase-in allows refiners to stagger their
investments to their economic advantage. Refineries which are expected to incur the lowest
costs for achieving lower gasoline sulfur levels can comply early and earn sulfur credits. These
credits can then be used to demonstrate compliance starting in 2017 by the refineries which are
expected to incur higher costs for reducing their gasoline sulfur levels allowing those refineries
to delay investments for lowering their gasoline sulfur. This phasing-in of the gasoline sulfur
standard will help spread out the preliminary design demands on the vendor companies which
license the desulfurization technology to refiners, spread out the detailed design demands on the
engineering companies which provide that service to refiners, spread out the permitting demands
on the states which must provide environmental permits to refiners, and spread out the demands
on the fabrication shops which construct the reactors and other major hardware which must be
installed at refineries to realize the gasoline sulfur reductions. For more on how the proposed
ABT provisions are expected to help with lead time, refer to Section 4.3.
Finally, the averaging provisions of the ABT program will provide additional flexibility
and help to reduce the costs of the gasoline sulfur program. The averaging provisions will allow
refiners to reduce the gasoline sulfur levels to under 10 ppm at their lower cost refineries to show
compliance or help to show compliance at higher cost or financially challenged refineries.
4.2.3.6 Implications of an Average Gasoline Sulfur Standard Less than 10 ppm
While there may be emissions motivations for further reducing gasoline sulfur levels,
there are practical reasons for proposing a 10-ppm annual average sulfur standard instead of a
more stringent standard, e.g., 5 ppm. The lower the sulfur standard, the more costly it is for
refiners to achieve the lower sulfur standard. We identified several reasons why the costs
increase so much for more deeply desulfurizing the gasoline pool.
As desulfurization severity increases, the operating and capital costs associated with
desulfurizing FCC naphtha also increases. FCC naphtha is very rich in high-octane olefins. As
the severity of desulfurization increases, more olefins are saturated, further sacrificing the octane
value of this stream and further increasing hydrogen consumption. Also, as desulfurization
severity increases, there is an increase in the amount of the removed sulfur (in the form of
hydrogen sulfide) which recombines with the olefins in the FCC naphtha, thus offsetting the
principal desulfurization reactions. There are means to deal with the recombination reactions;
however, this probably means even greater capital investments. For example, the most
expensive capital investment for an FCC postreater is a two stage desulfurization unit. A sulfur
standard less than 10 ppm would likely require more refiners to invest in a second stage for their
FCC postreater.
There are several other reasons which further increases the desulfurization cost for a
gasoline sulfur standard less than 10 ppm beyond the higher FCC postreater cost. Per Table 4-2,
other refinery streams contain a very modest amount of sulfur, yet a 5-ppm sulfur standard would
likely require desulfurization of some of these streams. For example, we believe that to comply
with a 5-ppm gasoline sulfur standard, most refiners would need to treat the butane blended into
the gasoline pool. Because refineries have different sulfur levels in their non-FCC streams based
on their feedstock sulfur levels and their configurations, those with higher sulfur levels in other
4-20
-------
refinery streams may have to desulfurize additional streams besides butane. Each additional
individual gasoline stream that requires desulfurization is incrementally a lot more expensive
than addressing the sulfur from the FCC unit because the amount of sulfur reduction is a lot
lower, but the capital costs are higher on a per-barrel basis for lower volume gasoline blendstock
streams. Furthermore, desulfurizing gasoline down to 5 ppm essentially removes the flexibility
offered by the 10-ppm gasoline sulfur standard with ABT program. Each U.S. refinery is in a
different position today, both technically and financially, relative to the other refineries. In
general, they are configured to handle the different crude oils they process and turn their crude
oil slate into a widely varying product slate to match their available markets. Those processing
heavier, sour crudes would have a more challenging time reducing gasoline sulfur under the
proposed Tier 3 program. Also, U.S. refineries vary greatly in size (atmospheric crude capacities
range from less than 5,000 to more than 500,000 barrels per day) and thus have different
economies of scale for adding capital to their refineries. As such, it is much easier for some
refineries to get their sulfur levels below 10 ppm than for others to reach 10 ppm. This allows
the ABT program to be used to reduce the cost of the proposed gasoline sulfur standard. If the
gasoline sulfur standard were to be 5 ppm, this would essentially end the ability of the refiners to
average sulfur reductions across their refineries thus significantly increasing the costs while
significantly reducing the desulfurization flexibility.
Going lower than 10 ppm would cause control costs to quickly escalate as more
challenged refineries would be forced into much larger investments. Our cost estimates for 5
ppm versus 10 ppm with averaging bears this out. We estimate the cost for a 10-ppm gasoline
sulfur standard (assuming intra-company credit trading) to be 0.89 eVgal compared to 1.38 eVgal
for the 5-ppm standard. The cost per sulfur reduction for the 10-ppm average standard is 0.89
eVgal for the 20 ppm sulfur reduction from Tier 2, or 0.045 eVppm-gal. The cost per sulfur
reduction for the 5-ppm standard is 0.49 eVgal for the 5-ppm sulfur difference from the 10-ppm
average standard, or 0.098 e7ppm-gal, which is over 2 times higher. Therefore, we believe that
an annual average standard of 10 ppm at the refinery gate with an ABT program appears to be
the point which properly balances feasibility with costs.
In much of Europe and Japan, the gasoline sulfur level is capped at 10 ppm. We,
however, are not considering a 10-ppm cap for the U.S. The U.S. gasoline distribution system
poses contamination challenges that make it difficult to set and enforce tight downstream per-
gallon sulfur standards. The U.S. gasoline distribution system poses contamination challenges
that make it difficult to set and enforce tight downstream sulfur standards. In Europe, Japan, and
California, finished petroleum products are generally shipped short distances directly from the
refinery to the terminal with limited susceptibility to contamination. The U.S. has the longest
and most complex gasoline distribution system in the world, making it harder to control sulfur
contamination than in other countries. Petroleum products are shipped long distances through
multi-product pipelines. Further, gasoline goes through the same pipelines and terminals back-
to-back with jet fuel (containing up to 2,000 ppm sulfur). Products are often in the custody of a
number of separate companies before reaching the terminal. This system is very effective at
delivering petroleum products to the bulk of the country, but pipeline transport inevitably
introduces the potential for sulfur contamination of the gasoline being shipped through pipelines.
Gasoline additives, needed to provide critical fuel performance characteristics (e.g., corrosion
control, demulsifiers), also contain varying levels of sulfur which contribute to the overall sulfur
content of gasoline. Therefore, we are proposing a 10-ppm average sulfur standard coupled with
4-21
-------
higher per gallon caps at both the refinery gate and at all points downstream, as currently exists
under the Tier 2 program. We believe this is the most prudent approach for lowering in-use
sulfur while maintaining flexibility considering cost and other factors. These per-gallon caps are
important in the context of an average sulfur standard to provide an upper limit on the sulfur
concentration that vehicles must be designed to tolerate. Since there are many opportunities for
sulfur to be introduced into gasoline downstream of the refinery, these caps also limit
downstream sulfur contamination and enable the enforcement of the gasoline sulfur standard in-
use. For more on our consideration of downstream caps, refer to Section 4.2.4.2.
4.2.4 Challenges with Lowering Today's Sulfur Caps
4.2.4.1 Impacts of Lowering the 80-ppm Refinery Cap
We considered lowering the 80-ppm cap standard that applies to refiners under the Tier 2
program. If we were to lower the cap standard, we analyzed lowering it to two different possible
sulfur levels, either 50 ppm or 20 ppm. If we lowered the refinery cap standard to 20 ppm, then
the averaging aspect of the ABT program would essentially not be available to refiners. That is
because, under a 20-ppm cap standard, we estimate that refiners would average about 10 ppm
sulfur. Thus, the compliance scenario if the cap standard were 20 ppm would essentially be the
same as the non-ABT case we analyzed. In this case, refiners would not have much of the
flexibility offered by the ABT program.
If the cap standard were to be lowered to 50 ppm, the final compliance scenario under the
Tier 3 fuels program would be somewhere between the ABT scenario that we analyzed and the
non-ABT scenario that we analyzed (probably much closer to the ABT case). According to EPA
batch data, there were 20 refineries that averaged between 40 and 80 ppm sulfur during 2009.
These refineries are using credits to show compliance with the Tier 2 30-ppm gasoline sulfur
standard. If the 80-ppm cap were to be reduced to 50 ppm, those refineries that were averaging
over 40 ppm would be forced to reduce their sulfur levels below the cap even if their financial
situation is more tenable compared to other refineries. However, even if the cap standard were to
remain at 80 ppm, most of the 20 refineries that averaged between 40 and 80 ppm under Tier 2
would have to lower their sulfur anyways because of the stringency of the proposed 10-ppm
sulfur standard. There would not be sufficient credits available to allow most of those refineries
remain at high gasoline sulfur levels. Our cost analysis, which assumes intra-company credit
trading, projects that only one refinery would remain just above 40 ppm when the fuels program
is fully phased in. For more on our cost analysis, refer to Chapter 5 of the draft RIA.
Another way that a more stringent cap would affect refiners would be to restrict the
ability of refiners to process high sulfur FCC naphtha when there is a short term shutdown of the
FCC postreater. If the FCC postreater goes down, the refinery would likely continue operating
the FCC unit and store up the high sulfur FCC naphtha. Since the FCC naphtha is too high in
sulfur to blend directly with gasoline, the refinery would have to either sell the material to other
refiners, or hydrotreat the stored up FCC naphtha along with the ongoing production of high
sulfur FCC naphtha once the FCC postreater was back online. If a stringent cap were in place,
the refiner would have little room for short term production of higher sulfur gasoline if it was
feeding a larger than normal quantity (stored and new production) of FCC naphtha to the FCC
postreater. Without this flexibility, the refiner may have to oversize the FCC postreater and FCC
4-22
-------
naphtha storage to ensure that, regardless of the higher feed volume needed to process the stored
material, the FCC naphtha desulfurization unit could continue to desulfurize the FCC naphtha
down to the required sulfur level that would result in 10 ppm sulfur in the gasoline pool. If the
cap were to be lowered, a 50-ppm cap standard would clearly continue to provide refiners with
some flexibility while a 20-ppm cap would not. Even if refiners planned to tolerate some higher
sulfur batches when hydrotreating stored FCC naphtha, it could not tolerate much volume of
higher sulfur batched because of the need to average 10 ppm over the calendar year. If such
outages happen very infrequently, then a small amount of credits could regularly be banked over
time that would allow for some longer term higher sulfur batches of gasoline as the stored FCC
naphtha was being hydrotreated. Alternatively, the averaging of sulfur credits would help
refiners with FCC naphtha hydrotreater outages. Thus, the flexibility of the ABT program
coupled with a higher cap standard would provide refiners with some flexibility to handle FCC
unit outages.
4.2.4.2 Downstream Sulfur Caps
The feasibility of complying with a downstream sulfur cap is dependent on the
differential between the refinery/importer gate sulfur cap and the downstream cap. This
differential must provide sufficient flexibility for worst-case situations when the potential
sources of sulfur addition downstream of the refinery/importer compound in a single batch of
gasoline that was introduced into the system at the refinery/importer gate sulfur cap.
We are proposing two potential options for the per-gallon downstream sulfur cap. Under
the first option, we are proposing to maintain the current 95-ppm downstream sulfur cap. This
option is associated with the proposed maintenance of the current 80-ppm refinery/importer gate
sulfur cap and is reflected in the draft regulatory text. Under the second option, we are
proposing that the downstream sulfur cap would be reduced to 65 ppm. This option is associated
with the proposed reduction in the refinery/importer gate sulfur cap to 50 ppm. Under both of
these options, we would be maintaining the current 15-ppm differential between the
refinery/importer gate sulfur cap and the downstream sulfur cap.
We are also requesting comment on the potential implementation of a downstream sulfur
cap as low as 25 ppm. This scenario is associated with a reduction of the refinery/importer gate
sulfur cap to as low as 20 ppm. Under this scenario, the differential between the
refinery/importer gate sulfur cap and the downstream sulfur cap might be a little as 5 ppm.
Under all of these potential approaches, the downstream sulfur cap would apply at all
locations downstream of the refinery or importer gate including the gasoline produced by
transmix processors and after the use of additives. The potential sources of sulfur addition
downstream of the refinery/importer gate and issues associated with the feasibility of meeting the
downstream sulfur caps under consideration are discussed in the following subsections.
4.2.4.2.1 Sulfur Addition Downstream of the Refinery and Importer Gate
The sulfur content of gasoline can increase downstream of the refinery/importer due to
contamination during distribution, the use of additives, and the disposition of transmix generated
during distribution.
4-23
-------
A small amount of sulfur contamination takes place during distribution as a result of the
shipment of gasoline over long distances by pipeline and other modes due to the sharing of the
same distribution assets with other higher-sulfur petroleum products, e.g., jet fuel. Steps can be
taken to limit sulfur contamination. However, it is an unavoidable feature of the efficient multi-
product distribution system in the U.S. We estimate that sulfur contamination of gasoline can be
limited to a worst case maximum of 3 or 4 ppm in the future, even for the most involved and
long-distance distribution pathways.
There are currently no direct regulatory controls on the sulfur content of gasoline
additives. The contribution to the sulfur content of finished gasoline from gasoline additives is
accommodated in the differential between the refinery gate and downstream sulfur caps. The
functional components of some gasoline additives such as silver corrosion inhibitors and
demulsifiers are inherently high in sulfur content. However, the contribution to the overall sulfur
content of the finished fuel is very limited. For example, silver corrosion inhibitors can contain
as much as 30 percent sulfur but because of very low treatment rates can add only 0.17 ppm to
the sulfur content of the finished fuel.33 At seldom used highest treatment rates, the use of
gasoline additives upstream of the consumer has the potential to add ~1 ppm to the sulfur content
of the finished fuel. Aftermarket additives that are added directly into the vehicle fuel tank also
have the potential to increase gasoline sulfur content. One particular aftermarket performance
and anti-wear additive can contribute ~2 ppm sulfur to the treated fuel.B34
Transmix is a necessary byproduct of the multi-product refined product pipeline
distribution system. Batches of different products are shipped in sequence in pipelines without
any physical barrier between the batches. Transmix is produced when the mixture at the
interface between two adjacent products cannot be cut into either batch. Transmix typically
accumulates at the end of pipeline systems far from refineries. There are two methods of
disposing of transmix. Most transmix is sent to transmix processing facilities for separation into
salable distillate and gasoline products through use of a simple distillation tower.
The other means of transmix disposal is for pipeline operators to blend small quantities
directly into batches of gasoline during shipping. This typically takes place at remote pipeline
locations where small volumes of transmix accumulate that would be difficult to consolidate and
ship to transmix processors. Pipeline operators that blend transmix into the gasoline in their
systems must ensure that the resulting gasoline meets all fuel quality specifications and the
endpoint of the blended gasoline does not exceed 437 °F.C35 This practice currently can add as
much as 3 to 5 ppm to the sulfur content of gasoline although we believe that the contribution is
typically less.
Transmix processing facilities do not handle sufficient volumes to support the installation
of currently-available desulfurization units. Therefore, the sulfur content of the products they
produce is predominantly governed by the sulfur content of the transmix they receive. In many
cases, transmix contains jet fuel which can have a sulfur content as high as 3,000 ppm. Due to
B Aftermarket additives are defined as additives sold to vehicle operators for direct addition to vehicle fuel tanks.
c 437 F is the maximum endpoint allowed for gasoline in the ASTM International specification for gasoline in
ASTMD4814.
4-24
-------
the overlapping distillation characteristics of jet fuel and gasoline, it is unavoidable that some jet
fuel in transmix will be present in the gasoline produced by transmix processors.
Transmix processors produce ~0.1 percent of all gasoline consumed in the U.S. The
small volume of transmix-derived gasoline along with the fact that such gasoline is typically
mixed with other gasoline before delivery to the end user, substantially limits the potential
impact on gasoline sulfur levels. Furthermore, data provided by the largest operator of transmix
processing facilities, shown in Figure 4-2, indicates that relatively few batches of the gasoline
they produce approach 80 ppm sulfur.36 Most batches are approximately 10 ppm above the
current 30-ppm refinery sulfur average. We anticipate that this 10-ppm differential would likely
continue under the proposed 10-ppm refinery average sulfur standard.
90
70
60
20
10
•
*
: .
A * * A*
;-*- T^
A A . *A4 i ^ •• A*
• • ** > £^ A4* A* «F
J .A AA ". 4SV 44 A... A»^ ^ A^
»lndianol3
• St. Louis
A Deepwater
X Greensboro
V Dorsey
Colton
10/14/2009 01/22/2010 05/02/2010 08/10/2010 11/18/2010 02/26/2011 06/06/2011 09/14/2011
Figure 4-2 Kinder Morgan Transmix Gasoline Product Sulfur Levels
4.2.4.2.2 Maintaining the Current 15-ppm Differential Between the Refinery
/Importer Gate and Downstream Sulfur Caps
Under both of the co-proposals for a downstream sulfur cap (95 ppm and 65 ppm), we
would be maintaining the current 15-ppm differential between the refinery/importer gate sulfur
cap and the downstream sulfur cap.
4-25
-------
The current 15-ppm differential was established under the Tier 2 program to
accommodate the sulfur contamination during distribution, the sulfur contribution from transmix
blending by pipeline operators, the sulfur contribution from the use of additives, and to enable
compliant gasoline to be produced by transmix processors. Transmix processors need to produce
gasoline sufficiently below the downstream sulfur cap to accommodate the addition of sulfur
from the use of additives and contamination during further distribution. Experience under the
Tier 2 program has shown that a 15-ppm differential is sufficient for downstream parties to
ensure compliance with the downstream sulfur cap
Our co-proposal to maintain the current 95-ppm downstream sulfur cap with an 80-ppm
refinery/importer gate sulfur cap represents no change from current requirements. As a result,
there would be no increased difficulty or additional costs associated with satisfying a 95-ppm
downstream sulfur cap beyond those that were already incurred under the Tier 2 program.
Our co-proposal to implement a 65-ppm downstream sulfur cap with a 50-ppm refinery
gate sulfur cap would also maintain the current 15-ppm differential between these sulfur caps
under the Tier 2 program. Since it is this differential that determines the difficulty in complying
with the downstream sulfur cap, we expect that there would be no operational changes and
additional costs for downstream parties associated with satisfying a 65-ppm cap downstream
sulfur cap beyond those that were already incurred under the Tier 2 program to comply with the
current 95-ppm downstream sulfur cap.
Under both of the co-proposals, the reduction in the refinery average sulfur standard may
make it somewhat easier to comply with the downstream sulfur cap given that most gasoline
produced would be at or near 10 ppm sulfur.
4.2.4.2.3 Potential Reduction in the Differential Between the Refinery/Importer
Gate and Downstream Sulfur Caps
We requested comment on the potential implementation of a refinery gate sulfur cap as
low as 20 ppm and a corresponding downstream sulfur cap as low as 25 ppm. This was driven
by vehicle manufacturers concerns about the potential impacts on emissions performance if
vehicles are exposed to gasoline above the proposed 10-ppm refinery average standard.37 As
discussed in Sections 1.2 and 5.2 of this draft RIA, the vehicle emissions benefits associated with
today's proposal are driven by the proposed reduction in the average sulfur content of gasoline
from 30 to 10 ppm. We believe that the potential benefits from further reductions in the sulfur
caps would be minimal. However, further reductions in the sulfur caps could have negative
impacts on refiners and downstream parties.
The potential impacts on refiners and additional costs associated with a lower
refinery/importer gate sulfur cap discussed in Section 4.2.4.1. A reduction in the differential
between the refinery gate sulfur standard and the downstream sulfur standard could also result in
negative impacts and additional costs to downstream entities. Reducing the current 15-ppm
differential between the refinery/importer gate and downstream sulfur caps could limit the ability
of transmix processors to continue to produce finished gasoline, limit the ability of pipeline
operators to continue to blend transmix in gasoline, and potentially require the direct regulation
4-26
-------
of gasoline additive sulfur content which might cause certain gasoline additives to be removed
from the market.
Some gasoline additive manufactures relate that it would not be technically possible to
reformulate their additives to meet a lower sulfur cap. Hence, the implementation of a sulfur cap
for gasoline additives could result in the withdrawal of some necessary and cost-effective
gasoline additives (e.g., corrosion inhibitors, and demulsifiers) from the market. Other additive
manufactures related that there would be significant costs in reformulating their additives to meet
a lower sulfur cap. Some additive manufactures related that they could not justify the cost of
reformulation and would need to cease manufacture.
Since gasoline additives may add as much as 3 ppm to the sulfur content of the finished
fuel, allowing for further sulfur contamination during distribution and for test variability means
that transmix processors must produce gasoline about 5 ppm below the downstream sulfur cap.
The sulfur levels in the transmix that processors must cope with would be reduced due to the
proposed reduction the gasoline sulfur requirements for refiners/importers. However, the
continued presence of high-sulfur jet fuel in transmix would continue to significantly influence
the sulfur content of the gasoline produced by transmix processors. Given these considerations,
a reduction in the differential between the refinery gate sulfur cap and the downstream sulfur cap
might require that the majority of the gasoline produced by transmix processors to be
desulfurized, something that is cost-prohibitive to do at transmix processing facilities today.
Other options for dealing with transmix, however, are just as impractical, including shipping it
back to refineries for reprocessing. Refiners are typically averse to accepting transmix into their
facilities for reprocessing due to technical, logistical, and economic constraints. In addition,
transmix would typically need to be shipped long distances from the ends of the product
distribution by truck to reach a refinery.
If pipeline operators were further limited in their ability to blend small amounts of
transmix into gasoline due to a reduction in the differential between the refinery/importer gate
and downstream sulfur caps, they could be compelled to install additional transmix storage and
shipping facilities at numerous remote locations to facilitate the shipment of small volumes of
transmix to transmix processors by truck. One major pipeline operator reported that it has over
100 locations on its system where transmix can be injected into gasoline, some of which do not
have tank truck access. Thus, a reduction in the differential between the refinery/importer gate
and downstream sulfur caps could result in substantial additional costs and potential changes to
transmix operating practices for the pipeline operators.
4.3 Lead Time Assessment
4.3.1 Engineering and Construction Analysis
Given the complexity of gasoline refining, numerous planning and action steps would be
required for refiners to complete the refinery changes needed to comply with the proposed Tier 3
sulfur standards. The steps required to implement these changes include: the completion of
scoping studies, financing, process design for new or revamped refinery units or subunits,
permitting, detailed engineering based upon the process design, field construction of the gasoline
4-27
-------
sulfur reduction facilities, and start-up and shakedown of the newly installed desulfurization
equipment.
We conducted a more thorough lead time analysis in which we sequenced the estimated
time to complete scoping studies, process design, permitting, detailed engineering, field
construction, and start-up and shakedown in advance of production based upon the methodology
used in our recent gasoline and diesel rules.
For the proposed Tier 3 gasoline sulfur program, we estimated refinery lead times
required for two general types of refinery projects: the construction of new grassroots FCC
postreaters and the revamp of existing pre and postreaters. For each refinery project, we
estimated lead times for scoping studies, process design, permitting, detailed engineering, field
construction, and start-up and shakedown. Estimated required lead times for scoping studies are
six months. Process design ranged from six months for desulfurization equipment revamping to
nine months for a grassroots postreater. Based on discussions with refiners, a review of the
permitting experience for Tier 2 and our current analysis, we estimate that permitting for
desulfurization equipment revamping and the construction of a grassroots postreater would take
9 months. However, we estimate the overall lead-times for Tier-3-related revamps to be
considerably shorter, as described below. The estimates for permitting time are consistent with
those of EPA's Office of Air Quality Planning and Standards (OAQPS) and our regional offices,
both of which have engaged in extensive dialog with potentially affected parties. A discussion of
the permitting implications of Tier 3 is contained in Section V.B of the preamble. Detailed
engineering efforts were estimated to require six months for desulfurization equipment
revamping and nine months for grassroots postreaters. Field construction was estimated to
require six months for revamped pre-and postreaters and 12 months for grassroots postreaters.
Start-up and shakedown processes were estimated to require six months for revamped FCC
treaters and 9 months for grassroots postreaters. There is some degree of overlap among each of
these steps as shown in Table 4-3.
To allow refiners to complete all these different steps and comply with the 10 ppm
average gasoline sulfur standard, assuming the Tier 3 proposal were to be finalized by the end of
2013, we would be providing three years of lead time. In addition to the three years of lead time,
the proposed rulemaking also provides additional flexibility provided by the ABT program,
small refinery delays, and hardship provisions. To support this timeline, we conducted several
analyses of the expected refinery lead time requirements associated with the proposed Tier 3
standards and found that refinery operators would have more than adequate time to implement
the required refinery charges. A justification for proposed timeline appears below.
Complying with Tier 3 is expected to involve some grassroots (new) FCC postreaters, but
mostly we believe that refiners will revamp existing FCC postreaters. Revamping of existing
FCC postreaters can be accomplished in approximately 2 years, or less (See Table 4-3)
Grassroots FCC postreaters are expected to require on average about three-years to install and
start-up (See Table 4-3). In comparision to FCC pretreaters, hydrocrackers and distillate
hydrotreaters, FCC postreaters are much less costly, low pressure units that take less time to
scope out, require shorter lead times for ordering the equipment, and less time to install.
Furthermore, the grassroots FCC postreaters to be installed for Tier 3 are expected to be in a
moderate to light desulfurization service because the refineries they will be installed in will
4-28
-------
already be complying with Tier 2 using an FCC pretreater. FCC naphtha from a refinery with an
FCC pretreater is expected to only contain about 100 ppm sulfur. To comply with Tier 3,
refiners installing these grassroots FCC postreaters would only need to desulfurize the FCC
naphtha down to 25 ppm (about a 75% reduction). In comparison, a single-stage FCC postreater
would have to desulfurze FCC naphtha from as high as 2400 ppm sulfur down to 25 ppm, a 99%
sulfur reduction. The more moderate desulfurization service of the grassroots FCC postreaters
needed to comply with Tier 3 would be expected to streamline the scoping and design work.
Table 4-3 Anticipated Compliance Timelines
It is useful to compare the proposed lead time for Tier 3 to what was provided for Tier 2.
In the case of the Tier 2 standard, we provided a three-year lead time along with an ABT
program and other flexibilities to ease compliance. Refiners, though, commented that the three
year timeline that we provided was not enough time. For the Tier 2 analysis, we assumed that
refiners would solely install low-pressure FCC postreaters, which we believe could be scoped
out, designed, installed and started up within a 3 year time period. However, many refiners
complied with Tier 2 by installing high-pressure FCC pretreaters which require long lead times
for the procurement of the required equipment. Furthermore, those refiners that did not install
high-pressure FCC pretreaters instead installed grassroots FCC postreaters, many of which were
designed for severe desulfurization service. An additional difference between Tier 3 and Tier 2
is that for Tier 3 we expect the installation of only 16 grassroots units, along with many revamps,
4-29
-------
but for Tier 2 virtually all refiners installed both grassroots FCC pretreaters and postreaters. The
demands on the desulfurization vendors for scoping studies, and on the E & C industry for
design and construction, and on the refiners to train their operations staff and start up the new
units, was a lot greater for Tier 2 than what we would expect for Tier 3. The total estimated
investment cost for Tier 2 versus Tier 3 also highlights the difference in investment demands
The total investment for Tier-2 desulfurization processing units was estimated to be about
$6.1 billion, while the total investment for Tier-3 desulfurization processing units is estimated to
be about $2.1 billion. This simple comparison indicates that the proposed Tier 3 lead time
should be adequate for refineries to obtain necessary permits, secure engineering and
construction (E&C) resources, install new desulfurization equipment and make all necessary
retrofits to meet the proposed sulfur standards.
We assessed the permitting situation in more detail working in conjunction with the
Office of Air Quality, Planning and Standards (OAQPS). On a refinery-by-refinery basis, we
provided OAQPS estimates of the additional heating demands for the new and revamped units
per the desulfurization vendor submissions. OAQPS was able to project which refineries would
likely trigger NOx, particulate matter and greenhouse gas emission permitting limits, which
would likely lengthen the permitting process as refiners would need to offset the projected
emission increases. As it turns out, only 2 of the 16 refineries which are projected to install
grassroots units were projected to exceed particular permitting limits, and these solely did so
based on the most conservative assumption that each would produce all the additional hydrogen
on site using hydrogen plants (as opposed to using existing reforming capacity) and produce the
electricity on site, to satisfy the needs of the new desulfurization equipment. When we provided
a second heat demand estimate to OAQPS which assumes that refiners purchase their hydrogen
and electricity from third parties, none of the refineries which we projected would install
grassroots units was projected to have emission increases which would require offsets. Thus,
many of the grassroots units that we project would be installed may end up with a streamlined
permitting processe.
The various flexibilities that the proposed Tier 3 rule provides to refiners provide refiners
additional time for complying. These flexibilities include the ABT program, the small refiner
delay provisions and the hardship provisions. The ABT program allows a refiner, either within
its own company or by purchasing credits on the open market, to delay higher investment cost
investments, such as the investments in grassroots FCC postreaters, which would provide
additional lead time for installing these units. This would occur if refiners would reduce the
sulfur levels of their gasoline through operational changes or revamps of their existing FCC
pretreaters and postreaters when the ABT Program begins in 2014. Potentially every refinery
with either an FCC pretreater or an FCC postreater may be capable of generating early credits.
Furthermore, we project that 66 refineries would revamp their existing FCC postreaters to
comply with Tier 3. Since revamps can be completed within two years or less, these refiners
could potentially begin generating early credits during 2016, or before if refiners begin each of
these revamps in early 2014. During the period between 2014 and 2017, these refineries which
reduce their gasoline sulfur levels below that required by Tier 2 would generate credits.
Refineries with higher cost capital investments, such as the grassroots FCC postreaters, could
then delay making those investments through the purchase of credits. We estimate that sufficient
credits could be generated early to allow many refineries to delay compliance until as late as
4-30
-------
2020. The quantitative early credit analysis that we conducted showed that if refiners with an
existing pretreater or postreater would generate early credits by lowering their gasoline sulfur
down to 20 ppm starting in 2014 and if revamps were started up in 2016, one year before the
program start date, that almost 6 times more credits would be available to offset the early credit
demand by the refiners installing grassroots postreater units, assuming that they start up those
units in 2018. Even if all grassroots postreaters were assumed to not start up until 2020, there
would be almost 4 times more early credits available to those refiners installing grassroots
postreaters assuming that the same early credit generation scenario would occur
Additional flexibility is also provided by the small refineries provisions which delays
compliance for the refineries which refine less than a net of 75,000 barrels of crude oil per day
until 2020. Three of the 16 FCC postreater grassroots units that we project will be installed
would be by small refineries. However, small refineries could also decide to comply early and
generate credits starting as early as 2014.
As in previous fuel programs, we are proposing hardship provisions to accommodate a
refiner's inability to comply with the proposed standard at the start of the Tier 3 program, and to
deal with unforeseen circumstances that may occur at any point during the program. These
provisions would be available to all refiners, small and non-small, though relief would be granted
on a case-by-case basis following a showing of certain requirements; primarily that compliance
through the use of credits was not feasible. We are proposing that any hardship waiver would
not be a total waiver of compliance; rather, a hardship waiver would be short-term relief that
would allow a refiner facing a hardship situation to, for example, receive additional time to
comply. This hardship provision would allow a refiner to seek a delay in the case that there was
insufficient time to comply.
Finally, we believe that in reality, less leadtime than shown in Table 4-3would actually be
necessary. We held discussions with many refiners during most of 2011, and so they have been
well aware of Tier 3 and are familiar with the likely requirements. During our subsequent
discussions with technology vendors and engineering firms, they explained to us that many
refiners have already initiated, and by now, likely completed their scoping studies. Thus, actual
time needed for designing, installing and starting of new desulfurization equipment for Tier 3
times would even be less than what we projected because many refineries may have already
completed required scoping studies in anticipation of the Tier-3 standards. Moreover, lead times
for those refineries that have yet to start the scoping process can also be expected to decrease,
since fewer refineries will be competing for the services of the desulfurization vendors.
4.3.2 Permitting Analysis
Our analysis found that GHG emission increases were the most common reason that
Prevention of Significant Deterioration (PSD) applicability would be triggered, followed by NOx
emissions. Specifically, 19 refineries appeared likely to have significant emissions for one or
more pollutants and thus would trigger major source New Source Review (NSR). Of these 19
refineries, 13 refineries would need permits for NAAQS-related pollutants.
4-31
-------
With respect to NAAQS-related pollutants, 6 of theselS refineries were predicted to
require both PSD and Nonattainment NSR permits. Of the remaining 7 refineries, 2 required
only a Nonattainment NSR permit while the remaining 5 refineries required a PSD permit.
In comparison, for the Tier 2 program, EPA expected the need for NAAQS-related NSR
permits might be widespread among refineries. For the proposed Tier 3 gasoline sulfur standard,
however, only about 10 refineries would need air permits that address NAAQS pollutants.
This number could be lower if those refineries apply emission controls, such as selective
catalytic reduction (SCR) for NOx, to reduce the emission increases below the significance level.
For refineries that do need a major source NSR permit for NAAQS pollutants, the permitting
process is expected to take 9-12 months. For an in depth assessment of stationary source
implications, refer to Section V.B of the preamble and a supporting memorandum in the
docket.38
4.3.3 Employment Constraint Analysis
As in prior rules, we also evaluated the capability of E&C industries to design and build
gasoline hydrotreaters as well as performing routine maintenance. This includes an employment
analysis. Two areas where it is important to consider the impact of the fuel proposed sulfur
standards are: 1) refiners' ability to procure design and construction services and 2) refiners'
ability to obtain the capital necessary for the construction of new equipment required to meet the
new quality specification. We evaluated the requirement for engineering design, and
construction personnel, in a manner consistent with the Tier 2 analysis, particularly for three
types of workers: front-end designers, detailed designers and construction workers, needed to
implement the refinery changes. We developed estimates of the maximum number of each of
these types of workers needed throughout the design and construction process and compared
those figures to the number of personnel currently employed in these areas.
The number of person-hours necessary to design and build individual pieces of refinery
equipment and the person-hours per piece of equipment were taken from Moncrief and
Ragsdale39. Their paper summarizes analyses performed in support of a National Petroleum
Council study of gasoline desulfurization, as well as other potential fuel quality changes. The
design and construction factors for desulfurization equipment are summarized in Table 4-4.
Table 4-4 Design and Construction Factors"
Gasoline Refiners
Number of New Pieces of Equipment per Refinery
Number of Revamped Pieces of Equipment per Refinery
60
15
Job Hours Per Piece of New Equipment"
Front End Design
Detailed Design
Direct and Indirect Construction
300
1,200
9,150
4-32
-------
Note:
" Revamped equipment estimated to require half as many hours per piece of equipment
Refinery projects will differ in complexity and scope. Even if all refiners desired to
complete their project by the same date, their projects would inevitably begin over a range of
months. Thus, two projects scheduled to start up at exactly the same time are not likely to
proceed through each step of the design and construction process at the same time. Second, the
design and construction industries will likely provide refiners with economic incentives to avoid
temporary peaks in the demand for personnel.
Applying the above factors, we projected the maximum number of personnel needed in
any given month for each type of job. The results are shown in Table 4-5. In addition to total
personnel required, the percentage of the U.S. workforce in these areas is also shown, assuming
that half of all projects occur in the Gulf Coast in Table 4-5. Very few refineries are expected to
require the full 45-month period to complete scoping studies, process design, permitting, detailed
engineering, field construction, and start-up/shakedown.
Table 4-5 Maximum Monthly Demand for Personnel
Front-End
Design
Detailed
Engineering
Construction
Tier 3 Gasoline Sulfur Program
Number of Workers
Percentage of Current Workforce51
202
11%
809
9%
6,012
4%
Note:
" Based on current employment in the U.S. Gulf Coast assuming half of all projects occur in the Gulf Coast
To meet the proposed Tier 3 sulfur standards, refiners are expected to invest $2.2 billion
between 2014 and 2019 and utilize approximately 1,000 front-end design and engineering jobs
and 6,000 construction jobs. The number of estimated jobs required is small relative to overall
number available in the U.S. job market. As such, we believe that four years is adequate lead
time for refineries to obtain necessary permits, secure E&C resources, install new desulfurization
equipment and make all necessary retrofits to meet the proposed sulfur standards.
4.3.4 ABT Impacts
We conducted a refinery-by-refinery analysis to determine the impacts on refinery E&C
demand of implementing the 10-ppm standard without an ABT program. The analysis suggests
that a greater number of refineries would need to make investments in refinery apparatus and
upgrades than would have under an ABT program. This would result in a greater demand on the
E&C industry. Moreover, the analysis also indicated that the demand upon the E&C industry
would be spread over a shorter period than with the ABT case. In particular, our refinery-by-
refinery analysis indicates that without an ABT program, 73 refineries would revamp existing
pre- and postreaters and 21 would install grassroots postreaters in order to meet the proposed
sulfur standards. The remaining 17 refineries are either already in compliance with the 10-ppm
standard or expected to comply with simple process changes. This is compared to 66 refineries
4-33
-------
that would revamp existing pre- and postreaters and 16 refineries that would install grassroots
postreaters in order to meet the proposed sulfur standards under an ABT program.
4-34
-------
References
1 Meyers, Robert A., Handbook of Petroleum Refining Processes, McGraw Hill, 1997.
2 40 CFR 80 Subpart H
3 Shorey, Scott W., AM - 99-55, Exploiting the Synergy Between FCC and Feed Pretreating Units to Improve
Refinery Margins and Produce Low-Sulfur Fuels, National Petroleum and Refiners Association's 1999 Annual
Meeting.
4 Conversation with Woody Shiflett, Advanced Refining Technologies, October 2011.
5 Barletta, Tony, Refiners must optimize FCC feed hydrotreating when producing low-sulfur gasoline, Oil and Gas
Journal, October 14, 2002.
6 Conversation with Woody Shiflett, Advanced Refining Technologies, October 2011.
7 Shorey, Scott W., AM - 99-55, Exploiting the Synergy Between FCC and Feed Pretreating Units to Improve
Refinery Margins and Produce Low-Sulfur Fuels, National Petroleum and Refiners Association's 1999 Annual
Meeting.
8 Brunei, Sylvette, On the hydrodesulfurization of FCC gasoline: a review, Applied Catalysis A: General 278
(2005) 143 - 172.
9 Leonard, Laura E., Recombination: A Complicating Issue in FCC Naphtha Desulfurization, Prepared for the
AIChE 2006 Spring National Meeting, April 26, 2006
Petroleum Refinery Process Economics, Maples, Robert E., PennWell Books, Tulsa, Oklahoma, 1993.
11 Nocca, J.L., et al, Cost-Effective Attainment of New European Gasoline Sulfur Specifications within Existing
Refineries, November 1998.
12 Prime G, A Sweet Little Process for Ultra-Low Sulfur FCC Gasoline without Heavy Octane Penalty, IFF
Industrial Division.
13 Debuisschert, Quentin, Prime G+ Update, 12th European FCC Conference - Grace Davidson Seminar, Seville
Spain, May 2004.
14 Beck, J.S., Advanced Catalyst Technology and Applications for Higher Quality Fuels and Fuels, Prepr. Pap. Am
Chem Soc., Div. Fuel Chem, 2004 49(2), 507.
15 McGihon, Ron, Exxon Mobil, FCC Naphtha Desulfurization - New Developments, Presentation at the 2009
Technology Conference, October 5 &6, Dubai, United Arab Emirates.
16 Ellis, E.S., Meeting the demands of low sulfur gasoline, Petroleum Technology Quarterly Spring 2002.
17 Successful Start-Up of New Scanfining Unit at Statoil's Mongstad Refinery, November 19, 2003.
4-35
-------
1 &
Greeley, J.P., Zaczepinski, S., Selective Cat Naphtha Hydrofining with Minimal Octane Loss, NPRA 1999
Annual Meeting (this document available from docket A-97-10).
19 Halbert, Thomas R., Technology Options for Meeting Low Sulfur Mogas Targets AM-00-11, Presented at the
2000 Annual Meeting of the National Petrochemical and Refiners Association, March 2000.
20 McGihon, Ron, Exxon Mobil, FCC Naphtha Desulfurization - New Developments, Presentation at the 2009
Technology Conference, October 5 &6, Dubai, United Arab Emirates.
21 Refining Processes 2004, Hydrocarbon Processing.
22 Upson, Lawrence L., Low-sulfur specifications cause refiners to look at hydrotreating options, Oil and Gas
Journal, December 8, 1997.
23 Krenzke, David L., Hydrotreating Technology Improvements for Low-Emissions Fuels AM-96-67, Presented at
the 1996 Annual Meeting of the National Petrochemical and Refiners Association, March 1996.
24 UOP SelectFining™ Process - New Technology for FCC Naphtha HDS, 2009
25 CDTECH, FCC Gasoline Sulfur Reduction, CDTECH, Sulfur 2000, Hart's Fuel and Technology Management,
Summer 1998.
26 Rock, Kerry J., Putman, Hugh, Global Gasoline Reformulation Requires New Technologies, Presented at Hart's
World Fuels Conference, San Francisco, March 1998.
Rock, Kerry L., et al, Improvements in FCC Gasoline Desulfurization via Catalytic Distillation, Presented at the
1998 NPRA Annual Meeting, March 1998.
28 Greenwood, Gil J., Next Generation Sulfur Removal Technology AM-00-12, Presented at the 2000 NPRA Annual
Meeting, March 2000.
29 Meier, Paul F., S Zorb Gasoline Sulfur Removal Technology - Optimized Design AM-04-14, Presented at the
2004 NPRA Annual Meeting, March 2004.
30 Printed Literature by Phillips Petroleum Shared with EPA September 1999.
31 Fatal, Raj, Advanced FCC Feed Pretreatment Technology and Catalysts Improves FCC Profitability AM-02-58,
Presented at the 2002 NPRA Annual Meeting, March 2002.
32 Conversation with Woody Shiflett of Advanced Refining Technologies October 2011.
33 Letter to Margo Oge, EPA, from Mike Ricca, Baker Hughes, July 25, 2011.
34 Letter to Caryn Muellerleine, EPA, from Richard Kelly, Marvel Oil Company, July 13, 2011.
35 The requirements for transmix blenders are contained in 40 CFR 80.84(d).
36 Graphs of transmix gasoline product sulfur levels at Kinder Morgan transmix processing facilities e-mail from
James Holland, Kinder Morgan, August 24, 2011.
4-36
-------
37 Letter from the Alliance of Automobile Manufacturers to Administrator Lisa Jackson, October 6, 2011.
38 Keller, P. (February, 2013). New Source Review Permitting Impact Analysis for Proposed Tier 3 Gasoline
Program. Memorandum to the docket.
39 Moncrief, Philip and Ralph Ragsdale, "Can the U.S. E&C Industry Meet the EPA's Low Sulfur Timetable,'
NPRA 2000 Annual Meeting, March 26-28. 2000, Paper No. AM-00-57.
4-37
-------
Chapter 5 Fuel Program Costs
5.1 Methodology
This chapter provides a summary of the methodology used and the results obtained from
our cost analyses of the proposed gasoline sulfur control. We start by summarizing the refinery
models used for our analysis. We then describe our detailed methodology for estimating the
sulfur control costs for our proposed sulfur program followed by the results. We present the
results from our energy and supply analyses for our proposed gasoline sulfur program. Finally,
we discuss and compare the results of several cost analyses for various sulfur programs.
5.1.1 Overview
When we began our planning for estimating the cost of additional reductions in gasoline
sulfur, we considered two different options. One option for estimating the costs would be to
utilize a linear programming (LP) model, while the second option would be to develop a
refinery-by-refinery cost model. While the LP refinery models are necessary and appropriate for
many analyses, they also have several important limitations of relevance here. When used to
model the cost of nationwide fuel control programs on the entire refining industry, LP models are
usually used to model groups of refineries in geographic regions called Petroleum
Administration for Defense Districts (PADDs). The LP refinery model averages the costs over
the refineries represented in the PADDs; however, the technology chosen by the refinery model
would normally be the lowest cost technology found by the refinery model. This may represent
an unreasonable choice of technologies for individual refineries because of how refineries are
configured and based on the sulfur control technologies installed for compliance with the Tier 2
gasoline sulfur program. While the choice of technologies can be limited based on an
approximate analysis of what mix of technologies would best suit the group of refineries
modeled in each PADD, this would only provide an approximate estimate of the cost incurred.
Based on the quality of input data to these LP models and the assumptions made for complying
with a regulatory requirement, LP refinery models may overestimate or underestimate the
program costs. For example, an LP refinery model would not be a sensible tool for estimating
the credit averaging and trading between refineries. This could be partially overcome by
iterating between PADD refinery model runs, thus estimating the number of credits traded
between PADDs and estimating the level of sulfur control in each PADD. However, the need to
make multiple runs per PADD for each case, coupled with the need to run multiple control cases
for different sulfur standards, would be very time consuming, costly and still would only result in
approximate estimates of the sulfur levels achieved and the cost incurred.
For this reason, EPA developed a refinery-by-refinery cost model which models the
capability for each refinery to revamp existing or install new sulfur control technologies
available to them to reduce their gasoline sulfur levels. Rather than start from scratch, we started
from a refinery-by-refinery cost model developed by APT (Mathpro) for EPA to estimate the
cost of benzene control under MSAT2. However, instead of using the representations of benzene
control technology contained in the model, we obtained information about gasoline
desulfurization and represented the cost of this desulfurization in the refinery-by-refinery cost
model.
5-1
-------
We believe the refinery-by-refinery cost model best estimates the cost of individual
refineries, especially when considering an averaging, banking and trading (ABT) program and
therefore is the best analysis tool for estimating nationwide costs. However, the refmery-by-
refmery cost model cannot estimate certain inputs necessary for estimating costs. Because the
refinery-specific information is not publicly available, it was necessary to find another way to
estimate this information. The inputs and outputs from LP refinery cost modeling provide this
needed information and it was utilized in the refinery-by-refinery cost model. The information
from LP refinery modeling used in the refinery-by-refinery cost model is described in Section
5.1.3.
Since the refinery-by-refinery cost model contains confidential business information for
each refinery, we could not publish the model or present some of the details of the model here.
Therefore, to ensure its viability the refinery-by-refinery cost model was subjected to peer review
by two refinery industry consultants. Our review of most of the suggested changes
recommended by the peer reviewers suggested that there would be little to no change in our
desulfurization cost estimate (some of the changes would increase the estimated costs, while
others would reduce the estimated costs). Also, we anticipate making other improvements to the
cost analysis conducted for the final rule, which would necessitate a second round of peer
review. Therefore, the peer review comments will be addressed prior to undertaking the cost
analysis for the final rulemaking along with the other changes that we will be making to our cost
analysis. The peer review comments are contained in two reports submitted to the docket.
The refinery-by-refinery cost model focuses on reducing sulfur from the FCC naphtha
because of its high sulfur content. To comply with the 30-ppm Tier 2 sulfur control program,
most refiners installed FCC naphtha hydrotreaters (referred to as FCC postreaters) or FCC feed
hydrotreaters (referred to as FCC pretreaters) to reduce that unit's sulfur contribution to their
gasoline pool. If refiners installed an FCC postreater under Tier 2, we modeled refiners
revamping those units. However, if refiners relied on FCC pretreaters to comply with Tier 2, we
assumed that grassroots FCC postreaters would have to be installed at those refineries to reduce
its gasoline pool down to 10 ppm. However, since adding grassroots FCC postreaters is
relatively expensive for the amount of sulfur reduction obtained, the ABT analysis we conducted
avoided many of these types of investments. Refineries with both pre and postreaters today
could achieve further gasoline sulfur reductions less than 10 ppm at a relatively low incremental
cost and sell the credits to those refiners who are operating refineries which would otherwise be
faced with grassroots postreater investments. In addition to addressing the sulfur in the FCC
naphtha, we believe that some refineries may need to reduce the sulfur in light straight run (LSR)
naphtha. Some refineries might also need to reduce sulfur in butane, although we don't expect
refiners to need to address butane sulfur unless they are pursuing a very stringent gasoline sulfur
standard, e.g., 5 ppm.
To better understand the desulfurization costs, we evaluated several different scenarios or
cases. For a 10-ppm average sulfur standard, we assessed the costs based on each refinery
achieving the 10-ppm standard with no averaging among refineries, an averaging program which
assumed intra-company transfers of sulfur credits, and an averaging program which assumed
nationwide transfers of sulfur credits. To provide credits for averaging and trading under the 10-
ppm average standard, we also evaluated refiners reducing their gasoline sulfur down to 5 ppm.
Since we had estimated costs for each refinery to get to 5 ppm sulfur, we also report out the cost
5-2
-------
for a 5 ppm average gasoline sulfur standard assuming no averaging between refineries. The
costs for the proposed sulfur program are based on a 10-ppm sulfur standard with intra-company
credit transfers. These different scenarios are summarized in Table 5-1.
Table 5-1 Sulfur Control Cases Evaluated for the Proposal
10-ppm Standard
No ABT Program
ABT Program with Intra-
Company Credit Transfers
(Proposed Rule Costs)
ABT Program with
Nationwide Credit Transfers
5 -ppm Standard
No ABT Program
N/A
N/A
5.1.2 LP Refinery Modeling Methodology and Results
Although we used the refinery-by-refinery cost model to estimate gasoline
desulfurization costs, certain input information was needed to estimate the costs with refmery-
by-refinery cost model, and without access to detailed refinery-specific information, we relied on
outputs from our LP refinery modeling. Perhaps the most important input is the cost for making
up the octane loss that occurs with desulfurization. Certain refinery operations information from
the LP refinery model was used for estimating the volume of gasoline produced in the refinery -
by-refinery model, including the utilization factors of individual refinery units, and the
percentage that straight run naphtha, FCC naphtha and hydrocrackate comprises of the feed
volume of their respective units.
LP refinery models are detailed mathematical representations of refineries. They are
used by individual refining companies to project how best to operate their refineries. They are
also used by government agencies, such as EPA and DOE, as well as by refining industry
associations and individual companies, to estimate the cost and supply impacts of fuel quality
changes. LP refinery models have been used for these purposes for decades and a certain
protocol has been established to conduct these studies.
Two different sets of refinery modeling runs from two different LP refinery models were
used as inputs into the refinery-by-refinery cost model. The refinery-by-refinery cost model
already contained the utilization factors and gasoline production volumes for individual refinery
units from the analysis conducted by Mathpro for the MSAT2, and we continued to use that
information for this cost analysis. The gasoline demand is expected to be fairly flat in the future,
so using the previous refinery modeling work for these inputs will likely have little impact on the
cost estimate. We plan on updating these inputs for the final rule to reflect more recent refinery
modeling work.
Additional refinery modeling was conducted using the Haverly GRTMPS refinery model.
The primary reason for conducting new LP refinery modeling analysis was to estimate the cost
of making up the octane loss associated with desulfurization as well as estimate how gasoline
qualities would be affected by the octane recovery to feed into the emissions inventory impact
analysis discussed in Chapter 7. While the gasoline demand and production volumes are not
5-3
-------
expected to change in the future, the cost of octane is expected to decrease dramatically due to
expected much larger use of ethanol under the RFS2 rulemaking.
The first step in conducting an LP refinery modeling analysis was the development of a
base case. The base case is a refinery modeling case that calibrates the refinery model based on
actual refinery unit capacity and input and output data. The base year for this study was the year
2000 for the Mathpro model and the year 2004 for the Haverly model. Because much of the
information available for establishing the base case is only available for PADDs of refineries, the
LP refinery modeling was conducted on a PADD-wide basis. Refinery capacity information
from the Oil and Gas Journal was aggregated by PADD and entered into the LP refinery model.
The feedstock volumes, including crude oil and gasoline blendstocks, were obtained from the
Energy Information Administration (EIA) and entered into each PADD's model. Similarly,
product volumes such as gasoline, jet fuel, and diesel fuel were obtained from EIA and entered
into the cost model. The environmental and ASTM fuel quality constraints in effect in the base
year were imposed on the products. This includes the Reformulated Gasoline program and the
500-ppm highway diesel fuel sulfur standard, and for the Haverly LP refinery modeling, the first
year of the Tier 2 gasoline sulfur standard. This information was input into the LP refinery cost
model for each PADD and each PADD model was run to model the U.S. refinery industry for the
base year. The gasoline quality for each PADD refinery model was then compared to the actual
gasoline quality for conventional and reformulated gasoline which is available from the RFG
database. Each model was calibrated to closely approximate the gasoline quality of each PADD.
The second step in modeling is the development of a reference case. The purpose of the
reference case is to model the refining industry operations and cost in a future year, which is the
year that the control program is modeled to be in effect (serving as a point of reference to the
control cases for estimating costs and other impacts). The reference year for the Mathpro LP
refinery modeling was 2012 while the reference year for the Haverly refinery modeling was 2017
and 2030. We developed two reference cases with the Haverly model to model different control
case scenarios. The reference case was created by starting with the base cases for each PADD
and adjusting each base case to model the future year, accounting for the changes between the
two years.
Two different types of adjustments were made to the base case refinery models to enable
modeling the refining industry for the reference case. First, the change in certain inputs such as
product volumes and energy prices need to be accounted for U.S. refinery gasoline, diesel fuel
and jet fuel demands are projected year-by-year by EIA in its Annual Energy Outlook (AEO);
the projections from the AEO for the reference case are used in the refinery modeling analysis.
The Mathpro LP refinery modeling relied on AEO 2006 while the Haverly LP refinery modeling
relied on AEO 2011. This growth in demand is used to project refinery production for each
PADD to meet that increased demand. This projected growth in U.S. refinery production was
entered into the reference case version of the LP refinery model. The utility and crude oil and
other feedstock prices which are projected by EIA for the future year being modeled were also
entered into the refinery model as well as the estimated product prices.
The second adjustment made to model the reference cases was the application of fuel
quality changes. Environmental programs which have been implemented or which will largely
be implemented by the time that the prospective fuels control program would take effect were
5-4
-------
modeled in the reference case. These fuel quality changes include limits such as the 30-ppm
average gasoline sulfur standard, 15-ppm caps on highway and nonroad diesel fuel and the
MSAT2 benzene control program, in addition to the environmental programs which were already
being modeled in the base case. This also included the fact that California gasoline was already
averaging 10 ppm sulfur or less as a result of prior changes to their predictive model used for
gasoline certification, well in advance of their 20 ppm cap on gasoline sulfur taking effect. As a
result, our Tier 3 gasoline standards are not proposed to apply in California. Thus, for this
analysis we only assumed further sulfur control on gasoline volumes produced by California
refineries for distribution outside of California. For the Mathpro refinery modeling, which was
conducted before the nonroad diesel fuel program and MSAT2 benzene control programs were
finalized, those fuels control programs were not modeled in the reference case. Also, the
implementation of EPAct required a large increase in the amount of ethanol to be blended into
gasoline to comply with the renewable fuels standard (RFS), but not RFS2. In its AEO 2006,
EIA projected that the volume of ethanol blended into gasoline exceeded the RFS required
amounts, resulting in 9.6 billion gallons of ethanol blended into gasoline by 2012. Other
provisions of EPAct that were modeled with both the Mathpro and Haverly models included a de
facto ban on MTBE and rescinding the RFG oxygenate requirement. The reference case unit
throughputs and gasoline blendstock volumes were used in the refinery-by-refinery cost model.
For the Haverly refinery modeling work, in addition to the EPAct provisions, the RFS2
renewable fuels volumes were modeled for 2017. For the 2017 reference case, 17.8 billon
gallons of ethanol were assumed to be blended into gasoline, and 3.9 billion gallons of renewable
and cellulosic diesel fuel and biodiesel were assumed to be blended into diesel fuel for the
control case. For gasoline, the ethanol volume beyond the E10 blendwall was assumed to be
blended as E15. For the 2030 reference case, we modeled 22.2 billion gallons of corn and
cellulosic ethanol, and 8.3 billion gallons of renewable diesel and biodiesel.
The third step in conducting the LP refinery modeling was to run the various control
cases. The control cases are created by applying a specific fuel control standard to each PADD
reference case. To single out a specific cost or other impact, the sole difference between the
control case and the reference case is the parameter change being studied.
For the Haverly modeling, a control case was run to model the octane loss associated
with desulfurization using 2017 as the year of analysis. Since we solely wanted to identify the
cost of recovering lost octane for the refinery-by-refinery modeling, this case was run by
reducing the octane value of the FCC naphtha by one octane number, and this was the sole
change relative to the reference case. The control case was run with capital costs evaluated at a
15 percent rate of return on investment (ROI) after taxes.A The octane cost estimated by the LP
cost model is 0.76 cents per octane number per gallon of FCC naphtha. Because the octane loss
associated with a specific technology may be lower or higher than 1 octane number, we scaled
the octane cost based on the relative estimated octane loss on the FCC naphtha (i.e., a 1A octane
loss of the FCC naphtha was estimated to cost 0.38 cents per gallon of FCC naphtha. Table 5-35
A Normally we conduct the refinery modeling assuming an after-tax 15% ROI and adjust the costs to reflect a
before-tax 7% ROI to report the costs. However, in this case because the new capital investments were so minimal,
we omitted the capital cost amortization adjustment because its effect on costs was judged to be negligible.
5-5
-------
at the end of this chapter summarizes the data output from the refinery modeling from which we
calculated the octane cost for using in the refinery-by refinery cost model.
It was necessary to estimate the gasoline qualities for estimating the emissions impact of
the proposed Tier 3 program. This was conducted in two separate steps. First it was necessary
to estimate the gasoline qualities of the 2017 and 2030 reference cases relative to the gasoline
qualities of a revised base case. The sole differences that we modeled between the 2005 revised
basecase and the 2017 and 2030 reference cases was the phase out of MTBE and the addition of
ethanol. For the 2005 revised basecase we modeled 1.7 billion gallons of MTBE and 4.1 billion
gallons of ethanol. For the 2017 and 2030 reference and control cases, we modeled 17.8 and
22.2 billion gallons of ethanol, respectively. In 2017, we estimated that approximately half the
gasoline would be 10 percent ethanol and the about the other half would be 15 percent. To
model the emissions impact of the different ethanol blends, we modeled two reference cases, one
with 100 percent E10 and the other with 100 percent El 5. These two ethanol cases were
modeled in 2030 and we used the results for 2017 as well. The gasoline qualities for the
reference and two ethanol cases are summarized in Table 5-57 to Table 5-61 in the appendix at
the end of this chapter. The changes in gasoline quality are summarized in Table 5-2. Because
of the tendency for the LP refinery model to shift gasoline blendstocks around resulting in odd
gasoline quality changes in individual PADDs, we solely used the national average change in
gasoline qualities and applied those changes for all E10 or El 5 gasoline for the emissions
analysis.
5-6
-------
Table 5-2 Difference in Gasoline Qualities between E10 and E15 Control Cases with the
Reference Case
PADD1
PADD2
PADD3
PADDs 4
&50C
USavg
minus CA
E200
E300
Aromatics
Olefins
E200
E300
Aromatics
Olefins
E200
E300
Aromatics
Olefins
E200
E300
Aromatics
Olefins
E200
E300
Aromatics
Olefins
E]
Summer
5.08
2.41
-2.88
-0.92
-3.29
-2.82
-0.17
-0.31
6.06
3.27
-4.93
-1.06
8.78
0.95
-2.92
-1.53
3.46
0.88
-2.58
-0.87
.0
Winter
2.16
-1.89
-1.20
-1.26
6.96
2.10
-2.70
-1.26
5.65
0.76
-5.86
-2.56
5.19
-0.10
-5.01
-1.06
4.65
0.04
-3.19
-1.59
E:
Summer
11.08
3.58
-5.95
-1.68
1.10
-4.08
-2.03
-1.55
12.03
3.54
-8.43
-2.13
11.29
1.58
-4.71
-1.87
8.65
1.08
-5.28
-1.80
L5
Winter
9.71
3.23
-3.99
-1.75
11.07
2.68
-3.22
-1.46
15.76
8.51
-9.89
-2.03
10.88
3.81
-7.62
-1.56
11.77
4.50
-5.65
-1.73
The second step for estimating gasoline qualities was to model the impact of
desulfurization on gasoline qualities. The total impact of desulfurization on gasoline qualities is
comprised of the reduction in gasoline sulfur, the associated reduction in olefins and the impacts
of recovering the lost octane. The sulfur reduction is fixed by the standard and the olefins
reduction is a function of the selectivity of the desulfurization technologies. We reviewed the
information that we had obtained for the gasoline desulfurization technologies and estimated that
desulfurizing gasoline from 30 ppm to 10 ppm would result in a 1 percent reduction in olefm
level. Since we estimated the cost of making up lost octane using the LP refinery model, we
used that case for estimating the impact of octane recovery on gasoline qualities. The gasoline
qualities for the reference case and the control case which reflects a 1 octane number loss in the
FCC naphtha pool are summarized in Table 5-47 to Table 5-51 at the end of this chapter. The
difference in gasoline qualities between the reference and control cases is summarized in Table
5-3. Because of the tendency for the LP refinery model to shift gasoline blendstocks around
resulting in odd gasoline quality changes in individual PADDs, we solely used the national
average change in gasoline qualities and applied those changes for all gasoline for the emissions
analysis. After we integrated the gasoline desulfurization information into the refinery-by-
refmery cost model, we estimated that desulfurizing gasoline from 30 ppm down to 10 ppm
would result in about a one-half reduction in FCC naphtha octane ((R+M)/2) number. To
estimate the changes in gasoline quality from a one-half octane number loss in FCC naphtha that
5-7
-------
we estimated, we divided the gasoline quality changes for one octane number in the FCC
naphtha by a factor of two resulting in half the changes in gasoline quality that we estimated for
a one octane number change in FCC naphtha. The second set of columns in Table 5-3
summarizes the gasoline quality changes that we estimated for reducing the gasoline sulfur levels
from 30 to 10ppm.B
Table 5-3 Differences in Gasoline Qualities Between the Control and Reference Cases
PADD1
PADD2
PADD3
PADD4&
50C
USavg
minus CA
E200
E300
Aromatics
Olefins
E200
E300
Aromatics
Olefins
E200
E300
Aromatics
Olefins
E200
E300
Aromatics
Olefins
E200
E300
Aromatics
Olefins
2017 minus ION in
FCC Naphtha
Summer
-0.20
-2.03
1.14
0.60
-0.19
-0.06
0.48
0.09
-1.68
-1.99
2.50
0.15
0.50
0.14
0.10
0.02
-1.55
-1.49
1.26
0.37
Winter
-0.27
-0.63
0.69
0.01
-0.04
-0.12
0.27
0.03
-0.06
-0.35
0.45
0.04
-0.69
-1.67
-0.32
-0.49
-2.54
-1.37
0.95
-0.23
Adjusted for 1/2 ON
Summer
-0.10
-1.01
0.57
0.30
-0.09
-0.03
0.24
0.04
-0.84
-1.00
1.25
0.07
0.25
0.07
0.05
0.01
-0.78
-0.75
0.63
0.18
Winter
-0.14
-0.32
0.34
0.01
-0.02
-0.06
0.13
0.01
-0.03
-0.17
0.23
0.02
-0.35
-0.84
-0.16
-0.25
-1.27
-0.68
0.48
-0.12
Since we completed the LP refinery modeling to estimate the cost for recovering the lost octane and the associated
changes in gasoline quality, we found that other Tier 3 refinery modeling studies did not show the same increase in
aromatics and decrease in E300 (see also 7.1.3.2). We then discovered that the LP refinery model that we have
licensed to use required some improvements in how the refinery model was characterizing both the light-cut and the
heavy-cut naphtha from the reformer streams to more accurately estimate the E300 and aromatics content of these
streams. We have subsequently worked with a contractor to make these improvements to the LP refinery model and
will reassess the changes in gasoline quality for the final rule analysis. Thus, while our modeling results shown in
Table 5.3 show a meaningful impact on aromatics and E300, we believe there will in fact be little or no change.
Note that these improvements are not expected to have any impact on the cost estimates made by the refinery model.
5-8
-------
5.1.3 Summary of Refinery-by-Refinery Model Methodology
The purpose of the refinery-by-refinery cost model is to project how each refinery would
reduce the sulfur in its gasoline pool to 10 ppm or lower and to estimate the cost for doing so.
To do this we created a U.S. refining industry refinery-by-refinery spreadsheet cost model using
inputs from an LP refinery model case to allow us to better understand the gasoline sulfur control
costs to individual refineries. This spreadsheet cost model also allowed us to model how costs
would be affected by an ABT program.
The building of the refinery-by-refinery model consisted of two major steps. The first
step was to estimate baseline operating conditions for each refinery. This involves estimating the
volumes and sulfur levels of the gasoline blendstocks that comprise each refinery's gasoline. We
chose to use information from 2009 for modeling the baseline operating conditions for the
refineries as it's the latest year we had data for refiner operations and yields. Additionally, EIA
projections indicated that gasoline demand is expected to be essentially flat between 2009 and
2017, alleviating the need to adjust refinery operating throughputs and yields for future changes
in gasoline demand.1'0 Because of these factors, the 2009 gasoline production volumes and
refinery operating conditions can reasonably be projected to be at the same level in 2017 (the
first year of implementation of the Tier 3 program) in estimating costs and refinery impacts. As
a final adjustment to our estimated gasoline volumes and sulfur levels, we calibrated the model
to actual refinery gasoline volume and sulfur levels to ensure our model's accuracy.
To estimate the cost for each refiner to lower its gasoline pool down to 10 ppm, we used
our refinery-by-refinery model to estimate the FCC naphtha volume, the sulfur level of the FCC
naphtha, and the amount of sulfur reduction needed in FCC naphtha to meet a 10-ppm sulfur
standard at each refinery. We also incorporated in our refinery-by-refinery model the impacts
that FCC pretreaters have on FCC naphtha yields and sulfur levels, as well as the impact of
refinery-specific crude oil sulfur levels on FCC naphtha yields. Similarly, we also used the
refinery-by-refinery cost model to estimate the volume levels of light straight run naphtha (LSR)
and natural gas liquids (NGL) that require additional hydrotreating, as well as butane volumes
that are directly blended into the gasoline pool.
The second step involves applying the various sulfur control technologies to each refinery
as necessary to meet the 10-ppm sulfur standard. We expect that the majority of the sulfur
reductions necessary to comply with a 10-ppm gasoline sulfur standard will come from reducing
the sulfur level in their FCC naphtha. Using our refinery-by-refinery model we also estimate that
a few refineries will add additional LSR/NGL hydrotreating capacity. We also evaluate each
refiner's cost to install new butane Merox extraction equipment to lower the sulfur level of
butane that is directly blended to the gasoline pool. Reducing the sulfur content of butane was
assumed necessary to meet a 5-ppm sulfur standard for our ABT cases. This assumption is
c Since we conducted the cost analysis for the proposed rulemaking, we put in place an additional round of
greenhouse emission reductions (2017 - 2025) for light duty cars and trucks that will reduce future gasoline
demand. When we model the costs for the final rulemaking, we will incorporate this reduction in gasoline demand
in our costs analyses.
5-9
-------
conservative as many refiners may already have this equipment, or may purchase low-sulfur
butanes that have already been treated by their supplier.
This allows us to generate a cost estimate for the sulfur control technology in each
refinery. The capital costs for installing the sulfur control technologies in each refinery were
evaluated based on a 7 percent return on investment (ROI) before taxes. In the following
sections, we present the various steps that were used in this refinery-by-refinery modeling
analysis.
5.1.3.1 Estimating Individual Refinery Gasoline Blendstock Volumes
In order to develop a baseline for our refinery-by-refinery analysis, it was necessary to
understand the sulfur levels and volumes of the various blendstocks which make up each
refinery's gasoline. Each refinery blends up its gasoline pool from the various gasoline
blendstocks that are produced from the refinery units installed at each refinery. However,
information on the volumes and sulfur levels of each gasoline blendstock produced by each
refinery is not publicly available, so it was necessary to estimate them. Estimating each
refinery's gasoline blendstock volumes was accomplished using actual 2009 refinery specific
throughput rates that we obtained from EIA for crude, FCC, cokers and hydrocracking units, and
published refinery unit capacity information for the other refinery units. We used this
information to estimate the extent that each refinery process unit is utilized, followed by a unit-
specific analysis for estimating how each refinery unit produces material for blending into
gasoline. After the unit-by-unit estimates are completed, we performed an overall check by
comparing our estimated gasoline volumes with reported gasoline volumes for each refinery,
using EPA's RFG database
The model requires the total gasoline volume and each gasoline blendstock volume for
each refinery as an input. Although the model does estimate this volume of gasoline produced
by each refinery based on the estimated volumes of each gasoline blendstock, we chose to use
actual 2009 gasoline production data reported by refiners as for the total gasoline volume for
each refinery in our cost calculations. To comply with the RFG program, refiners report gasoline
production volumes and sulfur levels for reformulated and conventional gasoline to EPA. We
used this data and imputed each refiner's 2009 total gasoline production and corresponding
sulfur levels into our model.
In the end, our completed refinery-by-refinery modeling estimates of gasoline produced
on a national basis, correlated very well with the actual refinery production volume in 2009, with
our estimated volumes having an overall error of approximately 0.5 percent relative to the
reported refinery production volumes. In order to minimize the impact of this error, we forced
the estimated total refinery gate gasoline volume to match actual reported 2009 gasoline
production volume across all the refineries. The unadjusted refinery-by-refinery estimates of
FCC and LSR gasoline volumes, however, are used for estimating necessary equipment
modifications and costs for sulfur removal. This is due to the fact that the refinery by refinery
models estimates for each refiner's FCC gasoline are likely to be very close to actual FCC
production, as we use actual refinery specific FCC charge throughput rates and account for the
effects of FCC feed pretreating on FCC gasoline yields. Additionally, the cost for treating FCC
5-10
-------
gasoline in our Tier 3 programs, comprise over 85 percent of the total costs, while LSR
comprises the bulk of the remaining costs.0
5.1.3.1.1 Principal Refinery Unit Volumes
To estimate the production volumes for each of the refinery's gasoline blendstocks, the
refinery-by-refinery model needs process capacity information. The Oil and Gas Journal (OGJ)
publishes and the EIA reports unit capacities for the principal refinery units for each refinery in
the U.S.2'3 We updated our database from these two sources to reflect capacity that was in place
in 2009, the base year for the model. Where differences between the two databases existed, we
used the information that was judged best overall from the two sources and entered it into the
refinery-by-refinery cost model. These unit capacities indicate the maximum throughput rate for
each individual unit, not the actual unit throughput rates for each facility, as this is proprietary
business information and not publicly available. In order to enhance our model, we obtained
from EIA the actual 2009 annual unit throughput rates for each refiner's crude and major
refinery units (FCC, cokers and hydrocracker units). With this information, the refmery-by-
refmery model was fine-tuned to reflect each refineries gasoline blendstocks yields. Our use of
this information significantly improved our model's ability to estimate FCC naphtha, as well as
other gasoline blendstocks that each refinery makes. The FCC, coker and hydrocracker unit's
throughputs versus actual capacity that we obtained from EIA for each domestic refinery on a
PADD average basis are listed in Table 5-4. This information is presented on a PADD average
to protect CBI.
Table 5-4 Process Capacity Utilization"
Total U.S.
PADD lb
PADD 2
PADD 3
PADDs 4/5
excluding California
Crude
Throughput
0.843
0.779
0.859
0.858
0.817
FCC
Throughput
0.840
0.754
0.814
0.880
0.794
Coker
Throughput
0.761
0.643
0.810
0.782
0.824
Hydrocracker
Throughput
0.768
0.707
0.774
0.637
0.903
aActual unit throughput rates as a fraction of maximum unit capacity on a PADD basis
bPADD 1 data includes Hovensa, VI
In the model, we also adjusted the refinery capacity information to account for refinery
expansions or refinery shutdowns that we were aware of and are scheduled to occur over the next
several years. Refinery expansions include those announced for WRB Refinery in Wood River
Illinois, the Valero Refinery in Norco Louisiana, and the Marathon Refinery in Garyville,
Louisiana. For these expansions, there is limited public data on which of the specific process
unit capacities would be increased, though each expansion project has information on the crude
D We did not account for any undercutting of the heavy FCC naphtha into jet and diesel fuel, nor did we account for
the removal of any pentanes that might be occurring in refineries to comply with stringent summertime RVP
standards, therefore our analysis is likely somewhat conservative and overestimates the costs.
5-11
-------
unit capacity increase. Since the data was limited, we increased all of the existing individual
process unit capacities by the fractional increase in crude oil unit capacity at each of the
expanding refineries. Refiners that we believe are permanently shutdown in PADD 1 were
removed from our analysis but, consistent with recent import/export trends, we allowed PADD 3
to supply any lost capacity to PADD 1 as a result of this lost production. PADD 1 refiners that
were presumed to be permanently shutdown are; Giant refinery located in Yorktown, Virginia,
Sunoco refinery in Westville, New Jersey, and Shell Oil refinery in Bakersfield, California.
5.1.3.1.2 Other Refinery Unit Volumes
The next step was to calculate actual unit throughput rates for the other refinery processes
that produce gasoline blendstocks. These units include alkylation, dimerization, polymerization,
isomerization, naphtha reforming. All of these processes feedstocks are primarily supplied by
the crude and FCC Units. Since this data is similarly not publicly available we tuned these units
to the EIA throughputs rates for crude and FCC units at each facility, with alkylation units
running at the same throughput rates as the FCC and the remaining units running at the crude oil
throughput ratesThe results of the capacity utilizations of these downstream units are
summarized in Table 5-5 below.
Table 5-5 Other Unit Process Capacity Utilization"
PADD 1
PADD 2
PADD 3
PADDs 4/5
excluding
California
Reformer
Throughput
0.774
0.859
0.858
0.817
Alkylation
Throughput
0.886
0.878
0.880
0.794
Isomerization
Throughput
0.931
0.859
0.858
0.714
Poly/Dimersol
Throughput
1.000
0.859
0.345
0.100
a Actual unit throughput rates as a fraction of maximum unit capacity on a PADD basis
With these inputs the refinery-by-refinery model now contained estimates of the
feedstock charge rates for all of the gasoline blendstock producing units, though estimating
refinery unit capacity and capacity utilization may or may not translate directly into the gasoline
blendstock volume produced by a specific refinery unit. This is because some refinery units may
also produce products other than gasoline blendstock. Additionally, some processes have
volume loss of feedstock due to process reactions and conversions that take place that increase or
decrease the density and therefore the volume of products. To take this into account, a gasoline
fraction yield factor has to be applied to each process to convert the process charge rate into the
yield of gasoline blendstocks. The process fractional yields that were used in our refinery by
refinery model were taken from our MSAT2 final rule LP refinery modeling work, which
represented the U.S. refining industry on a PADD basis. The FCC unit process yields of naphtha
blendstock are different for units with an FCC feed pretreater, versus those without feed
pretreating. In our modeling we accounted for this by adjusting yields and sulfur levels of FCC
units with pretreaters and those without a pretreater. The fractional yields of gasoline blendstock
for the major process units and the 2009 throughputs for each of these units used in our model
are summarized below in Table 5-6 and Table 5-7.
5-12
-------
Table 5-6 Gasoline Blendstock Fraction Yields Per Process Unit Charge
PADD 1
PADD2
PADD 3
PADDs 4/5 excluding
California
Crude
0.190
0.211
0.188
0.183
FCC Units
Average
0.560
0.570
0.554
0.565
Coker
0.234
0.234
0.239
0.234
Hydrocracker
0.369
0.311
0.212
0.276
Table 5-7 2009 Refinery Unit Throughputs (1,000 BPSD)
PADD 1
PADD 2
PADD 3
PADDs 4/5 excluding
California
Crude
1,624
3,193
7,262
1,363
FCC Units
652
1,017
2,604
263
Coker
<3a
322
1,043
130
Hydrocracker
<3a
223
500
94
" Since there are less than three refiners in this PADD with these units, the data was not reported to protect
CBI information.
The FCC unit produces significant volumes of naphtha, a gasoline blendstock. The
conversion percentage to naphtha is affected by the severity of the operation of the FCC unit. As
shown in Table 5-6 above, the portion of FCC feedstock converted to naphtha ranged from 55 to
57 percent across the various PADDs. The range among individual refineries can be quite large,
but we didn't have access to refinery specific data for this. However, as a group there is
expected to be differences between refineries with and without FCC pretreaters. Therefore,
rather than simply use the PADD average conversion of FCC feedstock to naphtha for all
refineries in a given PADD, the refinery-by-refinery model differentiates between refineries that
have an FCC feedstock pretreater and those that do not. We have also quantified the gasoline
blendstock fraction yield for FCC units that have both feed pretreater and postreater units.
Historically, refiners have installed FCC feed pretreaters for economic reasons, as
pretreaters increase FCC unit conversion to high value gasoline blendstock while decreasing the
production of low value light cycle oils and residual material from FCC units. FCC feed
pretreaters also have the benefit of reducing sulfur from the FCC feedstocks, resulting in the
production of lower sulfur FCC naphtha and ultimately lower sulfur gasoline. In developing our
refinery-by-refinery model, we quantified the impact FCC feed pretreating and postreating has
on FCC naphtha yields and sulfur levels based on our evaluation of information we received
from technology vendors. The results of this analysis are shown in Table 5-8 below.
5-13
-------
Table 5-8 FCC Unit Gasoline Blendstock Fraction Yields
PADD 1
PADD2
PADD 3
PADD 4/5 excluding
California
Average of All
FCC Units
0.560
0.570
0.554
0.565
FCC Units with
No Pretreater
0.558
0.533
0.520
0.548
FCC Units with a
Pretreater Only
0.638
0.648
0.630
0.642
FCC Units with a
Pretreater and
Postreater
0.607
0.617
0.600
0.612
5.1.3.1.2.1 Poly Gas and Alky late
For the polymerization and alkylation units the capacity of the unit coupled with its
estimated utilization rates listed in Table 5-5 is sufficient to establish the volume of gasoline
blendstock produced by these units. For example, a particular refinery unit in PADD 1 might
have a 10,000 barrel per day alkylation unit. If the alkylation units in PADD 1 are estimated to
be operating at 56 percent of its listed capacity in 2017, the alkylate production is projected to be
5,600 barrels per day at that refinery. Each of the refineries within a given PADD was assumed
to have the same utilization rate for any alkylation units.
5.1.3.1.2.2 Light Straight Run Naphtha
The remaining gasoline blendstocks, including light straight run naphtha (LSR), coker
naphtha and hydrocrackate cannot be estimated simply using the unit capacity and unit
utilization rate. In order to determine the volume of gasoline blendstock produced by each of
these units, additional steps are required. LSR naphtha is principally comprised of five- and six-
carbon hydrocarbons which come directly from crude oil. Thus the volume of LSR for each
refinery was based on the volume of crude oil processed by each refinery as determined in
Section 5.1.3.1.1, as well as the percentage of that crude oil that is LSR. The fraction of LSR in
each refinery's crude oil was estimated on a PADD average basis using the LP refinery model
since it is not available on a refinery-by-refinery basis. This percentage is based on the types and
quality of crude oil processed by all the refineries in each PADD from our LP model4. LSR as a
percentage of crude oil is estimated to vary from 4 to 5 percent across the PADDs. These PADD
level results are shown in Table 5-9 below.
Table 5-9 LSR as a Percentage of Crude Oil by PADD
LSR as a Percentage of Crude Oil
PADD 1
4.5%
PADD 2
5.0%
PADD 3
4.4%
PADDs 4/5a
4.4%
"Excluding California
After we calculated how much LSR is produced at each refinery we determined how
much of the LSR is used as a gasoline blendstock. LSR has several possible destinations that
vary from refinery to refinery. For each refinery, with the exception of those located in PADD 2,
a portion of the LSR is designated to be sold into the petrochemicals market where it is
processed into other hydrocarbon compounds. EIA publishes the volume of naphtha which is
sold into the petrochemicals market in each PADD5. This information is summarized in Table
5-14
-------
5-10 below. Since this information is not publicly available on a refinery-by-refinery basis, we
assumed that the volume of LSR naphtha sold into the petrochemicals market by each refinery is
proportional to the refinery's percentage of the total volume of crude oil processed in the PADD
in which the refinery is located. After accounting for the volume of LSR naphtha sold to the
petrochemicals market, the balance of LSR naphtha is used as a feedstock for the isomerization
unit if the refinery has one. If a refinery does not have an isomerization unit, all of the LSR not
sold to the petrochemical market is assumed to be used as a gasoline blendstock. Any volume of
LSR at a given refinery that exceeds the capacity of the isomerization unit at the facility is also
assumed to be used as a gasoline blendstock. However, if a refiner does not have enough naptha
hydrotreating capacity to process all of the refiner's LSR volume, we assumed that the refiner
would use excess capacity in their FCC postreater, to reduce the sulfur content of the LSR
blendstock.
Table 5-10 Refiner Sales of Naphtha in 2009 (1,000 BPSD)
Naphtha Sold to the Petrochemical Industry
Sales of Special Naphtha
PADD 1
12.2
0.8
PADD 2
22.4
0
PADD 3
161.2
31.4
PADD 4/5a
0
0
a Excluding California data
For further clarity on gasoline blendstock yields from the model, the gasoline blendstock
volumes of LSR and naphtha from the naphtha splitter overhead tower are adjusted to subtract
sales of these blendstocks that are sent to the petrochemicals market.E The values listed in Table
5-11 for LSR and naphtha splitter overhead are the volumes sent to gasoline, as a fraction of
crude throughput.
Table 5-11 PADD Average Gasoline Blendstock Yields per Fraction of Crude Input
LSR to Gasoline
Naphtha from Light Naphtha Splitter Overhead
PADD 1
0.0309
0.0021
PADD 2
0.0286
0.0081
PADD 3
0.0187
0.0072
PADD's4/5a
0.0381
0.0093
a Excluding California
In refineries with an isomerization unit, much of the LSR is processed into isomerate, the
product produced by the isomerization unit. The volume of isomerate produced is dependent on
the volume of feedstock processed by the isomerization unit up to its capacity. As described
above, all of the LSR that is not assumed to be sold into the petrochemical markets is assumed to
be sent to the isomerization unit, up to the maximum capacity of the isomerization unit. The
isomerization unit produces a blendstock with a slightly higher energy density and smaller
volume compared to the feedstock volume. To account for this effect, the volume of isomerate
produced is estimated to be 1.6 percent less than the volume of LSR feedstock to the
isomerization unit. Hydrocrackate and Coker Naphtha
E Naphtha Splitter towers separate the naphtha feed stream into a light and heavy streams, whereby the heavy stream
is typically reformer feedstock, while the light stream is blend stock lighter than reformer feed.
5-15
-------
The hydrocracker and coker units also produce some light naphtha material which is
blended into gasoline. Heavy naphtha is also produced in these units, which is feed to the
reformer, as discussed in the next section. The light naphtha material produced by the
hydrocracker and coker are termed light hydrocrackate and light coker naphtha, respectively.
Based on LP refinery modeling work done for the MSAT2 rule we estimated that the portion of
the feedstock processed by each of these units converted to light coker naphtha and light
hydrocrackate was 5 percent for coker units across all the PADDs, and ranges from 23 to 32
percent for hydrocracker units depending on the PADD. The light coker naphtha is poor in
quality and require hydrotreating to removes sulfur, olefins and other impurities, before sending
them to an isomerization unit, if a refiner has one. Table 5-12 below summarizes the percentage
of the feedstock to these units that is converted to light naphtha and blended into gasoline.
Table 5-12 Isomerization Unit Feed Rates by PADD
Light Coker Naphtha
(% of Coker Feed)
Light Hydrocrackate
(% of Hydrocracker Feed)
PADD 1
5.0%
28.7%
PADD 2
5.0%
32.0%
PADD3
5.0%
23.3%
PADD 4/5a
5.0%
27.2%
"Excluding California
5.1.3.1.2.3 Re for mate
The volume of reformate produced by the reformer was estimated based on the volume of
feed to the reformer as limited by each unit's capacity. The feed to the reformer comes from
various sources depending on the refinery configuration. For virtually all refineries, the heavy
part of the straight run naphtha from the atmospheric crude tower is sent to the reformer, while
the light naphtha is generally processed in the isomerization unit or blended directly into gasoline
as discussed above. Those refineries with a hydrocracker or a coker will send the heavy naphtha
from these units to the reformer as well. This reformate feed naphtha contains the six, seven,
eight and usually the nine carbon compounds from these various sources. In some cases, the six
carbon compounds are separated from the rest of the reformate feedstock to reduce the benzene
in the final reformate. The volume of the feed to the reformer is estimated based on a fractions
of the material processed in the atmospheric crude tower, hydrocracker and coker on a PADD by
PADD basis using information from the LP refinery model.
The fraction of crude oil that is fed to the reformer from the atmospheric crude tower
ranges from about 13 to 16 percent of the crude oil input depending on the PADD. About 18
percent of the material processed in the coker unit is estimated to end up as feedstock to the
reformer. The percentage of the feedstock processed in the hydrocracker that is fed to the
reformer ranges from 30 to 50 percent depending on the PADD in which the refinery is located.
The variance in the fraction of hydrocracker material sent to the reformer is due to the significant
flexibility that the hydrocracker has for producing either gasoline or diesel fuel. In certain
PADDs, such as PADD 4 and 5, there is a higher relative demand for diesel fuel compared to
gasoline so there is a lower conversion to naphtha than in other PADDs. The product from the
reformer experiences a volume decrease of about 18 percent relative to the volume of feed, due
to the conversion of straight chain and cyclical hydrocarbons to energy dense aromatics and
5-16
-------
other light products. This volume reduction and conversion to lighter products increases with the
severity and thus the conversion of the reformer unit. All the refineries in each PADD are
assumed to be operating their reformers at the same severity as estimated by the LP refinery
model. Each of the values discussed in this paragraph are shown on a PADD by PADD basis in
Table 5-13 below.
Table 5-13 Reformer Feed Rates and Volume Loss
Medium/Heavy Straight Run Naphtha
(% of Crude Input)
Medium/Heavy Coker Naphtha
(%ofCokerFeed)
Medium/Heavy Hydrocrackate
(% of Hydrocracker Feed)
Volume Loss in Reformer
PADD 1
13.8%
18.4%
35.4%
18%
PADD 2
16.2%
18.4%
43.4%
17%
PADD3
14.0%
18.4%
50.2%
18%
PADD 4/5a
13.6%
18.4%
33.3%
19%
"Excluding California
5.1.3.1.2.4 Purchased Blendstocks
Some gasoline blendstocks are purchased and blended into gasoline. The gasoline
blendstocks typically purchased include natural gasoline, alkylate, isooctene and ethanol. We
did not have information on the volume of these gasoline blendstocks purchased and blended
into gasoline by each refinery, so we again relied on the information from EIA, which reports the
consumption of these blendstocks on a PADD basis. The EIA information on the amount of
pentane plus, naphtha's and NGLs purchased in each PADD are listed in Table 5-14 below. Our
RFG database has each refiners amount of ethanol blended into RFG, but does not contain the
amount of ethanol that is splash blended into CG at terminals. We accounted for ethanol blended
into CG, as well as the purchase of other gasoline blendstocks, by assuming that each refinery
purchased a volume of any given gasoline blendstock purchased within their respective PADD
proportional to that refinery's crude oil consumption within the PADD. In the 2009 RFG
database, the ethanol volumes only averaged 2.7 percent of refiner's gasoline production, which
results in an over estimation of our refinery and program costs in this NPRM. In our NPRM
analysis, we did not include any desulfurization costs for Pentane plus and naphtha and lighter
blendstocks, since we did not know the extent that they were being treated today. However, the
addition of these blendstocks, results in very negligible increases in demand for additional
naphtha hydrotreating. We will evaluate whether we need to include any costs for treating these
streams in the final rule analysis. If there are costs, because the streams are so small, the costs
would be negligible.
Table 5-14 Refiner Purchases in 2009 (1,000 BPSD)C
Natural Gas Liquids
Naphtha's and lighter
Pentanes Plus
Ethanol b
PADD 1
17.5
24.6
0
57.3
PADD 2
115.9
39.3
41.8
66.9
PADD 3
272.4
53.7
94.9
77.3
PADD 4/5a
15.5
3.8
25.7
0.40
Notes:
5-17
-------
" Excluding California
* Ethanol from EPA RFG database, excluding volumes that are splash blended into conventional gasoline
c Natural Gas Liquids and Pentanes Plus are different names for the same hydrocarbon stream and we
inadvertently found two different volumes for the same hydrocarbon stream and added them both as inputs into
our refinery cost model. We will correct this in the final rulemaking analysis.Butane Volumes
To estimate the butane volumes in our refinery-by-refinery model we used an RVP
balance equation. This equation states that the product of the overall RVP and volume of the
gasoline pool is equal to the sum of the product of the RVP and volumes of the non-butane
components plus the product of the RVP and the volume of the butane blendstocks. This
equation can be rearranged to solve for the volume of butane blendstocks as shown in Equation
5-1 below.
Equation 5-1 RVP Butane Balance Equation
Butane = (A*D-B*D)/(C-A)
Where:
Butane = Volume of Butane added in each refinery inBPSD
A = Blended gasoline RVP average
B = Non-butane blendstock RVP average
C = Butane RVP
D = Volume of gasoline produced
The gasoline production volumes and RVP of the blended gasoline are reported to EPA
by refiners for each refinery and were used for the A and D terms in Equation 5-1. To calculate
the RVP of the butane used as gasoline blendstock we first had to consider the relative
proportion of isobutane versus n-butane being used as a gasoline blendstock as their RVP values
differ. This ratio was estimated on a PADD by PADD basis from the LP modeling work. We
then used a volume weighted average to calculate the RVP of the mixed butane stream blended
into gasoline in each PADD. The information for these calculations is shown in Table 5-15
below. The non-butane blendstock RVP was estimated by multiplying each individual gasoline
blendstock RVP times the gasoline blendstocks volume fraction of each refineries gasoline pool
(CG and RFG) using 2009 ethanol volumes and taking the sum of all of these values. The RVP
value for each of these streams is shown in Table 5-16 below. With this information we were
then able to estimate the volume of butane added to the gasoline blendstock at each refinery.
The annual volumes of butane added by refineries on a PADD level are listed in Table 5-17. The
volume of butane blended into gasoline at each individual refinery varies based on the annual
average gasoline RVP that the refinery produces (the RVP of CG and RFG gasoline are volume
weighted together), as well the variance in gasoline blendstock streams that a particular refinery
uses to produce CG and RFG gasoline.
Table 5-15 PADD Average Composition of Mixed Butanes Added to Gasoline
Isobutane %
N-butane %
Mixed Butane RVP (C), psi
PADD 1
96%
4%
71.376
PADD 2
32%
68%
58.192
PADD 3
53%
47%
62.518
PADD's4/5a
66%
34%
65.196
"Excluding California
5-18
-------
Table 5-16 PADD Average RVP's of Gasoline Blendstocks
LSR
Naphtha from Light Naphtha
Splitter Overhead
Reformate
FCC Naphtha
Coker Naphtha
Isomerate C5
Isomerate C6
Natural Gasoline (NGL)
Polymerization Gasoline
Light Hydrocrackate
Alkylate, C3
Alkylate, C4
Dimersol
Ethanol
PADD 1
12.0
3.0
4.5
4.6
13.0
13.0
7.2
12.6
2.8
9.2
3.6
3.2
5.8
10.7
PADD 2
12.0
3.0
6.6
4.6
13.0
13.0
7.2
12.6
2.8
9.2
3.6
3.2
5.8
10.7
PADD 3
12.0
3.0
5.0
4.6
13.0
13.0
7.2
12.6
2.8
9.2
3.6
3.2
5.8
10.7
PADD's4/5a
12.0
3.0
6.2
4.6
13.0
13.0
7.2
12.6
2.8
9.2
3.6
3.2
5.8
10.7
"Excluding California data
Table 5-17 PADD Average Gasoline Data
Non-butane blendstock RVP (B), psi
Gasoline Pool Volume (D), BPSD
Volume Butane Added, BPSD
Blended Gasoline RVP average (A), psi
PADD 1
6.4
738.2
21.4
8.5
PADD 2
6.3
1744
95.1
9
PADD 3
5.7
3487.8
143.1
8
PADD's4/5a
6.3
500.3
25.7
9.1
"Excluding California data
5.1.3.2 Calibrating the Blendstock Volumes in the Refinery-By-Refinery Model
After calculating gasoline volume estimates for each refinery in the refinery-by-refinery
cost model, we calibrated these values against their reported gasoline blendstock volumes.
Refiners report their production volumes for both conventional and reformulated gasoline to
EPA to comply with the gasoline reporting requirements. We used these reported volumes from
2009, along with LP modeling results from our MSAT2 Rule, to calibrate the refmery-by-
refmery model. Before making any adjustments, the refinery-by-refinery modeling estimates for
gasoline produced on a national basis correlated very well with the reported refinery production
volume in 2009, with volumes differing by less than 0.5 percent from actual production. In order
to eliminate this discrepancy we modified each refiner's yields in our refinery-by-refinery
analysis based on the 2009 data as reported by the refineries. The volume of each of the gasoline
blendstocks, excluding the light straight run (LSR) and FCC gasoline streams, were increased or
decreased proportionally in order to align the aggregated national finished gasoline production
volumes in our refinery-by-refinery model with the aggregated national finished gasoline
production volume reported by US refineries in 2009.
5-19
-------
In making adjustments to the refinery-by-refinery analysis to better align its volumes with
the reported gasoline volumes and sulfur levels, we did not make any changes to the production
volumes of LSR or FCC gasoline. The LSR and FCC gasoline volumes are left unchanged as
these volumes are based on actual refinery-specific FCC charge throughput rates. This accounts
for the effects of FCC feed pretreating on FCC gasoline yields and is therefore likely to
accurately reflect the production volumes for each individual refinery. Additionally, these
volumes are of central importance to our analysis as they are used for estimating the equipment
modifications necessary for complying with new Tier 3 sulfur standards and the costs associated
with additional sulfur removal.
With these calibrated volumes for each of the gasoline blendstocks, the refinery-by-
refmery model can now be used to estimate the sulfur level that refiners must achieve in the FCC
naphtha to meet the current sulfur limit under the Tier 2 standards. These volumes also allow us
to model what modifications to existing equipment and refinery operations will be required to
comply with the new Tier 3 sulfur standards.
5.1.3.3 Refinery Blendstock Sulfur Levels
After determining the volume of each gasoline blendstock stream, we next estimated the
sulfur level of each of the gasoline blendstocks for our modeling analysis using information we
collected from literature reviews and discussions with refinery consultants and technology
providers. We also considered the blendstock sulfur levels estimated for the MSAT2 rule and
the estimates derived from our refinery-by-refinery model to estimate the sulfur levels of the
blendstock streams. Establishing these sulfur levels is important as this sets a baseline for the
refinery-by-refinery model that represents our estimate for the current operations of each
refinery. This allows us to project what changes refiners would have to make in their refineries
to comply with the Tier 3 standards, and project the new investments and operating costs
associated with these changes. The following section contains further details on how the sulfur
content of each of the blendstocks was estimated. The results of this analysis can be found in
Table 5-18 at the end of this section.
The first stream we considered was the butanes that are used as a gasoline blendstock.
The butanes used as gasoline blendstock within a refinery come from a variety of sources. Much
of the butane used as a gasoline blendstock is distilled from the crude oil or other blendstock
streams within the refinery. Refiners remove the butanes from crude oil and sometimes gasoline
blendstocks which contain some butane (i.e., FCC naphtha, hydrocrackate) and then blend them
back into the gasoline up to the RVP or vapor/liquid limit applicable to the gasoline market that
the gasoline is being sold into. During the summer months refiners usually have excess butane
which cannot be blended into the gasoline pool because of the tighter RVP standards. Many
refiners store the excess butanes and then blend them back into gasoline in the winter months
when the volatility limits for gasoline are less stringent. Other sources of butanes used as
gasoline blendstocks are natural gas processers and crude oil drilling operations. The butanes
from these sources are produced in downstream units which separate the various hydrocarbon
components. Most of these downstream units "sweeten" the butanes using a Merox unit prior to
shipping them in pipelines or selling them directly to refiners. The sweetening process reacts the
hydrocarbon mercaptan compounds to disulfide compounds reducing their odor and corrosivity.
The sweetening process, however, does not lower the sulfur level. If the source natural gas well
5-20
-------
is very high in sulfur, the operator may need to use an extractive Merox treatment technology
which actually removes the sulfur from the butane stream. This treatment generally lowers the
sulfur level of the butanes to under 5ppm. Butanes that are blended into gasoline have a sulfur
limit of 30 ppm and those that are shipped through pipelines, regardless of their end use, have a
limit of 140 ppm. Furthermore, many refiners have Merox units on site that are capable of
removing sulfur from butanes that are either purchased or generated internally from refinery
units. We were, however, unable to evaluate existing butane Merox treating capacity at NGL
processers, crude drilling operations, or in refineries as there was no information available in the
OGJ, from EIA, or other publically available sources. Because we do not know the prevalence
of these units, we conservatively assumed in our baseline case that refiners are adding treated
butanes with a sulfur content of 10 ppm to their gasoline pool.F
For hydrocrackate, dimersol, and poly gas blendstock streams, we used the same sulfur
levels that we estimatd for our MSAT2 rulemaking. The sulfur levels for these streams are
inherently low due to the dynamics of process reactions in the hydrocracker, dimersol and
polymerization units. Furthermore, it is unlikely that refiners have altered these processes in
their refineries since our analysis for the MSAT2 rule was completed.
Alkylate blendstocks usually have a small amount of sulfur, usually less than 15 ppm.
The primary source of sulfur in alkylate is the sulfuric acid that is used as a catalyst in the
alkylation process. Finished product coalescers and knockout drums are used by refiners to
remove impurities, including sulfuric compounds, from the alkylate product as it leaves the
alkylation unit. This separation is imperfect, and a small quantity of the sulfuric compounds
which remain in the alkylate account for the majority of its sulfur content. Prior to the enactment
of the Tier 2 standards, the alkylate produced by most refineries contained 10 to 15 ppm sulfur
which assumes that there was some carryover of sulfuric compounds into the alkylate. Based on
our discussions with gasoline desulfurization technology vendors, however, refiners have
installed new acid coalescers and knock out drums in recent years. These new units improved
the removal of residual sulfuric compounds and can produce an alkylate blendstock with a 5-ppm
sulfur level. This adjustment by refiners seems to be a low cost method for reducing the sulfur
content of alkylate. For our refinery-by-refinery baseline analysis, we assumed that refiners have
already installed improved acid knockout drums and are currently producing a 5-ppm alkylate.
We also assumed that Hydrofluoric Acid (FTP) alkylation processes had the same alkylate yield
per feedstock throughput as a sulfuric acid alkylation unit in our refinery by refinery model. We
assumed that the sulfur level of alkylate from an HF units also averages 5 ppm sulfur, even
though HF processing units use hydrofluoric acid as the processing catalyst, instead of using
sulfuric acid.
The coker unit produces a gasoline blendstock with a significant amount of sulfur. These
units convert the heavy portion of crude oil, called residuals, into gasoline and diesel blendstocks
through the use of heat and pressure. The gasoline blendstock produced by the coker can contain
F After we completed our cost analysis, we met with UOP staff, including those who market their Merox technology
for removing mercaptans from gasoline streams. The UOP staff said that pretty much all butanes are already being
treated by Merox (or similar) extraction units. Thus, there would be no additional cost for treating butanes for
complying with Tier 3. We will update our cost analysis to reflect this for the final rule cost analysis.
5-21
-------
more than 3,000 ppm sulfur. This stream is normally split into two different streams. The
stream which contains the six to nine carbon hydrocarbons is processed in the naphtha
hydrotreater, which reduces the sulfur level of this blendstock to below 1 ppm. This stream is
then routed to the reformer for octane improvement. The five and six carbon hydrocarbon
portion of coker naphtha is called light coker naphtha and usually contains on the order of
several hundred ppm sulfur. Because of the instability of this stream due to its high olefm
content, it is generally processed by the naphtha hydrotreater and sent to the isomerization unit if
the refinery has one. After being processed in the hydrotreater, the sulfur content of this stream
is reduced to approximately 1 ppm. These treating pathways were assumed for each refinery in
the refinery-by-refinery baseline analysis.
Straight run naphtha is a gasoline blendstock which contains a moderate amount of
sulfur. Straight run naphtha is the product stream from the atmospheric crude oil tower with a
boiling point that falls within the boiling range of gasoline. The heaviest portion of straight run
naphtha is higher in sulfur relative to the lighter portion of the straight run naphtha. The heavy
portion of straight run naphtha is normally processed by the reformer in order to improve its
octane before being blended into gasoline. After this processing, the reformate has a sulfur level
of less than 1 ppm. The light straight run naphtha (LSR) contains the five and part of the six
carbon hydrocarbons and has on the order of 100 ppm sulfur before any hydrotreating. LSR that
is routed as feedstock to isomerization units has its sulfur lowered to 1 ppm by processing in the
naphtha hydrotreater. This hydrotreating is necessary to allow this material to be processed in
the isomerization unit, as the catalysts in these units require low sulfur feedstocks to function
properly. Some refiners, however, do not have isomerization units or they produce LSR volumes
that are greater than the capacity of their isomerization units. Even cases where there is
insufficient capacity in the isomerization units it is still desirable for refiners to hydrotreat as
much of the LSR as possible since it is more cost-effective to reduce the sulfur content of the
LSR than the FCC naphtha. Refiners can either hydrotreat this volume of LSR in the naphtha
hydrotreaters or in FCC naphtha postreaters.
Natural Gas Liquids (NGL) have a composition that is similar to LSR, as it is comprised
primarily of pentanes and hexanes. NGLs are produced from natural gas processers and crude
oil drilling operations and the sulfur content of the NGLs can vary depending on its source,
although we estimate that this stream averages about 100. While some of the NGLs are treated
to remove sulfur by the NGL producers before being purchased by the refineries we did not have
sufficient information to be able to determine the extent to which NGLs are treated before
arriving at the refinery. For the baseline case in our refinery-by-refinery model we assumed that
NGL liquids are purchased with a sulfur content of 100 ppm and hydrotreated based on capacity
availability at refineries in a similar manner as LSR. Based on the gasoline blendstock volumes
and hydrotreating capacities as discussed in the previous sections, we estimated in the baseline
case for our refinery-by-refinery analysis that refiners are hydrotreating 66 percent of the
volumes of LSR and NGLs produced and purchased for gasoline blendstock usage on a national
average basis. Our hydrotreating capacity evaluation for each refinery is discussed in more
detail in Section 5.1.3.4. As a result of the proposed sulfur standards under the Tier 3 program,
we anticipate that refiners will revamp existing hydrotreaters and add new hydrotreating capacity
to allow them to hydrotreat all of their LSR and NGL material.
5-22
-------
We also assumed that all ethanol blended into gasoline has a sulfur content of 5 ppm.
Ethanol produced at ethanol plants naturally has a negligible amount of sulfur. Before being
sold, however, a denaturant is added to the ethanol. This denaturant most commonly used is
natural gasoline, a C5 to C7 naphtha produced during natural gas processing. Natural gasoline
has a sulfur content that ranges anywhere from a few parts per million to a couple hundred parts
per million sulfur. We assumed that the natural gasoline used as an ethanol denaturant is not
hydrotreated and has an average sulfur level of 250 ppm. Ethanol contains 2 percent denaturant,
which results in denatured ethanol having a sulfur level of 5 ppm.
After determining the sulfur level for each of the gasoline blendstock streams as
discussed above we can use this information, along with the gasoline production volumes and
sulfur levels for the United States in 2009, to determine the sulfur level of the FCC naphtha
stream on a national average basis. To do this we used the following equation, referred to as
Equation 5-2 hereafter:
FCC Naphtha Sulfur ppm = [(A*B) - (C*D+E*F+G*H+I*J+K*L+M*N+O*P+Q*R+S*T)] / Z
Where:
A = Refinery Total Gasoline Yield, BPSD
B = Refinery Total Gasoline Sulfur level, ppm
C = Butane to Gasoline, BPSD
D = Butane Sulfur, ppm
E = Alkylate BPSD
F = Alkylate Sulfur, ppm
G= Reformate BPSD
H= Reformate Sulfur, ppm
I = Coker Naphtha, BPSD
J = Coker Naphtha Sulfur, ppm
K= Hydro-crackate BPSD
L= Hydro-crackate Sulfur, ppm
M= Light Straight Run (LSR) and Natural Gas Liquids (NGL), BPSD
N =LSR and NGL Sulfur, ppm
O= Dimersol, BPSD
P= Dimersol Sulfur, ppm
Q= Polymerization BPSD
R= Polymerization Sulfur, ppm
S= Ethanol, BPSD
T = Ethanol Sulfur, ppm
Z= FCC Gasoline Yield, BPSD
Equation 5-2 Calculating FCC Naphtha Sulfur Content for Refinery-By-Refinery Model
We used this equation to assess two cases; a baseline case where the 30-ppm Tier 2 sulfur
standards were fully implemented and a control case that reflects the proposed 10-ppm Tier 3
sulfur standards. The only terms in Equation 5-2 that change between the two cases are the
national average sulfur level and the sulfur levels of the LSR, NGL, and FCC naphtha streams.
The national average sulfur levels for the two cases were set at the sulfur limits under the Tier 2
5-23
-------
and Tier 3 programs — 30 ppm and 10 ppm, respectively. For the baseline case we assumed that
the sulfur level of the NGL and LSR streams was 34 ppm. This reflects our assessment of how
these streams are currently being handled as discussed earlier in this section. We estimate that
66 percent of the volume of NGL and LSR are hydrotreated before being blended into gasoline
and have a very low sulfur content of approximately 1 ppm. The remaining 34 percent are
untreated and have a sulfur content of approximately 100 ppm. For the Tier 3 control case we
assumed that all of the NGLs and LSR were hydrotreated and therefore had an average sulfur
content of 1 ppm. This information allowed us to solve Equation 5-2 for the FCC naphtha
content. The resulting FCC naphtha sulfur numbers, along with our estimation of the gasoline
blendstock sulfur levels and percent of total gasoline volume made up by each blendstock are
shown in Table 5-18 below.
Table 5-18 Sulfur Levels for Gasoline Blendstocks in the Refinery-By-Refinery Model
Gasoline Blendstocks
FCC Naphtha
Reformate
Alkylate
Isomerate
Butane
Light Straight Run Naphtha (LSR) and
Natural gas Liquids (NGL)
Hydrocrackate
Ethanol
Coker Naphtha
Other Gasoline Blendstocks
Total/Sulfur Average
Baseline Tier 2
Case
Percent
of Total
Volume
37.2
22.5
12.7
3.2
4.0
5.2
3.0
9.9
2.2
0.2
100
Sulfur
Levels
30
ppm
80a
0.5
5
0.5
10
34
8
5
0.5
10
30
Proposed Tier 3
Case Year 20 17
Percent
of Total
Volume
36.0
21.8
12.5
3.1
4.0
4.9
2.9
12.5
2.1
0.2
100
Sulfur
Levels
10
ppm
21a
0.5
5
0.5
10
1
8
5
0.5
10
10
Proposed Tier 3
Case Year 2030
Percent
of Total
Volume
35.0
21.2
12.1
3.1
3.8
4.8
2.8
15
2.0
0.2
100
Sulfur
Levels
10
ppm
21a
0.5
5
0.5
10
1
8
5
0.5
10
10
a These values are calculated using Equation 5-2; all other sulfur levels are assumed
The numbers in the table above represent national averages. While this is useful
information, it is insufficient for us to be able to model the implications of the proposed Tier 3
standards for an individual refinery. Each refinery has a unique combination of processing units
that will determine the cost and operational changes necessary for that refiner to comply with our
proposed sulfur limit. While each of these processing units may impact the cost for refiners to
lower the sulfur content of the gasoline they produce we believe these costs will be dominated by
the units responsible for the desulfurization of the FCC naphtha, and to a lesser extent the NGLs
and LSR. This is because these are the only streams we anticipate would see significant sulfur
reduction under the proposed Tier 3 sulfur standards. The units that are used to desulfurize these
streams include the FCC unit pre- and postreaters and the naphtha hydrotreaters. It is important,
5-24
-------
therefore, to have a good understanding of which of these units are in place in each refinery, as
well as the type and capacity of these units, in order to allow us to most accurately estimate the
cost of the Tier 3 sulfur standards to the refining industry. We used the above FCC naphtha
sulfur balance information as the basis of our vendor request for refiner modifications to FCC
postreaters under Tier 3. However, for the vendor requests, we used a preliminary model, where
the FCC naphtha levels under Tier 2 averaged 75 ppm, while FCC naphtha levels under Tier 3
averaged 25 ppm for 10-ppm sulfur gasoline, representing a 50 ppm sulfur reduction, close to the
same delta presented in the table above.0 The following section discusses our assessment of the
desulfurization equipment currently being used in refineries.
5.1.3.4 Assessment of Refineries' Existing Desulfurization Equipment
Since the desulfurization cost of the Tier 3 program is largely impacted by the cost of
lowering sulfur in FCC gasoline, it is important to understand what refiners are already doing to
lower the sulfur content of the FCC gasoline blendstock to meet the Tier 2 sulfur standards. This
was important to our analysis of the cost for each individual refiner to reduce the sulfur content
of their gasoline to meet the proposed Tier 3 sulfur standard. Refiners that already have an FCC
pretreater or postreater can revamp these units for a lower cost than installing grass roots units.
It was also important to determine which refineries have an FCC feed pretreater, since these units
increase the refineries FCC conversion and production of FCC naphtha and also lower the sulfur
level of the FCC naphtha. To compile this information we analyzed capacity information for
FCC naphtha pretreaters and postreaters for each refinery listed in the OGJ and the EIA database.
If one of the databases showed that a refinery had FCC pretreating and/or post-treating capacity,
while the other did not, we assumed that the refinery did have the units listed with a capacity as
reported.
For refineries that have FCC naphtha postreaters we next determined which vendor's
FCC naphtha desulfurization technology is installed in each refinery. To do this we conducted a
public database search using OGJ, company web postings and, other refinery publications. To
supplement this data we also had extensive discussions with many refiners to obtain confidential
data from many of them on type and capacity of the desulfurization technology currently
installed in their refineries as well as how their operations might be adjusted to meet the new
Tier 3 sulfur standards. The various FCC naphtha desulfurization technologies that we identified
as currently being used by refiners are CD Tech's Cd Hydro and CDHDS, Axens Prime G and
Prime G+, UOP's ISAL and Selectfming, Exxon's Scanfming I or II and Sinopec's S-Zorb. For
refiners that we could not find or obtain information on the type of desulfurization they were
using, Axens was chosen as the default as they have the largest market share of desulfurization
units in the U. S. To confirm the accuracy of our work we reviewed our assessments with one of
the main technology vendors. Our desulfurization technology selection assumptions were
G Because the technology vendors provided us with cost data only for the increment of reducing FCC naphtha sulfur
content from 75 ppm to 25 ppm and in some cases from 75 ppm to 10 ppm, we modeled all refineries, regardless of
their current sulfur level, using the same technology costs. In reality, those with finished gasoline sulfur levels
higher than 30 ppm would have slightly higher costs and those with finished gasoline sulfur levels lower than 30
ppm would have slightly lower costs. We are trying to obtain additional information that would enable us to adjust
our cost analysis to reflect actual refinery starting sulfur levels.
5-25
-------
adjusted based on feedback from the vendor. The aggregated results of this assessment are
summarized in Table 5-19 below.
Table 5-19 Postreater Technologies Used By Refineries
Refiners with Existing
Postreater
CD Tech
15
IFP/Axens
40
Scanfining
16
UOP ISAL
1
S-Zorb
4
The next step of our analysis was to determine which refineries use FCC feed
hydrotreating technology (pretreaters) in addition to post-treating units. FCC feed hydrotreating
was primarily installed at refineries not as a sulfur control technology, but because of the
economic benefits that can be obtained from hydro-treating FCC feed. Hydrotreating the FCC
feed increases the crackability of this stream by saturating the components with hydrogen
resulting in a higher paraffin content in the feed stream. Hydrotreating also removes FCC feed
impurities such as nitrogen, metals, con-carbon and sulfur, which improve FCC unit catalyst
effects. An additional benefit of FCC feed pretreating is that it reduces the sulfur content of the
FCC feedstock by 70 to 90 percent, resulting in the production of FCC naphtha with lower sulfur
levels than what would be produced using FCC feed that is not hydrotreated.
Our analysis indicates that approximately 53 refiners are currently using FCC feed
pretreaters. Of the 53 refineries with pretreaters, 35 of also have FCC postreaters installed to
comply with the Tier 2 gasoline sulfur standard. The technologies used by these 35 refineries are
shown in Table 5-20 below. FCC naphtha produced using only an FCC pretreater operating at
standard severity generally produces a gasoline with a sulfur content that exceeds the Tier 2
standards. According to information from vendors, the average FCC naphtha sulfur level of
refineries with an FCC feed pretreater operating at standard conditions without a postreater
ranges from 200 to 500 ppm. Further reductions in the sulfur level of the FCC naphtha are
possible using only an FCC pretreater by operating the pretreater at a higher severity or higher
pressure (if the unit is designed to do so). These high pressure FCC pretreating units were
designed to be able to run at a high severity to further increase the crackability of the FCC feed
and therefore increase the conversion rate of the FCC unit. These more severe conditions also
further reduce the sulfur level of the FCC naphtha. The naphtha produced from these units
operating with high severity or high pressure has an average sulfur content ranging from 75 to
100 ppm, allowing these refineries to produce gasoline that meets the Tier 2 sulfur standards.
Operating FCC pretreaters at the high severities necessary to meet the Tier 2 standards, however,
also results in increased operating cost, as the pretreater requires more frequent catalyst
changeouts, consumes more hydrogen, and operates higher temperatures than pretreaters
operating under standard conditions.
Table 5-20 Technologies Used By Refiners with FCC Pre and Postreaters
Refiners with FCC Pretreater
and Naphtha Postreater
CD Tech
9
IFP/Axens
20
Scanfining
5
UOP ISAL
0
S-Zorb
1
Our analysis also showed that there are several refineries that have an FCC unit but have
installed neither an FCC naphtha postreater nor an FCC feed pretreater. These are small
5-26
-------
refineries, or refineries that produce a refinery gate gasoline with a sulfur level below the Tier 2
cap of 80 ppm sulfur, but above the 30-ppm average. These refiners are relying on buying or
sharing sulfur credits from other refineries that are over-complying with Tier 2 and make
gasoline with a sulfur level less than 30 ppm.
Finally, some refineries do not have an FCC unit and therefore have not installed FCC
postreaters to comply with the Tier 2 sulfur standards. These refiners primarily use reformate,
alkylate, LSR, butanes, and pentanes to make gasoline. Since these blendstocks all have low
sulfur content this allows refiners to produce gasoline with a low enough sulfur content to meet
the Tier 2 sulfur standards.
A summary of the number refineries which fall into differing categories of how they are
complying with Tier 2 is shown in Table 5-21 below.
Table 5-21 Refinery FCC Naphtha Desulfurization Unit Characterization
FCC Treatment Units Installed
No FCC Unit
FCC Unit, No Pretreater or Postreater
FCC Unit With Postreater Only
FCC Unit With Pretreater Only
FCC Unit With Pretreater and Postreater
Number of Refineries
14
7
38
17
35
5.1.3.5 Crude Oil and FCC Feed Sulfur Levels
After we had determined the desulfurization technology in place at each refinery, we
sought to calculate the sulfur content of the FCC feedstock. Knowing this is important as it
allows us to determine how far the sulfur level of the FCC naphtha, and ultimately the gasoline,
produced at any given refinery can be reduced using the units currently in place at each refinery.
It also helps us understand the extent to which the existing hydrotreaters are being taxed to
comply with the Tier 2 gasoline sulfur standard. Some refineries may have excess capacity in
their FCC naphtha pretreater or postreaters that would allow them to produce gasoline that would
meet the proposed Tier 3 standards without having to revamp existing units or add grass roots
units. These refineries will have much lower cost impacts than refineries that have to make more
significant capital investments.
The sulfur level of the FCC feedstock is primarily dependent on the sulfur level of the
crude oil being processed by the refinery and whether or not the refinery has an FCC feed
pretreater. The first step, therefore, in determining the sulfur level of the FCC feedstock was to
input the crude sulfur level for each refinery into our refinery-by-refinery model. For this, we
obtained confidential business information (CBI) from EIA on the annual average crude sulfur
levels that each refinery processed in 2009. This data, which is reported to EIA for each refinery,
was used as the baseline crude sulfur level in our refinery-by-refinery analysis. Using this data,
we then determined what each refiners FCC feed charge rate sulfur level would be, using a
regression co-relation we built from data on crude sulfur levels and FCC feedstock material, as
discussed below. We assumed that refineries today are primarily processing heavy gas oils
5-27
-------
(HGO) and vacuum gas oils (VGO) produced from each refinery's crude and coker units. To
determine the volume of feedstock processed by the FCC units we assumed that after distillation
HGO makes up 20.5 percent (by volume) of the processed crude and VGO makes up an
additional 15.6 percent of the crude.6 Together, these two streams comprise the FCC feed.
The boiling point range that we assumed for VGO also contained some residual material,
representing FCC feed with residual content. This was done to reflect the imperfect distillation
cuts in crude towers and that some refiners use small amounts of residual material as FCC
feedstock. The balance of the residual material, however, was excluded from the feed to FCCs
since this material makes a poor feedstock due to its high aromatics, metals and concarbon
content. Each of these materials negatively affects the FCC gasoline conversion yields. Most
refiners today do not directly use residuals as feedstock to their FCC units, but instead send them
to be processed in coker units or use the residual material for fuel oil and asphalt production.
The boiling point ranges that we used for HGO and VGO are listed Table 5-22 below.
Table 5-22 Boiling Ranges of FCC Feedstocks
Heavy Gas Oil (HGO)
Vacuum Gas Oil (VGO)a
TBP Initial
600°F
800°F
TBP Final
800°F
1,000°F
a Contains some residual material
For our FCC feed sulfur regression, we used data obtained various crude oil assays that
we obtained from Jacobs Engineering for work that Jacobs conducted for us. We used data from
five specific crude types, including West Texas intermediate, Bonny Light, Saudi Heavy,
Alaskan North Slope, and Mayan, and three blended crude assays. The equation for this
regression, along with the estimated FCC feed sulfur contents for various crude oils are shown in
Equation 5-3 and Table 5-23 below.
Equation 5-3 FCC Feed Sulfur Content Based on Crude Sulfur Content
FCC Feed Sulfur Weight Percent = (Crude Sulfur Weight Percent)0'8 * 1.1858 + 0.0409
5-28
-------
Table 5-23 FCC Feed Sulfur For Various Crude Sulfur Levels
Crude Sulfur Level (Weight %)
0.11%
0.28%
0.85%
1.3%
1.4%
1.6%
2.8%
3.04%
FCC Feed Sulfur Level
(Weight %)
0.24%
0.47%
1.08%
1.50%
1.59%
1.77%
2.74%
2.93%
5.1.3.6 Impact of FCC Feed Pretreaters on FCC Feed Sulfur Levels
With FCC feed sulfur estimated, the next step in our analysis was to consider the impact
of pretreaters on the FCC feed sulfur levels for refineries that have these units. There are several
factors that must be considered to determine the impact of the pretreaters on the FCC sulfur
level, including the pressure at which the unit operates, the severity at which it is run, and
whether or not the FCC naphtha will be postreated.
To inform our understanding of how FCC pretreaters operate, we obtained guidance from
technology and catalyst providers. From these discussions we learned that the capability for
FCC pretreaters to remove sulfur from the gas oil feed varies significantly depending on the
pressure at which the unit operates. FCC pretreaters can generally be subdivided into high
pressure units (1400 psi and above), medium pressure units (900 to 1,400 psi), and low pressure
units (below 900 psi). High pressure FCC pretreaters are capable of removing about 90 percent
of the sulfur contained in the gas oil feedstock to the FCC unit, while low and medium pressure
units are capable of removing 65 to 80 percent of the feed sulfur. Information we received from
the vendors also indicated that refiners with both a pretreater and a postreater are producing FCC
naphtha that ranges from 200 to 450 ppm before being processed by the postreater. Having a
postreater allows these refineries to not have to operate their pretreaters at a high severity as the
sulfur will further be reduced to levels necessary to meet the applicable standards in the post-
treating units. Refiners with only a pretreater are making lower sulfur FCC naphtha in the 75 to
100 ppm range, according to vendor estimates.
With this information we used our refinery-by-refinery model to estimate the pretreater
desulfurization rates required to get FCC naphtha sulfur levels within the ranges specified. We
estimated that FCC units with a pretreater and a naphtha postreater are operating their pretreaters
at a severity which results in a 76 percent desulfurization of the FCC feed stream. This number
represents the national average. While the actual severity at which the pretreating units are run
varies on a refinery-by-refinery basis this average was used in our modeling for all refineries
with both pretreating and postreating units due to a lack of refinery-specific information. For
FCC units with a feed pretreater but no postreater we calculated the FCC naphtha sulfur level
required by refiner to make a refinery gate gasoline that meets the Tier 2 standard. To do this
calculation we used the gasoline yields from our refinery-by-refinery model along with the
gasoline blendstock sulfur levels discussed above and shown in Table 5-18. These calculations
showed that refiners with FCC feed pretreating units, but no postreaters, need to produce FCC
5-29
-------
naphtha that averages about 85 ppm on a national level. This sulfur level corresponded to these
refiners operating their pretreaters at a severity that results in a reduction of sulfur in the FCC
feed stream of approximately 91-92 percent. This number is close to the estimate we received
from the vendors for this category of refineries and therefore was used in our refinery-by-
refinery model to determine the FCC feed sulfur level for refiners with pretreaters.
After we have calculated the sulfur level of the FCC feed we must then take into
consideration the impact the FCC unit itself has on the sulfur level of the FCC naphtha. We
"7 o
reviewed several literature sources ' and found that the FCC naphtha sulfur level can be
accurately determined by dividing the FCC feed sulfur level by 20 for refineries with an FCC
feed pretreating unit. For refineries without an FCC feed pretreater, the FCC naphtha sulfur
levels can be calculated by dividing the desulfurized FCC feed sulfur level by 10. In these cases
the effect of the FCC unit itself on the sulfur level of the FCC naphtha is lower, as the FCC feed
has already been through a desulfurization process. These factors, when combined with the
sulfur levels of the FCC streams as discussed above, allow us to calculate the sulfur level of the
FCC naphtha before any post-treating operations on a refinery-by-refinery basis. The results of
this analysis are summarized in Table 5-24 below.
Table 5-24 FCC Naphtha Sulfur Levels for Various Refinery Configurations
No Pretreater or Postreater
Pretreater Only
Postreater Only
Pretreater and Postreater
PADD 1
N.A.a
N.A.a
1348
49
PADD 2
-<3b
73
1157
<3b
PADD 3
606
74
705
278
PADD 4/5
784
82
1036
230
" N.A. - not applicable, no units of this type in the PADD
* Since there are less than three refiners in this PADD with the described configuration, the data was
removed to protect potential CBI concerns.
This information, along with the information described in previous sections (e.g.,
gasoline blendstock volumes and sulfur levels, desulfurization equipment currently in place at
refineries, and crude oil sulfur levels) allows us to conduct the best analysis for the baseline case
for our refinery-by-refinery model. This baseline case reflects what we believe the current
operating conditions are at each refinery, including any modifications they have made to meet
the Tier 2 sulfur standards. The next step of our analysis was to project what further changes,
either to operations, adding new equipment or revamping existing units, each refiner would have
to make to meet the proposed Tier 3 standards. After these changes are estimated, we can then
estimate the cost associated with each of these changes, and ultimately the cost of the program.
5.1.4 Cost Inputs for the Sulfur Control Technologies
After we determined the sulfur levels of the gasoline blendstocks for each refinery and
the sulfur levels that these blendstocks would have to achieve to meet the proposed Tier 3 sulfur
standards, the next step in our refinery-by-refinery analysis was to project the changes to refinery
units and operations each refinery would have to make in order to comply with the proposed Tier
3 sulfur standard. The first step refiners would take for further gasoline sulfur control would be
to desulfurize the light straight run, natural gas liquids, and butane blendstocks. The costs to
reduce the sulfur content of these streams is relatively low and would therefore be the most cost
5-30
-------
effective way to further reduce the sulfur content of the finished gasoline. We also projected in
our analysis for RFS2 that 50 percent of all gasoline produced by refiners in 2017 would contain
15 percent ethanol. In 2030 we projected in our analysis for RFS 2 that almost all gasoline
contains 15 percent ethanol.H Because ethanol tends to be a relatively low sulfur blendstock
(assumed to be 5 ppm for our refinery-by-refinery model) increasing the amount of ethanol in the
gasoline pool lowers the overall sulfur content of the gasoline. Reducing the sulfur content of
the LSR, NGL, and butane streams and increasing the ethanol content, however, would be
insufficient to allow refiners to comply with the proposed Tier 3 standards. Refineries with an
FCC unit would still have to reduce the sulfur content of their FCC naphtha blendstock in order
to meet the Tier 3 standards.
For each refinery we considered two cases. In the first case each refinery had to meet the
proposed Tier 3 gasoline sulfur standard of 10 ppm. In order to meet this standard, as discussed
in 0, we determined that they would have to reduce the sulfur level of their FCC naphtha stream
to 25 ppm. We also considered a case where each refinery would reduce the sulfur level of their
gasoline to 5 ppm. We assumed that to meet a 5-ppm sulfur standard, refiners would desulfurize
the butane blendstock to lOppm in addition to further lowering sulfur in the FCC naphtha to 10
ppm. This information was used to help us determine which refineries might reduce the sulfur
level of their gasoline below our proposed 10-ppm standard to earn credits for our ABT
scenarios discussed in Sections 5.2.1.2 and 5.2.1.3.
Our refinery-by-refinery model assumed that reducing the sulfur content of the FCC
naphtha to 25 ppm and 10 ppm for the two cases discussed above would require that each
refinery that produces FCC naphtha have an FCC naphtha postreater. In our final rule we may
also investigate the ability of refineries with high pressure pretreaters to meet the required
standards without the addition of postreaters. For companies that already have an FCC naphtha
postreater we assumed that all that would be necessary to meet the proposed sulfur standards was
to revamp their existing FCC postreating units. We received cost information from several
vendors for revamping FCC postreating units and assumed a revamp cost for each refinery in
line with the cost projections quoted by the vendor of the technology already in place in their
refinery. We assumed that refineries with FCC units that currently do not have an FCC
postreater would have no choice other than to add a new grass roots FCC postreating unit. We
ultimately only received cost information from one vendor for the cost of adding a new grass
roots FCC postreating unit that fit the sulfur reduction requirements of the proposed Tier 3
program. We therefore assumed the cost for each refinery that would need to add a new FCC
post-treating unit would be in line with this estimate. More details on the costs used in our
refinery-by-refinery model for the desulfurization of LSR, NGL, and butanes as well as the new
FCC post-treating units and revamps can be found in the following sections.
H In our Tier 3 cost analysis, we inadvertently only accounted for the ethanol volumes that are contained in the 2009
RFG database, not those mandated by RFS 2. For our final rulemaking cost analyiss, we will adjust our model to
account for higher ethanol use in the future, which will slightly reduce our Tier 3 cost estimates, due to the dilution
effects of ethanol.
5-31
-------
5.1.4.1 FCC Naphtha Desulfurization Costs
To estimate the cost for revamping existing FCC postreating units or for adding new
postreating capacity, we contacted several technology vendors for cost estimates and reviewed
literature, including cost information provided for the Tier 2 rulemaking. Because no two
refineries are exactly the same, the cost for new FCC postreater units or revamps to existing units
will vary significantly from refinery to refinery. Some of the factors that have the most
significant impact on the cost of FCC postreaters are the technology that the refiner used to
comply with Tier 2, the volume of FCC naphtha, the sulfur content of the FCC unit feed and the
level of desulfurization in the existing postreater, and the location of the refinery. Based on
feedback from vendors we considered three categories of FCC postreaters based on whether the
FCC naphtha (the feed for the existing Tier 2 postreater) contained low (0 - 400 ppm) medium
(400 - 1,200 ppm) or high (>1,200 ppm) levels of sulfur. Specific cost factors applicable to
estimating unit revamps or grass units were also taken into account as described below. The
following sections discuss in greater detail how the cost estimates we received from the vendors
were used in our refinery-by-refinery analysis.
5.1.4.1.1 Cost to Revamp Existing FCC Naphtha Postreaters
We obtained information from several technology providers for the revamp costs of
existing FCC postreaters. One of the technology providers, however, declined to provide us with
information applicable for Tier 3 sulfur control; they merely provided us with historical
information for controlling sulfur from pre-Tier 2 uncontrolled sulfur levels. As a result, their
capital and operating costs were extraordinarily high relative to the rest of the cost information
we received from the other technology providers that was specific to Tier 3. In addition, the
information they provided did not contain sufficient detail to enable us to adjust it to be of
relevance for Tier 3. Given the magnitude of their cost estimate, we also believe it is likely that
their cost estimate represented not only a grass roots FCC postreater, but also significant refinery
investment in other refinery processes such as FCC feed pretreating, coker unit expansion, etc.
Thus, we were unable to use this information to estimate the revamp costs for FCC post treating
for the refineries that employ that company's technology, and our peer reviewers concurred. The
other technology providers submitted information applicable to Tier 3, but not necessarily
covering all of the scenarios refineries would experience. Therefore, as discussed below, we
used and adjusted the information we were provided as necessary to apply it across all the
refineries. Some of the submitted information only had cost information for our medium sulfur
(400 - 1,200 ppm) FCC feed case. Another technology provider did not provide cost estimates
for producing an FCC naphtha with a sulfur level of 25 ppm, corresponding to a finished
gasoline with 10 ppm sulfur, therefore, we needed to interpolate their cost information. Because
the cost information provided by the technology providers was labeled CBI, this cost information
cannot be listed individually, however we aggregated the cost information we received for FCC
postreater revamps to meet 10 ppm and 5 ppm sulfur levels in gasoline. The aggregated
information is summarized in Table 5-26 and Table 5-27.
One of the vendors we contacted for a cost estimate for FCC naphtha desulfurization
technology provided information for several potential FCC postreater revamp cases. The first
case was a no capital costs case where refiners made no equipment modifications, but rather
solely made operational changes using their existing equipment installed for Tier 2. The second
5-32
-------
case we requested was one where refiners would incur only minor capital costs and was intended
to be used for analyzing program options with moderate octane costs. The third case we
requested was one where refiners were willing to incur greater capital costs in order to minimize
operating costs and octane loss. The majority of the vendors only supplied cost estimates for the
third case, which included adding an additional catalyst reactor bed to the existing FCC
postreater unit. This ensures that refiners will be able to run their existing FCC postreater at 4 to
5 year catalyst cycle lengths, which is a critical feature for FCC unit operations.
The costs for the FCC postreater revamps submitted by one of the vendors, however
showed that for low (0 - 400 ppm) and medium (400 - 1,200 ppm) sulfur FCC naphtha sulfur
levels, second case, with low capital costs, resulted in the lowest cents per gallon costs for
meeting the proposed 10-ppm Tier 3 standards. According to this vendor these cases also had a
4 to 5 year catalyst cycle length, equivalent to the higher capital cost cases even though a second
stage reactor was not required. We therefore assumed that refineries using this vendor's
technology would choose the minor capital cost pathway for meeting the 10-ppm Tier 3 standard
when they had low or medium sulfur levels in their FCC feed. The high capital cost cases for
producing gasoline to meet a 5 ppm sulfur standard from low and medium sulfur FCC feeds were
found to have the lowest cost on a cents per gallon basis and were therefore selected by our
model for these cases.
One vendor only submitted information for postreater revamp cost estimates for FCC
naphtha in the 400 - 1,200 ppm sulfur category that produced a 5-ppm sulfur gasoline. In our
refinery-by-refinery model, however, we had multiple refineries with FCC feed sulfur levels in
the 0 - 400 and > 1,200 ppm categories that use this vendor's postreating technology. In order to
apply this vendor's cost estimate to cases of low (0 - 400 ppm) and high (>1,200 ppm) sulfur
feed categories we adjusted this vendors 400 - 1,200 ppm postreater revamp cost based on the
cost differentials between the three FCC naphtha sulfur levels in the other vendors' revamp
estimates. We similarly derived a postreater revamp cost estimate to produce a 10 ppm gasoline
for this vendor using cost differentials between the 10 and 5 ppm cases from other vendors. For
refineries currently employing technology by other vendors for which we had no specific cost
information, we used an average of all of the vendors' estimates to represent FCC postreater
revamp costs for refiners using this particular technology in our refinery-by-refiner model.
After we had determined cost estimates for the FCC postreater revamps based on
information from the vendors the next step was to scale these costs based on the size of the FCC
postreating unit present in each refinery. The vendor estimates submitted for revamp costs were
based on various FCC postreater design volumes ranging from 10,000 BPSD to 45,000 BPSD
depending on the base unit size used by the vendor. To determine how to apply these vendor
costs to each refinery, we first calculated each refinery's maximum FCC naphtha production.
The maximum production was derived by assuming each refiner runs its FCC unit at its
maximum nameplate throughput capacity (barrels per stream day) with the FCC naphtha yield
rates discussed in Section 5.1.3.1.2. We then increased the size of the FCC postreater by 7.5
percent above each refinery's maximum FCC gasoline production rate as an over design factor to
account for excess capacity that refiners generally design into their unit for processing additional
flows of FCC naphtha (i.e, rerunning high sulfur FCC naphtha batches either from that refinery
or from neighboring refineries). After sizing the FCC postreater that would be required for each
refinery we then scaled the costs given by the vendors using the six-tenths rule as shown below.
5-33
-------
This is a "rule of thumb" cost estimating tool commonly used for cost estimating by the refining
and petrochemical industries for estimating the cost of a process unit based on a similar unit of
differing size.
Cost to Revamp an FCC Unit= A * (B/C)0'6
Where:
A = Cost Estimate Received from the Vendor
B = Size of the FCC Unit in the Refinery
C = Size of the FCC Unit in the Vendor's Estimate
Equation 5-4 Six-Tenths Rule for Estimating Capital Cost
We also adjusted the costs submitted by the technology providers based on the location of
each refinery. We assumed that each vendor's estimate was based on revamping an FCC
postreater in PADD 3 (Gulf Coast), which is the lowest cost region for installing new capital in
refineries. The cost for refineries that are not located in PADD 3 were adjusted upwards based
on a ratio of the cost of refinery capital projects in the PADD in which they are located relative
to PADD 3. An additional factor was applied to account for the "offsite" costs that are incurred
when installing new capital in refineries. When vendors provide a cost estimate for their
technology, this estimated cost is called the inside battery limits (ISBL) cost and it is solely for
the immediate unit of interest. However, refiners may need to install peripheral capital to
support the new unit, such as electrical switchgear, a control room, storage for feed, intermediate
or unit products, and longer than anticipated pipeline runs - these costs are usually considered
Outside battery limit (OSBL) costs, or offsite costs. In some cases, OSBL costs may include
hydrogen and sulfur plant costs, although, for our analysis, we separately estimated the cost for
providing additional hydrogen and for processing the removed sulfur and included this cost in
our cost analysis. We estimate that the grassroots offsite cost factors for very modest levels of
desulfurization inherent in this analysis are in the 1.2 to 1.3 range, and those for these very
simple revamps are in the 1.1 to 1.15 range (the ISBL costs are multiplied by these factors to
derive a total cost). These cost factors, as well as the utility prices that we used in our refinery -
by-refinery cost model, are shown in Table 5-25 below:
Table 5-25 Cost Factors for Various PADDs
Capital Cost Factor
Natural Gas ($/MMBTU)
Electricity (e7kW-hr)
Steam ($/l, 000 Ib)
Offsite Capital Cost Factor -
New Units
Offsite Capital Cost Factor -
Unit Revamps
PADD 1
1.5
7.88
7.98
12.9
1.26
1.13
PADD 2
1.3
7.86
5.47
12.6
1.26
1.13
PADD 3
1.0
5.64
5.78
9.20
1.20
1.10
PADD 4
1.4
7.06
4.69
11.6
1.30
1.15
PADD5a
1.2
7.44
8.36
11.9
1.30
1.15
a Excluding California
The volume-weighted cost estimates for revamping FCC postreaters across the entire
refining industry as calculated by our refinery-by-refinery model are shown in Table 5-26 and
Table 5-27 below. These costs are aggregated cost estimates for the FCC revamp costs used in
5-34
-------
our refinery-by-refinery model. In our model, we paired vendor cost data with refineries that are
already using that particular vendor's technology for their FCC postreating units. We further
tailored the information provided by the vendors to match the specific refinery configuration to
the extent possible. We assumed that the data provided by the vendors includes the cost for
complying with the applicable stationary emission standards and that any incidental costs for
permitting, if necessary, are negligible and covered by the offsite factor. The information that
we received from the vendors and the individual refinery capital costs, however, cannot be
shown due to CBI concerns.
Table 5-26 Revamp Cost for a 30,000 BPSD FCC Postreater for 10-ppm Standard
FCC Feed Sulfur Level*
Capital Cost ($/B ISBL)
Hydrogen (scf/bbl)
Fuel Gas (kBTU/bbl)
Electricity (kWh/bbl)
Octane Loss (R+M)/2
Olefm Decrease (vol%)
Catalyst Cost ($/B)
Steam (Ib/bbl)
0 - 400 ppm
235
36.8
15.5
0.14
0.56
2.67
0.01
0.61
400 -1,200 ppm
382
25.9
15.4
0.27
0.46
2.14
0.01
30.88
>l,200ppm
505
25.9
8.51
0.52
0.42
1.97
0.01
0.61
Volume Weighted
Average
265
31.3
14.4
0.223
0.492
2.39
0.010
13.44
a $/B = dollars per barrel, scf/bbl = standard cubic feet per barrel; kBTU/bbl = thousand BTU per barrel; kWh/bbl =
kilowatt-hours per barrel; (R+M)/2 = (research octane + motor octane)/2; vol% = volume percent; $/B = dollars per
barrel; Ib/bbl = pounds of steam per barrel of feed.
b Of the refineries that are expected to revamp their FCC naphtha hydrotreater for the no ABT case, 36 have FCC
naphtha sulfur levels in the 0 - 400 ppm range, 20 have FCC naphtha sulfur levels in the 400 - 1200 ppm range and
13 have FCC naphtha sulfur levels greater than 1200 ppm.
Table 5-27 Revamp Cost for a 30,000 BPSD FCC Postreater for 5 ppm Standard"
FCC Feed Sulfur Level
Capital Cost ($/B ISBL)
Hydrogen (scf/bbl)
Fuel Gas (KBTU/bbl)
Electricity (kWh/bbl)
Octane Loss (R+M)/2
Olefm Decrease
Catalyst Cost ($/B)
Steam (Ib/bbl)
0 - 400 ppm
400
36.2
22.7
0.30
0.60
2.81
0.00
23.61
400 -1,200 ppm
578
52.6
14.6
0.52
0.77
3.38
0.01
30.88
>1,200 ppm
600
56.3
13.2
0.52
0.90
3.96
0.00
0.61
Volume Weighted
Average
444
38
17.2
0.42?
0.62
2.37
0.004
26.9
a Assumes every refinery is complying with a 5 ppm gasoline sulfur standard
We also found that there were 11 refineries that had an existing FCC postreaters that
were not sized large enough to process their maximum FCC naphtha production volume. For
these refineries we assigned additional capital costs to debottleneck the existing first stage
reactor in order to increase the postreater capacity so that it could accommodate maximum FCC
naphtha production. For each refinery with an existing unit that could not process more than 70
percent of our estimate of a refiner's maximum FCC naphtha production we added capital costs
to revamp and expand the first stage to increase its capacity to allow the postreater to process
5-35
-------
100 percent of its maximum FCC naphtha rate. For the capital costs for this debottlenecking we
used 35 percent of the cost of a new grass roots unit (discussed below) for the volume of the
expansion. We once again used the six-tenths rule to adjust the capital cost for the volume
expansion needed versus the cost for the 30,000 BPSD grass roots treater used for technology
vendor estimates.
5.1.4.1.2 Cost for Grassroots FCC Postr eaters
While all refineries that already have FCC postreaters should be able to meet the
proposed Tier 3 standards by revamping their existing postreaters, refineries that do not currently
have an FCC postreater would have to add a grass roots FCC postreater. To determine the cost
of building grass roots FCC postreating units at a refinery we similarly requested cost estimates
from vendors. Only one of the vendors that supplied FCC postreating equipment submitted
information on the cost of a grass roots FCC postreating unit for desulfurizing FCC naphtha with
a feed sulfur content of 100 ppm. Based on the calculation methodology shown in Equation 5-2,
we estimate that refineries that require a grass roots postreater will already have an FCC feed
sulfur level that averages between 85 and 100 ppm as these refineries already have FCC feed
pretreaters. The other grassroots vendor estimate we received, as well as those we received for
Tier 2, represented a grass roots postreater with an FCC feed sulfur content of about 800 ppm.
These estimates were deemed to be not representative of the costs to refineries that would be
installing grass roots postreating units as the capital, hydrogen, and other operating costs would
be much higher for an FCC feed sulfur of 800 ppm vs. 100 ppm. We did not consider this other
vendor's cost estimate for a grass roots postreater and therefore relies on a single vendor's cost
estimate for grass roots FCC postreating units for the Tier 3 program. In our FRM analysis we
attempt to incorporate additional information from other vendors and literature sources. This one
vendor cost estimate seemed reasonable relative to the other cost data that we have for higher
levels of desulfurization.
The vendor estimate submitted for a grass roots postreater was based on a postreater with
a capacity of 30,000 BPSD capable of producing an FCC naphtha with a sulfur level of 10 ppm,
corresponding to a gasoline sulfur level of 5 ppm. To scale the cost submitted by the technology
vendor to be applicable to a specific refinery, we used a similar methodology to that which was
used for postreater revamps. We first determined the appropriate size for each unit based on
each refiners maximum FCC naphtha production rate, adding 7.5 percent above each refinery's
maximum FCC gasoline production rate as an over design factor. We then used the six-tenths
rule (Equation 5-4) to scale the cost reported by the vendor up or down as appropriate based on
the relative volume of the grass roots unit required by the refinery and the size on which the
vendor's cost estimate was based. We once again assumed that the capital cost from the
technology vendor was representative of a refinery in PADD 3 complying with the applicable
stationary emission standards. We then adjusted the cost based on the cost of refinery capital
projects in the PADD in which they are located relative to PADD 3. Finally, we used a new unit
offsite adjustment factor as listed in Table 5-25 to determine the final cost of a grass roots FCC
postreater for each refinery. The costs to produce an FCC naphtha with a sulfur level of 25 ppm
(corresponding to a 10-ppm gasoline) were estimated based on the grass roots postreater unit that
makes FCC naphtha for the 5-ppm standard. These costs are summarized in Table 5-28 and
Table 5-29 below.
5-36
-------
Table 5-28 Cost for a 30,000 BPSD Grass Roots FCC Postreater for 10-ppm Standard
FCC Feed Sulfur Level
Capital Cost ($/B ISBL)
Hydrogen (scf/bbl)
Fuel Gas (KBTU/bbl)
Electricity (kWh/bbl)
Octane Loss (R+M)/2
Olefin Decrease
Catalyst Cost ($/B)
Steam (Ib/bbl)
100 ppm
1500
62.1
4.95
1.06
0.55
2.65
0.04
20.0
Table 5-29 Cost for a 30,000 BPSD Grass Roots FCC Postreater for 5-ppm Standard
FCC Feed Sulfur Level
Capital Cost ($/B ISBL)
Hydrogen (scf/bbl)
Fuel Gas (kBTU/bbl)
Electricity (kWh/bbl)
Octane Loss (R+M)/2
Olefin Decrease
Catalyst Cost ($/B)
Steam (Ib/bbl)
100 ppm
1500
133
9
1.06
1.00
5.15
0.04
20.0
5.1.4.2 Light Straight Run and Natural Gas Liquids Desulfurization Costs
Another action refiners may need to take to reduce the sulfur content of their gasoline is
to desulfurize their light straight run naphtha (LSR) and natural gas liquids (NGL) blendstocks.
While these blendstocks have lower sulfur contents than the FCC naphtha, they may be cheaper
to desulfurize for refineries that are not already treating these streams. Many refineries can
desulfurize some or all of these blendstocks using existing excess capacity in their naphtha
hydrotreaters or FCC naphtha postreaters. Further, as opposed to hydrotreating FCC naphtha
which contains olefms, the LSR and NGL blendstocks contain no olefms and therefore,
hydrotreating them does not result in octane loss and has a lower hydrogen consumption. The
combination of the potential for using excess capacity in existing units and low operating costs
result in the relatively low desulfurization costs for the LSR and NGL blendstocks. From our
discussions with refiners, several refineries indicated that they would install new standalone
hydrotreaters for processing LSR and NGL blendstocks, though it is unclear which other
refineries will have to add equipment to desulfurize LSR and NGL. For our Tier 3 cost
estimates, however, we have conservatively used our refinery-by-refinery model to estimate the
costs for other refiners to hydrotreat all refiner volumes of LSR and NGL by installing additional
hydrotreating equipment. To determine the cost to desulfurize the LSR and NGL blendstocks we
first had to determine the volume of blendstock that requires desulfurization. Our determination
of the quantity of LSR and NGL used as gasoline blendstock at each refinery is discussed in
Section 5.1.3.1.2. From this total we then subtracted the volume of LSR processed in the
5-37
-------
isomerization unit. Because the isomerization units require a low sulfur feedstock we assume all
the feed to this unit is treated by the naphtha hydrotreater.
The next step in our assessment of the desulfurization costs of the LSR and NGL
blendstocks was to estimate the extent to which these blendstocks are already being treated at
refineries to meet the existing Tier 2 sulfur standards. Based on our discussion with refining
consultants, vendors and refiners, it appears that in response to the Tier 2 standards refiners have
altered their operations to use excess capacity in their FCC naphtha postreaters and naphtha
hydrotreaters to reduce the sulfur content of LSR and NGL blendstocks. Since information on
the extent to which these streams are currently being hydrotreated is not publicly available we
estimated these volumes using the capacities of the FCC postreaters and reformer feed
hydrotreaters under normal refiner crude throughputs and yields from the refinery-by-refinery
model.
We evaluated each refinery's capacity to hydrotreat LSR and NGL using existing
equipment by first determining the volume that can be processed in their naphtha hydrotreaters.
We assumed that every refinery's production volume of straight run and coker naphthas, as
determined in Section 5.1.3.1, are processed in the refinery's reformer feed hydrotreater and that
any capacity up to 85 percent of the unit's maximum capacity can be used to treat LSR or NGL.
If a refinery had insufficient excess capacity in their naphtha hydrotreater to treat all of the LSR
and NGL volumes we next determined if there was excess capacity in that refinery's FCC
postreater. We allowed LSR and NGL to be processed using excess FCC postreating capacity in
refineries where the capacity of the FCC postreater exceeds 120 percent of that required to
process a refinery's maximum FCC naphtha yield as determined by the refinery-by-refinery
model. Only several refiners had excess FCC postreating capacity available for the treating of
LSR or NGL feedstocks, as the capacity of most FCC postreaters was less than 120 percent of
the maximum FCC naphtha production rate. We assumed that refiners are currently using any
excess hydrotreating capacity in their naphtha hydrotreating and FCC postreating units to
desulfurize LSR and NGL in response to the Tier 2 sulfur standards. To meet the proposed Tier
3 sulfur standard 10 ppm and the 5 ppm level we considered for the ABT cases, we assumed that
refiners would desulfurize all of the LSR and NGL blendstocks. We assumed that LSR and
NGL purchased by refiners would be hydrotreated in addition to that processed by the
isomerization unit and that blended directly into the gasoline. To determine which of the
refineries in our refinery-by-refinery model would need to add hydrotreating capacity, we
evaluated each refinery based on their maximum crude processing rate and yields. We used our
model to increase each refinery's yield of coker naphtha, reformer naphtha, isomerate, LSR and
NGL to correspond to a crude utilization rate of 100 percent. We then used these maximum
production volumes to evaluate whether or not each refinery could process all of the LSR and
NGL using excess naphtha hydrotreater and FCC postreating capacity. If a refinery did not have
sufficient excess hydrotreating capacity for all of the LSR and NGL in these units we assumed
the refinery would have to either revamp their existing equipment or add new hydrotreating
capacity. If the additional capacity needed at any given refinery exceeded the existing naphtha
hydrotreater capacity by less than 30 percent we assumed the necessary capacity could be added
by revamping the existing unit. (For our FRM, we will also evaluate if refiners could also use the
caustic Merox extraction process to lower the sulfur level in LSR and NGL blendstocks, and if
so, are they already doing that today). If, however, the additional capacity required exceeded the
existing reformer feed hydrotreater capacity by more than 30 percent we assumed the refinery
5-38
-------
would install a new stand-alone hydrotreater to desulfurize the excess LSR and NGL. Based on
available capacity in our refinery-by-refinery model and 2009 crude throughput data, we
estimated that refiners are already hydrotreating 66 percent of the LSR and NGL that are directly
blended into gasoline (excluding LSR processed in the isomerization units). The results of this
assessment are shown in Table 5-30 below.
Table 5-30 Refineries Adding Hydrotreating Capacity for LSR and NGL
Number of
Refineries
New Hydrotreater
(No FCC Unit)
4
New Hydrotreater
(with FCC Unit)
11
Revamped
Hydrotreater
(No FCC Unit)
1
Revamped
Hydrotreater
(with FCC Unit)
7
We conservatively evaluated the capital cost required for hydrotreater revamps and new
units by assuming that refiners will size their hydrotreater equipment needs to treat all production
volumes of LSR and NGL based on each refinery's maximum crude run rate. The operating
costs used in our refinery-by-refinery model, however, are based on LSR and NGL blendstock
rates from the models yields at the 2009 operational crude throughputs as discussed in Section
5.1.3.1. Sizing the equipment this way allows each refiner to have excess hydrotreater capacity
utilization, which is beneficial in the event of process unit shutdowns and to reprocess
blendstocks from abnormal operations.
Our estimate for the cost of adding a new hydrotreater at a refinery was obtained from
Gary and Handework's Petroleum Refining Technology and Economics, page 182-183, Curve C,
Table 9.1, 30,000 BSD unit. The capital cost for a grass roots hydrotreater listed by this source
was for a hydrotreater with a capacity of 30,000 BPSD and was based on 1999 dollars. We
multiplied this cost by 1.534 to determine the equivalent cost in 2010 dollars based on the
relative increase in the Nelson Refining Construction index from 1999 to 2010 (listed as 1497
and 2296 respectively). We used the six-tenths rule to scale the capital cost listed in Petroleum
Refining Technology and Economics to those of differing capacities based on relative size of the
desired unit. We assumed a hydrogen consumption of 40 SCF/Bbl for the processing of LSR
and NGL blendstocks which we obtained from the Jacobs Refining LP modeling database for
naphtha hydrotreating as this information was not presented in the literature source. For
refineries that only required a revamp of existing units we assumed a capital cost equivalent to
40 percent of the cost of a new hydrotreating unit of equal size. We assumed equivalent
operating costs for new hydrotreating units and revamped units. The capital and operating costs
for these hydrotreating units that were used in our model is shown in Table 5-31 below.
5-39
-------
Table 5-31 Capital and Operating Costs for LSR and NGL Hydrotreaters
Capital ($/BBL ISBL)
Hydrogen (SCF/BBL)
Fuel Gas (kBTU/BBL)
Electricity (kWh/BBL)
Octane Loss (R+M)/2
Change in Olefins (vol%)
Steam (Ib/BBL)
Catalyst Cost ($/BBL)
New Hydrotreating Units
1380
40
100
2.0
0.0
0.0
6.0
0.03
Revamped Hydrotreating Units
550
40
100
2.0
0.0
0.0
6.0
0.03
5.1.4.3 Butane Desulfurization Costs
After considering the desulfurization of LSR and NGL we then estimated the cost to each
refinery for desulfurizing the butane that is blended into gasoline. For the baseline case without
ABT where each refinery had to meet the 10-ppm sulfur standard we did not assume any cost
associated with butane desulfurization, as we assumed that any refiners likely to reduce the
sulfur content of the butane stream already had the necessary equipment. To estimate the costs
for each refinery to produce a 5-ppm gasoline for our ABT cases, however, we assumed
refineries would reduce the sulfur content of any butane blended into gasoline using extractive
desulfurization units (such as Merox or Merichem) which extract mercaptan sulfur (the bulk of
sulfur species in butanes) from butane volumes that are added to each refiner's gasoline pool.
We accounted for butane treating cost in our refinery-by-refinery model by adding capital to
install Merox treating units to each refinery. This is likely a conservative estimate based on our
lack of information on which refineries may already have these units. We set up the model to
size the Merox treater based on each refinery's maximum addition of butane volume to the
gasoline pool, using the amount that is blended in the winter, as winter grade gasoline has the
highest seasonal gasoline RVP limits and therefore allows for the highest levels of butane
blending. As with some of the LSR and NGL desulfurization units, it is unclear if additional
butane desulfurization units will be added to meet the Tier 3 standards. As a conservative
estimate, though, we have included the cost of adding these units to our program costs.
To calculate the volume of butanes added to winter grade gasoline we used the refmery-
by-refmery model estimate of the full year addition of butane to each refiners CG and RFG
gasoline grades which was based on the EPA annual gasoline database. The winter addition of
butane volumes was assumed to be 66 percent of the total annual volume of butanes added to a
refiner's gasoline pool on a yearly basis. The capital cost we used was for a 10,000 BPSD
Merox unit and was listed in 1995 dollars.9 We multiplied this cost by 1.64 to determine the
equivalent cost in 2010 dollars based on the relative increase in the Nelson Refining
Construction index from 1995 to 2010. We then multiplied the maximum winter daily butane
volume by 1.08 to account for an over design factor to size the equipment. We next used the six-
tenths rule to scale the capital cost listed in Petroleum Refining Technology and Economics to
those of differing capacities based on relative size of the desired unit in BPSD divided by 10,000
(the size of the unit the cost is based on). Finally, we applied offsite and PADD location factor
adjustments as listed in Table 5-25. For the operating cost we assumed that each unit had an
annual fixed operating cost of 6.7 percent of the total installed capital cost. To this we added an
5-40
-------
operating cost of 0.030 per gallon based on an operating cost figure from literature listed in 1995
dollars and scaled to 2010 dollars using appropriate factors. Operating costs were presumed to
represent all process energy and utility costs, as well as costs associated with purchasing caustic.
Table 5-32 Merox Treating cost for Sulfur Removal from Butane
Size, BPSD
10,000
Capital costs, $ MM
3.28
Operating Costs, 0/gal
0.03
5.1.5 Overall Cost Methodology
5.1.5.1 Sulfur Costs without ABT Program
While we are proposing an ABT program as described in Section V.E.I of the preamble,
we also evaluated a scenario with no averaging, banking, or trading between refineries. Under
this scenario, every individual refinery would comply with the 10-ppm annual average sulfur
standard, and none would lower their annual average sulfur below 10 ppm. This evaluation
provided a reference point for determining the impacts of an ABT program on compliance costs.
In complying with the 10-ppm sulfur standard, our refinery-by-refinery analysis
concluded that some refineries would do so with revamps to existing equipment while others
would need to install new grassroots equipment. Those refineries whose average sulfur levels
are already at or below the 10-ppm standard would not need to do anything. Because the most
recently available data was for 2009 when some phase-ins under Tier 2 were still effective, we
made adjustments to ensure that refinery-specific sulfur levels represented levels after all Tier 2
phase-ins had expired. For instance, if any refinery exhibited an annual average sulfur level
above the 80-ppm cap based on 2009 data, we assumed that their current annual average sulfur
level was 30 ppm. Further discussion of our refinery-by-refinery analysis can be found in
Section 5.1.3.
To determine the impacts on cost of the 10-ppm sulfur standard without ABT, we
volume-weighted the refinery-specific 0/gal costs calculated according to the refinery-by-
refinery modeling using 2009 gasoline production data for each refinery. Likewise, we
determined nationwide capital costs by summing the individual revamp and/or grassroots
equipment costs across all refineries. In these calculations, we assumed that all early credit
generation and phase-ins had expired, consistent with our approach to the ABT analysis. The
results of our analysis of costs under a scenario without ABT are provided in Section 5.2.1.1
below.
5.1.5.2 Sulfur Costs with ABT Program
As described in Section V.E.I of the preamble, we are proposing an ABT program that is
designed to ease the overall burden on the industry while still achieving the 10-ppm annual
average sulfur standard for the nation as a whole. Under the proposed ABT program, refineries
that can reduce sulfur below 10 ppm at a relatively low cost can generate credits which can then
be sold to refineries for whom the cost of attaining the 10 ppm sulfur standard would be
relatively high. The net effect of this credit trading would be expected to reduce the overall cost
of the program.
5-41
-------
To estimate the impact that the ABT program could have on nationwide average fuel
costs, we began with the refinery-by-refinery costs without the benefit of ABT described in
Section 5.2.1.1 for sulfur reductions down to 10 ppm, where each and every refinery is assumed
to get to 10 ppm. For the ABT analysis we then looked at the projected costs for refineries to get
down to 5 ppm (also described in Section 5.2.1.1.). We then determined the lowest cost option
among three alternatives for each refinery:
1. The refinery reduces its sulfur to 10 ppm.1
2. The refinery reduces its sulfur to 5 ppm and generates credits (in ppm-gal) for the
increment between 10 ppm and 5 ppm.
3. The refinery does not lower sulfur, but instead obtains credits to comply with the
10-ppm standard.
A fourth category applied to refineries whose average gasoline sulfur levels are already below 10
ppm. All such refineries were assumed to generate credits for the increment between 10 ppm
and their current sulfur level.
This process resulted in an ordering of refineries by their cost to reduce sulfur on a per-
ppm basis. Those with the lowest costs were assumed to make the investment. To optimize the
nationwide average costs under an ABT program, we determined the credit price at which the
total number of credits generated was equal to the total number of credits consumed. This
resulted in a list of refineries by their compliance strategy starting from the lowest-cost refinery
and increasing to the highest-cost refinery necessary to achieve the cumulative sulfur reduction
necessary to average 10 ppm across the gasoline pool. As discussed below, we did this under
two different trading assumptions: one where we assumed sulfur credits would be traded freely
across companies nationwide, and our primary, more conservative case where we only allowed
"trading" across refineries within a company.
To simplify the modeling of how an ABT program might operate, we focused on the
circumstances that refineries would face in the longer term, specifically after 2020. This
approach meant that the ABT program modeling did not consider the impact on gasoline sulfur
levels of delayed compliance for small refiners and small volume refineries, nor did it consider
the generation and use of any early sulfur credits. Moreover, our ABT modeling considered only
gasoline sold for use outside of California, and only gasoline produced by domestic refineries
(not imports).
Since the cost information available for this NPRM was limited to the costs of reducing
sulfur to either 10 ppm or 5 ppm (under which FCC gasoline sulfur levels would be reduced to
either 25 or 10 ppm, respectively), we were not able to estimate refinery-specific costs of
reducing sulfur to other levels. As a result, our ABT modeling could not account for scenarios in
which a refinery makes some capital investments to lower sulfur to some interim level, such as
20 ppm, and then obtains credits in order to demonstrate compliance with the 10-ppm standard.
1 No refinery compliance margins were included in this analysis.
5-42
-------
Our ABT analysis also could not account for credit generation at sulfur levels other than 5 ppm.
Our ABT analysis, then, most likely underestimates the cost savings that could occur due to ABT
since the greatest efficiencies are achieved when every refinery has the option of using any
combination of capital investments and credits generation or use. For the final rule, we may
investigate methods for expanding our ABT analysis to examine these types of scenarios.
We evaluated two ABT scenarios designed to bound the likely outcomes. In the first
scenario, we assumed perfect nationwide credit trading in which all credit generators make their
credits available to any refinery that needs them, regardless of their ownership. As described in
Section V.E.I of the preamble, the proposed ABT program is designed such that it could operate
in this way. However, such perfect nationwide trading may not be realistic. Under Tier 2 today,
there is still a considerable amount of inter-company trading occurring, but a significant fraction
of Tier 2 sulfur credits are bought and sold within companies/ Under Tier 3, it will be more
difficult to generate credits, and also more difficult to make up for deficits. Consequently, we
also investigated a second ABT scenario in which trading between companies does not occur and
averaging would only occur within companies that own more than one refinery. Under this
second scenario, individual companies might decide to bank credits for their own use, declining
to make credits that they generate at one of their facilities available to other companies. They
might do this, for instance, to address unplanned equipment downtime or other circumstances
that could make future compliance more difficult. While we do anticipate some trading to occur
between companies, we believe that this second, more limited scenario more closely represents
how our proposed ABT program might actually operate. As a result, the within-company credit
trading scenario is the primary scenario that we used to estimate overall costs for our proposed
program. It represents a somewhat conservative scenario with respect to estimation of costs.
The results of both scenarios are described in Sections 5.2.1.2 and 5.2.1.3 below.
Since an ABT program would allow some refineries to continue producing gasoline with
an average sulfur level above 10 ppm, we also investigated whether any areas of the country
might experience inordinately high sulfur levels as a result, and if so whether those higher sulfur
levels might be problematic in terms of either vehicle performance or local air quality. While we
were not able to model the distribution of all gasoline from the point of production to the point of
consumption, we did compare average sulfur levels at each refinery's location to the location of
nearby ozone nonattainment areas (based on the designation status with respect to the 1997
ozone NAAQS). A discussion of our analysis of sulfur levels by refinery and location can be
found in Section 5.2.1 below.
1 Based on Tier 2 sulfur compliance data, of the 26 companies that purchased sulfur credits in 2010, eight bought
credits only from their own company, and another five bought credits both from their own company and from other
companies. The remainder bought credits only from other companies.
5-43
-------
5.2 Estimated Tier 3 Sulfur Control Costs
5.2.1 Sulfur Cost Results
5.2.1.1 Cost of the Sulfur Program without ABT
Without ABT, all refineries would be required to meet the 10-ppm annual average sulfur
standard. We estimate that 13 refineries are already at or below 10 ppm, and thus their costs of
compliance would be zero. The remaining refineries would incur eVgal costs at various levels
depending on their particular configurations and the steps they would need to take. While we
estimate that the nationwide average cost of compliance would be 0.97 eVgal under this scenario,
two refineries would incur costs above 6 eVgal. The distribution of costs is shown in Figure 5-1
as a function of the number of refineries.
50 -
45 -
40 -
I/I .,!-
QJ 3b ~
Ol
.E 30 -
*S
£ 25 -
O
5 20
I*'
10 -
5 -
0,
1
_ •
||
0.00 0.50 1.
Nationwide average = 0.97 c/gal
1
JULljL__JL^__-__
00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 6.50 7.00 7.50
cents/gallon
Figure 5-1 Distribution of Compliance Costs by Refinery for 10 ppm without ABT
Program
In order to enable evaluation of the costs of the program with ABT, we also needed to
estimate the costs for each refiner to reach 5 ppm. The distribution of costs for this scenario is
shown below.
5-44
-------
40
35
30
25
£ 20
O
^
-------
The nationwide average cost of compliance with the 10-ppm sulfur standard would be
K
reduced from 0.97 eVgal without ABT to 0.89 eVgal under company-only trading. However,
there would continue to be a significant variation in costs by refinery as illustrated in Figure 5-3.
Nationwide average = 0.89 c/gal
-0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5
cents/gallon
Figure 5-3 Distribution of Compliance Costs by Refinery for 10-ppm Average Standard
with ABT Program and Intra-Company Credit Transfers
In this figure, refineries with negative costs are those whose sulfur levels were already
below 10 ppm and who would generate and sell credits. While there are several refineries with
costs above 3 eVgal in this scenario, in practice we would expect such refineries to incur lower
costs by pursuing some combination of lowering their gasoline sulfur to some level above 10
ppm and also purchasing credits to cover the remaining sulfur reductions needed to reach 10
ppm. The analysis we conducted for this draft RIA was unable to capture such nuances, but we
will be investigating them for the final rule.
Figure 5-4 illustrates how the cumulative distribution of costs by volume would change
under this scenario. In general, the distribution would shift downward slightly in comparison to
a scenario in which there was a 10 ppm average standard but no ABT program.
K Approximately 55% of this desulfurization cost is comprised of the variable cost, and the majority of that is due to
the cost of recovering the octane lost when the hydrotreater unit saturates the octane-rich olefms contained in the
FCC naphtha. Another 30% of the cost is due to the capital cost amortized over the volume of gasoline. Finally,
about 15% of the cost is due to the fixed operating cost which includes the maintenance of the new equipment and
taxes.
5-46
-------
6 -
5 -
"
c
OJ
u
Distribution with no ABT
Distribution with company-only trading
20000 40000 60000 80000 100000
Cumulative gasoline volume (mill gallons)
120000
Figure 5-4 Distribution of Costs by Volume for 10-ppm Average Standard with ABT
Program and Intra-Company Credit Transfers
Since we estimate that 85 out of 111 refineries would meet the 10-ppm sulfur standard
without the use of credits under the intra-company trading scenario, only about 5 percent of
gasoline would continue to have an annual average sulfur level above 10 ppm. These refineries
tend to be smaller than average. The distribution of sulfur levels under this scenario is shown in
Figure 5-5.
5-47
-------
100 -
90
80 -
£ 70 -
'i-
O)
c 60~
Ol
•6 50^
| 40 -
1 30
20 •
10 -
n ~
1
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85
Sulfur (ppm)
Figure 5-5 Distribution of Sulfur Levels by Refinery with ABT Program and Intra-
Company Credit Transfers
Of the 5 percent of gasoline volume that would have average sulfur levels above 10 ppm
under this scenario, nearly half would be below 20 ppm, and a majority would be below 30 ppm.
Moreover, as noted above, we would expect refineries with the highest sulfur levels to pursue
some combination of lowering their gasoline sulfur to some level above 10 ppm and also
purchasing credits to cover the remaining sulfur reductions needed to attain the 10-ppm standard.
We also investigated whether the existence of refineries producing gasoline with average
sulfur levels higher than 10 ppm after 2020 might contribute to local areas where the average
sulfur level is higher than 10 ppm, and whether this might be problematic in terms of either
vehicle performance/emissions or local air quality. To do this, we compared the location of
refineries with sulfur levels higher than 10 ppm to the location of other refineries serving similar
areas. We also compared the location of refineries to areas that have historically been in
nonattainment for ozone.
As shown above in Table 5-33, we estimated that eight refineries would have annual
average sulfur levels above 10 ppm under company-only credit trading. Three of these refineries
serve areas that are, at a minimum, several hundred miles away from historical ozone
nonattainment areas. The remaining five refineries with average sulfur levels above 10 ppm
supply areas that are also supplied by other refineries whose average sulfur levels would be 10
ppm or lower. As a result, gasoline with higher sulfur levels would likely be diluted by gasoline
with lower sulfur levels, and we would not expect any ozone non-attainment area to receive
gasoline with an average sulfur level higher than 20 ppm.
5-48
-------
5.2.1.3 Cost of Sulfur Control Program with ABT and Nationwide Credit
Transfers
Although we believe that credit trading will occur primarily between refineries within the
same company, we also investigated a scenario in which credit transfers would occur between all
refineries regardless of location or company. For this scenario, we determined that more than 60
percent of refineries would be involved in either generating or consuming credits. Table 5-34
summarizes our estimate of the amount of credit trading that would occur under this scenario.
Table 5-34 Impacts of ABT Program with Nationwide Credit Transfers
Number of refineries whose sulfur is already below 10
ppm, and which generate credits
Number of refineries which lower their sulfur to 5 ppm
and generate credits
Number of refineries that do not lower their sulfur but
instead consume credits
Number of refineries that lower sulfur to 10 ppm and
neither generate nor consume credits
12
34
25
40
The nationwide average cost of compliance with the 10-ppm sulfur standard would be reduced
from 0.97 eVgal without any ABT program to 0.79 eVgal under nationwide trading. There would
continue to be some variation in costs by refinery as illustrated in Figure 5-6, though the
distribution would not be as wide as under the intra-company credit transfer scenario.
01
_Q
45
40
35
25
15
10
ll
Nationwide average = 0.79 c/gal
-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5
cents/gallon
Figure 5-6 Distribution of Refinery Compliance Costs with ABT Program and Nationwide
Credit Transfers
5-49
-------
Figure 5-7 illustrates how the cumulative distribution of costs by volume would change
under this scenario. In comparison to a scenario wherein credit trading occurs only between
refineries within the same company, the downward shift in the distribution would be more
pronounced.
Distribution with nationwide trading
20000 40000 60000 80000 100000
Cumulative gasoline volume (mill gallons)
120000
Figure 5-7 Distribution of Compliance Costs by Volume with ABT Program and
Nationwide Credit Transfers
Under a scenario in which credit transfers occur nationwide, we estimate that 25 out of
111 refineries, or about 23 percent, would not lower their sulfur levels and would instead obtain
credits. These refineries tend to be smaller than average, and represent about 17 percent of
gasoline volume. The distribution of sulfur levels under this scenario is shown in Figure 5-8.
5-50
-------
100
90
80
S 70
'i-
O)
c 60
Ol
£ so
Ol
.Q
E
3
40
30
20
10
0
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85
Sulfur (ppm)
Figure 5-8 Distribution of Sulfur Levels by Refinery with ABT Program and Nationwide
Credit Transfers
Of the 17 percent of gasoline volume that would have average sulfur levels above 10 ppm
under this scenario, the vast majority would be below 30 ppm.
As for the previous scenario under which credit transfers would occur only within
companies, we investigated the localized impact of higher sulfur levels under a scenario in which
credit transfers occur nationwide and thus there are more refineries with average sulfur levels
above 10 ppm. Under this scenario, most refineries with average sulfur levels higher than 10
ppm supply areas that are also supplied by refineries with sulfur levels at 10 ppm or below. We
expect the dilution of higher sulfur with lower sulfur in such areas to result in annual average
sulfur levels no higher than about 20 ppm. Of those refineries projected to have average sulfur
levels higher than 10 ppm under this scenario and which are the primary or only suppliers of
gasoline to a particular area, none of the affected areas have historically been in nonattainment
for ozone.
5.3 Other Cost Studies
Other cost studies were recently conducted to estimate the cost of additional reduction in
gasoline sulfur. We evaluated each of these studies and compare them to our own cost analysis.
The International Council for Clean Transportation (ICCT) retained Mathpro in October
2011 to study the cost of a 10 ppm average gasoline sulfur standard as well as a 1 psi reduction
in RVP.10 Since the lower RVP standard was modeled as a separate step from the low sulfur
standard, we were able to isolate the gasoline sulfur reduction costs from the low RVP costs.
5-51
-------
ICCT's estimated cost for a 10 ppm average gasoline sulfur standard is 0.8 cents per
gallon which reflects the capital costs amortized assuming a before-tax 7 percent rate of return
on investment. This cost reflects an assumption that the capital cost for revamps of FCC
postreaters is 30 percent of the capital costs for a grassroots FCC postreater . Mathpro also
analyzed costs assuming that the capital cost for revamps of FCC postreaters are 50 percent of a
grassroots FCC postreaters, which is 1.1 cents per gallon. ICCT's cost estimate for complying
with a 10 ppm average gasoline sulfur standard is very close to ours.
In 2008 The Alliance retained Mathpro to use its LP refinery cost model to estimate the
costs of what they termed National Clean Gasoline (NCG) in PADDs 1, 2 and 3 (generally
speaking, this is the part of the U.S. east of the Rocky Mountains).11 Achieving NCG would
entail reducing gasoline sulfur to 5 ppm under a 10-ppm cap standard and the reduction of
gasoline RVP to 7 psi. For the low-RVP standard, a 1-psi waiver was allowed for conventional
gasoline, but not for current low-RVP areas. The study also evaluated two sensitivity cases
which increases the stringency of the distillation index (DI) from 1250 to 1200. The Alliance
study also evaluated crude oil price as a second sensitivity case, evaluating crude oil prices at
$51/bbland$125/bbl.
The Alliance studied three different cases. The first case applied the 10-ppm sulfur cap
to RFG. The second case applied the 10-ppm sulfur cap and the 7.0-psi low-RVP standard to
RFG as well as 7.0- and 7.8-psi low-RVP gasoline. The third case applies the 10-ppm sulfur cap
and 7.0-psi RVP standard to all RFG and CG. Of these three cases, the first case is most relevant
because applying the fuels changes to RFG solely applies the 10-ppm sulfur cap to RFG and
does not involve any changes in RVP. However, the 10-ppm sulfur cap standard studied by the
Alliance is still 5 ppm more stringent than the 10-ppm average standard that we are proposing.
The Alliance cost estimate for Case 1 is 1.6 cents per gallon for RFG in PADDs 1, 2 and
3. This cost estimate is based on amortizing the capital costs on a 10 percent after-tax return on
investment (ROI). We adjusted the cost estimate to amortize the capital costs based on a before
tax 7 percent ROI and adjusted the costs to 2010 dollars which increases the costs to 1.75 cents
per gallon. The 1.75 eVgal cost estimate is based on a crude oil price of $51/bbl. The Alliance
estimated the cost of a 10-ppm sulfur cap standard on RFG assuming that crude oil is priced at
$125/bbl. At the $125/bbl crude oil price, the Alliance study estimates that it costs 2.50 eVgal to
require that RFG comply with a 10-ppm sulfur cap standard. Adjusting the Alliance costs to
reflect a 7 percent before tax ROI and 2010 dollars increases the Alliance costs based on a
$125/bbl crude oil price to 2.69 e7gal.
For our cost analysis we analyzed the cost of sulfur control assuming that crude oil is
priced at $91.8/bbl. We can interpolate between the Alliance costs based on $51 and $125 per
barrel crude oil prices, which results in a single cost which is 2.3 cents per gallon. We also
estimated a cost for refiners lowering their gasoline sulfur to 5 ppm using the refmery-by-
refmery cost model and our cost is 1.38 eVgal.
In response to the Alliance study, API retained Baker and O'Brien (BOB) in 2010 to
study the cost of additional gasoline sulfur control and an increase in RVP using a refmery-by-
refmery cost approach with BOB's Prism model.12 The Prism model is largely a spreadsheet
cost model with blending optimization. The primary case analyzed by the API study is the cost
5-52
-------
of reducing gasoline sulfur to an average of 10 ppm and reducing gasoline RVP to 7.0 psi
without a 1-psi waiver for blending 10 percent ethanol. The study also analyzed three other
sensitivity cases: 1) a 5-ppm average gasoline sulfur standard with 7 psi RVP limit on
conventional gasoline without a 1-psi waiver; 2) a 10-ppm average gasoline sulfur standard and a
7.8-psi RVP limit on conventional gasoline without a 1-psi waiver; and 3) a 10-ppm average
gasoline sulfur standard with a 7.8-psi RVP limit on all conventional gasoline with a 1-psi
waiver.
In an addendum to its fuels study report released in 2011, API contracted with Baker and
O'Brien to study a sensitivity case 4, which is a sulfur only case, using its PRISM refinery
model. From our understanding of the study, the study parameters seemed to be about the same
as the original study, except that API solely studied a 10 ppm average gasoline sulfur standard
(not including any RVP control), the same sulfur standard that we are proposing. However, API
also assumed that a 20 ppm cap standard would also be in place which would not allow the
application of an averaging, banking and trading (ABT) program to optimize refinery
investments and minimize overall costs.
API made a series of conclusions based on the study. Perhaps the most important
difference with the original study is that API concluded that not a single refinery would shut
down as a result of the proposed 10 ppm gasoline sulfur control standard, even though API did
not study the flexibilities of an ABT program and used excessively high capital costs for a
grassroots FCC postreater (see below). Like the original study, API did not report average costs,
but reported the marginal costs for the cost study. Marginal costs reflect the cost of the program
to the refinery or refineries which would incur the highest costs, assuming that the highest cost
refineries would set the price (or in this case, the price increase) of gasoline. The report
concluded that marginal costs after the imposition of a 10 ppm gasoline sulfur program would
increase the price of gasoline by 6 to 9 cents per gallon in most markets. API did not define how
its statement "in most markets" would apply to the US gasoline supply. API also did not provide
any justification why it assumed that the refineries that would experience the highest
desulfurization cost under Tier 3 would also be the same refineries which sets the gasoline price
in the gasoline market today.
Although API did not provide an average gasoline desulfurization cost in its report, we
could calculate an average cost based on the gasoline volume and total annual costs provided.
The total cost reported in the report for the 10 ppm average gasoline sulfur standard is
$2390MM/yr and the non-California gasoline volume is 7343 thousand barrels per day. This
results in an average per-gallon desulfurization cost of $0.89/bbl or 2.12 c/gal. The difference
between the average cost and marginal cost (price increase) that API is projecting is profit. API
is projecting that the oil industry would profit from 10 ppm low sulfur standard by the roughly 4
to 7 cents per gallon difference between the average cost and the two marginal price values.
That per-gallon profit translates into $4 to $8 billion dollars per year in profit.
The average cost of the 10 ppm average gasoline sulfur standard was calculated using
API's methodology for amortizing capital investments. To facilitate a fairer comparison
between the API cost study and our cost study, we adjusted the API costs to be on a similar basis
as our costs. We adjusted the API costs to reflect a before-tax 7 percent return on investment
(ROI) for capital invested for the hydrotreaters and hydrogen plants instead of the after-tax 10
5-53
-------
percent ROI used by API. This lowered the API estimated costs from 2.12 c/gal to 1.58 c/gal.
API's 1.58 cents per gallon cost is still higher than our 0.89 c/gal cost with an ABT program that
assumes intercompany trading of credits, and higher than our 0.97 c/gal for the case which
assumes no ABT program. Thus comparing "apples-to-apples" to the to the extent possible,
API's 1.58 c/gal estimated cost for complying with a 10 ppm average gasoline sulfur standard
compares very favorably with our own cost estimates, and not at all near the 25 c/gal value that
was sometimes quoted from the first API study. The remaining cost difference between our
estimated costs and those by API are the capital cost assumptions that API used, as discussed
below. While little detail is is provided by API about what hardware comprises their
desulfurization units, the inside battery limits (ISBL) and total capital costs for the FCC
postreaters and FCC pretreaters are provided in API's report. API's FCC pretreaters capital
costs are consistent with the capital costs that we have used for this unit. However, the FCC
postreater costs used by API are much higher than what we used and have been used in the past
by others. API's capital cost for a grassroots FCC postreater is $228 million for a 35,000 bbl/day
unit, or $6540 per/bbl per day. API's capital cost includes the outside battery limit (OSBL)
costs. In contrast, the ISBL capital cost that we used for a grassroots FCC postreater is
$1500/bbl-day for a 30,000 bbl/day grassroots unit, which increases to $1875/bbl/day when the
offsite costs are added on. Thus, the API capital costs are about 3 1A times higher than the capital
costs that we are using for a grassroots FCC postreater. To check our capital costs, we found
other capital cost estimates to which we could compare our costs, including the capital costs used
by the National Petroleum Council when it studied the cost of gasoline desulfurization prior to
Tier 2.L Table 5-35 contains a cost comparison of ISBL, and ISBL and OSBL FCC postreater
capital costs.
Table 5-35 Capital Cost Comparison
Technology
ISBL Capital
Cost ($/bbl/day)
ISBL and OSBL
Capital Cost
($/bbl/day)
EPA
(Tier 3)
1500
1875
API
-
6540
Mathpro
(ICCT)
-
1800
Jacobs
2440
3538
Exxon
Scanfining
(NPC)
1045
1360
IFF
Prime G
(NPC)
1410
1833
CDTech
(NPC)
960
1248
Phillips
S-Zorb
(NPC)
860
1118
Table 5-35 shows that, compared to the average of the rest of the capital cost estimates,
the API capital cost for FCC postreater is about four times higher. Compared to the next highest
L The capital costs that the vendors provided to the National Petroleum Council were adjusted from year 2000
dollars, which was the year that the NPC analysis was conducted, to year 2011 dollars, which is the base year of the
Tier 3 analysis. To make this adjustment we used the Nelson-Farrar capital cost index which increased the NPC
capital costs by 55% over the year 2000 costs.
5-54
-------
cost estimate, which is the FCC postreater capital cost from the Jacobs data base in the Haverly
refinery cost model that we use,M the API capital costs are almost two times higher.
An important distinction must be made with respect to the severity of desulfurization for
the capital cost comparison made in Table 5-27. For complying with the Tier 2 gasoline sulfur
standard (Jacobs and NPC costs), a typical refinery would have installed an FCC postreater to
desulfurize the FCC naphtha from about 800 ppm down to about 75 ppm, a 725 ppm, or a 91
percent sulfur reduction. In the case of a grassroots postreater that would be installed for Tier 3,
the postreater would treat FCC naphtha already low in sulfur due to the pretreater installed
before the FCC unit (these refineries are currently complying with Tier 2 using an FCC
pretreater). Thus, the new grassroots FCC postreater would only have to reduce the FCC naphtha
from 100 ppm to 25 ppm, a much smaller 75 ppm or 75 percent sulfur reduction. A grassroots
FCC postreater installed for Tier 2 would typically remove 10 times more sulfur than one
installed for Tier 3. This is important because a significant portion of the FCC postreater capital
cost is devoted to avoiding the recombination reactions which occur when hydrogen sulfide
concentrations are high and react with the olefms contained in the FCC naphtha. Thus, a
grassroots FCC postreater installed for Tier 3 would be expected to be significantly lower in
capital cost compared to a Tier 2 FCC postreater. API's costs are based solely on Tier 2
compliance costs, which is one reason why their costs are so high. API obtained either estimated
installation costs or actual installation costs (API did not specify) for FCC postreaters for
installation in 5 different refineries for complying with the Tier 2 30 ppm gasoline sulfur
standard. The postreater capital cost information which reflected cost information from the years
2003 to 2005 was adjusted upward to reflect mid-year 2009 capital costs using the Nelson-Farrar
index and normalized to reflect a 35 thousand barrel per day unit. This resulted in an average
ISBL cost of $144.5 million for installing a Tier 2 compliant FCC postreater. After discussing
this capital cost estimate with several refiners who built several of the units in recent years, those
refiners felt that the estimated capital costs that API had calculated were too low, and one refiner
thought that the estimated capital costs should be doubled. Based on the information provided
by that one refiner, API doubled its estimated capital costs for a 35K bbl/day FCC postreater to
$228.8 million.
Another way to assess the API capital cost for the FCC postreaters is to compare it to the
FCC pretreater cost that API is using. FCC pretreaters are much higher pressure units and use
more expensive metallurgy than FCC postreaters and, for these two reasons, are much more
expensive than FCC postreaters on a per-barrel basis. However, API's FCC postreater capital
costs are about 50 percent more expensive than its FCC pretreater capital costs, which is
inconsistent with the design requirements of the units. API acknowledged this inconsistency, but
did not take steps to correct it.
API's estimated range of capital cost for revamping an FCC postreater is also higher than
our range of capital cost for revamping an FCC postreater, when assessing the revamped costs as
M The installed capital cost for an FCC postreater from the Jacobs data base was adjusted to current year dollars.
This estimated installed capital cost is several years old and may not represent Jacobs current cost estimate for a
FCC postreater.
5-55
-------
a percentage of the capital cost for a grassroots unit. API estimates that revamping an FCC
postreater would cost 30 to 70 percent of the capital cost for a grassroots FCC unit. Our capital
cost estimate for revamping an FCC naphtha postreaters from 17 to 50 percent of the capital cost
for a grassroots FCC postreater, however, most of the revamps are estimated to cost at the lower
end of that range.
The Emissions Control Technology Association (ECTA) retained personnel within
Navigant Economics to study the costs of a 10 ppm average gasoline sulfur standard and assess
the ICCT and API cost studies.13 The authors made a number of conclusions. After reviewing
both the ICCT and API studies, the authors found that a primary difference in estimated costs
between the two studies was the capital costs. The authors contacted vendor companies that
license FCC postreater technologies and surveyed the companies to find out what the capital
costs are for a FCC postreater. As a result of the survey, the report authors concluded that API's
capital costs were too high, and those used in the ICCT study were about right. The authors
found that Baker and O'Brien has a history of exaggerating the economic impacts of EPA rules,
citing the costs and other impacts of its analysis of the 2007 on-highway heavy-duty proposed
rulemaking. The authors concluded that the impact of a 10 ppm gasoline sulfur standard on the
average refining cost would likely be closer to the 1 cent per gallon estimate by the ICCT study.
Furthermore, the report's authors also pointed out that the marginal cost analysis conducted by
API did not consider the proposed averaging banking and trading (ABT) program that we were
expected to propose, which would reduce the marginal costs of the Tier 3 proposed rule. Because
API's addendums to its original report came out many months after its original report, we
originally assessed the most similar case to our proposal from API's original study. This was
sensitivity case 3, which studied a 10-ppm average gasoline sulfur standard with a 7.8-psi RVP
limit on all conventional gasoline with a 1-psi waiver. One of the most important conclusions by
API with respect to sensitivity case 3 was the projected 623 thousand barrels per day (about 6.5
percent of total gasoline demand) reduction in U.S. gasoline production due to reduced blending
of light hydrocarbons in response to RVP control and the closure of 4 refineries caused by the
investment hurdle of complying with both the sulfur and RVP standards evaluated. Because of
API's significant projected impacts, which differed from those made by the Alliance and by
ECTA in their studies, we looked closely at this projected impact. We found that this projection
for a large shortfall in supply is similar to previous projections made by API for earlier
rulemakings. Baker and O'brien made a very dire projection for diesel fuel supply when
analyzing the final highway and proposed nonroad diesel fuel sulfur regulations for the proposed
nonroad diesel rulemaking back in 2003.14 API's projection was that 12 refineries would be
shutdown and that U.S. refiners would exit the diesel market resulting in a shortage of diesel fuel
supply of 639 thousand barrels per day and to make up the shortfall, the distillate market would
need to be supplied by imports. Since that projection was made back in 2003 and the highway
diesel fuel program is fully implemented and the nonroad diesel fuel program is mostly
implemented, we can look at the impacts to see how well API's projection played out.
First, with respect to imports, the distillate import/export market was fairly stable at about
200 thousand barrels per day of net imports during the period 2003 to 2005, before the highway
and nonroad programs started. If we add the expected increase in imports estimated by API
caused by the ULSD highway rule, API projected the imports to increase to 549 thousand barrels
per day in 2006. When the temporary compliance option of the ULSD highway rule ended in
2010 and nonroad diesel (excluding locomotive and marine diesel because those requirements
5-56
-------
don't start until 2012) must comply with the ULSD nonroad diesel rule in 2010, API estimated
that imports will increase to 829 thousand barrels per day including the baseline imports.
However, what actually occurred is that imports not only did not increase, but they actually
decreased to the point that the U.S. became net exporters of 428 thousand barrels per day of
distillate.1516 Thus, API was off by over 1200 thousand barrels per day in its estimate of
distillate production in 2010 by U.S. refiners. Figure 5-9 summarizes the imports and exports of
diesel fuel and gasoline during the period when the clean fuels regulations were being phased in.
2003 2004 2005 2006 200
•NetGasoline Imports
•Net Distillate Imports
Net Gasoline and
Distillate Imports
•API Distillate Import
Projection
Figure 5-9 Petroleum Product Imports and Exports During the Implementation of Clean
Fuels Regulations
One reason why API was so far off in its projection that U.S. distillate supply would
decrease significantly is that it had projected that 12 refineries would shut down due to the
highway and nonroad rulemakings, but, in fact, there were very few refinery shutdowns.17 18
Between 2003 and 2011 when the highway and nonroad diesel fuel ULSD programs were
phasing in (as well as Tier 2, MSAT2, RFS1 and RFS2), there was a total of 5 refinery
shutdowns of refineries which produce transportation fuels. However, also during this time
period there were reactivations of three previously shutdown transportation fuel producing
refineries. Considering both the shutdown refineries and reactivated refineries, there was a net
shutdown of two transportation fuel producing refinery closures. It is unclear why those 5
transportation fuel producing refineries shutdown, but there is no evidence to suggest they were
caused by the diesel rules. There are many factors which cause refinery shutdowns.
To understand these factors it is necessary to revisit the history of the U.S. refining
industry.19 During the second crude oil embargo which occurred during the Carter
Administration, subsidies were established that incentivized the construction of many small
refineries. After crude oil prices dropped, those subsidies ended and many of those refineries
5-57
-------
were not economically competitive with larger refineries. Over time the number of refineries
decreased and the remaining refineries were expanded to supply the U.S. market. This is part of
a rationalization that most any industry experiences as the industry matures. Figure 5-10 and
Table 5-36 summarize the changes in the number of refineries, the average size of refineries, and
the total production capacity of the U.S. refining industry based on data from the Energy
Information Administration (EIA).
350
300
eg
Q.
ra
u
>• >•
>- TO
Ol T3
0.' -D
cc. -o
01
on
000
Figure 5-10 Refining Industry Statistics
Table 5-36 U.S. Refinery Industry Statistics
Year
Net Change in the Number of
Operable Refineries
Change in Average Refinery
Size (kbbl/day)
Change in Total U.S. Operating
Refinery Capacity (MMbbl/day)
1982 - 1992
-102
+18,400
-1.2
1993-2002
-50
+31,800
+1.6
2003 -2011
-2
+11,000
+0.4
Figure 5-10 and Table 5-36 show that there were many refinery closures during the 1980s
and 1990s, although over time the number of refinery closures was diminishing. Therefore, the
refinery closures which occurred during the 2003 to 2011 time frame could very well be due to
this rationalization process reflecting a maturing industry. In fact, examining the trend, the
number of net refinery shutdowns is virtually flat during the 2003 to 2011 time frame relative to
the previous periods. This is clearly a rationalization of the refining industry because as
refineries were closing, the average refinery size increased over each of the periods. Finally, the
total U.S. refinery capacity decreased and increased back to its 1982 levels during the 1980s and
5-58
-------
early 90s (decrease was likely due to the corporate average fuel economy standards), but showed
a modest increase from the late 1990's through 2011.
5.4 Projected Energy Impacts and Impacts on Permitting
Our refinery-by-refinery model was also used to determine the impact the Tier 3
standards would have on the energy related CO2 emissions and permitting of existing refineries.
While the Tier 3 proposal will reduce emissions from vehicles, the addition of grass roots units
and revamping of existing units which we project will happen as a result of the Tier 3 sulfur
standards are likely to result in some increased emissions of regulated air pollutants at refineries.
Refinery projects designed to meet the new fuel standards could trigger preconstruction air
permitting requirements under the Clean Air Act and EPA's New Source Review (NSR)
regulations. To address this concern, we used our refinery-by-refinery model to estimate the
likely process and equipment changes that may be required to meet the Tier 3 gasoline standards.
This information was submitted to EPA's Office of Air Quality Planning and Standards
(OAQPS) to provide the inputs that are necessary for the modeling and analysis of the refinery
emissions and permitting impacts of the Tier 3 fuel standards.
Using our refinery-by-refinery model we generated refinery-specific estimates of the
increased energy, hydrogen, and gasoline octane demands that we estimate will result from the
proposed Tier 3 standards. We also estimated the increase in sulfur plant recovery unit (SRU)
loading/operations for the 111 U.S. refineries that we modeled in our analysis. Energy demand
includes fuel that is needed to generate refinery process heat, steam and electricity. Hydrogen
demand is associated with increased hydrotreating of Fluid Catalytic Cracking (FCC) naphtha
and light straight run (LSR) streams. Increased gasoline octane demand results from refineries
having to replace octane that is lost due to increased FCC naphtha hydrotreating. Increased SRU
loading results from the increased fuel desulfurization and associated H2S generation. All of
these incremental demands will be referred to as "demands" in the following sections. We used
our refinery-by-refinery model to calculate the increase in these various demands for several
scenarios where sulfur averaging, banking, and trading (ABT) was not allowed and each refinery
had to meet the lOppm standard, as well as scenarios that allowed ABT between refineries
owned by the same parent company to minimize the cost of compliance with the Tier 3
standards.
5.4.1 Emissions Impacts of Different Production Volumes
In addition to considering scenarios with and without ABT we also considered the
impacts on emissions and permitting of different gasoline production volumes for each refinery.
In the first case, called the normal case, we considered the incremental demands for each refinery
assuming no change in gasoline production volume. We also considered a case, called the
maximum demand case, where each refinery maximized gasoline production based on currently
existing refinery capacity and equipment.
5.4.1.1 Normal Case
The normal case was estimated using each refinery's predicted yields of FCC naphtha
and LSR from our refinery-by-refinery model, along with each refinery's total gasoline
5-59
-------
production volume from EPA's RFG database. For each refinery the refinery-by-refinery model
generated specific Tier 3 demands for hydrogen, steam, fuel gas, electricity and gasoline octane
based on the desulfurization technology used by each refinery for any FCC postreating and LSR
hydrotreating. To determine the FCC postreating demands the model considers each refinery's
volume of FCC naphtha under normal operations, the FCC naphtha sulfur level at the refinery
prior to postreating, and the process requirements of the FCC postreater technology used by that
refinery. The demands are calculated by multiplying the FCC naphtha volume by the demands
from the use of the associated FCC postreating technology. Table 5-26 through Table 5-29 show
the FCC postreater technology demand averages as applied to refineries on a national basis for
the 5 and 10 ppm gasoline sulfur standards. Note that the demands vary significantly with the
FCC naphtha sulfur level prior to postreating.
Similarly, the normal case demands for any LSR blendstocks that require additional
hydrotreating as a result of the Tier 3 standards were determined based on each refinery's yield
of LSR blendstock under normal operations and the demands for the additional LSR
hydrotreating. These demands, on a national average basis, are listed in Table 5-31. The normal
case demands for FCC postreating and LSR hydrotreating were then summed to determine the
increase in energy, hydrogen, and octane demand. To determine the additional sulfur removed
from gasoline we first calculated the difference between the current gasoline sulfur level of the
gasoline produced at each refinery according to their compliance reports to EPA and the
proposed Tier 3 standard. This difference was multiplied by the refinery's gasoline production
volume and divided by the number of days of operation to calculate the additional sulfur removal
level at each refinery. This sulfur removal information was then used to determine the increase
in SRU loading on a fractional basis by dividing the additional sulfur removal as a result of the
Tier 3 standards (in tons of sulfur per day) by the refineries SRU process capacity.
5.4.1.2 Maximum Case
We also considered a second demand case, called the maximum case, in which we
calculated the demands that result from the Tier 3 standards if each refinery maximizes gasoline
production based on currently existing refinery capacity. For this case we first determined each
refiners FCC unit process capacity utilization rate in the normal case. The annual FCC unit
feedstock charge rate for each refinery as reported in the 2009 EIA data was divided by the FCC
unit design capacity as reported in the Oil and Gas Journal (OGJ) to calculate the capacity
utilization rate for the normal case. These normal capacity utilization rates were then scaled up
to reflect maximum capacity utilization rates, and further adjusted using an overdesign factor.
For refineries projected to meet the proposed Tier 3 standards by revamping existing
FCC postreating units we assumed that their maximum gasoline production rate was equal to the
rate produced running the FCC unit at 92% of the refinery's maximum FCC design capacity.
There were several refineries that are currently operating their FCC unit greater than a 92%
capacity utilization rate. We assumed that these refineries were already operating at their
maximum annual capacity utilization rate. For refineries projected to install a new FCC
postreater we similarly assumed that the new unit would be scaled to process the output of the
FCC unit operating with a 92% utilization rate. For new FCC postreaters, however, we increased
the results by 15% as an overdesign factor and adjusted the results accordingly. A similar sizing
5-60
-------
approach was taken for refineries we projected would revamp or add new LSR hydrotreating
capacity to comply with the proposed Tier 3 standards.
The results represent a "maximum" annual gasoline production case for each refinery
under the Tier 3 standards based on each refiners FCC unit design capacity. These cases
represent a scenario where each refinery projects emissions based on the maximum achievable
annual production rate for their existing processing units. These cases reflect each refiner's
potential emissions impacts as a result of the proposed Tier 3 standards, rather than the existing
Tier 2 standards, when operating at maximum FCC rates as opposed to normal operation more
indicative of national gasoline demand.
5.4.2 Refinery Demand Sourcing
After determining the increased demands for each refinery as a result of the proposed
Tier 3 regulations we next developed cases for each refinery demand scenario that represented
different options for sourcing these demands. Some refiners may choose to produce all of the
required hydrogen and electricity, as well as make up for all of the lost octane at their refinery.
Others may choose to purchase some or all of the hydrogen, electricity, and high octane
blendstocks that they would need to comply with the Tier 3 standards from external suppliers.
These decisions have a significant impact on the emissions and permitting impacts of the
proposed Tier 3 regulations. In order to bound all possible scenarios we considered both high
and low impact cases for each refinery demand scenario. In the high impact scenarios we
assume that each refinery produces all of the required hydrogen, and electricity needs while
making up for any gasoline octane loss at their refinery. In the low impact scenarios we assume
that all the necessary hydrogen, electricity, and high octane blendstocks are purchased from an
external supplier.
In both cases we assumed that fuel gas demands would increase to meet the increased
thermal demands at the refinery. In the high impact scenarios the refinery's fuel gas needs
would be further increased to produce the needed hydrogen and electricity while replacing the
lost octane in their gasoline. We consulted literature sources to determine the conversion factors
90
from MBTU fuel gas to 1,000 standard cubic feet (scf) of hydrogen and 1,000 pounds of
steam21 that are typical for refineries. We also assumed a standard conversion efficiency from
fuel gas to electricity for our modeling. These conversion factors are shown in Table 5-37
below.
Table 5-37 Fuel Gas Required to Produce Hydrotreater Utilities
Utility
Hydrogen (1,000 SCF)
Steam (1,000 Ibs)
Electricity (1 kWh)
Fuel Gas Required (M BTU)
239
1530
5.1
We also had to determine the fuel gas demands that would be required to make up for lost
octane in our high impacts scenario. For this analysis we assumed that all lost octane in the
gasoline that results from increased FCC postreating to meet the proposed Tier 3 standards
would be recovered by running the reformers at the refineries at a higher severity as opposed to
5-61
-------
sourcing it from ethanol or other means. This would increase the octane of the reformate and
offset the octane losses from the FCC naphtha, resulting in no change in the octane of the overall
gasoline pool. Running the reformer at a higher severity requires a higher reactor processing
temperature and an increased volumetric rate for the pumps and compressors. These changes
further increase the fuel gas and electricity demands of the refinery. We estimated these
demands using information we obtained from our Jacob's refinery LP modeling analysis. The
estimates we used for the fuel gas and electricity demands for increasing gasoline octane are
shown in Table 5-38 below. These demand increases were only used for the high impact
scenarios. For the low impact scenarios we assumed that lost octane was recovered by
purchasing high octane blendstocks resulting in no fuel gas or electricity demand increases at the
refinery.
Table 5-38 Energy Required to Increase Gasoline Octane
Utility
Fuel Gas (M BTU)
Electricity (kWh)
Amount Required to Increase
Octane3 1 Point (per barrel gasoline)
1.87
0.057
Note:
a(R + M)/2 method
5.4.3 Refinery Demand Impacts
Using our refinery-by-refinery model, along with the technology vendor data for new and
revamped FCC postreating and LSR hydrotreating data (shown in Table 5-26 through Table 5-29
and Table 5-31) we were able to calculate the increases in refinery demands as a result of our
proposed Tier 3 regulations for each of the various scenarios outlined in the previous sections.
This information is summarized in Table 5-39 and Table 5-40 below. The refineries have been
identified by randomly assigned numbers to protect confidential business information (CBI).
This information was submitted to our sister office, OAQPS, to serve as the basis for
their emissions and permitting analysis of the Tier 3 regulations. Based on the data provided,
OAQPS determined that under the proposed Tier 3 gasoline sulfur standard primary ABT
scenario "high case" at most only about 19 refineries would be expected to need New Source
Review (NSR) permits. This equates to approximately 17% of the 111 refineries projected to be
affected by the proposed Tier 3 standards. Of these 19 refineries, about 13 refineries would need
permits for NAAQS-related pollutants and their precursors and for GHGs. In addition to these
13 refineries, the analysis estimated that 6 other refineries may require a permit addressing only
GHG emissions. A technical memorandum describing the OAQPS analysis and results is in the
public docket for this proposal.22
5-62
-------
Table 5-39 Tier 3 Refinery Energy, Hydrogen and Sulfur Plant Demand Increases (10 ppm,
No ABT Cases)
Refinery
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Demand Estimates
Fuel Gas Demands
Low Impact Cases
(Million BTU/Yr)
Normal
Volumes
117,423
0
2,239,110
173,672
46,380
4,433
60,923
23,941
107,543
25,924
13,984
76,663
51,432
0
16,876
831,027
31,964
150,568
27,087
43,909
165,539
14,886
83,222
190,870
702,808
20,503
0
3,145,454
0
64,080
0
62,002
1,693,620
16,574
10,469
337,961
5,175
0
Maximum
Volumes
117,423
0
2,739,637
173,672
58,279
7,291
69,776
29,792
135,777
29,812
16,250
92,786
51,432
0
16,876
1,244,900
31,964
150,568
44,061
46,106
212,719
16,690
83,222
210,535
702,808
22,001
0
3,528,532
0
66,959
0
62,002
1,948,219
24,879
10,469
411,915
5,175
0
High Impact Cases
(Million BTU/Yr)
Normal
Volumes
243,241
0
2,566,562
569,009
261,528
14,523
199,605
134,995
222,775
146,179
78,855
251,175
168,509
6,444
89,463
899,341
82,442
311,900
152,736
143,861
343,091
38,395
272,663
206,560
798,012
67,175
0
3,223,899
0
165,279
0
159,918
1,832,841
93,456
12,340
365,743
29,180
2,907
Maximum
Volumes
243,241
0
3,140,286
569,009
328,624
23,888
228,609
167,988
281,262
168,106
91,630
303,997
168,509
6,444
89,463
1,347,235
82,442
311,900
248,452
151,058
440,875
43,048
272,663
227,841
798,012
72,082
0
3,616,532
0
172,703
0
159,918
2,108,370
140,290
12,340
445,776
29,180
2,907
Sulfur Plant Production
Sulfur
Production
Increase
(Tons
Sulfur/Day)
0.28
0.00
0.00
0.43
0.13
0.04
0.10
0.17
0.20
0.10
0.09
0.43
0.47
0.00
0.07
0.26
2.19
0.22
0.26
0.37
0.17
0.01
0.41
0.33
0.10
0.90
0.00
0.31
0.00
0.15
0.00
0.35
0.27
0.15
0.04
0.15
0.13
0.00
Sulfur Plant
Capacity
Increase
(Percent of
Existing
Facility)
0.02
0.00
0.00
0.03
0.04
1.96
0.03
0.07
0.18
0.08
0.05
0.07
1.18
0.00
0.13
0.22
0.96
0.03
a
0.08
0.04
0.02
0.12
0.29
0.09
0.60
0.00
0.07
0.00
0.01
0.00
3.46
0.03
0.21
0.96
a
3.30
0.00
Hydrogen
Hydrogen
Demand
Increase
(million
scf/year)
426
0
863
1,628
655
42
571
338
390
366
198
719
482
27
219
218
206
547
383
412
538
96
780
50
256
192
0
201
0
414
0
400
444
234
6
89
73
12
5-63
-------
Refinery
Number
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
Demand Estimates
Fuel Gas Demands
Low Impact Cases
(Million BTU/Yr)
21,773
19,171
177,093
3,288
41,705
3,609,133
27,946
6,846
1,416,139
21,918
7,024
2,170,692
0
3,945,526
102,473
40,155
21,405
26,807
91,964
0
52,160
109,050
23,773
9,233
136,167
54,151
360,695
8,113
26,289
89,955
30,259
9,369
37,188
1,079,231
15,845
38,525
0
14,588
16,875
139,558
0
64,517
2,009,982
58,475
21,773
23,883
203,657
4,683
43,146
4,265,613
41,431
8,054
1,416,139
41,206
7,191
2,170,692
0
3,984,109
124,044
51,299
26,758
28,515
107,033
0
62,683
109,050
23,773
9,655
175,790
60,794
402,762
8,589
34,157
103,448
33,479
20,299
37,188
1,299,822
17,618
40,291
0
14,588
20,135
139,558
0
76,071
2,009,982
60,441
High Impact Cases
(Million BTU/Yr)
71,336
91,207
636,542
10,773
107,567
3,905,818
105,829
38,603
1,532,550
45,403
18,116
2,349,131
0
4,269,863
264,305
108,185
120,695
70,005
518,563
0
108,050
357,285
36,151
30,251
282,068
165,033
390,346
20,926
67,806
507,237
99,139
61,521
121,841
1,167,948
51,915
101,764
0
47,795
43,524
289,093
0
65,729
2,280,122
191,583
71,336
113,621
732,024
15,343
111,284
4,616,262
156,897
45,416
1,532,550
85,358
18,547
2,349,131
0
4,311,617
319,941
138,209
150,884
74,463
603,534
0
129,848
357,285
36,151
31,632
364,148
185,278
435,870
22,152
88,100
583,323
109,690
133,298
121,841
1,406,672
57,724
106,429
0
47,795
51,932
289,093
0
77,500
2,280,122
198,026
Sulfur Plant Production
Sulfur
Production
0.58
0.04
0.30
0.03
0.52
0.66
0.06
0.05
0.25
0.02
0.19
0.61
0.00
0.61
0.46
0.11
0.16
0.04
0.00
0.00
0.10
0.62
0.04
0.04
0.08
0.24
0.05
0.04
0.19
0.20
0.14
0.20
1.31
0.09
0.03
0.09
0.00
0.03
0.13
0.16
0.00
0.07
0.24
0.26
Sulfur Plant
Capacity
0.93
0.01
0.03
0.97
0.26
0.22
a
0.55
0.03
0.01
2.16
0.23
0.00
0.03
a
0.09
0.13
0.13
0.03
0.00
0.03
0.11
0.00
1.96
0.01
0.12
0.19
0.39
0.17
0.02
0.14
0.03
0.44
0.14
0.00
0.04
0.00
0.00
0.70
0.02
0.00
0.10
0.05
a
Hydrogen
Hydrogen
Demand
204
214
1,288
31
269
946
221
97
371
80
45
569
0
1,034
662
236
302
144
1,299
0
189
1,022
52
87
494
221
95
52
170
1,271
284
103
349
283
149
215
0
137
109
507
0
234
729
548
5-64
-------
Refinery
Number
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
Demand Estimates
Fuel Gas Demands
Low Impact Cases
(Million BTU/Yr)
2,008,357
0
23,472
20,570
97,698
4,171
147,823
786,104
786,104
25,608
35,993
4,151,496
1,232
2,457
28,695
0
46,468
2,591,010
0
32,339
0
74,538
5,400
139,029
0
8,314
1,485,414
122,762
27,548
2,021,700
0
27,526
24,731
112,353
5,743
147,823
984,013
984,013
25,608
47,565
4,231,505
2,692
3,308
28,695
0
50,408
2,708,025
0
57,437
0
74,538
5,668
139,029
0
8,879
1,704,414
122,762
33,475
High Impact Cases
(Million BTU/Yr)
2,173,451
0
132,354
115,988
363,505
27,391
306,331
850,725
850,725
53,152
36,669
4,606,346
8,090
13,417
74,012
0
152,246
2,804,001
0
111,894
0
154,710
35,458
455,507
11,540
17,223
1,607,521
279,765
155,338
2,187,891
0
155,214
139,454
418,031
37,710
306,331
1,064,903
1,064,903
53,152
48,459
4,695,121
17,675
18,060
74,012
0
165,155
2,930,635
0
198,734
0
154,710
37,218
455,507
11,540
18,393
1,844,524
279,765
188,757
Sulfur Plant Production
Sulfur
Production
0.48
0.00
0.27
0.00
0.35
0.08
0.13
0.30
0.41
0.03
0.05
0.65
0.01
0.02
0.52
0.00
0.26
0.83
0.00
0.12
0.00
0.00
0.00
0.25
0.05
0.00
0.32
0.22
0.10
Sulfur Plant
Capacity
0.09
0.00
0.14
0.00
0.10
0.12
0.01
0.18
1.47
0.08
a
0.08
0.03
a
a
0.00
0.66
0.13
0.00
0.06
0.00
0.00
0.00
0.02
a
0.00
0.05
0.06
0.07
Hydrogen
Hydrogen
Demand
526
0
332
291
752
46
495
206
206
56
131
1,307
14
33
185
0
436
679
0
221
0
162
59
1,303
48
30
389
515
389
The refinery did not have published information on the capacity of the existing sulfur plant.
5-65
-------
Table 5-40 Tier 3 Refinery Energy, Hydrogen and Sulfur Plant Demand Increase
(10 ppm, ABT Cases)
Refinery
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Demand Estimates
Fuel Gas Demands
Low Impact Cases
(Million BTU/Yr)
Normal
Volumes
20,503
64,080
323,504
89,955
0
426,001
0
26,807
117,423
5,400
1,525,387
8,113
7,024
74,538
0
2,591,010
177,093
0
19,171
26,289
2,239,110
83,222
147,823
0
2,273,706
0
107,543
16,875
150,568
0
30,259
0
139,558
9,369
0
0
109,050
0
Maximum
Volumes
22,001
66,959
388,955
103,448
0
692,966
0
28,515
117,423
5,668
1,525,387
8,589
7,191
74,538
0
2,708,025
203,657
0
23,883
34,157
2,739,637
83,222
147,823
0
2,751,864
0
135,777
20,135
150,568
0
33,479
0
139,558
20,299
0
0
109,050
0
High Impact Cases
(Million BTU/Yr)
Normal
Volumes
67,175
229,725
476,338
507,237
0
627,257
0
70,005
243,241
35,458
1,558,831
20,926
18,116
154,710
0
2,804,001
636,542
0
91,207
67,806
2,566,562
272,663
306,331
2,907
2,323,556
0
222,775
43,524
311,900
0
99,139
0
289,093
61,521
6,444
0
357,285
0
Maximum
Volumes
72,082
240,045
572,710
583,323
0
1,020,346
0
74,463
243,241
37,218
1,558,831
22,152
18,547
154,710
0
2,930,635
732,024
0
113,621
88,100
3,140,286
272,663
306,331
2,907
2,812,199
0
281,262
51,932
311,900
0
109,690
0
289,093
133,298
6,444
0
357,285
0
Sulfur Plant Production
Sulfur
Production
Increase
(Tons
Sulfur/Day)
0.90
0.22
0.00
0.20
0.00
0.29
0.00
0.04
0.28
0.00
0.51
0.04
0.19
0.00
0.00
0.83
0.30
0.00
0.04
0.19
0.00
0.41
0.13
0.00
0.50
0.00
0.20
0.13
0.22
0.00
0.14
0.00
0.16
0.20
0.00
0.00
0.62
0.00
Sulfur Plant
Capacity
Increase
(Percent of
Existing
Facility)
0.60
0.02
0.00
0.02
0.00
a
0.00
0.13
0.02
0.00
1.28
0.39
2.16
0.00
0.00
0.13
0.03
0.00
0.01
0.17
0.00
0.12
0.01
0.00
0.08
0.00
0.18
0.70
0.03
0.00
0.14
0.00
0.02
0.03
0.00
0.00
0.11
0.00
Hydrogen
Hydrogen
Demand
Increase
(million
scf/day)
192
467
528
1,271
0
696
0
144
426
59
64
52
45
162
0
679
1,288
0
214
170
863
780
495
12
96
0
390
109
547
0
284
0
507
103
27
0
1,022
0
5-66
-------
Refinery
Number
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
Demand Estimates
Fuel Gas Demands
Low Impact Cases
(Million BTU/Yr)
10,469
210,772
173,672
21,773
3,145,454
337,961
3,945,526
97,698
1,416,139
16,876
433,258
4,433
58,475
2,170,692
8,497
31,964
336,635
32,339
38,525
54,151
0
23,773
190,870
3,288
0
2,008,357
2,457
41,705
46,468
1,079,231
3,071
4,123,377
786,104
37,188
0
40,155
1,485,414
122,762
102,473
0
3,609,133
43,909
1,693,620
0
10,469
269,511
173,672
21,773
3,528,532
411,915
3,984,109
112,353
1,416,139
16,876
526,468
7,291
60,441
2,170,692
8,497
31,964
420,834
57,437
40,291
60,794
0
23,773
210,535
4,683
0
2,021,700
3,308
43,146
50,408
1,299,822
3,325
4,123,377
984,013
37,188
0
51,299
1,704,414
122,762
124,044
0
4,265,613
46,106
1,948,219
0
High Impact Cases
(Million BTU/Yr)
12,340
319,729
569,009
71,336
3,223,899
365,743
4,269,863
363,505
1,532,550
89,463
637,943
14,523
191,583
2,349,131
12,247
82,442
495,673
111,894
101,764
165,033
0
36,151
206,560
10,773
0
2,173,451
13,417
149,509
152,246
1,167,948
7,294
4,213,781
850,725
121,841
0
108,185
1,607,521
279,765
264,305
0
3,905,818
143,861
1,832,841
0
12,340
408,832
569,009
71,336
3,616,532
445,776
4,311,617
418,031
1,532,550
89,463
775,189
23,888
198,026
2,349,131
12,247
82,442
619,649
198,734
106,429
185,278
0
36,151
227,841
15,343
0
2,187,891
18,060
154,676
165,155
1,406,672
7,896
4,213,781
1,064,903
121,841
0
138,209
1,844,524
279,765
319,941
0
4,616,262
151,058
2,108,370
0
Sulfur Plant Production
Sulfur
Production
0.04
0.07
0.43
0.58
0.31
0.15
0.61
0.35
0.25
0.07
0.14
0.04
0.26
0.61
0.00
2.19
0.19
0.12
0.09
0.24
0.00
0.04
0.33
0.03
0.00
0.48
0.02
0.63
0.26
0.09
0.00
0.38
0.30
1.31
0.00
0.11
0.32
0.22
0.46
0.00
0.66
0.37
0.27
0.00
Sulfur Plant
Capacity
0.96
0.00
0.03
0.93
0.07
a
0.03
0.10
0.03
0.13
0.09
1.96
a
0.23
0.00
0.96
0.15
0.06
0.04
0.12
0.00
0.00
0.29
0.97
0.00
0.09
a
0.31
0.66
0.14
0.00
0.03
0.18
0.44
0.00
0.09
0.05
0.06
a
0.00
0.22
0.08
0.03
0.00
Hydrogen
Hydrogen
Demand
6
379
1,628
204
201
89
1,034
752
371
219
707
42
548
569
16
206
550
221
215
221
0
52
50
31
0
526
33
304
436
283
17
174
206
349
0
236
389
515
662
0
946
412
444
0
5-67
-------
Refinery
Number
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
Demand Estimates
Fuel Gas Demands
Low Impact Cases
(Million BTU/Yr)
60,923
4,171
4,151,496
8,314
5,175
96,113
988,379
0
9,233
21,918
14,886
35,993
202,851
786,104
25,924
1,232
369,152
0
28,695
828,747
0
13,984
62,002
260,661
194,048
0
428,992
52,160
0
69,776
5,743
4,231,505
8,879
5,175
113,326
1,480,617
0
9,655
41,206
16,690
47,565
261,880
984,013
29,812
2,692
432,912
0
28,695
828,747
0
16,250
62,002
391,286
194,048
0
479,023
62,683
0
High Impact Cases
(Million BTU/Yr)
199,605
27,391
4,606,346
17,223
29,180
98,454
1,109,185
0
30,251
45,403
38,395
36,669
506,929
850,725
146,179
8,090
543,551
0
74,012
965,965
11,540
78,855
159,918
383,805
294,359
0
481,425
108,050
0
228,609
37,710
4,695,121
18,393
29,180
116,085
1,661,587
0
31,632
85,358
43,048
48,459
654,442
1,064,903
168,106
17,675
637,434
0
74,012
965,965
11,540
91,630
159,918
576,142
294,359
0
537,572
129,848
0
Sulfur Plant Production
Sulfur
Production
0.10
0.08
0.65
0.00
0.13
0.13
0.31
0.00
0.04
0.02
0.01
0.05
0.15
0.41
0.10
0.01
0.30
0.00
0.52
0.13
0.06
0.09
0.35
0.17
0.07
0.00
0.08
0.10
0.00
Sulfur Plant
Capacity
0.03
0.12
0.08
0.00
3.30
0.17
0.26
0.00
1.96
0.01
0.02
a
0.01
1.47
0.08
0.03
0.15
0.00
a
0.11
a
0.05
3.46
0.25
0.00
0.00
0.27
0.03
0.00
Hydrogen
Hydrogen
Demand
571
46
1,307
30
73
543
412
0
87
80
96
131
1,146
206
366
14
603
0
185
412
48
198
400
426
349
0
179
189
0
" The refinery did not have published information on the capacity of the existing sulfur plant.
5-68
-------
Chapter 5 Appendix
LP Refinery Modeling Output Tables
5-69
-------
Table 5-41 Volume and Cost Information Used for Estimating the Cost of Octane
FCC Naphtha
Volume (kbbl)
Obj Function
($000)
Cost
PADD
Summer
Winter
Total
Reference
Control
Difference
($/bbl)
(c/gal)
PADD1
33,001
33,692
66,693
10,021,915
9,961,128
PADD 2
90,996
137,617
228,613
19,715,175
19,662,097
PADD 3
248,651
302,644
551,294
51,232,848
51,072,893
PADD 4/5
20,296
29,719
50,015
11,926,884
11,913,432
USA
392,943
503,671
„ __
92,896,822
92,609,550
-287,272
-0.3203965
-0.7628488
5-70
-------
Table 5-42 PADD 1 Unit Capacity and Throughput Volumes from LP Refinery Modeling
(Thousand bbl/day)
PADD1
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Is o octane
DHT- Total
DHT 2nd RCT- Total
DHT Arom Saturation
NHT- Total Fd
CGH- Generic
CGH-OlefinSat'n
FCCUFdHDT
LSR Splitter
LSRBzSaturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant -Total MSCF
Lube Unit
Sulfur Plant
MeroxJet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2009
1,260
492
0
0
0
490
0
22
0
44
0
0
223
0
0
32
7
40
29
0
5
0
334
0
0
274
119
119
44
0
0
0
0
32,600
3
0
64
146
1,434
0
0
13
23
0
0
0
2017 Refcase
Summer
1,982
788
110
157
0
490
0
69
0
153
0
18
281
0
29
32
126
40
47
0
1
0
299
240
0
571
158
0
1,459
87
0
1
3
33
0
0
974
53
1
0
0
11
23
0
0
0
Winter
2,791
1,159
120
93
0
291
0
69
0
292
0
35
337
0
36
32
149
20
44
0
1
0
530
460
0
650
158
0
1,459
51
0
1
3
33
0
0
1,007
53
2
0
0
13
23
0
0
0
Control Case Minus 1
ON
Summer
1,982
788
109
157
0
490
21
69
0
153
0
18
284
0
25
32
133
40
49
0
1
0
299
240
0
590
158
0
1,459
100
0
1
3
33
0
0
974
53
1
0
0
11
23
0
0
0
Winter
2,792
1,159
120
93
0
291
0
69
0
292
0
35
337
0
36
32
158
20
44
0
1
0
530
460
0
651
158
0
1,459
50
0
1
3
33
0
0
1,008
53
2
0
0
13
23
0
0
0
Control Case Relative
to Ref Case
Summer
0
0
-1
0
0
0
21
0
0
0
0
0
3
0
-4
0
7
0
3
0
0
0
0
0
0
19
0
0
0
13
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Winter
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
9
0
0
0
0
0
0
0
0
0
0
0
0
-1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
5-71
-------
Table 5-43 PADD 2 Unit Capacity and Throughput Volumes from LP Refinery Modeling
(Thousand bbl/day)
PADD 2
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4Alkylation
Dimersol
Cat Poly
Isooctane
DHT- Total
DHT 2nd RCT- Total
DHT Arom Saturation
NHT - Total Fd
CGH- Generic
CGH-OlefinSat'n
FCCU Fd HOT
LSR Splitter
LSR BzSaturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant - Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2009
3,612
1,452
0
0
0
1,144
0
234
0
375
0
0
807
0
0
15
157
156
81
3
4
0
1,056
0
6
1,038
282
282
521
0
0
0
0
32,600
0
0
799,200
66
1,434
0
0
18
34
0
0
0
2017 Refcase
Summer
2,996
1,326
167
255
0
852
500
216
0
280
0
34
516
0
42
65
168
126
81
0
0
0
691
512
0
818
127
0
521
71
0
1
3
0
0
0
1,004
24
4
0
0
16
34
0
0
0
Winter
3,831
1,660
171
301
0
1,036
594
216
0
329
0
40
529
0
54
60
125
78
40
0
0
0
577
528
0
929
176
0
521
100
0
1
3
36
0
0
1,066
24
5
0
0
18
34
0
0
0
Control Case Minus 1
ON
Summer
2,996
1,326
169
255
0
852
500
216
0
280
0
34
540
0
42
65
170
126
81
0
0
0
691
512
0
823
133
0
521
75
0
1
3
0
0
0
985
24
4
0
0
16
34
0
0
0
Winter
3,840
1,664
175
302
0
1,039
596
216
0
329
0
40
542
0
54
63
141
78
40
0
0
0
577
529
0
932
177
0
521
100
0
1
3
36
0
0
1,058
24
5
0
0
18
34
0
0
0
Control Case Relative
to RefCase
Summer
0
0
2
0
0
0
0
0
0
0
0
0
24
0
0
0
2
0
0
0
0
0
0
0
0
5
5
0
0
4
0
0
0
0
0
0
-19
0
0
0
0
0
0
0
0
0
Winter
10
4
4
1
0
3
2
0
0
0
0
0
13
0
0
3
16
0
0
0
0
0
0
0
0
3
1
0
0
0
0
0
0
0
0
0
-7
0
0
0
0
0
0
0
0
0
5-72
-------
Table 5-44 PADD 3 Unit Capacity and Throughput Volumes from LP Refinery Modeling
(Thousand bbl/day)
PADD 3
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT- Total
DHT 2nd RCT- Total
DHT Arom Saturation
NHT- Total Fd
CGH- Generic
CGH-OlefinSat'n
FCCUFdHDT
LSR Splitter
LSR BzSaturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant - Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer- Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2009
8,735
3,963
0
0
0
2,808
0
523
0
1,350
51
0
1,765
0
0
50
199
295
290
19
23
0
2,606
0
4
2,284
905
905
1,375
0
0
0
0
32,600
28
3
1,569,400
1,169
1,434
0
0
146
318
0
0
0
2017 Refcase
Summer
5,576
2,587
534
839
0
2,808
1,636
1,416
144
625
0
74
524
0
381
133
47
235
290
0
0
0
2,080
1,592
0
1,177
173
0
1,559
1
0
10
30
263
0
0
5,345
506
15
0
0
146
318
0
0
0
Winter
5,706
2,647
522
728
0
2,411
1,207
1,416
144
648
51
79
410
0
383
133
0
148
267
0
0
0
2,032
1,565
0
1,041
18
0
1,559
1
0
1
3
263
0
0
5,487
506
13
0
0
146
318
0
0
0
Control Case Minus 1
ON
Summer
5,589
2,593
541
839
0
2,808
1,636
1,397
144
628
0
74
613
0
385
133
75
235
290
0
0
0
2,086
1,599
0
1,206
196
0
1,559
1
0
10
30
263
0
0
5,217
506
15
0
0
146
318
0
0
0
Winter
5,706
2,647
523
730
0
2,422
1,218
1,397
144
649
49
79
454
0
383
133
0
148
270
0
0
0
2,038
1,570
0
1,049
26
0
1,559
1
0
1
3
263
0
0
5,414
506
13
0
0
146
318
0
0
0
Control Case Relative
to Ref Case
Summer
13
6
8
0
0
0
0
-19
0
2
0
0
89
0
4
0
28
0
0
0
0
0
5
6
0
30
23
0
0
0
0
0
0
0
0
0
-129
0
0
0
0
0
0
0
0
0
Winter
0
0
0
3
0
11
11
-19
0
2
-2
0
44
0
0
0
0
0
3
0
0
0
6
6
0
8
8
0
0
0
0
0
0
0
0
0
-73
0
0
0
0
0
0
0
0
0
5-73
-------
Table 5-45 PADD 4 and PADD 5OC Unit Capacity and Throughput Volumes from LP
Refinery Modeling (Thousand bbl/day)
PADD 4 and 5OC
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT- Total
DHT 2nd RCT- Total
DHT Arom Saturation
NHT- Total Fd
CGH- Generic
CGH-OlefinSat'n
FCCUFdHDT
LSR Splitter
LSR BzSaturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant - Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer- Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2009
1,796
599
0
0
0
344
0
91
0
143
56
0
0
287
0
19
43
37
38
1
10
0
404
0
0
326
66
66
89
0
0
0
0
32,600
0
0
373,900
0
1,434
0
0
0
0
0
0
0
2017 Refcase
Summer
1,518
559
71
66
0
224
132
144
0
133
41
17
0
214
0
19
2
19
38
1
0
0
492
331
0
426
0
0
89
100
0
10
30
14
0
0
518
0
2
0
0
0
0
0
0
0
Winter
1,558
541
71
70
0
224
140
194
0
133
20
16
0
211
0
19
0
36
38
0
0
0
483
331
0
435
0
0
89
100
0
4
11
28
0
0
619
0
2
0
0
0
0
0
0
0
Control Case Minus 1
ON
Summer
1,518
559
71
66
0
224
132
144
0
133
41
17
0
236
0
19
2
19
38
1
0
0
492
331
0
434
0
0
89
100
0
1
3
14
0
0
499
0
2
0
0
0
0
0
0
0
Winter
1,558
541
68
70
0
224
140
194
0
133
20
16
0
248
0
19
0
36
38
0
0
0
483
331
0
434
0
0
89
100
0
1
3
28
0
0
603
0
2
0
0
0
0
0
0
0
Control Case Relative
to Ref Case
Summer
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Winter
0
0
-3
0
0
0
0
0
0
0
0
0
0
37
0
0
0
0
0
0
0
0
0
0
0
-1
0
0
0
0
0
-3
-8
0
0
0
-16
0
0
0
0
0
0
0
0
0
5-74
-------
Table 5-46 U.S. (except CA) Unit Capacity and Throughput Volumes from LP Refinery
Modeling (Thousand bbl/day)
PADDs 1 -4 and 5OC
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT- Total
DHT 2nd RCT- Total
DHT Arom Saturation
NHT- Total Fd
CGH- Generic
CGH-OlefinSat'n
FCCUFdHDT
LSR Splitter
LSRBzSaturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant - Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2009
15,403
6,505
0
0
0
4,786
0
869
0
1,912
107
0
2,794
287
0
116
406
528
438
23
42
0
4,400
0
10
3,921
1,373
1,373
2,029
0
0
0
0
130,400
31
3
2,742,564
1,380
5,736
0
0
177
375
0
0
0
2017 Refcase
Summer
12,071
5,260
882
1,317
0
4,373
2,267
1,845
144
1,190
41
143
1,321
214
452
250
343
420
455
1
1
0
3,562
2,676
0
2,992
459
0
3,628
259
0
22
66
310
0
0
7,841
582
22
0
0
173
375
0
0
0
Winter
13,886
6,007
885
1,192
0
3,962
1,941
1,895
144
1,401
71
170
1,276
211
473
245
275
281
389
0
1
0
3,622
2,885
0
3,056
352
0
3,628
252
0
7
20
359
0
0
8,178
582
22
0
0
177
375
0
0
0
Control Case Minus 1
ON
Summer
12,084
5,266
890
1,317
0
4,373
2,288
1,826
144
1,193
41
144
1,437
236
452
249
380
420
458
1
1
0
3,567
2,682
0
3,054
487
0
3,628
276
0
13
39
310
0
0
7,675
582
22
0
0
173
375
0
0
0
Winter
13,897
6,012
886
1,196
0
3,976
1,954
1,876
144
1,403
70
171
1,333
248
473
247
299
281
392
0
1
0
3,628
2,891
0
3,066
361
0
3,628
251
0
4
12
359
0
0
8,084
582
22
0
0
177
375
0
0
0
Control Case Relative
to RefCase
Summer
13
6
9
0
0
0
20
-20
0
2
0
0
115
0
1
-1
37
0
3
0
0
0
5
6
0
53
28
0
0
17
0
0
0
0
0
0
-147
0
0
0
0
0
0
0
0
0
Winter
11
5
1
3
0
13
12
-20
0
2
-2
0
57
37
0
2
25
0
3
0
0
0
6
6
0
10
9
0
0
-1
0
-3
-8
0
0
0
-95
0
0
0
0
0
0
0
0
0
5-75
-------
Table 5-47 PADD 1 Gasoline Qualities Estimated by LP Refinery Modeling
Gasoline Qualities
and Volume
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur (ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volumes (kbbl/day)
2017 Ref Case
Summer
CG
4.84
254.7
23.6
45.9
81.8
9.11
124
208
335
1145
137.4
20.8
0.42
9.17
11.2
4.23
1,693,907
RFC
4.96
261.9
22.0
45.5
75.8
7.00
137
209
362
1195
148.8
22.0
0.42
11.48
13.8
5.06
169,974
Pool
4.85
255.3
23.5
45.9
81.2
8.92
125
208
338
1150
138.4
20.9
0.42
9.38
11.4
4.31
1,863,881
Winter
CG
4.87
255.7
22.2
48.7
81.5
12.15
104
202
336
1100
120.1
21.5
0.50
8.61
11.1
4.19
345,786
RFC
4.82
258.1
24.9
52.2
79.1
11.93
106
195
347
1092
120.1
21.2
0.63
8.30
13.7
5.12
240,886
Pool
4.85
256.7
23.3
50.2
80.5
12.06
105
199
341
1097
120.1
21.3
0.55
8.48
12.2
4.57
586,672
2017 minus ION in FCC naphtha
Summer
CG
4.84
255.0
23.4
45.3
81.0
9.11
124
209
339
1152
137.6
22.1
0.45
9.09
11.2
4.22
1,709,571
RFC
4.95
261.7
21.2
46.2
76.5
7.00
137
208
359
1187
148.6
22.0
0.51
11.35
13.8
5.07
1,118,250
Pool
4.88
257.6
22.6
45.7
79.2
8.28
129
209
347
1166
141.9
22.0
0.47
9.98
12.2
4.56
2,827,821
Winter
CG
4.87
255.7
22.6
48.4
80.8
12.17
104
203
339
1105
120.1
22.3
0.51
8.74
11.1
4.19
1,623,409
RFC
4.81
258.2
24.9
52.0
78.6
11.94
106
196
349
1095
120.1
21.7
0.63
8.15
13.7
5.12
1,130,922
Pool
4.85
256.7
23.6
49.9
79.9
12.07
105
200
344
1101
120.1
22.0
0.56
8.50
12.2
4.57
2,754,331
Table 5-48 PADD 2 Gasoline Qualities Estimated by LP Refinery Modeling
Gasoline Qualities
and Volume
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur (ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volumes (kbbl/day)
2017 Ref Case
Summer
CG
4.91
259.9
27.4
56.2
82.9
9.03
124
187
330
1077
134.2
21.2
0.77
8.66
11.2
4.17
1,799,149
RFC
4.97
262.4
20.3
52.4
80.0
7.00
137
195
343
1133
146.4
21.2
0.88
5.44
13.8
5.06
334,626
Pool
4.92
260.3
26.3
55.6
82.4
8.72
126
188
332
1086
136.1
21.2
0.79
8.16
11.6
4.31
2,133,775
Winter
CG
4.89
259.1
25.2
58.0
81.7
11.55
108
184
335
1048
120.1
22.1
0.49
9.10
11.2
4.18
1,710,734
RFC
4.84
258.4
22.7
59.2
82.9
11.47
109
181
330
1036
120.1
20.1
0.67
8.50
12.0
4.48
139,111
Pool
4.88
259.1
25.1
58.1
81.8
11.55
108
183
335
1047
120.1
21.9
0.50
9.06
11.3
4.20
1,849,845
2017 minus ION in FCC naphtha
Summer
CG
4.91
260.0
26.5
56.0
82.8
9.03
124
188
331
1079
134.3
21.8
0.76
8.83
11.2
4.16
1,799,149
RFC
4.97
262.1
18.9
52.4
80.2
7.00
137
195
342
1132
146.4
21.2
0.88
5.08
13.8
5.07
334,626
Pool
4.92
260.3
25.3
55.5
82.4
8.72
126
189
332
1088
136.2
21.7
0.78
8.25
11.6
4.31
2,133,775
Winter
CG
4.88
259.3
25.2
57.9
81.6
11.56
108
184
336
1049
120.1
22.4
0.50
9.14
11.2
4.18
1,710,734
RFC
4.83
258.4
22.4
59.3
82.9
11.47
109
181
330
1036
120.1
20.0
0.65
8.40
12.0
4.48
139,111
Pool
4.88
259.2
25.0
58.0
81.7
11.55
108
183
335
1048
120.1
22.2
0.51
9.08
11.3
4.20
1,849,845
5-76
-------
Table 5-49 PADD 3 Gasoline Qualities Estimated by LP Refinery Modeling
Gasoline Qualities
and Volume
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur (ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volumes (kbbl/day)
2017 Ref Case
Summer
CG
5.06
259.5
28.0
59.0
83.6
9.07
124
182
327
1057
133.1
20.8
0.62
13.42
11.2
4.15
1,276,784
RFC
4.90
254.7
27.2
59.4
88.2
7.00
137
181
306
1053
144.0
6.5
0.61
1.72
13.7
5.20
488,041
Pool
5.01
258.2
27.8
59.1
84.9
8.49
128
181
321
1056
136.1
16.9
0.61
10.18
11.9
4.44
1,764,825
Winter
CG
4.99
254.8
28.0
59.6
85.1
11.45
109
180
320
1024
120.1
15.0
0.58
10.82
11.2
4.23
1,400,642
RFC
4.98
255.8
28.0
61.4
85.3
11.33
110
177
319
1013
120.1
15.2
0.52
10.90
13.7
5.17
497,125
Pool
4.98
255.1
28.0
60.1
85.2
11.42
109
179
320
1021
120.1
15.1
0.56
10.84
11.8
4.47
1,897,767
2017 minus ION in FCC naphtha
Summer
CG
5.02
258.6
28.0
59.5
84.0
9.07
124
180
325
1052
132.9
19.8
0.61
13.07
11.2
4.16
1,276,784
RFC
5.00
261.2
27.2
51.8
80.0
7.00
137
196
343
1137
146.6
18.3
0.61
3.17
13.7
5.07
488,041
Pool
5.01
259.3
27.8
57.4
82.9
8.49
128
185
330
1075
136.7
19.4
0.61
10.33
11.9
4.42
1,764,825
Winter
CG
4.99
255.0
28.0
59.3
84.6
11.47
109
181
322
1028
120.1
15.6
0.58
10.90
11.2
4.22
1,400,642
RFC
4.96
255.6
28.0
62.0
85.4
11.29
110
175
319
1009
120.1
15.3
0.61
10.82
13.7
5.18
497,125
Pool
4.98
255.2
28.0
60.0
84.8
11.42
109
179
321
1023
120.1
15.5
0.58
10.88
11.8
4.47
1,897,767
Table 5-50 PADD 4 and 5OC Gasoline Qualities Estimated by LP Refinery Modeling
Gasoline Qualities
and Volume
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur (ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volumes (kbbl/day)
2017 Ref Case
Summer
CG
4.96
263.3
25.0
58.5
85.5
9.34
122
182
318
1048
131.8
24.4
0.84
5.70
11.3
4.15
701,850
RFC
0.00
0.0
0.0
0.0
0.0
0.00
0
0
0
0
0.0
0.0
0.00
0.00
0.0
0.00
0
Pool
4.96
263.3
25.0
58.5
85.5
9.34
122
182
318
1048
131.8
24.4
0.84
5.70
11.3
4.15
701,850
Winter
CG
4.95
261.2
28.0
58.6
84.7
11.52
108
182
322
1031
120.0
23.8
1.09
5.47
11.1
4.08
734,023
RFC
0.00
0.0
0.0
0.0
0.0
0.00
0
0
0
0
0.0
0.0
0.00
0.00
0.0
0.00
0
Pool
4.95
261.2
28.0
58.6
84.7
11.52
108
182
322
1031
120.0
23.8
1.09
5.47
11.1
4.08
734,023
2017 minus ION in FCC naphtha
Summer
CG
4.96
263.5
24.6
59.0
85.7
9.34
122
181
317
1045
131.6
24.5
0.86
5.72
11.3
4.14
701,850
RFC
0.00
0.0
0.0
0.0
0.0
0.00
0
0
0
0
0.0
0.0
0.00
0.00
0.0
0.00
0
Pool
4.96
263.5
24.6
59.0
85.7
9.34
122
181
317
1045
131.6
24.5
0.86
5.72
11.3
4.14
701,850
Winter
CG
4.95
261.1
28.0
57.9
83.0
11.55
108
184
329
1042
120.1
23.4
1.10
4.98
11.3
4.18
744,898
RFC
0.00
0.0
0.0
0.0
0.0
0.00
0
0
0
0
0.0
0.0
0.00
0.00
0.0
0.00
0
Pool
4.95
261.1
28.0
57.9
83.0
11.55
108
184
329
1042
120.1
23.4
1.10
4.98
11.3
4.18
744,898
5-77
-------
Table 5-51 U.S. (except CA) Gasoline Qualities Estimated by Refinery Modeling
Gasoline Qualities
and Volume
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur (ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volumes (kbbl/day)
2017 Ref Case
Summer
CG
4.93
258.6
26.1
54.0
83.1
9.10
124
192
329
1090
134.6
21.4
0.63
9.55
11.2
4.18
5,471,690
RFC
4.93
258.5
24.0
54.7
83.3
7.00
137
190
328
1104
145.7
14.1
0.67
4.65
13.8
5.13
992,641
Pool
4.93
258.6
25.8
54.1
83.1
8.78
126
191
329
1092
136.3
20.3
0.64
8.80
11.6
4.33
6,464,331
Winter
CG
4.93
257.8
26.4
57.9
83.3
11.56
108
184
328
1041
120.1
20.0
0.63
9.00
11.2
4.18
4,191,185
RFC
4.91
256.9
26.3
58.5
83.2
11.52
108
182
328
1038
120.1
17.6
0.57
9.81
13.4
5.05
877,122
Pool
4.93
257.6
26.4
58.0
83.3
11.55
108
184
328
1041
120.1
19.6
0.62
9.14
11.6
4.33
5,068,307
2017 minus ION in FCC naphtha
Summer
CG
4.92
258.6
25.7
53.9
82.9
9.10
124
192
330
1091
134.6
21.8
0.64
9.50
11.2
4.18
5,487,354
RFC
4.96
261.7
22.3
48.7
78.0
7.00
137
202
352
1165
147.7
20.9
0.60
8.21
13.8
5.07
1,940,917
Pool
4.93
259.4
24.8
52.5
81.6
8.55
127
195
336
1111
138.1
21.5
0.63
9.16
11.9
4.41
7,428,271
Winter
CG
4.91
257.4
25.5
55.4
82.3
11.71
107
189
333
1059
120.1
20.8
0.60
8.91
11.2
4.19
5,479,683
RFC
4.85
257.5
25.6
55.4
80.8
11.72
107
189
339
1066
120.1
19.8
0.62
8.92
13.6
5.08
1,767,158
Pool
4.90
257.4
25.5
55.4
82.0
11.72
107
189
334
1061
120.1
20.5
0.61
8.91
11.8
4.41
7,246,841
5-78
-------
Table 5-52 PADD 1 Unit Capacity and Throughput Volumes from LP Refinery Modeling
(Thousand bbl/day)
PADD1
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT - Total
DHT 2nd RCT - Total
DHT Arom Saturation
NHT - Total Fd
CGH - Generic
CGH-OlefinSat'n
FCCU Fd HOT
LSR Splitter
LSR Bz Saturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant - Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2009
1,260
492
0
0
0
490
0
22
0
44
0
0
223
0
0
32
7
40
29
0
5
0
334
0
0
274
119
119
44
0
0
0
0
0
3
0
63,700
146
1,434
0
0
13
23
0
0
0
Reference 2005 in
2030
Summer
1,156
507
104
185
0
678
387
52
0
22
0
3
138
0
36
53
51
40
125
0
1
0
431
0
0
263
161
0
460
1
0
21
63
33
0
0
645
53
1
0
0
13
23
0
0
0
Winter
1,716
738
62
226
0
827
467
82
0
94
0
11
223
0
36
0
95
20
68
0
31
0
742
0
0
397
158
0
460
51
0
1
3
33
0
0
935
53
1
0
0
13
23
0
0
0
100%E10Casein
2030
Summer
757
332
70
156
0
565
324
52
0
22
0
3
90
0
36
32
32
40
94
0
1
0
300
0
0
165
159
0
396
1
0
19
57
33
0
0
538
53
1
0
0
13
23
0
0
0
Winter
1,267
550
50
191
0
691
391
82
0
82
0
10
154
0
36
0
32
20
20
0
32
0
572
0
0
286
158
0
396
24
0
1
3
33
0
0
788
53
1
0
0
13
23
0
0
0
E10 Case Relative to
RefCase
Summer
-400
-175
-34
-29
0
-113
-63
0
0
0
0
0
-48
0
0
-21
-19
0
-31
0
0
0
-130
0
0
-98
-2
0
-63
0
0
-2
-6
0
0
0
-107
0
0
0
0
0
0
0
0
0
Winter
-449
-187
-12
-36
0
-136
-76
0
0
-12
0
-1
-70
0
0
0
-63
0
-48
0
1
0
-170
0
0
-111
0
0
-63
-27
0
0
0
0
0
0
-147
0
0
0
0
0
0
0
0
0
1 00% E1 5 Case in
2030
Summer
757
332
72
156
0
565
324
52
0
22
0
3
134
0
36
32
17
40
77
0
8
0
310
0
0
152
158
0
396
1
0
10
30
33
0
0
518
53
1
0
0
13
23
0
0
0
Winter
1,267
550
47
191
0
691
391
82
0
82
0
10
154
0
36
0
8
20
20
0
32
0
572
0
0
286
158
0
396
40
0
1
3
33
0
0
795
53
1
0
0
13
23
0
0
0
E15Case Relative to
RefCase
Summer
-400
-175
-32
-29
0
-113
-63
0
0
0
0
0
-3
0
0
-21
-34
0
-48
0
7
0
-121
0
0
-111
-3
0
-63
0
0
-11
-33
0
0
0
-127
0
0
0
0
0
0
0
0
0
Winter
-449
-187
-15
-36
0
-136
-76
0
0
-12
0
-1
-70
0
0
0
-87
0
-48
0
1
0
-170
0
0
-111
0
0
-63
-11
0
0
0
0
0
0
-140
0
0
0
0
0
0
0
0
0
5-79
-------
Table 5-53 PADD 2 Unit Capacity and Throughput Volumes from LP Refinery Modeling
(Thousand bbl/day)
PADD 2
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT - Total
DHT 2nd RCT - Total
DHT Arom Saturation
NHT - Total Fd
CGH - Generic
CGH-OlefinSat'n
FCCU Fd HOT
LSR Splitter
LSR Bz Saturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant - Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2009
3,612
1,452
0
0
0
1,144
0
234
0
375
0
0
807
0
0
15
157
156
81
3
4
0
1,056
0
6
1,038
282
282
521
0
0
0
0
0
0
0
799,200
66
1,434
0
0
18
34
0
0
0
Reference 2005 in
2030
Summer
2,638
1,092
162
248
0
864
503
234
0
153
0
19
393
0
42
50
157
121
81
0
0
0
827
0
0
621
242
0
521
100
0
2
5
36
0
0
1,149
24
4
0
0
16
34
0
0
0
Winter
4,141
1,770
137
332
0
1,144
653
234
0
395
0
48
613
0
54
0
300
78
71
0
29
0
644
0
0
1,007
233
0
521
100
0
10
30
36
0
0
1,149
24
5
0
0
18
34
0
0
0
100%E10Casein
2030
Summer
2,494
1,026
136
204
0
666
6
234
0
129
0
16
376
0
42
27
157
78
79
0
0
0
874
0
0
597
1
0
650
93
0
1
3
36
0
0
1,197
24
4
0
0
16
34
0
0
0
Winter
3,972
1,711
144
319
0
1,080
628
234
0
375
0
45
515
0
54
27
129
92
81
0
3
0
610
0
0
966
183
0
650
100
0
1
3
36
0
0
1,224
24
5
0
0
18
34
0
0
0
E10 Case Relative to
RefCase
Summer
-145
-66
-26
-44
0
-197
-496
0
0
-24
0
-3
-18
0
0
-23
0
-44
-2
0
0
0
47
0
0
-24
-241
0
129
-7
0
-1
-2
0
0
0
48
0
0
0
0
0
0
0
0
0
Winter
-169
-58
7
-13
0
-64
-24
0
0
-21
0
-3
-98
0
0
27
-171
14
10
0
-26
0
-34
0
0
-41
-49
0
129
0
0
-9
-27
0
0
0
75
0
0
0
0
0
0
0
0
0
1 00% E1 5 Case in
2030
Summer
2,206
905
107
166
0
569
0
234
0
86
0
10
329
0
42
10
100
78
40
0
0
0
762
0
0
495
0
0
521
77
0
1
3
36
0
0
1,086
24
3
0
0
16
34
0
0
0
Winter
3,830
1,656
120
302
0
1,031
595
234
0
355
0
43
496
0
54
15
0
78
65
0
8
0
581
0
0
924
179
0
521
71
0
1
3
36
0
0
1,086
24
5
0
0
18
34
0
0
0
E15Case Relative to
RefCase
Summer
-432
-186
-55
-82
0
-295
-503
0
0
-67
0
-8
-64
0
0
-40
-57
-44
-40
0
0
0
-66
0
0
-125
-242
0
0
-23
0
-1
-2
0
0
0
-63
0
-1
0
0
0
0
0
0
0
Winter
-311
-114
-18
-31
0
-113
-58
0
0
-40
0
-5
-116
0
0
14
-300
0
-6
0
-21
0
-63
0
0
-83
-54
0
0
-29
0
-9
-27
0
0
0
-63
0
0
0
0
0
0
0
0
0
5-80
-------
Table 5-54 PADD 3 Unit Capacity and Throughput Volumes from LP Refinery Modeling
(Thousand bbl/day)
PADD 3
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2S04 Alkylation
Dimersol
Cat Poly
Isooctane
DHT - Total
DHT 2nd RCT - Total
DHT Arom Saturation
NHT - Total Fd
CGH - Generic
CGH-OlefinSat'n
FCCU Fd HOT
LSR Splitter
LSR Bz Saturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant - Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2009
8,735
3,963
0
0
0
2,808
0
667
0
1,350
51
0
1,765
0
0
50
199
295
290
19
23
0
2,606
0
4
2,284
905
905
1,375
0
0
0
0
0
28
3
1,569,400
1,169
1,434
0
0
0
0
0
0
0
Reference 2005 in
2030
Summer
7,741
3,514
635
679
0
2,075
6
1,797
0
992
0
120
1,517
0
265
50
130
148
234
0
0
0
2,159
0
0
1,648
1
0
1,844
1
0
327
980
263
0
0
5,529
440
15
0
0
96
211
0
0
0
Winter
7,111
3,210
427
675
0
2,154
1,153
1,797
0
853
0
103
1,140
0
261
0
58
148
235
0
0
0
1,867
0
0
1,198
1
0
1,844
1
0
76
227
263
0
0
5,529
440
13
0
0
94
208
0
0
0
100%E10Casein
2030
Summer
7,477
3,381
624
690
0
2,103
161
2,020
0
933
0
113
1,130
0
261
50
229
148
252
0
0
0
1,726
0
0
1,474
26
0
1,850
100
0
91
273
263
0
0
5,931
440
14
0
0
94
208
0
0
0
Winter
7,036
3,264
471
612
0
1,901
435
2,020
0
881
0
107
904
0
261
0
50
148
197
0
0
0
1,689
0
0
1,076
1
0
1,630
1
0
10
31
263
0
0
5,931
440
14
0
0
94
208
0
0
0
E10 Case Relative to
RefCase
Summer
-264
-133
-11
10
0
28
155
223
0
-59
0
-7
-387
0
-4
0
99
0
18
0
0
0
-434
0
0
-174
25
0
6
99
0
-236
-707
0
0
0
402
0
-1
0
0
-2
-3
0
0
0
Winter
-75
54
44
-63
0
-252
-719
223
0
28
0
3
-236
0
0
0
-7
0
-38
0
0
0
-178
0
0
-122
0
0
-213
0
0
-65
-196
0
0
0
402
0
1
0
0
0
0
0
0
0
1 00% E1 5 Case in
2030
Summer
7,203
3,275
612
607
0
1,820
0
1,937
0
886
0
107
1,149
0
261
50
134
148
197
0
0
0
1,872
0
0
1,828
1
0
1,562
1
0
330
989
263
0
0
5,670
440
13
0
0
94
208
0
0
0
Winter
6,648
3,022
423
571
0
1,794
848
1,937
0
769
0
93
863
0
261
0
0
148
179
0
0
0
1,675
0
0
1,300
1
0
1,559
1
0
272
815
263
0
0
5,670
440
12
0
0
94
208
0
0
0
E15Case Relative to
RefCase
Summer
-538
-239
-23
-73
0
-255
-6
140
0
-106
0
-13
-368
0
-4
0
4
0
-37
0
0
0
-288
0
0
180
0
0
-282
0
0
3
9
0
0
0
141
0
-1
0
0
-2
-3
0
0
0
Winter
-464
-188
-4
-104
0
-360
-305
140
0
-85
0
-10
-277
0
1
0
-58
0
-56
0
0
0
-192
0
0
102
0
0
-285
0
0
196
588
0
0
0
141
0
0
0
0
0
0
0
0
0
5-81
-------
Table 5-55 PADD 4 and 5OC Unit Capacity and Throughput Volumes from LP Refinery
Modeling (Thousand bbl/day)
PADD 4 and 5OC
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2S04 Alkylation
Dimersol
Cat Poly
Isooctane
DHT - Total
DHT 2nd RCT - Total
DHT Arom Saturation
NHT - Total Fd
CGH - Generic
CGH-OlefinSat'n
FCCU Fd HOT
LSR Splitter
LSR Bz Saturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant - Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2009
1,796
599
0
0
0
344
0
91
0
143
56
0
0
287
0
19
43
37
38
1
10
0
404
0
0
326
66
66
89
0
0
0
0
0
0
0
373,900
0
1,434
0
0
0
0
0
0
0
Reference 2005 in
2030
Summer
1,513
655
92
113
0
392
218
141
0
162
0
19
0
129
0
45
78
37
74
0
0
0
575
0
0
313
78
0
139
100
0
2
5
28
0
0
705
0
2
0
0
0
0
0
0
0
Winter
1,547
632
57
115
0
382
222
191
0
162
0
19
0
201
0
0
2
37
74
1
0
0
578
0
0
349
74
0
189
100
0
10
30
28
0
0
785
0
2
0
0
0
0
0
0
0
100%E10Casein
2030
Summer
1,418
607
82
104
0
360
202
141
0
137
7
16
0
95
0
40
0
37
64
1
0
0
528
0
0
292
66
0
139
89
0
1
3
28
0
0
679
0
2
0
0
0
0
0
0
0
Winter
1,442
578
53
105
0
345
204
191
0
137
3
16
0
73
0
8
0
37
64
1
0
0
527
0
0
297
60
0
189
88
0
1
3
28
0
0
826
0
2
0
0
0
0
0
0
0
E10 Case Relative to
RefCase
Summer
-95
-48
-10
-9
0
-32
-15
0
0
-25
7
-3
0
-33
0
-5
-78
0
-10
0
0
0
-48
0
0
-21
-11
0
0
-11
0
-1
-2
0
0
0
-26
0
0
0
0
0
0
0
0
0
Winter
-104
-55
-4
-10
0
-37
-18
0
0
-25
3
-3
0
-128
0
8
-2
0
-10
0
0
0
-51
0
0
-53
-14
0
0
-12
0
-9
-27
0
0
0
40
0
0
0
0
0
0
0
0
0
1 00% E1 5 Case in
2030
Summer
1,400
588
75
101
0
344
182
141
0
133
26
17
0
69
0
40
3
37
60
1
0
0
518
0
0
288
65
0
139
85
0
1
3
28
0
0
688
0
2
0
0
0
0
0
0
0
Winter
1,400
560
39
100
0
331
196
191
0
129
8
15
0
27
0
6
0
37
60
1
0
0
507
0
0
287
60
0
177
100
0
1
3
28
0
0
809
0
2
0
0
0
0
0
0
0
E15Case Relative to
RefCase
Summer
-113
-67
-17
-12
0
-47
-35
0
0
-29
26
-2
0
-60
0
-5
-75
0
-14
1
0
0
-57
0
0
-25
-12
0
0
-15
0
-1
-2
0
0
0
-18
0
0
0
0
0
0
0
0
0
Winter
-146
-73
-18
-14
0
-51
-26
0
0
-33
8
-3
0
-174
0
5
-2
0
-14
0
0
0
-71
0
0
-62
-14
0
-12
0
0
-9
-27
0
0
0
23
0
0
0
0
0
0
0
0
0
5-82
-------
Table 5-56 U.S. (except CA) Unit Capacity and Throughput Volumes from LP Refinery
Modeling (Thousand bbl/day)
PADDs 1 -4 and 5OC
Refinery Unit
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT - Total
DHT 2nd RCT - Total
DHT Arom Saturation
NHT - Total Fd
CGH - Generic
CGH-OlefinSat'n
FCCU Fd HOT
LSR Splitter
LSR Bz Saturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant - Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2009
15,403
6,505
0
0
0
4,786
0
1,013
0
1,912
107
0
2,794
287
0
116
406
528
438
23
42
0
4,400
0
10
3,921
1,373
1,373
2,029
0
0
0
0
0
31
3
1,238,369
1,380
5,736
0
0
127
268
0
0
0
Reference 2005 in
2030
Summer
13,049
5,768
993
1,225
0
4,008
1,114
2,223
0
1,329
0
159
2,048
129
344
199
415
346
513
0
1
0
3,993
0
0
2,844
481
0
2,964
202
0
351
1,054
359
0
0
8,029
516
21
0
0
124
268
0
0
0
Winter
14,515
6,350
683
1,348
0
4,506
2,494
2,303
0
1,505
0
181
1,976
201
351
0
455
283
448
1
60
0
3,831
0
0
2,952
466
0
3,014
252
0
97
290
359
0
0
8,399
516
21
0
0
125
265
0
0
0
100%E10Casein
2030
Summer
12,146
5,346
912
1,153
0
3,694
694
2,446
0
1,222
7
147
1,596
95
340
149
418
302
488
1
1
0
3,428
0
0
2,527
252
0
3,035
283
0
112
337
359
0
0
8,346
516
20
0
0
123
265
0
0
0
Winter
13,717
6,104
718
1,227
0
4,018
1,658
2,526
0
1,475
3
178
1,572
73
351
35
212
297
363
1
35
0
3,398
0
0
2,625
402
0
2,866
213
0
13
40
359
0
0
8,769
516
22
0
0
125
265
0
0
0
E10 Case Relative to
RefCase
Summer
-903
-422
-81
-72
0
-314
-420
223
0
-108
7
-12
-453
-33
-4
-49
2
-44
-25
0
0
0
-565
0
0
-317
-229
0
71
81
0
-239
-717
0
0
0
317
0
-1
0
0
-2
-3
0
0
0
Winter
-798
-246
35
-122
0
-489
-837
223
0
-30
3
-3
-404
-128
0
34
-244
14
-85
0
-25
0
-432
0
0
-327
-64
0
-148
-39
0
-83
-250
0
0
0
370
0
0
0
0
0
0
0
0
0
1 00% E1 5 Case in
2030
Summer
11,567
5,100
865
1,030
0
3,298
506
2,363
0
1,127
26
137
1,613
69
340
132
253
302
374
1
8
0
3,462
0
0
2,763
225
0
2,619
164
0
342
1,025
359
0
0
7,962
516
19
0
0
123
265
0
0
0
Winter
13,145
5,788
629
1,163
0
3,846
2,030
2,443
0
1,334
8
161
1,513
27
351
20
8
282
324
1
40
0
3,336
0
0
2,797
398
0
2,653
212
0
275
824
359
0
0
8,360
516
20
0
0
125
265
0
0
0
E15Case Relative to
RefCase
Summer
-1,482
-668
-128
-195
0
-710
-608
140
0
-202
26
-23
-436
-60
-4
-67
-162
-44
-139
1
7
0
-531
0
0
-81
-257
0
-345
-38
0
-10
-29
0
0
0
-67
0
-2
0
0
-2
-3
0
0
0
Winter
-1,370
-562
-54
-185
0
-660
-464
140
0
-171
8
-20
-463
-174
1
20
-447
0
-124
0
-20
0
-495
0
0
-155
-68
0
-360
-40
0
178
534
0
0
0
-38
0
-1
0
0
0
0
0
0
0
5-83
-------
Table 5-57 PADD 1 Gasoline Qualities for the E10 and E15 Cases relative to the Reference Case
Gasoline Qualities and
Volume for PADD 1
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur(ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volumes (kbbl/day)
2005 in 2030
Summer
CG
5.18
257.70
27.5
45.8
76.6
7.92
131.2
208.5
358.5
1180.7
143.8
27.9
0.58
11.0
0.50
0.74
1,399,252
RFC
5.09
259.91
23.3
50.8
81.4
6.54
140.0
198.2
336.6
1141.2
149.5
20.2
0.57
10.4
4.89
2.56
1,387,051
Pool
5.14
258.80
25.4
48.3
79.0
7.23
135.6
203.3
347.6
1161.0
146.6
24.1
0.58
10.7
2.68
1.65
2,786,303
Winter
CG
5.11
255.81
22.4
44.6
75.8
12.61
101.4
210.9
362.1
1147.0
119.1
27.0
0.61
10.6
0.00
0.15
1,338,953
RFC
4.93
252.05
16.9
62.4
89.2
13.19
97.7
174.5
301.3
971.4
109.7
14.8
0.56
16.8
4.94
2.50
1,368,062
Pool
5.02
253.91
19.6
53.6
82.6
12.90
99.6
192.5
331.4
1058.2
114.3
20.8
0.59
13.8
2.50
1.34
2,707,015
E10
Summer
CG
5.05
257.59
23.7
54.8
81.6
9.27
122.6
190.1
336.0
1090.3
133.4
22.0
0.51
9.1
10.00
3.74
1,760,831
RFC
5.04
260.94
23.6
51.1
81.2
7.00
137.1
197.5
337.6
1135.7
146.9
20.0
0.57
10.9
10.00
3.69
1,121,463
Pool
5.04
258.89
23.7
53.3
81.4
8.39
128.2
193.0
336.7
1107.9
138.7
21.2
0.53
9.8
10.00
3.72
2,882,294
Winter
CG
5.02
255.60
20.9
53.3
78.1
13.78
94.0
193.0
351.6
1071.7
109.7
20.3
0.52
10.2
10.00
3.77
1,698,176
RFC
4.91
255.72
18.0
59.4
84.6
13.39
96.5
180.5
322.1
1008.4
109.7
18.5
0.58
16.0
10.00
3.77
1,106,878
Pool
4.98
255.65
19.8
55.7
80.7
13.62
95.0
188.1
340.0
1046.8
109.7
19.6
0.54
12.5
10.00
3.77
2,805,054
E15
Summer
CG
4.93
259.60
28.0
59.8
81.5
8.27
129.0
179.8
336.3
1069.1
137.1
18.5
0.73
8.5
15.00
5.57
1,736,796
RFC
4.93
260.61
28.0
58.7
84.1
7.00
137.1
182.1
324.3
1076.0
144.3
17.6
0.89
9.8
15.00
5.55
1,204,323
Pool
4.93
260.01
28.0
59.4
82.6
7.75
132.3
180.7
331.4
1072.0
140.0
18.1
0.79
9.0
15.00
5.56
2,941,119
Winter
CG
4.87
257.91
12.6
60.9
83.0
13.29
97.1
177.5
329.5
1007.7
109.7
18.8
0.46
7.9
15.00
5.61
1,619,601
RFC
4.82
255.72
17.6
66.7
89.8
12.92
99.5
165.7
298.3
944.7
109.7
14.0
0.55
17.9
15.00
5.65
1,125,767
Pool
4.85
257.01
14.7
63.3
85.8
13.14
98.1
172.7
316.7
981.9
109.7
16.8
0.50
12.0
15.00
5.63
2,745,368
Table 5-58 PADD 2 Gasoline Qualities for the E10 and E15 Cases relative to the Reference Case
Gasoline Qualities and
Volume for PADD 2
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur(ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volumes (kbbl/day)
2005 in 2030
Summer
CG
5.18
257.79
27.4
45.7
76.6
7.91
131.2
208.6
358.7
1181.5
143.9
27.9
0.58
11.5
0.42
0.40
1,772,351
RFC
5.02
258.55
9.7
53.7
85.0
7.00
137.1
192.3
320.2
1102.5
146.0
18.9
0.56
3.8
10.00
3.73
329,642
Pool
4.97
259.05
20.7
56.1
84.6
8.73
126.0
187.4
322.3
1073.5
135.9
22.2
0.54
8.8
8.80
3.27
1,759,558
Winter
CG
4.99
255.20
25.0
48.8
79.7
13.97
92.8
202.3
344.5
1090.7
110.3
24.7
0.58
13.0
0.00
0.00
1,685,250
RFC
4.75
251.15
16.9
68.4
93.7
12.70
100.9
162.3
280.8
918.9
110.3
12.3
0.57
6.8
10.00
3.84
326,672
Pool
4.95
254.54
23.7
51.9
82.0
13.76
94.1
195.8
334.2
1062.8
110.3
22.7
0.58
12.0
1.62
0.62
2,011,922
E10
Summer
CG
4.99
258.09
23.3
53.4
81.7
9.28
122.6
192.9
335.6
1098.0
133.9
22.5
0.51
8.8
10.00
3.74
1,802,319
RFC
5.03
259.13
17.9
49.7
82.2
7.00
137.1
200.4
333.0
1139.7
147.4
19.5
0.56
6.7
10.00
3.72
342,519
Pool
4.99
258.25
22.4
52.8
81.7
8.92
124.9
194.1
335.2
1104.7
136.0
22.0
0.52
8.5
10.00
3.73
2,144,838
Winter
CG
4.87
256.71
24.1
58.9
83.9
13.31
97.0
181.7
325.4
1015.8
110.3
20.4
0.50
10.7
10.00
3.76
1,751,081
RFC
4.89
256.01
24.1
59.0
84.9
13.31
97.0
181.5
320.8
1010.7
110.3
18.2
0.55
11.2
10.00
3.77
339,433
Pool
4.88
256.59
24.1
58.9
84.1
13.31
97.0
181.6
324.6
1015.0
110.3
20.0
0.51
10.7
10.00
3.76
2,090,514
E15
Summer
CG
4.94
260.15
28.0
57.5
80.3
8.28
128.9
184.6
341.6
1088.7
137.8
20.3
0.55
7.3
15.00
5.56
1,841,582
RFC
4.93
259.67
28.0
55.8
81.3
7.00
137.1
188.0
337.2
1106.7
145.3
19.2
0.76
7.0
15.00
5.57
342,519
Pool
4.94
260.07
28.0
57.2
80.5
8.08
130.2
185.1
340.9
1091.6
139.0
20.2
0.58
7.2
15.00
5.56
2,184,101
Winter
CG
4.83
258.46
26.8
62.7
84.2
13.07
98.5
173.9
323.8
993.4
110.3
20.0
0.67
10.5
15.00
5.60
1,751,081
RFC
4.81
256.91
26.7
64.8
86.7
12.93
99.4
169.6
312.8
970.6
110.3
16.9
0.65
10.6
15.00
5.63
339,433
Pool
4.83
258.20
26.8
63.0
84.6
13.05
98.6
173.2
322.0
989.7
110.3
19.5
0.67
10.6
15.00
5.60
2,090,514
5-84
-------
Table 5-59 PADD 3 Gasoline Qualities for the E10 and E15 Cases relative to the Reference Case
Gasoline Qualities and
Volume for PADD 3
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur(ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volumes (kbbl/day)
2005 in 2030
Summer
CG
5.18
258.31
27.6
44.8
75.9
7.94
131.1
210.4
361.9
1189.8
144.0
29.2
0.59
11.5
0.90
0.67
1,257,764
RFC
5.16
258.54
24.8
51.4
82.7
6.54
140.0
197.0
330.7
1131.7
149.3
19.7
0.58
1.1
5.00
2.54
480,771
Pool
5.18
258.37
26.8
46.6
77.8
7.55
133.5
206.7
353.3
1173.7
145.5
26.6
0.59
8.6
2.03
1.19
1,738,535
Winter
CG
5.13
256.15
25.7
43.7
75.5
12.47
102.3
212.6
363.6
1154.8
120.1
27.6
0.59
11.5
0.00
0.00
1,400,629
RFC
5.18
254.74
22.4
56.5
83.3
11.65
107.5
186.6
328.2
1049.1
120.1
20.0
0.58
9.8
4.00
2.50
489,720
Pool
5.14
255.78
24.7
45.8
76.8
12.34
103.2
208.4
357.6
1137.5
120.1
26.1
0.58
11.0
0.63
0.40
1,890,349
E10
Summer
CG
5.07
258.18
24.1
54.0
81.1
8.85
125.3
191.6
338.1
1100.8
135.9
22.4
0.52
9.4
10.00
3.73
1,306,897
RFC
5.11
259.96
24.9
49.0
80.8
7.00
137.1
201.8
339.4
1150.5
147.6
19.5
0.51
2.7
10.00
3.71
479,961
Pool
5.08
258.66
24.3
52.7
81.0
8.35
128.5
194.4
338.4
1114.2
139.1
21.6
0.52
7.6
10.00
3.73
1,786,858
Winter
CG
5.07
256.90
22.8
50.7
76.8
12.02
105.2
198.4
357.6
1110.6
120.1
20.8
0.50
8.4
10.00
3.75
1,492,196
RFC
5.00
255.26
23.1
53.7
79.8
11.83
106.4
192.3
343.9
1080.5
120.1
18.4
0.57
8.5
10.00
3.78
508,849
Pool
5.05
256.48
22.9
51.4
77.6
11.97
105.5
196.9
354.1
1102.9
120.1
20.2
0.52
8.5
10.00
3.76
2,001,045
E15
Summer
CG
4.95
260.34
27.8
58.8
80.6
7.85
131.7
181.9
340.3
1083.6
139.7
19.1
0.73
7.9
15.00
5.55
1,306,897
RFC
4.95
260.35
26.0
58.3
83.1
7.00
137.1
182.8
328.9
1082.8
144.4
15.6
0.90
2.8
15.00
5.55
499,552
Pool
4.95
260.35
27.3
58.6
81.3
7.61
133.1
182.2
337.2
1083.4
141.0
18.1
0.77
6.5
15.00
5.55
1,806,449
Winter
CG
4.91
258.80
13.1
59.9
83.4
11.43
108.9
179.7
327.5
1029.9
120.1
17.9
0.40
7.5
15.00
5.59
1,433,678
RFC
4.84
254.37
21.6
66.5
90.9
11.00
111.6
166.2
293.7
959.7
120.1
11.0
0.59
13.5
15.00
5.68
488,895
Pool
4.89
257.67
15.3
61.6
85.3
11.32
109.6
176.2
318.9
1012.0
120.1
16.2
0.45
9.0
15.00
5.61
1,922,573
5-85
-------
Table 5-60 PADD 4 and 5OC Gasoline Qualities for the E10 and E15 Cases relative to the Reference Case
Gasoline Qualities and
Volume for PADDs 4 &
50C
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur(ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volumes (kbbl/day)
2005 in 2030
Summer
CG
5.06
258.41
24.8
53.3
90.7
8.07
130.3
193.0
294.6
1069.1
140.4
20.4
1.17
9.7
0.00
0.01
889,157
RFC
0.00
0.00
0.0
0.0
0.0
0.00
0.0
0.0
0.0
0.0
0.0
0.0
0.00
0.0
0.00
0.00
0
Pool
5.06
258.41
24.8
53.3
90.7
8.07
130.3
193.0
294.6
1069.1
140.4
20.4
1.17
9.7
0.00
0.01
889,157
Winter
CG
5.00
253.46
24.4
59.2
92.2
13.00
99.0
181.1
287.8
979.5
111.9
20.3
1.12
8.8
0.00
0.00
733,803
RFC
0.00
0.00
0.0
0.0
0.0
0.00
0.0
0.0
0.0
0.0
0.0
0.0
0.00
0.0
0.00
0.00
0
Pool
5.00
253.46
24.4
59.2
92.2
13.00
99.0
181.1
287.8
979.5
111.9
20.3
1.12
8.8
0.00
0.00
733,803
E10
Summer
CG
4.95
258.55
28.0
62.1
91.6
9.08
123.8
175.1
290.3
1001.4
131.9
17.4
1.19
8.1
10.00
3.73
718,402
RFC
0.00
0.00
0.0
0.0
0.0
0.00
0.0
0.0
0.0
0.0
0.0
0.0
0.00
0.0
0.00
0.00
0
Pool
4.95
258.55
28.0
62.1
91.6
9.08
123.8
175.1
290.3
1001.4
131.9
17.4
1.19
8.1
10.00
3.73
718,402
Winter
CG
4.89
254.99
28.0
64.4
92.1
12.96
99.2
170.5
288.3
948.5
110.3
15.3
1.18
7.8
10.00
3.78
762,467
RFC
0.00
0.00
0.0
0.0
0.0
0.00
0.0
0.0
0.0
0.0
0.0
0.0
0.00
0.0
0.00
0.00
0
Pool
4.89
254.99
28.0
64.4
92.1
12.96
99.2
170.5
288.3
948.5
110.3
15.3
1.18
7.8
10.00
3.78
762,467
E15
Summer
CG
4.89
260.44
28.0
64.6
92.2
8.08
130.2
170.0
287.5
992.8
136.4
15.7
1.19
7.8
15.00
5.55
718,402
RFC
0.00
0.00
0.0
0.0
0.0
0.00
0.0
0.0
0.0
0.0
0.0
0.0
0.00
0.0
0.00
0.00
0
Pool
4.89
260.44
28.0
64.6
92.2
8.08
130.2
170.0
287.5
992.8
136.4
15.7
1.19
7.8
15.00
5.55
718,402
Winter
CG
4.80
256.45
25.8
70.1
96.0
12.59
101.5
158.9
270.5
899.4
110.3
12.7
0.73
7.3
15.00
5.64
762,467
RFC
0.00
0.00
0.0
0.0
0.0
0.00
0.0
0.0
0.0
0.0
0.0
0.0
0.00
0.0
0.00
0.00
0
Pool
4.80
256.45
25.8
70.1
96.0
12.59
101.5
158.9
270.5
899.4
110.3
12.7
0.73
7.3
15.00
5.64
762,467
5-86
-------
Table 5-61 U.S. (except CA) Gasoline Qualities for the E10 and E15 Cases relative to the Reference Case
Gasoline Qualities and
Volume for PADDs 1 -
4&50C
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur(ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volumes (kbbl/day)
2005 in 2030
Summer
CG
5.16
257.99
27.1
46.8
78.8
7.95
131.0
206.4
348.7
1164.5
143.3
27.0
0.68
11.1
0.48
0.49
5,318,524
RFC
5.10
259.40
21.6
51.4
82.2
6.61
139.5
197.0
332.9
1133.3
148.9
19.9
0.57
7.4
5.68
2.73
2,197,464
Pool
5.10
258.71
24.5
50.4
81.5
7.78
132.1
199.0
336.2
1131.2
142.9
23.8
0.64
9.6
3.69
1.73
7,173,553
Winter
CG
5.06
255.37
24.4
47.8
79.3
13.07
98.5
204.3
346.2
1106.9
115.4
25.5
0.67
11.4
0.00
0.04
5,158,635
RFC
4.96
252.52
18.1
62.0
88.5
12.77
100.4
175.4
304.3
981.0
112.1
15.6
0.56
13.7
5.49
2.70
2,184,454
Pool
5.03
254.52
22.5
51.7
81.9
13.00
98.9
196.4
334.5
1072.0
114.5
22.6
0.64
12.1
1.53
0.77
7,343,089
E10
Summer
CG
5.02
258.01
24.2
55.1
82.8
9.15
123.4
189.4
330.5
1083.8
134.0
21.7
0.60
8.9
10.00
3.74
5,588,449
RFC
5.06
260.38
22.9
50.4
81.3
7.00
137.1
199.1
337.3
1140.1
147.2
19.8
0.55
8.1
10.00
3.70
1,943,943
Pool
5.03
258.62
23.9
53.9
82.4
8.60
126.9
191.9
332.2
1098.3
137.4
21.2
0.59
8.7
10.00
3.73
7,532,392
Winter
CG
4.97
256.20
23.3
55.8
81.4
13.07
98.5
187.9
336.7
1048.3
112.7
19.8
0.60
9.6
10.00
3.76
5,703,920
RFC
4.93
255.65
20.4
57.9
83.4
12.97
99.2
183.8
327.5
1027.6
112.5
18.4
0.57
13.2
10.00
3.77
1,955,160
Pool
4.96
256.06
22.6
56.3
81.9
13.04
98.7
186.9
334.3
1043.0
112.6
19.4
0.59
10.5
10.00
3.76
7,659,080
E15
Summer
CG
4.93
260.06
28.0
59.4
82.3
8.15
129.7
180.6
332.7
1069.2
137.8
18.9
0.73
7.9
15.00
5.56
5,603,677
RFC
4.93
260.39
27.5
58.1
83.4
7.00
137.1
183.2
327.6
1082.8
144.5
17.4
0.87
7.6
15.00
5.55
2,046,394
Pool
4.93
260.15
27.8
59.1
82.6
7.84
131.7
181.3
331.3
1072.8
139.6
18.5
0.77
7.8
15.00
5.56
7,650,071
Winter
CG
4.86
258.11
19.0
62.5
85.3
12.65
101.2
174.4
319.1
994.1
112.6
18.1
0.55
8.5
15.00
5.60
5,566,827
RFC
4.82
255.59
20.2
66.3
89.5
12.44
102.5
166.5
299.7
953.0
112.4
13.8
0.58
15.5
15.00
5.66
1,954,095
Pool
4.85
257.45
19.3
63.5
86.4
12.59
101.5
172.3
314.1
983.4
112.6
17.0
0.56
10.4
15.00
5.62
7,520,922
5-87
-------
References
1 Annual Energy Outlook 2006, Energy Information Administration, Department of Energy
2 Worldwide Report, Oil and Gas Journal, www.ogionline.com. December 22, 2005.
3 Petroleum Supply Annual 2004, Volume 1, Table 38. Energy Information Administration, Department of Energy.
4 Petroleum Supply Annual 2004, Volume 1, Tables 22-25. Energy Information Administration, Department of
Energy.
5 Petroleum Supply Annual 2005, Volume 1, Table 17. Energy Information Administration, Department of Energy.
6 Petroleum Refining Technology and Economics 4th Edition. James Gary and Glenn Handerwork. Page 65, Table
4.7
7 The Challenges & Opportunities of 10 ppm Sulfur Gasoline. Jay Ross, Delphine Largeteau, Marc Laborde, Larry
Wisdom. NPRA Annual Meeting, AM-11-57, March, 2011, page 6
8 The Benefits of Cat Feed Hydrotreating and the Impact of Feed Nitrogen on Catalyst Stability. Brian Moyse at
Haldor Topso. NPRA Annual Meeting, AM-10-167, March 2010, Page 2
9 Handbook of Petroleum Refining Processes, Third Edition. Robert Meyers, page 11.39
10 Refining Economics of a National Low Sulfur, Low RVP Gasoline Standard, Performed for The International
Council for Clean Transportation by Mathpro, October 25, 2011.
1: Refining Economics of a National Clean Gasoline Standard for PADDs 1 -3; for The Alliance of Automobile
Manufacturers by Mathpro, June 27, 2008.
12 Potential Supply and Cost Impacts of Lower Sulfur, Lower RVP Gasoline; prepared for The American Petroleum
Institute by Baker and O'Brien; July 2011.
13 Schink, George R., Singer, Hal J., Economic Analysis of the Implications of Implementing EPA's Tier 3 Rules,
prepared for the Emissions Control Technology Association, June 14, 2012.
14 An Assessment of the Impact of Nonroad Diesel Fuel Sulfur Regulation on Distillate Fuel Production and
Availability in the U.S., prepared for the American Petroleum Institute by Baker and O'Brien, July 2003.
15 Petroleum Supply Annual 2003 - 2010; Table 24, Imports of Crude Oil and Petroleum Products, Energy
Information Administration.
16 Petroleum Supply Annual 2003 - 2010; Table 31, Exports of Crude Oil and Petroleum Products by PAD
Destination.
17 Petroleum Supply Annual 2011, Refinery Capacity Report, Table 13 Refineries Permanently Shutdown by PAD
District Between January 1, 1990 and January 1, 2010; Energy Information Administration.
18 Petroleum Supply Annual 2003 - 2011, Refinery Capacity Report, Table 11 New, Shutdown and Reactivated
Refineries During 2010; Energy Information Administration.
19 Petroleum and Other Liquids; Number and Capacity of Petroleum Refineries, Data, Total Number of Operable
Refineries and Total Number of Operating Refineries, 1990 - 2010; Energy Information Administration.
-------
20 Gary, I, & Handewerk, G. (2001). Petroleum Refining: Technology and Economics. (4th ed.). CRC Press, p 256.
21 Gary, I, & Handewerk, G. (2001). Petroleum Refining: Technology and Economics. (4th ed.). CRC Press, p 337.
22 Keller, P. (2013, February). New Source Review Permitting Impact Analysis for Proposed Tier 3 Gasoline
Program. Memorandum to the docket.
5-89
-------
Chapter 6 Health and Environmental Effects Associated with
Exposure to Criteria and Toxic Pollutants
6.1 Health Effects of Criteria and Toxic Pollutants
6.1.1 Particulate Matter
6.1.1.1 Background
Particulate matter (PM) is a highly complex mixture of solid particles and liquid droplets
distributed among numerous atmospheric gases which interact with solid and liquid phases.
Particles range in size from those smaller than 1 nanometer (10~9 meter) to over 100 micrometer
(|im, or 10"6 meter) in diameter (for reference, a typical strand of human hair is 70 um in
diameter and a grain of salt is about 100 jam). Atmospheric particles can be grouped into several
classes according to their aerodynamic and physical sizes, including ultrafine particles (<0.1
|im), accumulation mode or 'fine' particles (< 1 to 3 jim), and coarse particles (>1 to 3 jim). For
regulatory purposes, fine particles are measured as PIVb.s and inhalable or thoracic coarse
particles are measured as PMio-2.5, corresponding to their size (diameter) range in micrometers
and referring to total particle mass under 2.5 and between 2.5 and 10 micrometers, respectively.
The EPA currently has standards that measure PM2.5 and PMi0.A
Particles span many sizes and shapes and consist of hundreds of different chemicals.
Particles are emitted directly from sources and are also formed through atmospheric chemical
reactions; the former are often referred to as "primary" particles, and the latter as "secondary"
particles. Particle pollution also varies by time of year and location and is affected by several
weather-related factors, such as temperature, clouds, humidity, and wind. A further layer of
complexity comes from particles' ability to shift between solid/liquid and gaseous phases, which
is influenced by concentration and meteorology, especially temperature.
Fine particles are produced primarily by combustion processes and by transformations of
gaseous emissions (e.g., sulfur oxides (SOx), nitrogen oxides (NOx) and volatile organic
compounds (VOCs)) in the atmosphere. The chemical and physical properties of PM2.5 may vary
greatly with time, region, meteorology and source category. Thus, PM2.5 may include a complex
mixture of different components including sulfates, nitrates, organic compounds, elemental
carbon and metal compounds. These particles can remain in the atmosphere for days to weeks
and travel through the atmosphere hundreds to thousands of kilometers.1
A Regulatory definitions of PM size fractions, and information on reference and equivalent methods for measuring
PM in ambient air, are provided in 40 CFR Parts 50, 53, and 58.
6-1
-------
6.1.1.2 Health Effects of PM
This section provides a summary of the health effects associated with exposure to
ambient concentrations of PM.B The information in this section is based on the information and
conclusions in the Integrated Science Assessment (ISA) for Particulate Matter (December 2009)
prepared by EPA's Office of Research and Development (ORD).C
The ISA concludes that ambient concentrations of PM are associated with a number of
adverse health effects.0 The ISA characterizes the weight of evidence for different health effects
associated with three PM size ranges: PM2.5, PMio-2.5, and UFPs. The discussion below
highlights the ISA's conclusions pertaining to these three size fractions of PM, considering
variations in health effects associated with both short-term and long-term exposure periods.
6.1.1.2.1 Effects Associated with Short-term Exposure to PM2.5
The ISA concludes that cardiovascular effects and premature mortality are causally
associated with short-term exposure to PM2.5.2 It also concludes that respiratory effects are
likely to be causally associated with short-term exposure to PM2.5, including respiratory
emergency department (ED) visits and hospital admissions for chronic obstructive pulmonary
disease (COPD), respiratory infections, and asthma; and exacerbation of respiratory symptoms in
asthmatic children.
6.1.1.2.2 Effects Associated with Long-term Exposure to PM2.5
The ISA concludes that there are causal associations between long-term exposure to
PM2.5 and cardiovascular effects, such as the development/progression of cardiovascular disease
(CVD), and premature mortality, particularly from cardiovascular causes.3 It also concludes that
long-term exposure to PM2.5 is likely to be causally associated with respiratory effects, such as
reduced lung function growth, increased respiratory symptoms, and asthma development. The
ISA characterizes the evidence as suggestive of a causal relationship for associations between
long-term PM2.5 exposure and reproductive and developmental outcomes, such as low birth
weight and infant mortality. It also characterizes the evidence as suggestive of a causal
relationship between PM2.s and cancer incidence, mutagenicity, and genotoxicity.
B Personal exposure includes contributions from many different types of particles, from many sources, and in many
different environments. Total personal exposure to PM includes both ambient and nonambient components and
collectively these components may contribute to adverse health effects.
c The ISA is available at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=216546
D The ISA evaluates the health evidence associated with different health effects, assigning one of five "weight of
evidence" determinations: causal relationship, likely to be a causal relationship, suggestive of a causal relationship,
inadequate to infer a causal relationship, and not likely to be a causal relationship. For definitions of these levels of
evidence, please refer to Section 1.5 of the ISA.
6-2
-------
6.1.1.2.3 Effects Associated with PM'10-2.5
The ISA summarizes evidence related to short-term exposure to PMio-2.5. PMio-2.5 is the
fraction of PMio particles that is larger than PM2.s.4 The ISA concludes that available evidence
is suggestive of a causal relationship between short-term exposures to PMio-2.5 and
cardiovascular effects. It also concludes that the available evidence is suggestive of a causal
relationship between short-term exposures to PMio-2.5 and respiratory effects, including
respiratory-related ED visits and hospitalizations. The ISA also concludes that the available
literature suggests a causal relationship between short-term exposures to PMio-2.5 and mortality.
Data are inadequate to draw conclusions regarding health effects associated with long-term
exposure to PMio-2.5.5
6.1.1.2.4 Effects Associated with Ultrafine Particles
The ISA concludes that the evidence is suggestive of a causal relationship between short-
term exposures to UFPs and cardiovascular effects, including changes in heart rhythm and
vasomotor function (the ability of blood vessels to expand and contract).6
The ISA also concludes that there is suggestive evidence of a causal relationship between
short-term UFP exposure and respiratory effects. The types of respiratory effects examined in
epidemiologic studies include respiratory symptoms and asthma hospital admissions, the results
of which are not entirely consistent. There is evidence from toxicological and controlled human
exposure studies that exposure to UFPs may increase lung inflammation and produce small
asymptomatic changes in lung function. Data are inadequate to draw conclusions regarding
health effects associated with long-term exposure to UFPs.7
6.1.2 Ozone
6.1.2.1 Background
Ground-level ozone pollution is typically formed through reactions involving VOCs and
NOx in the lower atmosphere in the presence of sunlight. These pollutants, often referred to as
ozone precursors, are emitted by many types of pollution sources such as highway and nonroad
motor vehicles and engines, power plants, chemical plants, refineries, makers of consumer and
commercial products, industrial facilities, and smaller area sources.
The science of ozone formation, transport, and accumulation is complex. Ground-level
ozone is produced and destroyed in a cyclical set of chemical reactions, many of which are
sensitive to temperature and sunlight. When ambient temperatures and sunlight levels remain
high for several days and the air is relatively stagnant, ozone and its precursors can build up and
result in more ozone than typically occurs on a single high-temperature day. Ozone and its
precursors can be transported hundreds of miles downwind of precursor emissions, resulting in
elevated ozone levels even in areas with low VOC or NOx emissions.
The highest levels of ozone are produced when both VOC and NOx emissions are present
in significant quantities on clear summer days. Relatively small amounts of NOx enable ozone
to form rapidly when VOC levels are relatively high, but ozone production is quickly limited by
removal of the NOx. Under these conditions NOx reductions are highly effective in reducing
6-3
-------
ozone while VOC reductions have little effect. Such conditions are called "NOx-limited."
Because the contribution of VOC emissions from biogenic (natural) sources to local ambient
ozone concentrations can be significant, even some areas where man-made VOC emissions are
relatively low can be NOx-limited.
Ozone concentrations in an area also can be lowered by the reaction of nitric oxide (NO)
with ozone, forming nitrogen dioxide (NO2); as the air moves downwind and the cycle continues,
the NO2 forms additional ozone. The importance of this reaction depends, in part, on the relative
concentrations of NOx, VOC, and ozone, all of which change with time and location. When
NOx levels are relatively high and VOC levels relatively low, NOx forms inorganic nitrates (i.e.,
particles) but relatively little ozone. Such conditions are called "VOC-limited." Under these
conditions, VOC reductions are effective in reducing ozone, but NOx reductions can actually
increase local ozone under certain circumstances. Even in VOC-limited urban areas, NOx
reductions are not expected to increase ozone levels if the NOx reductions are sufficiently large.
Rural areas are usually NOx-limited, due to the relatively large amounts of biogenic VOC
emissions in such areas. Urban areas can be either VOC- or NOx-limited, or a mixture of both,
in which ozone levels exhibit moderate sensitivity to changes in either pollutant.
6.1.2.2 Health Effects of Ozone
Exposure to ambient ozone contributes to a wide range of adverse health effects.E These
health effects are well documented and are critically assessed in the EPA ozone air quality
criteria document (ozone AQCD) and EPA staff paper.8'9 We are relying on the data and
conclusions in the ozone AQCD and staff paper, regarding the health effects associated with
ozone exposure.
Ozone-related health effects include lung function decrements, respiratory symptoms,
aggravation of asthma, increased hospital and emergency room visits, increased asthma
medication usage, and a variety of other respiratory effects. Cellular-level effects, such as
inflammation of lungs, have been documented as well. In addition, there is suggestive evidence
of a contribution of ozone to cardiovascular-related morbidity and highly suggestive evidence
that short-term ozone exposure directly or indirectly contributes to non-accidental and
cardiopulmonary-related mortality, but additional research is needed to clarify the underlying
mechanisms causing these effects. In a recent report on the estimation of ozone-related
premature mortality published by the National Research Council (NRC), a panel of experts and
reviewers concluded that short-term exposure to ambient ozone is likely to contribute to
premature deaths and that ozone-related mortality should be included in estimates of the health
benefits of reducing ozone exposure.10 People who appear to be more susceptible to effects
associated with exposure to ozone include children, asthmatics and the elderly. Those with
greater exposures to ozone, for instance due to time spent outdoors (e.g., children and outdoor
workers), are also of concern.
E Human exposure to ozone varies over time due to changes in ambient ozone concentration and because people
move between locations which have notable different ozone concentrations. Also, the amount of ozone delivered to
the lung is not only influenced by the ambient concentrations but also by the individuals breathing route and rate.
6-4
-------
Based on a large number of scientific studies, EPA has identified several key health
effects associated with exposure to levels of ozone found today in many areas of the country.
Short-term (1 to 3 hours) and prolonged exposures (6 to 8 hours) to ambient ozone
concentrations have been linked to lung function decrements, respiratory symptoms, increased
hospital admissions and emergency room visits for respiratory problems.11'12'13'14'15'16
Repeated exposure to ozone can increase susceptibility to respiratory infection and lung
inflammation and can aggravate preexisting respiratory diseases, such as asthma.17'18'1 '20'21
Repeated exposure to sufficient concentrations of ozone can also cause inflammation of the lung,
impairment of lung defense mechanisms, and possibly irreversible changes in lung structure,
which over time could affect premature aging of the lungs and/or the development of chronic
respiratory illnesses, such as emphysema and chronic bronchitis.22'23'24'25
Children and outdoor workers tend to have higher ozone exposure because they typically
are active outside, working, playing and exercising, during times of day and seasons (e.g., the
summer) when ozone levels are highest.26 For example, summer camp studies have reported
statistically significant reductions in lung function in children who are active outdoors.27'28'29'30'
31,32,33,34 Further, children are more at risk of experiencing health effects from ozone exposure
than adults because their respiratory systems are still developing. These individuals (as well as
people with respiratory illnesses, such as asthma, especially asthmatic children) can experience
reduced lung function and increased respiratory symptoms, such as chest pain and cough, when
exposed to relatively low ozone levels during prolonged periods of moderate exertion.35'36'37'38
6.1.3 Nitrogen Oxides and Sulfur Oxides
6.1.3.1 B ackground
Sulfur dioxide (802), a member of the sulfur oxide (SOx) family of gases, is formed from
burning fuels containing sulfur (e.g., coal or oil), extracting gasoline from oil, or extracting
metals from ore. Nitrogen dioxide (NO2) is a member of the nitrogen oxide (NOx) family of
gases. Most NO2 is formed in the air through the oxidation of nitric oxide (NO) emitted when
fuel is burned at a high temperature. SO2 andNO2 and their gas phase oxidation products can
dissolve in water droplets and further oxidize to form sulfuric and nitric acid which react with
ammonia to form sulfates and nitrates, both of which are important components of ambient PM.
The health effects of ambient PM are discussed in Section 6.1.1.2. NOx along with VOCs are
the two major precursors of ozone. The health effects of ozone are covered in Section 6.1.2.2.
6.1.3.2 Health Effects of Sulfur Oxides
This section provides an overview of the health effects associated with SO2. Additional
information on the health effects of SO2 can be found in the EPA Integrated Science Assessment
for Sulfur Oxides.39 Following an extensive evaluation of health evidence from epidemiologic
and laboratory studies, the U.S. EPA has concluded that there is a causal relationship between
respiratory health effects and short-term exposure to SO2. The immediate effect of SO2 on the
respiratory system in humans is bronchoconstriction. Asthmatics are more sensitive to the effects
of SO2 likely resulting from preexisting inflammation associated with this disease. In laboratory
studies involving controlled human exposures to SO2, respiratory effects have consistently been
observed following 5-10 min exposures at SO2 concentrations > 0.4 ppm in asthmatics engaged
6-5
-------
in moderate to heavy levels of exercise, with more limited evidence of respiratory effects among
exercising asthmatics exposed to concentrations as low as 0.2-0.3 ppm. A clear concentration-
response relationship has been demonstrated in these studies following exposures to 862 at
concentrations between 0.2 and 1.0 ppm, both in terms of increasing severity of respiratory
symptoms and decrements in lung function, as well as the percentage of asthmatics adversely
affected.
In epidemiologic studies, respiratory effects have been observed in areas where the mean
24-hour SC>2 levels range from 1 to 30 ppb, with maximum 1 to 24-hour average SC>2 values
ranging from 12 to 75 ppb. Important new multicity studies and several other studies have found
an association between 24-hour average ambient SO2 concentrations and respiratory symptoms
in children, particularly those with asthma. Generally consistent associations also have been
observed between ambient 862 concentrations and emergency department visits and
hospitalizations for all respiratory causes, particularly among children and older adults (> 65
years), and for asthma. A limited subset of epidemiologic studies has examined potential
confounding by copollutants using multipollutant regression models. These analyses indicate
that although copollutant adjustment has varying degrees of influence on the 862 effect
estimates, the effect of SC>2 on respiratory health outcomes appears to be generally robust and
independent of the effects of gaseous and particulate copollutants, suggesting that the observed
effects of SC>2 on respiratory endpoints occur independent of the effects of other ambient air
pollutants.
Consistent associations between short-term exposure to SC>2 and mortality have been
observed in epidemiologic studies, with larger effect estimates reported for respiratory mortality
than for cardiovascular mortality. While this finding is consistent with the demonstrated effects
of 862 on respiratory morbidity, uncertainty remains with respect to the interpretation of these
associations due to potential confounding by various copollutants. The U.S. EPA has therefore
concluded that the overall evidence is suggestive of a causal relationship between short-term
exposure to 862 and mortality. Significant associations between short-term exposure to 862 and
emergency department visits and hospital admissions for cardiovascular diseases have also been
reported. However, these findings have been inconsistent across studies and do not provide
adequate evidence to infer a causal relationship between SO2 exposure and cardiovascular
morbidity.
6.1.3.3 Health Effects of Nitrogen Oxides
Information on the health effects of NC>2 can be found in the EPA Integrated Science
Assessment (ISA) for Nitrogen Oxides.40 The EPA has concluded that the findings of
epidemiologic, controlled human exposure, and animal toxicological studies provide evidence
that is sufficient to infer a likely causal relationship between respiratory effects and short-term
NO2 exposure. The ISA concludes that the strongest evidence for such a relationship comes from
epidemiologic studies of respiratory effects including symptoms, emergency department visits,
and hospital admissions. Based on both short- and long-term studies, the ISA concludes that
associations of NO2 with respiratory health effects are stronger among a number of groups; these
include individuals with preexisting pulmonary conditions (e.g., asthma or COPD), children and
older adults. The ISA also draws two broad conclusions regarding airway responsiveness
following NO2 exposure. First, the ISA concludes that NO2 exposure may enhance the
6-6
-------
sensitivity to allergen-induced decrements in lung function and increase the allergen-induced
airway inflammatory response following 30-minute exposures of asthmatics to NC>2
concentrations as low as 0.26 ppm. Second, exposure to NC>2 has been found to enhance the
inherent responsiveness of the airway to subsequent nonspecific challenges in controlled human
exposure studies of asthmatic subjects. Small but significant increases in non-specific airway
hyperresponsiveness were reported following 1-hour exposures of asthmatics to 0.1 ppm NC>2.
Enhanced airway responsiveness could have important clinical implications for asthmatics since
transient increases in airway responsiveness following NC>2 exposure have the potential to
increase symptoms and worsen asthma control. Together, the epidemiologic and experimental
data sets form a plausible, consistent, and coherent description of a relationship between NO2
exposures and an array of adverse health effects that range from the onset of respiratory
symptoms to hospital admission.
Although the weight of evidence supporting a causal relationship is somewhat less certain
than that associated with respiratory morbidity, NC>2 has also been linked to other health
endpoints. These include all-cause (nonaccidental) mortality, hospital admissions or emergency
department visits for cardiovascular disease, and decrements in lung function growth associated
with chronic exposure.
6.1.4 Health Effects of Carbon Monoxide
Information on the health effects of carbon monoxide (CO) can be found in the EPA
Integrated Science Assessment (ISA) for Carbon Monoxide.41 The ISA concludes that ambient
concentrations of CO are associated with a number of adverse health effects.17 This section
provides a summary of the health effects associated with exposure to ambient concentrations of
COG
Human clinical studies of subjects with coronary artery disease show a decrease in the
time to onset of exercise-induced angina (chest pain) and electrocardiogram changes following
CO exposure. In addition, epidemiologic studies show associations between short-term CO
exposure and cardiovascular morbidity, particularly increased emergency room visits and
hospital admissions for coronary heart disease (including ischemic heart disease, myocardial
infarction, and angina). Some epidemiologic evidence is also available for increased hospital
admissions and emergency room visits for congestive heart failure and cardiovascular disease as
a whole. The ISA concludes that a causal relationship is likely to exist between short-term
exposures to CO and cardiovascular morbidity. It also concludes that available data are
inadequate to conclude that a causal relationship exists between long-term exposures to CO and
cardiovascular morbidity.
F The ISA evaluates the health evidence associated with different health effects, assigning one of five "weight of
evidence" determinations: causal relationship, likely to be a causal relationship, suggestive of a causal relationship,
inadequate to infer a causal relationship, and not likely to be a causal relationship. For definitions of these levels of
evidence, please refer to Section 1.6 of the ISA.
G Personal exposure includes contributions from many sources, and in many different environments. Total personal
exposure to CO includes both ambient and nonambient components; and both components may contribute to adverse
health effects.
6-1
-------
Animal studies show various neurological effects with in-utero CO exposure. Controlled
human exposure studies report inconsistent neural and behavioral effects following low-level CO
exposures. The ISA concludes the evidence is suggestive of a causal relationship with both
short- and long-term exposure to CO and central nervous system effects.
A number of epidemiologic and animal toxicological studies cited in the ISA have
evaluated associations between CO exposure and birth outcomes such as preterm birth or cardiac
birth defects. The epidemiologic studies provide limited evidence of a CO-induced effect on
preterm births and birth defects, with weak evidence for a decrease in birth weight. Animal
toxicological studies have found associations between perinatal CO exposure and decrements in
birth weight, as well as other developmental outcomes. The ISA concludes these studies are
suggestive of a causal relationship between long-term exposures to CO and developmental
effects and birth outcomes.
Epidemiologic studies provide evidence of effects on respiratory morbidity such as
changes in pulmonary function, respiratory symptoms, and hospital admissions associated with
ambient CO concentrations. A limited number of epidemiologic studies considered copollutants
such as ozone, SO2, and PM in two-pollutant models and found that CO risk estimates were
generally robust, although this limited evidence makes it difficult to disentangle effects attributed
to CO itself from those of the larger complex air pollution mixture. Controlled human exposure
studies have not extensively evaluated the effect of CO on respiratory morbidity. Animal studies
at levels of 50-100 ppm CO show preliminary evidence of altered pulmonary vascular
remodeling and oxidative injury. The ISA concludes that the evidence is suggestive of a causal
relationship between short-term CO exposure and respiratory morbidity, and inadequate to
conclude that a causal relationship exists between long-term exposure and respiratory morbidity.
Finally, the ISA concludes that the epidemiologic evidence is suggestive of a causal
relationship between short-term exposures to CO and mortality. Epidemiologic studies provide
evidence of an association between short-term exposure to CO and mortality, but limited
evidence is available to evaluate cause-specific mortality outcomes associated with CO exposure.
In addition, the attenuation of CO risk estimates which was often observed in copollutant models
contributes to the uncertainty as to whether CO is acting alone or as an indicator for other
combustion-related pollutants. The ISA also concludes that there is not likely to be a causal
relationship between relevant long-term exposures to CO and mortality.
6.1.5 Health Effects of Air Toxics
6.1.5.1 Benzene
The EPA's IRIS database lists benzene as a known human carcinogen (causing leukemia)
by all routes of exposure, and concludes that exposure is associated with additional health
effects, including genetic changes in both humans and animals and increased proliferation of
bone marrow cells in mice.42'43'44 EPA states in its IRIS database that data indicate a causal
relationship between benzene exposure and acute lymphocytic leukemia and suggest a
relationship between benzene exposure and chronic non-lymphocytic leukemia and chronic
6-8
-------
lymphocyte leukemia. EPA's IRIS documentation for benzene also lists a range of 2.2 x 10~6 to
7.8 x 10~6 as the unit risk estimate (URE) for benzene.H'45 The International Agency for
Research on Carcinogens (IARC) has determined that benzene is a human carcinogen and the
U.S. Department of Health and Human Services (DHHS) has characterized benzene as a known
1 • 46 47
human carcinogen. '
A number of adverse noncancer health effects including blood disorders, such as
preleukemia and aplastic anemia, have also been associated with long-term exposure to
benzene.48'49 The most sensitive noncancer effect observed in humans, based on current data, is
the depression of the absolute lymphocyte count in blood.50'51 EPA's inhalation reference
concentration (RfC) for benzene is 30 |ig/m3. The RfC is based on suppressed absolute
lymphocyte counts seen in humans under occupational exposure conditions. In addition, recent
work, including studies sponsored by the Health Effects Institute (HEI), provides evidence that
biochemical responses are occurring at lower levels of benzene exposure than previously
known.52'53'54'55 EPA's IRIS program has not yet evaluated these new data. EPA does not
currently have an acute reference concentration for benzene. The Agency for Toxic Substances
and Disease Registry (ATSDR) Minimal Risk Level (MRL) for acute exposure to benzene is 29
|ig/m3 for 1-14 days exposure. 6>I
6.1.5.2 1,3-Butadiene
EPA has characterized 1,3-butadiene as carcinogenic to humans by inhalation.57'58 The
IARC has determined that 1,3-butadiene is a human carcinogen and the U.S. DHHS has
characterized 1,3-butadiene as a known human carcinogen.5 '60'61 There are numerous studies
consistently demonstrating that 1,3-butadiene is metabolized into genotoxic metabolites by
experimental animals and humans. The specific mechanisms of 1,3-butadiene-induced
carcinogenesis are unknown; however, the scientific evidence strongly suggests that the
carcinogenic effects are mediated by genotoxic metabolites. Animal data suggest that females
may be more sensitive than males for cancer effects associated with 1,3-butadiene exposure;
there are insufficient data in humans from which to draw conclusions about sensitive
subpopulations. The URE for 1,3-butadiene is 3 x io~5 per jig/m3.62 1,3-butadiene also causes a
variety of reproductive and developmental effects in mice; no human data on these effects are
available. The most sensitive effect was ovarian atrophy observed in a lifetime bioassay of
female mice.63 Based on this critical effect and the benchmark concentration methodology, an
RfC for chronic health effects was calculated at 0.9 ppb (approximately 2 |ig/m3).
6.1.5.3 Ethanol
EPA is planning to develop an assessment of the health effects of exposure to ethanol, a
compound which is not currently listed on EPA's IRIS database. Extensive health effects data
are available for ingestion of ethanol, while data on inhalation exposure effects are sparse. In
H A unit risk estimate is defined as the increase in the lifetime risk of an individual who is exposed for a lifetime to 1
ug/m3 benzene in air.
1A minimal risk level (MRL) is defined as an estimate of the daily human exposure to a hazardous substance that is
likely to be without appreciable risk of adverse noncancer health effects over a specified duration of exposure.
6-9
-------
developing the assessment, EPA is evaluating pharmacokinetic models as a means of
extrapolating across species (animal to human) and across exposure routes (oral to inhalation) to
better characterize the health hazards and dose-response relationships for low levels of ethanol
exposure in the environment.
6.1.5.4 Formaldehyde
In 1991, EPA concluded that formaldehyde is a carcinogen based on nasal tumors in
animal bioassays.64 An Inhalation Unit Risk for cancer and a Reference Dose for oral noncancer
effects were developed by the Agency and posted on the Integrated Risk Information System
(IRIS) database. Since that time, the National Toxicology Program (NTP) and International
Agency for Research on Cancer (IARC) have concluded that formaldehyde is a known human
carcinogen.65'66'67
The conclusions by IARC and NTP reflect the results of epidemiologic research
published since 1991 in combination with previous animal, human and mechanistic evidence.
Research conducted by the National Cancer Institute reported an increased risk of
nasopharyngeal cancer and specific lymphohematopoietic malignancies among workers exposed
to formaldehyde.68'69'70 A National Institute of Occupational Safety and Health study of garment
workers also reported increased risk of death due to leukemia among workers exposed to
formaldehyde.71 Extended follow-up of a cohort of British chemical workers did not report
evidence of an increase in nasopharyngeal or lymphohematopoietic cancers, but a continuing
statistically significant excess in lung cancers was reported.72 Finally, a study of embalmers
reported formaldehyde exposures to be associated with an increased risk of myeloid leukemia
but not brain cancer.73
Health effects of formaldehyde in addition to cancer were reviewed by the Agency for
Toxics Substances and Disease Registry in 199974 and supplemented in 2010,75 and by the World
Health Organization.76 These organizations reviewed the literature concerning effects on the
eyes and respiratory system, the primary point of contact for inhaled formaldehyde, including
sensory irritation of eyes and respiratory tract, pulmonary function, nasal histopathology, and
immune system effects. In addition, research on reproductive and developmental effects and
neurological effects were discussed.
EPA released a draft Toxicological Review of Formaldehyde - Inhalation Assessment
through the IRIS program for peer review by the National Research Council (NRC) and public
comment in June 2010.77 The draft assessment reviewed more recent research from animal and
human studies on cancer and other health effects. The NRC released their review report in April
•yo
2011 (http://www.nap.edu/catalog.php?record_id=13142). The EPA is currently revising the
draft assessment in response to this review.
6.1.5.5 Acetaldehyde
Acetaldehyde is classified in EPA's IRIS database as a probable human carcinogen,
based on nasal tumors in rats, and is considered toxic by the inhalation, oral, and intravenous
routes.79 The URE in IRIS for acetaldehyde is 2.2 x 10'6 per |ig/m3.80 Acetaldehyde is
reasonably anticipated to be a human carcinogen by the U.S. DHHS in the 12th Report on
6-10
-------
O 1 OO
Carcinogens and is classified as possibly carcinogenic to humans (Group 2B) by the IARC. '
EPA is currently conducting a reassessment of cancer risk from inhalation exposure to
acetaldehyde.
The primary noncancer effects of exposure to acetaldehyde vapors include irritation of
the eyes, skin, and respiratory tract.83 In short-term (4 week) rat studies, degeneration of
QA O C
olfactory epithelium was observed at various concentration levels of acetaldehyde exposure. '
Data from these studies were used by EPA to develop an inhalation reference concentration of 9
|ig/m3. Some asthmatics have been shown to be a sensitive subpopulation to decrements in
functional expiratory volume (FEV1 test) and bronchoconstriction upon acetaldehyde
o/r
inhalation. The agency is currently conducting a reassessment of the health hazards from
inhalation exposure to acetaldehyde.
6.1.5.6 Acrolein
EPA most recently evaluated the toxicological and health effects literature related to
acrolein in 2003 and concluded that the human carcinogenic potential of acrolein could not be
determined because the available data were inadequate. No information was available on the
carcinogenic effects of acrolein in humans and the animal data provided inadequate evidence of
carcinogen!city.87 The IARC determined in 1995 that acrolein was not classifiable as to its
oo
carcinogenicity in humans.
Lesions to the lungs and upper respiratory tract of rats, rabbits, and hamsters have been
observed after subchronic exposure to acrolein.89 The Agency has developed an RfC for acrolein
of 0.02 |ig/m3 and an RfD of 0.5 |ig/kg-day.90 EPA is considering updating the acrolein
assessment with data that have become available since the 2003 assessment was completed.
Acrolein is extremely acrid and irritating to humans when inhaled, with acute exposure
resulting in upper respiratory tract irritation, mucus hypersecretion and congestion. The intense
irritancy of this carbonyl has been demonstrated during controlled tests in human subjects, who
suffer intolerable eye and nasal mucosal sensory reactions within minutes of exposure.91 These
data and additional studies regarding acute effects of human exposure to acrolein are
summarized in EPA's 2003 IRIS Human Health Assessment for acrolein.92 Studies in humans
indicate that levels as low as 0.09 ppm (0.21 mg/m3) for five minutes may elicit subjective
complaints of eye irritation with increasing concentrations leading to more extensive eye, nose
and respiratory symptoms. Acute exposures in animal studies report bronchial hyper-
responsiveness. Based on animal data (more pronounced respiratory irritancy in mice with
allergic airway disease in comparison to non-diseased mice93) and demonstration of similar
effects in humans (e.g., reduction in respiratory rate), individuals with compromised respiratory
function (e.g., emphysema, asthma) are expected to be at increased risk of developing adverse
responses to strong respiratory irritants such as acrolein. EPA does not currently have an acute
reference concentration for acrolein. The available health effect reference values for acrolein
have been summarized by EPA and include an ATSDR MRL for acute exposure to acrolein of
7 |ig/m3 for 1-14 days exposure; and Reference Exposure Level (REL) values from the
California Office of Environmental Health Hazard Assessment (OEHHA) for one-hour and 8-
hour exposures of 2.5 |ig/m3 and 0.7 |ig/m3, respectively.94
6-11
-------
6.1.5.7 PAN
PAN (peroxy acetyl nitrate) has not been evaluated by EPA's IRIS program. Information
regarding the potential carcinogenicity of PAN is limited. As noted in the EPA air quality
criteria document for ozone and related photochemical oxidants, cytogenetic studies indicate that
PAN is not a potent mutagen, clastogen (a compound that can cause breaks in chromosomes), or
DNA-damaging agent in mammalian cells either in vivo or in vitro. Some studies suggest that
PAN may be a weak bacterial mutagen at concentrations much higher than exist in present urban
atmospheres.95
Effects of ground-level smog causing intense eye irritation have been attributed to
photochemical oxidants, including PAN.96 Animal toxicological information on the inhalation
effects of the non-ozone oxidants has been limited to a few studies on PAN. Acute exposure to
levels of PAN can cause changes in lung morphology, behavioral modifications, weight loss, and
susceptibility to pulmonary infections. Human exposure studies indicate minor pulmonary
function effects at high PAN concentrations, but large inter-individual variability precludes
definitive conclusions.97
6.1.5.8 Polycyclic Organic Matter
The term polycyclic organic matter (POM) defines a broad class of compounds that
includes the polycyclic aromatic hydrocarbon compounds (PAHs). One of these compounds,
naphthalene, is discussed separately below. POM compounds are formed primarily from
combustion and are present in the atmosphere in gas and paniculate form. Cancer is the major
concern from exposure to POM. Epidemiologic studies have reported an increase in lung cancer
in humans exposed to diesel exhaust, coke oven emissions, roofing tar emissions, and cigarette
smoke; all of these mixtures contain POM compounds.9899 Animal studies have reported
respiratory tract tumors from inhalation exposure to benzo[a]pyrene and alimentary tract and
liver tumors from oral exposure to benzo[a]pyrene.100 In 1997 EPA classified seven PAHs
(benzo[a]pyrene, benz[a]anthracene, chrysene, benzo[b]fluoranthene, benzo[k]fluoranthene,
dibenz[a,h]anthracene, and indeno[l,2,3-cd]pyrene) as Group B2, probable human
carcinogens.101 Since that time, studies have found that maternal exposures to PAHs in a
population of pregnant women were associated with several adverse birth outcomes, including
low birth weight and reduced length at birth, as well as impaired cognitive development in
preschool children (3 years of age).102'103 These and similar studies are being evaluated as a part
of the ongoing IRIS assessment of health effects associated with exposure to benzo[a]pyrene.
6.1.5.9 Naphthalene
Naphthalene is found in small quantities in gasoline and diesel fuels. Naphthalene
emissions have been measured in larger quantities in both gasoline and diesel exhaust compared
with evaporative emissions from mobile sources, indicating it is primarily a product of
combustion. Acute (short-term) exposure of humans to naphthalene by inhalation, ingestion, or
dermal contact is associated with hemolytic anemia and damage to the liver and the nervous
system.104 Chronic (long term) exposure of workers and rodents to naphthalene has been
reported to cause cataracts and retinal damage.105 EPA released an external review draft of a
reassessment of the inhalation carcinogenicity of naphthalene based on a number of recent
6-12
-------
animal carcinogen!city studies.106 The draft reassessment completed external peer review.107
Based on external peer review comments received, a revised draft assessment that considers all
routes of exposure, as well as cancer and noncancer effects, is under development. The external
review draft does not represent official agency opinion and was released solely for the purposes
of external peer review and public comment. The National Toxicology Program listed
naphthalene as "reasonably anticipated to be a human carcinogen" in 2004 on the basis of
bioassays reporting clear evidence of carcinogenicity in rats and some evidence of
carcinogenicity in mice.108 California EPA has released a new risk assessment for naphthalene,
and the IARC has reevaluated naphthalene and re-classified it as Group 2B: possibly
carcinogenic to humans.109
Naphthalene also causes a number of chronic non-cancer effects in animals, including
abnormal cell changes and growth in respiratory and nasal tissues.110 The current EPA IRIS
assessment includes noncancer data on hyperplasia and metaplasia in nasal tissue that form the
basis of the inhalation RfC of 3 jig/m3.111 The ATSDR MRL for acute exposure to naphthalene
is 0.6 mg/kg/day.
6.1.5.10 Other Air Toxics
In addition to the compounds described above, other compounds in gaseous hydrocarbon
and PM emissions from vehicles will be affected by this proposal. Mobile source air toxic
compounds that would potentially be impacted include ethylbenzene, propionaldehyde, toluene,
and xylene. Information regarding the health effects of these compounds can be found in EPA's
IRIS database/
6.1.6 Traffic-associated health effects
In addition to health concerns resulting from specific air pollutants, a large number of
studies have examined the health status of populations near major roadways. These studies
frequently have employed exposure metrics that are not specific to individual pollutants, but
rather reflect the large number of different pollutants found in elevation near major roads.
In this section of the RIA, information on health effects associated with air quality near
major roads or traffic in general is summarized. Generally, the section makes use of publications
that systematically review literature on a given health topic. In particular, this section makes
frequent reference of a report of by the Health Effects Institute (HEI) Panel on the Health Effects
of Traffic-Related Air Pollution, published in 2010 as a review of relevant studies.1^112 Other
systematic reviews of relevant literature are cited were appropriate.
1 U.S. EPA Integrated Risk Information System (IRIS) database is available at: www.epa.gov/iris
K It should be noted that there are no peer reviewed EPA-authored reviews of traffic-related health studies. The HEI
panel primarily used epidemiology studies for inferring whether there was sufficient evidence of a causal association
exists between a particular health effect and traffic-related air pollution, In its weight-of-evidence determinations,
the panel also placed "considerable weight" on controlled human exposure studies. However, it restricted
consideration of other lexicological studies to whether or not the studies provided "general mechanistic support" for
the inferences of causality made on the basis of epidemiology.
6-13
-------
6.1.6.1 Populations near major roads
Numerous studies have estimated the size and demographics of populations that live near
major roads. Other studies have estimated the number of schools near major roads, and the
populations of students in such schools.
Every two years, the U.S. Census Bureau's American Housing Survey (AHS) has
reported whether housing units are within 300 feet of an "airport, railroad, or highway with four
or more lanes." The 2009 survey reports that over 22 million homes, or 17 percent of all housing
units in the U.S., were located in such areas. Assuming that populations and housing units are in
the same locations, this corresponds to a population of more than 50 million U.S. residents in
close proximity to high-traffic roadways or other transportation sources. According to the
Central Intelligence Agency's World Factbook, in 2010, the United States had 6,506,204 km or
roadways, 224,792 km of railways, and 15,079 airports. As such, highways represent the
overwhelming majority of transportation facilities described by this factor in the AHS.
The AHS reports are published every two years. As such, trends in the AHS can be
reported to describe whether a greater or lesser proportion of homes are located near major roads
over time. Figure 6-1 depicts trends in the number and proportion of homes located near major
transportation sources, which generally indicate large roadways. As the figure indicates, since
2005, there has been a substantial increase in the number and percentage of homes located near
major transportation sources. As such, the population in close proximity to these sources, which
may be affected by near-road air quality and health concerns, appears to have increased over
time.
6-14
-------
1 18 +
o
I
I 16 +
14 --
12 --
10
18.0%
16.0%
14.0%
£
^
12.0% "§
-i— i
w
10.0% H
8.0% o>
6.0%
4.0%
2.0%
0.0%
i iMiiiinn Housing Units
•Near4+ Lane Highway,
Railroad, or Airport
1997 1999 2001 2003 2005 2007 2009
Yearof American Housing Survey
Figure 6-1 Trends in Populations Near Large Highways, Railroads, and Airports
Furthermore, according to data from the 2008 American Time Use Survey (ATUS),
conducted by the Bureau of Labor Statistics (BTS), Americans spend more than an hour
traveling each day, on average.113 Although the ATUS does not indicate their mode of travel, the
majority of trips undertaken nationally is by motor vehicle.114 As such, daily travel activity
brings nearly all residents into a high-exposure microenvironment for part of the day.
6.1.6.2 Premature mortality
The HEI panel report concluded that evidence linking traffic-associated air pollution with
premature mortality from all causes was "suggestive but not sufficient" to infer a causal
relationship. This conclusion was based largely on several long-term studies that "qualitatively"
examined whether or not someone was exposed to traffic-associated air pollution. In addition,
based on several short-term studies of exposure, the panel concluded that there was evidence that
there was "suggestive but not sufficient" evidence to infer a causal relation between traffic-
related exposure and cardiovascular mortality.
6.1.6.3 Cardiovascular effects
6.1.6.3.1 Cardiac physiology
Exposure to traffic-associated pollutants has been associated with changes in cardiac
physiology, including cardiac function. One common measure of cardiac function is heart rate
6-15
-------
variability (HRV), an indicator of the heart's ability to respond to variations in stress, reflecting
the nervous system's ability to regulate the heart.L Reduced HRV is associated with adverse
cardiovascular events, such as myocardial infarction, in heart disease patients. The HEI panel
concluded that available evidence provides evidence for a causal association between exposure
to traffic-related pollutants and reduced control of HRV by the nervous system. Overall, the
panel concluded that the evidence was "suggestive but not sufficient" to infer a causal relation
between traffic-related pollutants and cardiac function. Studies suggest that the HRV changes
from traffic-related air pollution result in changes to heart rhythms, which can lead to
arrhythmia.115'116
6.1.6.3.2 Heart attack and atherosclerosis
The HEI panel concluded that epidemiologic evidence of the association between traffic-
related pollutants and heart attacks and atherosclerosis was "suggestive but not sufficient" to
infer a causal association. In addition, the panel concluded that the toxicology studies they
reviewed provided "suggestive evidence that exposure to traffic emissions, including ambient
and laboratory-generated [PM] and diesel- and gasoline-engine exhaust, alters cardiovascular
function." The panel noted there are few studies of human volunteers exposed to real-world
traffic mixture, which were not entirely consistent. The panel notes that the studies provide
consistent evidence for exposure to PM and impaired cardiovascular responses. In addition to
the HEI study, several other reviews of available evidence conclude that there is evidence
supporting a causal association between traffic-related air pollution and cardiovascular
disease.11
A number of mechanisms for cardiovascular disease are highlighted in the HEI and AHA
report, including modified blood vessel endothelial function (e.g, the ability to dilate),
atherosclerosis, and oxidative stress. The HEI review cites "two well executed studies" in which
hospitalization for acute myocardial infarction (i.e., heart attack) were associated with traffic
exposures and a prospective study finding higher rates of arterial hardening and coronary heart
disease near traffic.
6.1.6.4 Respiratory effects
6.1.6.4.1 Asthma
Pediatric asthma and asthma symptoms are the effects that have been evaluated by the
largest number of studies in the epidemiologic literature on the topic. In general, studies
consistently show effects of residential or school exposure to traffic and asthma symptoms, and
the effects are frequently statistically significant. Studies have employed both short-term and
long-term exposure metrics, and a range of different respiratory measures. HEI Special Report
17 (HEI Panel on the Health Effects of Traffic-Related Air Pollution, 2010) concluded that there
L The autonomic nervous system (ANS) consists of sympathetic and parasympathetic components. The sympathetic
ANS signals body systems to "fight or flight." The parasympathetic ANS signals the body to "rest and digest." In
general, HRV is indicative of parasympathetic control of the heart.
6-16
-------
is sufficient evidence for a causal association between exposure to traffic-related air pollution
and exacerbation of asthma symptoms in children.
While there is general consistency in studies examining asthma incidence in children, the
available studies employ different definitions of asthma (e.g., self-reported vs. hospital records),
methods of exposure assessment, and population age ranges. As such, the overall evidence,
while supportive of an association between traffic exposure and new onset asthma, are less
consistent than for asthma symptoms. The HEI report determined that there is "sufficient" or
"suggestive" evidence of a causal relationship between exposure to traffic-related air pollution
and incident (new onset) asthma in children (HEI Panel on the Health Effects of Traffic-Related
Air Pollution, 2010). A recent meta-analysis of studies on incident asthma and air pollution in
general, based on studies dominated by traffic-linked exposure metrics, also concluded that
available evidence that exposures is consistent with a effect of exposure on asthma incidence
(Anderson et al., 2011). The study reported excess main risk estimates for different pollutants
ranging from 7-16 percent per 10 |ig/m3 of long-term exposure (random effects models). Other
qualitative reviews (Salam et al., 2008; Braback and Forsberg, 2009) conclude that available
evidence is consistent with the hypothesis that traffic-associated air pollutants are associated with
incident asthma.
6.1.6.4.2 Chronic obstructive pulmonary disease (COPD)
The HEI panel reviewed available studies examining COPD in the context of traffic-
associated air pollution. Because of how the panel selected studies for inclusion in review, there
were only two studies that they used to review the available evidence. Both studies reported
some positive associations, but not for all traffic metrics. The small number of studies and lack
of consistency across traffic metrics led the panel to conclude that there is insufficient evidence
for traffic-associated air pollution causing COPD.
6.1.6.4.3 Allergy
There are numerous human and animal experimental studies that provides strongly
suggestive evidence that traffic-related air pollutants can enhance allergic responses to common
allergens.118'119'120 However, in its review of 16 epidemiologic studies that address traffic-related
air pollution's effect on allergies, the HEI expert panel (HEI, 2010) reported that only two such
studies showed consistently positive associations. As a result, despite the strong experimental
evidence, the panel concluded that there is "inadequate/insufficient" evidence of an association
between allergy and traffic-associated air pollution. As noted above, the HEI panel considered
toxicological evidence only based on whether or not they provide mechanistic support for
observations and inferences derived from epidemiology.
6.1.6.4.4 Lung function
There are numerous measurements of breathing (spirometry) that indicate the presence or
degree of airway disease, such as asthma and chronic obstructive pulmonary disease (COPD).
Forced vital capacity (FVC) is measured when a patient maximally fills their lungs and then
blows their hardest in completely exhaling. The peak expiratory flow (PEF) is the maximum air
flow achievable during exhalation. The forced expiratory volume in the first second of
6-17
-------
exhalation is referred to as FEVi. FEVi and PEF reflect the function of the large airways. FVC
and FEVi, along with their ratio (FVC/FEVi) are used to classify airway obstruction in asthma
and COPD. Measurements of air flow at various times during forced exhalation, such as 25
percent, 50 percent, and 75 percent, are also used. The flow at 75 percent of forced exhalation
) reflects the status of small airways, which asthma and COPD affect.
The FIEI panel concluded that the available literature suggests that long-term exposure to
traffic-related air pollution is associated with reduced lung function in adolescents and young
adults and that lung function is lower in populations in areas with high traffic-related air
pollutant levels. However, the panel noted the difficulty of disentangling traffic-specific
exposures from urban air pollution in general. The studies reviewed that were more specifically
oriented toward traffic were not consistent in their findings. As a result, the panel found that the
evidence linking lung function and traffic exposure is "inadequate and insufficient" to infer a
causal relationship.
6. 1 .6.5 Reproductive and developmental effects
Several studies have reported associations between traffic-related air pollution and
adverse birth outcomes, such as preterm birth and low birth weight. At the time of the FIEI
review, the panel concluded that evidence for adverse birth outcomes being causally associated
with traffic-related exposures was "inadequate and insufficient." Only four studies met the
panel's inclusion criteria, and had limited geographic coverage. One study provided evidence of
small but consistently increased risks using multiple exposure metrics. No studies were at the
time available that examined traffic-specific exposures and congenital abnormalities. Since then,
several studies investigating birth outcomes have been published, but no new systematic reviews.
One new meta-analysis of air pollution and congenital abnormalities has been published, though
none of the reviewed studies includes traffic-specific exposure information.
The HEI panel also reviewed toxicological studies of traffic-related air pollutants and
fertility. While numerous studies examining animal or human exposure and sperm count have
been published, the panel concluded that the generally high exposure concentrations employed in
the studies limited the applicability to typical ambient concentrations. Because there was no
overlap in the effects studied by epidemiology and toxicology studies, no synthesis review of the
combined literature was undertaken.
Since the HEI panel's publication, a systematic review and meta-analysis of air pollution
and congenital abnormalities was published.121 In that review, only one study directly included
nearby traffic in its exposure analysis. As such, there are so systematic reviews that specifically
address traffic's impact on congenital abnormalities.
6.1.6.6 Cancer
6. 1. 6. 6. 1 Childhood cancer
A number of studies examining various types of childhood cancer have been published
with mixed results. The HEI panel concluded that the available epidemiologic evidence was
"inadequate and insufficient" to infer a causal relationship between traffic-related air pollution
and childhood cancer. An earlier review article on the topic noted that studies reporting positive
6-18
-------
effects tended to be small, while those with null effects tended to be larger, suggesting the
potential for publication bias in the available literature.122
6.1.6.6.2 Adult cancer
Several studies have examined the risk of adult lung cancers in relation to exposure to
traffic-related air pollutants. The HEI panel evaluated four such studies, and rated the available
evidence as "inadequate and insufficient" to infer a causal relation for non-occupational lung
cancer.
6.1.6.7 Neurological effects
The HEI panel found that current toxicologic and epidemiologic literature on the
neurotoxicity of traffic-related air pollution was inadequate for their evaluation. The panel noted
that there were a number of toxicologic studies of traffic-associated pollutants, but found them to
have diverse exposure protocols, animal models, and endpoints, making them unsuitable for
systematic evaluation.
6.2 Environmental Effects of Criteria and Toxic Pollutants
6.2.1 Visibility Degradation
Visibility can be defined as the degree to which the atmosphere is transparent to visible
light.123 Visibility impairment is caused by light scattering and absorption by suspended
particles and gases. Visibility is important because it has direct significance to people's
enjoyment of daily activities in all parts of the country. Individuals value good visibility for the
well-being it provides them directly, where they live and work, and in places where they enjoy
recreational opportunities. Visibility is also highly valued in significant natural areas, such as
national parks and wilderness areas, and special emphasis is given to protecting visibility in these
areas. For more information on visibility see the final 2009 PM ISA. 24
EPA is pursuing a two-part strategy to address visibility impairment. First, EPA
developed the regional haze program (64 FR 35714) which was put in place in July 1999 to
protect the visibility in Mandatory Class I Federal areas. There are 156 national parks, forests
and wilderness areas categorized as Mandatory Class I Federal areas (62 FR 38680-38681, July
18, 1997). These areas are defined in CAA section 162 as those national parks exceeding 6,000
acres, wilderness areas and memorial parks exceeding 5,000 acres, and all international parks
which were in existence on August 7, 1977. Second, EPA has concluded that PM2.5 causes
adverse effects on visibility in other areas that are not protected by the Regional Haze Rule,
depending on PM2.5 concentrations and other factors that control their visibility impact
effectiveness such as dry chemical composition and relative humidity (i.e., an indicator of the
water composition of the particles). EPA revised the PM2.5 standards in December 2012 and
established a target level of protection that is expected to be met through attainment of the
existing secondary standards for PM2.5. Figure 6-2 shows the location of the 156 Mandatory
Class I Federal areas.
6-19
-------
Produced by NPS Air Resources Division
* Rainbow Lake.WI and Brad well Bay, FL are Class 1 Areas
where visibility is not an important air quality related value
Figure 6-2 Mandatory Class I Federal Areas in the U.S.
6.2.1.1 Vi sibility Monitoring
In conjunction with the U.S. National Park Service, the U.S. Forest Service, other Federal
land managers, and State organizations in the U.S., the U.S. EPA has supported visibility
monitoring in national parks and wilderness areas since 1988. The monitoring network was
originally established at 20 sites, but it has now been expanded to 110 sites that represent all but
one of the 156 Mandatory Federal Class I areas across the country (see Figure 6-2). This long-
term visibility monitoring network is known as IMPROVE (Interagency Monitoring of Protected
Visual Environments).
IMPROVE provides direct measurement of fine particles that contribute to visibility
impairment. The IMPROVE network employs aerosol measurements at all sites, and optical and
scene measurements at some of the sites. Aerosol measurements are taken for PMio and PM2.5
mass, and for key constituents of PM2.s, such as sulfate, nitrate, organic and elemental carbon,
soil dust, and several other elements. Measurements for specific aerosol constituents are used to
calculate "reconstructed" aerosol light extinction by multiplying the mass for each constituent by
its empirically-derived scattering and/or absorption efficiency, with adjustment for the relative
humidity. Knowledge of the main constituents of a site's light extinction "budget" is critical for
source apportionment and control strategy development. In addition to this indirect method of
assessing light extinction, there are optical measurements which directly measure light extinction
or its components. Such measurements are made principally with either a nephelometer to
measure light scattering, some sites also include an aethalometer for light absorption, or at a few
sites using a transmissometer, which measures total light extinction. Scene characteristics are
typically recorded using digital or video photography and are used to determine the quality of
visibility conditions (such as effects on color and contrast) associated with specific levels of light
extinction as measured under both direct and aerosol-related methods. Directly measured light
6-20
-------
extinction is used under the IMPROVE protocol to cross check that the aerosol-derived light
extinction levels are reasonable in establishing current visibility conditions. Aerosol-derived
light extinction is used to document spatial and temporal trends and to determine how changes in
atmospheric constituents would affect future visibility conditions.
Annual average visibility conditions (reflecting light extinction due to both anthropogenic
and non-anthropogenic sources) vary regionally across the U.S. Visibility is typically worse in
the summer months and the rural East generally has higher levels of impairment than remote
sites in the West. Figures 9-9 through 9-11 in the PM ISA detail the percent contributions to
particulate light extinction for ammonium nitrate and sulfate, EC and OC, and coarse mass and
1 7S
fine soil, by season.
6.2.2 Particulate Matter Deposition
Particulate matter contributes to adverse effects on vegetation and ecosystems, and to
soiling and materials damage. These welfare effects result predominately from exposure to
excess amounts of specific chemical species, regardless of their source or predominant form
(particle, gas or liquid). The following characterizations of the nature of these environmental
effects are based on information contained in the 2009 PM ISA and the 2005 PM Staff Paper as
well as the Integrated Science Assessment for Oxides of Nitrogen and Sulfur- Ecological
Criteria.126'127'128
6.2.2.1 Deposition of Nitrogen and Sulfur
Nitrogen and sulfur interactions in the environment are highly complex. Both nitrogen
and sulfur are essential, and sometimes limiting, nutrients needed for growth and productivity.
Excesses of nitrogen or sulfur can lead to acidification, nutrient enrichment, and eutrophication
of aquatic ecosystems.129
The process of acidification affects both freshwater aquatic and terrestrial ecosystems.
Acid deposition causes acidification of sensitive surface waters. The effects of acid deposition
on aquatic systems depend largely upon the ability of the ecosystem to neutralize the additional
acid. As acidity increases, aluminum leached from soils and sediments, flows into lakes and
streams and can be toxic to both terrestrial and aquatic biota. The lower pH concentrations and
higher aluminum levels resulting from acidification make it difficult for some fish and other
aquatic organisms to survive, grow, and reproduce. Research on effects of acid deposition on
forest ecosystems has come to focus increasingly on the biogeochemical processes that affect
uptake, retention, and cycling of nutrients within these ecosystems. Decreases in available base
cations from soils are at least partly attributable to acid deposition. Base cation depletion is a
cause for concern because of the role these ions play in acid neutralization, and because calcium,
magnesium and potassium are essential nutrients for plant growth and physiology. Changes in
the relative proportions of these nutrients, especially in comparison with aluminum
concentrations, have been associated with declining forest health.
At current ambient levels, risks to vegetation from short-term exposures to dry deposited
particulate nitrate or sulfate are low. However, when found in acid or acidifying deposition, such
particles do have the potential to cause direct leaf injury. Specifically, the responses of forest
6-21
-------
trees to acid precipitation (rain, snow) include accelerated weathering of leaf cuticular surfaces,
increased permeability of leaf surfaces to toxic materials, water, and disease agents; increased
leaching of nutrients from foliage; and altered reproductive processes—all which serve to
weaken trees so that they are more susceptible to other stresses (e.g., extreme weather, pests,
pathogens). Acid deposition with levels of acidity associated with the leaf effects described
above are currently found in some locations in the eastern U.S.130 Even higher concentrations of
acidity can be present in occult depositions (e.g., fog, mist or clouds) which more frequently
impacts higher elevations. Thus, the risk of leaf injury occurring from acid deposition in some
areas of the eastern U.S. is high. Nitrogen deposition has also been shown to impact ecosystems
in the western U.S. A study conducted in the Columbia River Gorge National Scenic Area
(CRGNSA), located along a portion of the Oregon/Washington border, indicates that lichen
communities in the CRGNSA have shifted to a higher proportion of nitrophilous species and the
nitrogen content of lichen tissue is elevated.131 Lichens are sensitive indicators of nitrogen
deposition effects to terrestrial ecosystems and the lichen studies in the Columbia River Gorge
clearly show that ecological effects from air pollution are occurring.
Some of the most significant detrimental effects associated with excess nitrogen
deposition are those associated with a condition known as nitrogen saturation. Nitrogen
saturation is the condition in which nitrogen inputs from atmospheric deposition and other
sources exceed the biological requirements of the ecosystem. The effects associated with
nitrogen saturation include: (1) decreased productivity, increased mortality, and/or shifts in plant
community composition, often leading to decreased biodiversity in many natural habitats
wherever atmospheric reactive nitrogen deposition increases significantly above background and
critical thresholds are exceeded; (2) leaching of excess nitrate and associated base cations from
soils into streams, lakes, and rivers, and mobilization of soil aluminum; and (3) fluctuation of
ecosystem processes such as nutrient and energy cycles through changes in the functioning and
species composition of beneficial soil organisms.132
In the U.S. numerous forests now show severe symptoms of nitrogen saturation. These
forests include: the northern hardwoods and mixed conifer forests in the Adirondack and
Catskill Mountains of New York; the red spruce forests at Whitetop Mountain, Virginia, and
Great Smoky Mountains National Park, North Carolina; mixed hardwood watersheds at Fernow
Experimental Forest in West Virginia; American beech forests in Great Smoky Mountains
National Park, Tennessee; mixed conifer forests and chaparral watersheds in southern California
and the southwestern Sierra Nevada in Central California; the alpine tundra/subalpine conifer
forests of the Colorado Front Range; and red alder forests in the Cascade Mountains in
Washington.
Excess nutrient inputs into aquatic ecosystems (i.e. streams, rivers, lakes, estuaries or
oceans) either from direct atmospheric deposition, surface runoff, or leaching from nitrogen
saturated soils into ground or surface waters can contribute to conditions of severe water oxygen
depletion; eutrophication and algae blooms; altered fish distributions, catches, and physiological
states; loss of biodiversity; habitat degradation; and increases in the incidence of disease.
Atmospheric deposition of nitrogen is a significant source of total nitrogen to many
estuaries in the United States. The amount of nitrogen entering estuaries that is ultimately
attributable to atmospheric deposition is not well-defined. On an annual basis, atmospheric
6-22
-------
nitrogen deposition may contribute significantly to the total nitrogen load, depending on the size
and location of the watershed. In addition, episodic nitrogen inputs, which may be ecologically
important, may play a more important role than indicated by the annual average concentrations.
Estuaries in the U.S. that suffer from nitrogen enrichment often experience a condition known as
eutrophication. Symptoms of eutrophication include changes in the dominant species of
phytoplankton, low levels of oxygen in the water column, fish and shellfish kills, outbreaks of
toxic alga, and other population changes which can cascade throughout the food web. In
addition, increased phytoplankton growth in the water column and on surfaces can attenuate light
causing declines in submerged aquatic vegetation, which serves as an important habitat for many
estuarine fish and shellfish species.
Severe and persistent eutrophication often directly impacts human activities. For
example, losses in the nation's fishery resources may be directly caused by fish kills associated
with low dissolved oxygen and toxic blooms. Declines in tourism occur when low dissolved
oxygen causes noxious smells and floating mats of algal blooms create unfavorable aesthetic
conditions. Risks to human health increase when the toxins from algal blooms accumulate in
edible fish and shellfish, and when toxins become airborne, causing respiratory problems due to
inhalation. According to a NOAA report, more than half of the nation's estuaries have moderate
to high expressions of at least one of these symptoms - an indication that eutrophication is well
developed in more than half of U.S. estuaries.1 3
6.2.2.2 Deposition of Heavy Metals
Heavy metals, including cadmium, copper, lead, chromium, mercury, nickel and zinc,
have the greatest potential for impacting forest growth.134 Investigation of trace metals near
roadways and industrial facilities indicate that a substantial load of heavy metals can accumulate
on vegetative surfaces. Copper, zinc, and nickel have been documented to cause direct toxicity
to vegetation under field conditions. Little research has been conducted on the effects associated
with mixtures of contaminants found in ambient PM. While metals typically exhibit low
solubility, limiting their bioavailability and direct toxicity, chemical transformations of metal
compounds occur in the environment, particularly in the presence of acidic or other oxidizing
species. These chemical changes influence the mobility and toxicity of metals in the
environment. Once taken up into plant tissue, a metal compound can undergo chemical changes,
exert toxic effects on the plant itself, accumulate and be passed along to herbivores or can re-
enter the soil and further cycle in the environment. Although there has been no direct evidence
of a physiological association between tree injury and heavy metal exposures, heavy metals have
been implicated because of similarities between metal deposition patterns and forest decline.
This hypothesized relationship/correlation was further explored in high elevation forests in the
northeastern U.S. These studies measured levels of a group of intracellular compounds found in
plants that bind with metals and are produced by plants as a response to sublethal concentrations
of heavy metals. These studies indicated a systematic and significant increase in concentrations
of these compounds associated with the extent of tree injury. These data strongly imply that
metal stress causes tree injury and contributes to forest decline in the northeastern United
States.135 Contamination of plant leaves by heavy metals can lead to elevated soil levels. Trace
metals absorbed into the plant frequently bind to the leaf tissue, and then are lost when the leaf
drops. As the fallen leaves decompose, the heavy metals are transferred into the soil.136'137 Upon
entering the soil environment, PM pollutants can alter ecological processes of energy flow and
6-23
-------
nutrient cycling, inhibit nutrient uptake, change ecosystem structure, and affect ecosystem
biodiversity. Many of the most important effects occur in the soil. The soil environment is one
of the most dynamic sites of biological interaction in nature. It is inhabited by microbial
communities of bacteria, fungi, and actinomycetes. These organisms are essential participants in
the nutrient cycles that make elements available for plant uptake. Changes in the soil
environment that influence the role of the bacteria and fungi in nutrient cycling determine plant
and ultimately ecosystem response.138
The environmental sources and cycling of mercury are currently of particular concern due
to the bioaccumulation and biomagnification of this metal in aquatic ecosystems and the potent
toxic nature of mercury in the forms in which is it ingested by people and other animals.
Mercury is unusual compared with other metals in that it largely partitions into the gas phase (in
elemental form), and therefore has a longer residence time in the atmosphere than a metal found
predominantly in the particle phase. This property enables mercury to travel far from the
primary source before being deposited and accumulating in the aquatic ecosystem. The major
source of mercury in the Great Lakes is from atmospheric deposition, accounting for
approximately eighty percent of the mercury in Lake Michigan.139'140 Over fifty percent of the
mercury in the Chesapeake Bay has been attributed to atmospheric deposition.141 Overall, the
National Science and Technology Council identifies atmospheric deposition as the primary
source of mercury to aquatic systems.142 Forty-four states have issued health advisories for the
consumption offish contaminated by mercury; however, most of these advisories are issued in
areas without a mercury point source.
Elevated levels of zinc and lead have been identified in streambed sediments, and these
elevated levels have been correlated with population density and motor vehicle use.143'144 Zinc
and nickel have also been identified in urban water and soils. In addition, platinum, palladium,
and rhodium, metals found in the catalysts of modern motor vehicles, have been measured at
elevated levels along roadsides.145 Plant uptake of platinum has been observed at these locations.
6.2.2.3 Deposition of Polycyclic Organic Matter
Polycyclic organic matter (POM) is a byproduct of incomplete combustion and consists
of organic compounds with more than one benzene ring and a boiling point greater than or equal
to 100 degrees centigrade.146 Polycyclic aromatic hydrocarbons (PAHs) are a class of POM that
contains compounds which are known or suspected carcinogens.
Major sources of PAHs include mobile sources. PAHs in the environment may be
present as a gas or adsorbed onto airborne parti culate matter. Since the majority of PAHs are
adsorbed onto particles less than 1.0 jim in diameter, long range transport is possible. However,
studies have shown that PAH compounds adsorbed onto diesel exhaust particulate and exposed
to ozone have half lives of 0.5 to 1.0 hours.147
Since PAHs are insoluble, the compounds generally are particle reactive and accumulate
in sediments. Atmospheric deposition of particles is believed to be the major source of PAHs to
the sediments of Lake Michigan.148'149 Analyses of PAH deposition in Chesapeake and
Galveston Bay indicate that dry deposition and gas exchange from the atmosphere to the surface
water predominate.150'151 Sediment concentrations of PAHs are high enough in some segments
6-24
-------
of Tampa Bay to pose an environmental health threat. EPA funded a study to better characterize
the sources and loading rates for PAHs into Tampa Bay.152 PAHs that enter a water body
through gas exchange likely partition into organic rich particles and can be biologically recycled,
while dry deposition of aerosols containing PAHs tend to be more resistant to biological
recycling.153 Thus, dry deposition is likely the main pathway for PAH concentrations in
sediments while gas/water exchange at the surface may lead to PAH distribution into the food
web, leading to increased health risk concerns.
Trends in PAH deposition levels are difficult to discern because of highly variable
ambient air concentrations, lack of consistency in monitoring methods, and the significant
influence of local sources on deposition levels.154 Van Metre et al. noted PAH concentrations in
urban reservoir sediments have increased by 200-300 percent over the last forty years and
correlate with increases in automobile use.155
Cousins et al. estimate that more than ninety percent of semi-volatile organic compound
(SVOC) emissions in the United Kingdom deposit on soil.156 An analysis of PAH concentrations
near a Czechoslovakian roadway indicated that concentrations were thirty times greater than
background.157
6.2.2.4 Materials Damage and Soiling
The effects of the deposition of atmospheric pollution, including ambient PM, on
materials are related to both physical damage and impaired aesthetic qualities. The deposition of
PM (especially sulfates and nitrates) can physically affect materials, adding to the effects of
natural weathering processes, by potentially promoting or accelerating the corrosion of metals,
by degrading paints, and by deteriorating building materials such as concrete and limestone.
Only chemically active fine particles or hygroscopic coarse particles contribute to these physical
effects. In addition, the deposition of ambient PM can reduce the aesthetic appeal of buildings
and culturally important articles through soiling. Particles consisting primarily of carbonaceous
compounds cause soiling of commonly used building materials and culturally important items
such as statues and works of art.
6.2.3 Plant and Ecosystem Effects of Ozone
There are a number of environmental or public welfare effects associated with the
presence of ozone in the ambient air.158 In this section we discuss the impact of ozone on plants,
including trees, agronomic crops and urban ornamentals.
The Air Quality Criteria Document for Ozone and related Photochemical Oxidants notes
that, "ozone affects vegetation throughout the United States, impairing crops, native vegetation,
and ecosystems more than any other air pollutant."159 Like carbon dioxide (CO2) and other
gaseous substances, ozone enters plant tissues primarily through apertures (stomata) in leaves in
a process called "uptake."160 Once sufficient levels of ozone (a highly reactive substance), or its
reaction products, reaches the interior of plant cells, it can inhibit or damage essential cellular
components and functions, including enzyme activities, lipids, and cellular membranes,
disrupting the plant's osmotic (i.e., water) balance and energy utilization patterns.161'162 If
enough tissue becomes damaged from these effects, a plant's capacity to fix carbon to form
6-25
-------
carbohydrates, which are the primary form of energy used by plants, is reduced,163 while plant
respiration increases. With fewer resources available, the plant reallocates existing resources
away from root growth and storage, above ground growth or yield, and reproductive processes,
toward leaf repair and maintenance, leading to reduced growth and/or reproduction. Studies
have shown that plants stressed in these ways may exhibit a general loss of vigor, which can lead
to secondary impacts that modify plants' responses to other environmental factors. Specifically,
plants may become more sensitive to other air pollutants, more susceptible to disease, insect
attack, harsh weather (e.g., drought, frost) and other environmental stresses. Furthermore, there
is evidence that ozone can interfere with the formation of mycorrhiza, essential symbiotic fungi
associated with the roots of most terrestrial plants, by reducing the amount of carbon available
for transfer from the host to the symbiont.16 '165
This ozone damage may or may not be accompanied by visible injury on leaves, and
likewise, visible foliar injury may or may not be a symptom of the other types of plant damage
described above. When visible injury is present, it is commonly manifested as chlorotic or
necrotic spots, and/or increased leaf senescence (accelerated leaf aging). Because ozone damage
can consist of visible injury to leaves, it can also reduce the aesthetic value of ornamental
vegetation and trees in urban landscapes, and negatively affects scenic vistas in protected natural
areas.
Ozone can produce both acute and chronic injury in sensitive species depending on the
concentration level and the duration of the exposure. Ozone effects also tend to accumulate over
the growing season of the plant, so that even lower concentrations experienced for a longer
duration have the potential to create chronic stress on sensitive vegetation. Not all plants,
however, are equally sensitive to ozone. Much of the variation in sensitivity between individual
plants or whole species is related to the plant's ability to regulate the extent of gas exchange via
leaf stomata (e.g., avoidance of ozone uptake through closure of stomata)166'167'168 Other
resistance mechanisms may involve the intercellular production of detoxifying substances.
Several biochemical substances capable of detoxifying ozone have been reported to occur in
plants, including the antioxidants ascorbate and glutathione. After injuries have occurred, plants
may be capable of repairing the damage to a limited extent.169
Because of the differing sensitivities among plants to ozone, ozone pollution can also
exert a selective pressure that leads to changes in plant community composition. Given the range
of plant sensitivities and the fact that numerous other environmental factors modify plant uptake
and response to ozone, it is not possible to identify threshold values above which ozone is
consistently toxic for all plants. The next few paragraphs present additional information on
ozone damage to trees, ecosystems, agronomic crops and urban ornamentals.
Assessing the impact of ground-level ozone on forests in the United States involves
understanding the risks to sensitive tree species from ambient ozone concentrations and
accounting for the prevalence of those species within the forest. As a way to quantify the risks to
particular plants from ground-level ozone, scientists have developed ozone-exposure/tree-
response functions by exposing tree seedlings to different ozone levels and measuring reductions
in growth as "biomass loss." Typically, seedlings are used because they are easy to manipulate
and measure their growth loss from ozone pollution. The mechanisms of susceptibility to ozone
6-26
-------
within the leaves of seedlings and mature trees are identical, though the magnitude of the effect
may be higher or lower depending on the tree species.170
Some of the common tree species in the United States that are sensitive to ozone are
black cherry (Primus serotina\ tulip-poplar (Liriodendron tulipifera), and eastern white pine
(Pinus strobus). Ozone-exposure/tree-response functions have been developed for each of these
tree species, as well as for aspen (Populus tremuliodes), and ponderosa pine (Pinusponderosd).
Other common tree species, such as oak (Quercus spp.j and hickory (Carya spp.), are not nearly
as sensitive to ozone. Consequently, with knowledge of the distribution of sensitive species and
the level of ozone at particular locations, it is possible to estimate a "biomass loss" for each
species across their range.
Ozone also has been conclusively shown to cause discernible injury to forest trees.171'172
In terms of forest productivity and ecosystem diversity, ozone may be the pollutant with the
greatest potential for regional-scale forest impacts. Studies have demonstrated repeatedly that
ozone concentrations commonly observed in polluted areas can have substantial impacts on plant
function.173'174
Because plants are at the base of the food web in many ecosystems, changes to the plant
community can affect associated organisms and ecosystems (including the suitability of habitats
that support threatened or endangered species and below ground organisms living in the root
zone). Ozone impacts at the community and ecosystem level vary widely depending upon
numerous factors, including concentration and temporal variation of tropospheric ozone, species
composition, soil properties and climatic factors.1? In most instances, responses to chronic or
recurrent exposure in forested ecosystems are subtle and not observable for many years. These
injuries can cause stand-level forest decline in sensitive ecosystems.176'177'178 It is not yet
possible to predict ecosystem responses to ozone with much certainty; however, considerable
knowledge of potential ecosystem responses has been acquired through long-term observations
in highly damaged forests in the United States.
Air pollution can have noteworthy cumulative impacts on forested ecosystems by
affecting regeneration, productivity, and species composition.179 In the U.S., ozone in the lower
atmosphere is one of the pollutants of primary concern. Ozone injury to forest plants can be
diagnosed by examination of plant leaves. Foliar injury is usually the first visible sign of injury
to plants from ozone exposure and indicates impaired physiological processes in the leaves.180
However, not all impaired plants will exhibit visible symptoms.
Laboratory and field experiments have also shown reductions in yields for agronomic
crops exposed to ozone, including vegetables (e.g., lettuce) and field crops (e.g., cotton and
wheat). The most extensive field experiments, conducted under the National Crop Loss
Assessment Network (NCLAN) examined 15 species and numerous cultivars. The NCLAN
results show that "several economically important crop species are sensitive to ozone levels
I O 1
typical of those found in the United States." In addition, economic studies have shown
reduced economic benefits as a result of predicted reductions in crop yields associated with
observed ozone levels.182'183'184
6-27
-------
Urban ornamentals represent an additional vegetation category likely to experience some
degree of negative effects associated with exposure to ambient ozone levels. It is estimated that
more than $20 billion (1990 dollars) are spent annually on landscaping using ornamentals, both
by private property owners/tenants and by governmental units responsible for public areas.185
This is therefore a potentially costly environmental effect. However, in the absence of adequate
exposure-response functions and economic damage functions for the potential range of effects
relevant to these types of vegetation, no direct quantitative analysis has been conducted.
6.2.3.1 Recent Ozone Visible Foliar Injury Data for the U. S
In the U.S. the national-level visible foliar injury indicator is based on data from the U.S.
Department of Agriculture (USDA) Forest Service Forest Inventory and Analysis (FIA)
program. As part of its Phase 3 program, formerly known as Forest Health Monitoring, FIA
examines ozone injury to ozone-sensitive plant species at ground monitoring sites in forest land
across the country. For this indicator, forest land does not include woodlots and urban trees.
1 %.f\ 1 R7
Sites are selected using a systematic sampling grid, based on a global sampling design. ' At
each site that has at least 30 individual plants of at least three ozone-sensitive species and enough
open space to ensure that sensitive plants are not protected from ozone exposure by the forest
canopy, FIA looks for damage on the foliage of ozone-sensitive forest plant species. Because
ozone injury is cumulative over the course of the growing season, examinations are conducted in
July and August, when ozone injury is typically highest. Monitoring of ozone injury to plants by
the USD A Forest Service has expanded over time from monitoring sites in 10 states in 1994 to
nearly 1,000 monitoring sites in 41 states in 2002.
There is considerable regional variation in ozone-related visible foliar injury to sensitive
plants in the U.S. The U.S. EPA has developed an environmental indicator based on data from
the USDA FIA program which examines ozone injury to ozone-sensitive plant species at ground
monitoring sites in forest land across the country. The data underlying the indicator in Figure
6-3 are based on averages of all observations collected in 2002, the latest year for which data are
publicly available at the time the study was conducted, and is broken down by U.S. EPA Region.
Ozone damage to forest plants is classified using a subjective five-category biosite index based
on expert opinion, but designed to be equivalent from site to site. Ranges of biosite values
translate to no injury, low or moderate foliar injury (visible foliar injury to highly sensitive or
moderately sensitive plants, respectively), and high or severe foliar injury, which would be
expected to result in tree-level or ecosystem-level responses, respectively.188'189
The highest percentages of observed high and severe foliar injury, those which are most
likely to be associated with tree or ecosystem-level responses, are primarily found in the Mid-
Atlantic and Southeast regions. In EPA Region 3 (which comprises the States of Pennsylvania,
West Virginia, Virginia, Delaware, Maryland and Washington D.C.), 12 percent of ozone-
sensitive plants showed signs of high or severe foliar damage, and in Region 2 (States of New
York, New Jersey), and Region 4 (States of North Carolina, South Carolina, Kentucky,
Tennessee, Georgia, Florida, Alabama, and Mississippi) the values were 10 and 7 percent,
respectively. The sum of high and severe ozone injury ranged from 2 to 4 percent in EPA
Region 1 (the six New England States), Region 7 (States of Missouri, Iowa, Nebraska and
Kansas), and Region 9 (States of California, Nevada, Hawaii and Arizona). The percentage of
sites showing some ozone damage was about 45 percent in each of these EPA Regions.
6-28
-------
R*gion1
(54 lifts)
|42sit»s)
(111 MM)
R*gion4
(227s1*s)
R»jion5
IH «•
(59 srtes)
R*gion7
(63s«s)
R*gion8
(72 Silas)
R*gion9
(80 SMS)
RajtalO
Digm ol injury:
Non« Low Modtntt Higti | S*v*ra
Ptrctnt ol monitoring litti n tacti caltgory:
68! 167 " '\-\'' "
619 | 214 7l||7.1 :-i
55.» 180 144 I [7.2J-!i
753 'O'(7s|t£;
756 | 183 J6^
949 H&1
85.7 EiUie
1000
763 12SJ88[^
1MO
kxstad in 41 states ,~
:Totals may not add to 100% dm to Q
round tnQ
Djtt umn: USM Forts! Semct.
2006 ft®
ab
Figure 6-3 Ozone Injury to Forest Plants in U.S. by EPA Regions, 2002
6.2.3.2 Indicator Limitations
The categories for the biosite index are subjective and may not necessarily be directly
related to biomass loss or physiological damage to plants in a particular area. Ozone may have
other adverse impacts on plants (e.g., reduced productivity) that do not show signs of visible
foliar injury.190 The presence of diagnostic visible ozone injury on indicator plants does provide
evidence that ozone is having an impact in an area. However, absence of ozone injury in an area
does not necessarily mean that there is no impact from ozone exposure.
Field and laboratory studies were reviewed to identify the forest plant species in each
region that are sensitive to ozone air pollution and exhibit diagnostic injury. Other forest plant
species, or even genetic variants of the same species, may not show symptoms at ozone levels
that cause effects on the selected ozone-sensitive species.
Because species distributions vary regionally, different ozone-sensitive plant species
were examined in different parts of the country. These target species could vary with respect to
ozone sensitivity, which might account for some of the apparent differences in ozone injury
among regions of the U.S. Ozone damage to foliage may be reduced under conditions of low
soil moisture, but most of the variability in the index (70 percent) was explained by ozone
191
concentration.
6-29
-------
Though FIA has extensive spatial coverage based on a robust sample design, not all
forested areas in the U.S. are monitored for ozone injury. Even though the biosite data have been
collected over multiple years, most biosites were not monitored over the entire period, so these
data cannot provide more than a baseline for future trends.
6.2.4 Environmental Effects of Air Toxics
Emissions from producing, transporting and combusting fuel contribute to ambient levels
of pollutants that contribute to adverse effects on vegetation. Volatile organic compounds
(VOCs), some of which are considered air toxics, have long been suspected to play a role in
vegetation damage.192 In laboratory experiments, a wide range of tolerance to VOCs has been
observed.193 Decreases in harvested seed pod weight have been reported for the more sensitive
plants, and some studies have reported effects on seed germination, flowering and fruit ripening.
Effects of individual VOCs or their role in conjunction with other stressors (e.g., acidification,
drought, temperature extremes) have not been well studied. In a recent study of a mixture of
VOCs including ethanol and toluene on herbaceous plants, significant effects on seed production,
leaf water content and photosynthetic efficiency were reported for some plant species. 94
Research suggests an adverse impact of vehicle exhaust on plants, which has in some
cases been attributed to aromatic compounds and in other cases to nitrogen oxides.195'196'197 The
impacts of VOCs on plant reproduction may have long-term implications for biodiversity and
survival of native species near major roadways. Most of the studies of the impacts of VOCs on
vegetation have focused on short-term exposure and few studies have focused on long-term
effects of VOCs on vegetation and the potential for metabolites of these compounds to affect
herbivores or insects.
6-30
-------
References
1 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009.
2 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009. Section 2.3.1.1.
3 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009. Section 2.3.1.2.
4 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009. Section 2.3.4.
5 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009. Table 2-6.
6 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009. Section 2.3.5.1.
7 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009. Table 2-6.
8 U.S. EPA. (2006;. Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). EPA/600/R-
05/004aF-cF. Washington, DC: U.S. EPA.
9 U.S. EPA. (200?;. Review of the National Ambient Air Quality Standards for Ozone: Policy Assessment of
Scientific and Technical Information, OAQPS Staff Paper. EPA-452/R-07-003. Washington, DC, U.S. EPA. Docket
EP A-HQ-O AR-2011-0135
10National Research Council (NRC), (2008). Estimating Mortality Risk Reduction and Economic Benefits from
Controlling Ozone Air Pollution. The National Academies Press: Washington, D.C.
11 Bates, D.V., Baker-Anderson, M., Sizto, R. (1990). Asthma attack periodicity: a study of hospital emergency
visits in Vancouver. Environ. Res., 57,51-70.
12 Thurston, G.D., Ito, K., Kinney, P.L., Lippmann, M. (1992). A multi-year study of air pollution and respiratory
hospital admissions in three New York State metropolitan areas: results for 1988 and 1989 summers. J. Exposure
Anal. Environ. Epidemiol, 2,429-450.
13 Thurston, G.D., Ito, K., Hayes, C.G., Bates, D.V., Lippmann, M. (1994). Respiratory hospital admissions and
summertime haze air pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ. Res., 65,
271-290.
14Lipfert, F.W., Hammerstrom, T. (1992). Temporal patterns in air pollution and hospital admissions. Environ. Res.,
59,374-399.
15Burnett, R.T., Dales, R.E., Raizenne, M.E., Krewski, D., Summers, P.W., Roberts, G.R., Raad-Young, M.,
Dann,T., Brook, J. (1994). Effects of low ambient levels of ozone and sulfates on the frequency of respiratory
admissions to Ontario hospitals. Environ. Res., 65, 172-194.
16U.S. EPA. (2006/ Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). EPA/600/R-
05/004aF-cF. Washington, DC: U.S. EPA. Docket EP A-HQ-O AR-2011-0135
17U.S. EPA. (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). EPA/600/R-
05/004aF-cF. Washington, DC: U.S. EPA.
18Devlin, R. B., McDonnell, W. F., Mann, R., Becker, S., House, D. E., Schreinemachers, D., Koren, H. S. (1991).
Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular and biochemical changes in the lung.
Am. J. Respir. CellMol. Biol., 4, 72-81.
6-31
-------
19Koren, H. S., Devlin, R. B., Becker, S., Perez, R., McDonnell, W. F. (1991). Time-dependent changes of markers
associated with inflammation in the lungs of humans exposed to ambient levels of ozone. Toxicol. Pathol, 19, 406-
411.
20Koren, H. S., Devlin, R. B., Graham, D. E., Mann, R., McGee, M. P., Horstman, D. H., Kozumbo, W. I, Becker,
S., House, D. E., McDonnell, W. F., Bromberg, P. A. (1989). Ozone-induced inflammation in the lower airways of
human subjects. Am. Rev. Respir. Dis., 39, 407-415.
21 Schelegle, E.S., Siefkin, A.D., McDonald, RJ. (1991). Time course of ozone-induced neutrophilia in normal
humans. Am. Rev. Respir. Dis., 143,1353-1358.
22U.S. EPA. (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants. EPA600-P-93-004aF.
Washington. D.C.: U.S. EPA.
23Hodgkin, J.E., Abbey, D.E., Euler, G.L., Magie, A.R. (1984). COPD prevalence in nonsmokers in high and low
photochemical air pollution areas. Chest, 86, 830-838.
24Euler, G.L., Abbey, D.E., Hodgkin, J.E., Magie, A.R. (1988). Chronic obstructive pulmonary disease symptom
effects of long-term cumulative exposure to ambient levels of total oxidants and nitrogen dioxide in California
Seventh-day Adventist residents. Arch. Environ. Health, 43, 279-285.
25 Abbey, D.E., Petersen, F., Mills, P.K., Beeson, W.L. (1993). Long-term ambient concentrations of total
suspended particulates, ozone, and sulfur dioxide and respiratory symptoms in a nonsmoking population. Arch.
Environ. Health, 48, 33-46.
26 U.S. EPA. (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). EPA/600/R-
05/004aF-cF. Washington, DC: U.S. EPA.
27U.S. EPA. (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). EPA/600/R-
05/004aF-cF. Washington, DC: U.S. EPA.
28 Avol, E.L., Trim, S. C., Little, D.E., Spier, C.E., Smith, M. N., Peng, R.-C., Linn, W.S., Hackney, J.D., Gross,
K.B., D'Arcy, J.B., Gibbons, D., Higgins, I.T.T. (1990 June). Ozone exposure and lung function in children
attending a southern California summer camp. Paper no. 90-150.3. Paper presented at the 83rd annual meeting and
exhibition of the Air & Waste Management Association, Pittsburgh, PA.
29Higgins, I. T.T., D'Arcy, J. B., Gibbons, D. I., Avol, E. L., Gross, K.B. (1990). Effect of exposures to ambient
ozone on ventilatory lung function in children. Am. Rev. Respir. Dis., 141, 1136-1146.
30Raizenne, M.E., Burnett, R.T., Stern, B., Franklin, C.A., Spengler, J.D. (1989). Acute lung function responses to
ambient acid aerosol exposures in children. Environ. Health Perspect, 79,179-185.
31 Raizenne, M.; Stern, B.; Burnett, R.; Spengler, J. (1987 June). Acute respiratory function and transported air
pollutants: observational studies. Paper no. 87-32.6. Paper presented at the 80th annual meeting of the Air Pollution
Control Association, New York, NY.
32 Spektor, D. M., Lippmann, M. (1991). Health effects of ambient ozone on healthy children at a summer camp. In:
Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment: papers from an
international conference', March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management Association;
pp. 83-89. (A&WMA transaction series no. TR-19).
33 Spektor, D. M., Thurston, G.D., Mao, J., He, D., Hayes, C., Lippmann, M. (1991). Effects of single- and multiday
ozone exposures on respiratory function in active normal children. Environ. Res, 55,107-122.
34 Spektor, D. M., Lippman, M., Lioy, P. J., Thurston, G. D., Citak, K., James, D. J., Bock, N., Speizer, F. E., Hayes,
C. (1988). Effects of ambient ozone on respiratory function in active, normal children. Am. Rev. Respir. Dis., 137,
313-320.
35U.S. EPA. (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). EPA/600/R-
05/004aF-cF. Washington, DC: U.S. EPA.
36Hazucha, M. J., Folinsbee, L. J., Seal, E., Jr. (1992). Effects of steady-state and variable ozone concentration
profiles on pulmonary function. Am. Rev. Respir. Dis., 146, 1487-1493.
6-32
-------
37Horstman, D.H., Ball, B.A., Folinsbee, L.J., Brown, I, Gerrity, T. (1995). Comparison of pulmonary responses of
asthmatic and nonasthmatic subjects performing light exercise while exposed to a low level of ozone. Toxicol. Ind.
Health., 11(4), 369-85.
38Horstman, D.H.,; Folinsbee, L.J., Ives, P.J., Abdul-Salaam, S., McDonnell, W.F. (1990). Ozone concentration and
pulmonary response relationships for 6.6-hour exposures with five hours of moderate exercise to 0.08, 0.10, and
0.12 ppm.Am. Rev. Respir. Dis., 142, 1158-1163.
39 U.S. EPA (2008). Integrated Science Assessment (ISA) for Sulfur Oxides - Health Criteria (Final Report).
EPA/600/R-08/047F. Washington, DC,: U.S.EPA. Retrieved onMarch 19, 2009 from
http://cfpub.epa. gov/ncea/cfm/recordisplav. cfm?deid= 198843.
40 U.S. EPA (2008). Integrated Science Assessment for Oxides of Nitrogen -Health Criteria (Final Report).
EPA/600/R-08/071. Washington, DC,: U.S.EPA. Retrieved onMarch 19, 2009 from
http ://cfpub .epa. gov/ncea/cfm/recordisplav. cfm?deid= 194645.
41 U.S. EPA, (2010). Integrated Science Assessment for Carbon Monoxide (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-09/019F, 2010.
http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=218686.
42U.S. EPA. (2000). Integrated Risk Information System File for Benzene. This material is available electronically
at: http://www.epa.gov/iris/subst/0276.htm.
43 International Agency for Research on Cancer. (1982). IARC monographs on the evaluation of carcinogenic risk of
chemicals to humans, Volume 29, Some industrial chemicals and dyestuffs, International Agency for Research on
Cancer, World Health Organization, Lyon, France 1982.
44 Irons, R.D.; Stillman, W.S.; Colagiovanni, D.B.; Henry, V.A. (1992). Synergistic action of the benzene metabolite
hydroquinone on myelopoietic stimulating activity of granulocyte/macrophage colony-stimulating factor in vitro,
Proc. Natl. Acad. Sci. 89:3691-3695.
45 U.S. EPA. (2000). Integrated Risk Information System File for Benzene. This material is available electronically
at: http://www.epa.gov/iris/subst/0276.htm.
46 International Agency for Research on Cancer (IARC). 1987. Monographs on the evaluation of carcinogenic risk
of chemicals to humans, Volume 29, Supplement 7, Some industrial chemicals and dyestuffs, World Health
Organization, Lyon, France.
47
U.S. Department of Health and Human Services National Toxicology Program. (2011). 12th Report on
Carcinogens available at: http://ntp.niehs.nih.gov/?objectid=03C9AF75-ElBF-FF40-DBA9EC0928DF8B15.
48Aksoy, M. (1989). Hematotoxicity and carcinogenicity of benzene. Environ. Health Perspect. 82:193-197.
EP A-HQ-O AR-2011-0135
49 Goldstein, B.D. (1988). Benzene toxicity. Occupational medicine. State of the Art Reviews. 3:541-554.
50Rothman, N., G.L. Li, M. Dosemeci, W.E. Bechtold, G.E. Marti, Y.Z. Wang, M. Linet, L.Q. Xi, W. Lu, M.T.
Smith, N. Titenko-Holland, L.P. Zhang, W. Blot, S.N. Yin, and R.B. Hayes. (1996). Hematotoxicity among Chinese
workers heavily exposed to benzene. Am. J. Ind. Med. 29: 236-246.
51 U.S. EPA (2002). Toxicological Review of Benzene (Noncancer Effects). Environmental Protection Agency,
Integrated Risk Information System (IRIS), Research and Development, National Center for Environmental
Assessment, Washington DC. This material is available electronically at http://www.epa.gov/iris/subst/0276.htm.
52Qu, O.; Shore, R.; Li, G.; Jin, X.; Chen, C.L.; Cohen, B.; Melikian, A.; Eastmond, D.; Rappaport, S.; Li, H.; Rupa,
D.; Suramaya, R.; Songnian, W.; Huifant, Y.; Meng, M.; Winnik, M.;Kwok, E.; Li, Y.; Mu, R.; Xu, B.; Zhang,
X.; Li, K. (2003). HEI Report 115, Validation & Evaluation of Biomarkers in Workers Exposed to Benzene in
China.
53 Qu, Q., R. Shore, G. Li, X. Jin, L.C. Chen, B. Cohen, et al. (2002). Hematological changes among Chinese
workers with a broad range of benzene exposures. Am. J. Industr. Med. 42: 275-285.
54Lan, Qing, Zhang, L., Li, G., Vermeulen, R., et al. (2004). Hematotoxically in Workers Exposed to Low Levels of
Benzene. Science 306: 1774-1776.
6-33
-------
55 Turtletaub, K.W. and Mani, C. (2003). Benzene metabolism in rodents at doses relevant to human exposure from
Urban Air. Research Reports Health Effect Inst. Report No. 113.
56 U.S. Agency for Toxic Substances and Disease Registry (ATSDR). (2007). Toxicological profile for benzene.
Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service.
http://www.atsdr.cdc.gov/ToxProfiles/tp3.pdf.
57 U.S. EPA. (2002). Health Assessment of 1,3-Butadiene. Office of Research and Development, National Center
for Environmental Assessment, Washington Office, Washington, DC. Report No. EPA600-P-98-001F. This
document is available electronically at http://www.epa.gov/iris/supdocs/buta-sup.pdf.
58U.S. EPA. (2002) "Full IRIS Summary for 1,3-butadiene (CASRN 106-99-0)" Environmental Protection Agency,
Integrated Risk Information System (IRIS), Research and Development, National Center for Environmental
Assessment, Washington, DC http://www.epa.gov/iris/subst/0139.htm.
59 International Agency for Research on Cancer (IARC). (1999). Monographs on the evaluation of carcinogenic risk
of chemicals to humans, Volume 71, Re-evaluation of some organic chemicals, hydrazine and hydrogen peroxide
and Volume 97 (in preparation), World Health Organization, Lyon, France.
60 International Agency for Research on Cancer (IARC). (2008). Monographs on the evaluation of carcinogenic risk
of chemicals to humans, 1,3-Butadiene, Ethylene Oxide and Vinyl Halides (Vinyl Fluoride, Vinyl Chloride and
Vinyl Bromide) Volume 97, World Health Organization, Lyon, France.
61 NTP. (2011). Report on Carcinogens, Twelfth Edition. Research Triangle Park, NC: U.S. Department of Health
and Human Services, Public Health Service, National Toxicology Program. 499 pp.
62 U.S. EPA. (2002). "Full IRIS Summary for 1,3-butadiene (CASRN 106-99-0)" Environmental Protection Agency,
Integrated Risk Information System (IRIS), Research and Development, National Center for Environmental
Assessment, Washington, DC http://www.epa.gov/iris/subst/0139.htm.
63Bevan, C.; Stadler, J.C.; Elliot, G.S.; et al. (1996). Subchronic toxicity of 4-vinylcyclohexene in rats and mice by
inhalation. Fundam. Appl. Toxicol. 32:1-10.
64 EPA. Integrated Risk Information System. Formaldehyde (CASRN 50-00-0)
http://www.epa.gov/iris/subst/0419/htm
65 National Toxicology Program, U.S. Department of Health and Human Services (HHS), 12th Report on
Carcinogens, June 10, 2011
66 IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 88 (2006): Formaldehyde, 2-
Butoxyethanol and l-tert-Butoxypropan-2-ol
67 IARC Mongraphs on the Evaluation of Carcinogenic Risks to Humans Volume 100F (2012): Formaldehyde
68 Hauptmann, M..; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Blair, A. 2003. Mortality from lymphohematopoetic
malignancies among workers in formaldehyde industries. Journal of the National Cancer Institute 95: 1615-1623.
69Hauptmann, M..; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Blair, A. 2004. Mortality from solid cancers among
workers in formaldehyde industries. American Journal of Epidemiology 159: 1117-1130.
70Beane Freeman, L. E.; Blair, A.; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Hoover, R. N.; Hauptmann, M. 2009.
Mortality from lymphohematopoietic malignancies among workers in formaldehyde industries: The National Cancer
Institute cohort. J. National Cancer Inst. 101: 751-761.
71 Pinkerton, L. E. 2004. Mortality among a cohort of garment workers exposed to formaldehyde: an update.
Occup. Environ. Med. 61: 193-200..
72 Coggon, D, EC Harris, J Poole, KT Palmer. 2003. Extended follow-up of a cohort of British chemical workers
exposed to formaldehyde. J National Cancer Inst. 95:1608-1615.
73 Hauptmann, M,; Stewart P. A.; Lubin J. H.; Beane Freeman, L. E.; Hornung, R. W.; Herrick, R. F.; Hoover, R. N.;
Fraumeni, J. F.; Hayes, R. B. 2009. Mortality from lymphohematopoietic malignancies and brain cancer among
embalmers exposed to formaldehyde. Journal of the National Cancer Institute 101:1696-1708.
6-34
-------
74 ATSDR. 1999. lexicological Profile for Formaldehyde, U.S. Department of Health and Human Services (HHS),
July 1999.
75 ATSDR. 2010. Addendum to theToxicological Profile for Formaldehyde. U.S. Department of Health and Human
Services (HHS), October 2010.
76IPCS. 2002. Concise International Chemical Assessment Document 40. Formaldehyde. World Health
Organization.
77 EPA (U.S. Environmental Protection Agency). 2010. lexicological Review of Formaldehyde (CAS No. 50-00-0)
- Inhalation Assessment: In Support of Summary Information on the Integrated Risk Information System (IRIS).
External Review Draft. EPA/635/R-10/002A. U.S. Environmental Protection Agency, Washington DC [online].
Available: http://cfpub.epa.gov/ncea/irs_drats/recordisplay.cfm?deid=223614
78 NRC (National Research Council). 2011. Review of the Environmental Protection Agency's Draft IRIS
Assessment of Formaldehyde. Washington DC: National Academies Press.
http ://books. nap. edu/openbook.php?record_id= 13142
79U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. Research and Development,
National Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/iris/subst/0290.htm.
80U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. This material is available
electronically at http://www.epa.gov/iris/subst/0290.htm.
81 NTP. (2011). Report on Carcinogens, Twelfth Edition. Research Triangle Park, NC: U.S. Department of Health
and Human Services, Public Health Service, National Toxicology Program. 499 pp.
82 International Agency for Research on Cancer (IARC). (1999). Re-evaluation of some organic chemicals,
hydrazine, and hydrogen peroxide. IARC Monographs on the Evaluation of Carcinogenic Risk of Chemical to
Humans, Vol 71. Lyon, France.
83U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. This material is available
electronically at http://www.epa.gov/iris/subst/0290.htm.
84 U.S. EPA. (2003). Integrated Risk Information System File of Acrolein. Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/iris/subst/0364.htm.
85 Appleman, L.M., R.A. Woutersen, and V.J. Feron. (1982). Inhalation toxicity of acetaldehyde in rats. I. Acute and
subacute studies. Toxicology. 23: 293-297.
86 Myou, S.; Fujimura, M; Nishi K.; Ohka, T.; and Matsuda, T. (1993). Aerosolized acetaldehyde induces
histamine-mediated bronchoconstriction in asthmatics. Am. Rev. Respir.Dis. 148(4 Pt 1): 940-943.
87 U.S. EPA. (2003). Integrated Risk Information System File of Acrolein. Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available at
http://www.epa.gov/iris/subst/0364.htm.
88 International Agency for Research on Cancer (IARC). (1995). Monographs on the evaluation of carcinogenic risk
of chemicals to humans, Volume 63. Dry cleaning, some chlorinated solvents and other industrial chemicals, World
Health Organization, Lyon, France.
89 U.S. EPA. (2003). Integrated Risk Information System File of Acrolein. Office of Research and Development,
National Center for Environmental Assessment, Washington, DC. This material is available at
http://www.epa.gov/iris/subst/0364.htm.
90 U.S. EPA. (2003). Integrated Risk Information System File of Acrolein. Office of Research and Development,
National Center for Environmental Assessment, Washington, DC. This material is available at
http://www.epa.gov/iris/subst/0364.htm.
6-35
-------
91 U.S. EPA. (2003). Toxicological review of acrolein in support of summary information on Integrated Risk
Information System (IRIS) National Center for Environmental Assessment, Washington, DC. EPA/635/R-03/003. p.
10. Available online at: http://www.epa.gov/ncea/iris/toxreviews/0364tr.pdf.
92 U.S. EPA. (2003). Toxicological review of acrolein in support of summary information on Integrated Risk
Information System (IRIS) National Center for Environmental Assessment, Washington, DC. EPA/635/R-03/003.
Available online at: http://www.epa.gov/ncea/iris/toxreviews/0364tr.pdf.
93 Morris JB, Symanowicz PT, Olsen JE, et al. (2003). Immediate sensory nerve-mediated respiratory responses to
irritants in healthy and allergic airway-diseased mice. J Appl Physiol 94(4): 1563-1571.
94 U.S. EPA. (2009). Graphical Arrays of Chemical-Specific Health Effect Reference Values for Inhalation
Exposures (Final Report). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-09/061, 2009.
http://cfpub.epa. gov/ncea/cfm/recordisplay.cfm?deid=211003
95U.S. EPA. (2006). Air quality criteria for ozone and related photochemical oxidants (Ozone CD). Research
Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-05/004aF-cF.3v. page 5-
78 Available at http://cfpub.epa.gov/ncea/.
96U.S. EPA Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. Environmental
Protection Agency, Washington, D.C., EPA 600/R-05/004aF-cF, 2006. page 5-63. This document is available in.
This document may be accessed electronically at: http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_cd.html
97 U.S. EPA Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. Environmental
Protection Agency, Washington, D.C., EPA 600/R-05/004aF-cF, 2006. page 5-78. This document is available in.
This document may be accessed electronically at: http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_cd.html
98 Agency for Toxic Substances and Disease Registry (ATSDR). (1995). Toxicological profile for Polycyclic
Aromatic Hydrocarbons (PAHs). Atlanta, GA: U.S. Department of Health and Human Services, Public Health
Service. Available electronically athttp://www.atsdr.cdc.gov/ToxProfiles/TP.asp?id=122&tid=25.
99 U.S. EPA (2002). Health Assessment Document for Diesel Engine Exhaust. EPA/600/8-90/057F Office of
Research and Development, Washington DC. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=29060.
100 International Agency for Research on Cancer (IARC). (2012). Monographs on the Evaluation of the
Carcinogenic Risk of Chemicals for Humans, Chemical Agents and Related Occupations. Vol. 100F. Lyon, France.
101 U.S. EPA (1997). Integrated Risk Information System File of indeno(l,2,3-cd)pyrene. Research and
Development, National Center for Environmental Assessment, Washington, DC. This material is available
electronically at http://www.epa.gov/ncea/iris/subst/0457.htm.
102 Perera, P.P.; Rauh, V.; Tsai, W-Y.; et al. (2002). Effect of transplacental exposure to environmental pollutants on
birth outcomes in a multiethnic population. Environ Health Perspect. 111: 201 -205.
103 Perera, P.P.; Rauh, V; Whyatt, R.M.; Tsai, W.Y.; Tang, D.; Diaz, D.; Hoepner, L.; Barr, D.; Tu, Y.H.; Camann,
D.; Kinney, P. (2006). Effect of prenatal exposure to airborne polycyclic aromatic hydrocarbons on
neurodevelopment in the first 3 years of life among inner-city children. Environ Health Perspect 114: 1287-1292.
104 U. S. EPA. 1998. Toxicological Review of Naphthalene (Reassessment of the Inhalation Cancer Risk),
Environmental Protection Agency, Integrated Risk Information System, Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/iris/subst/0436.htm.
105 U. S. EPA. 1998. Toxicological Review of Naphthalene (Reassessment of the Inhalation Cancer Risk),
Environmental Protection Agency, Integrated Risk Information System, Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/iris/subst/0436.htm.
106U. S. EPA. (1998). Toxicological Review of Naphthalene (Reassessment of the Inhalation Cancer Risk),
Environmental Protection Agency, Integrated Risk Information System, Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/iris/subst/0436.htm.
6-36
-------
107 Oak Ridge Institute for Science and Education. (2004). External Peer Review for the IRIS Reassessment of the
Inhalation Carcinogenicity of Naphthalene. August 2004.
http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=84403.
108 NTp (2011). Report on Carcinogens, Twelfth Edition. Research Triangle Park, NC: U.S. Department of Health
and Human Services, Public Health Service, National Toxicology Program. 499 pp.
109 International Agency for Research on Cancer (IARC). (2002). Monographs on the Evaluation of the
Carcinogenic Risk of Chemicals for Humans. Vol. 82. Lyon, France.
110U. S. EPA. (1998). Toxicological Review of Naphthalene, Environmental Protection Agency, Integrated Risk
Information System, Research and Development, National Center for Environmental Assessment, Washington, DC.
This material is available electronically at http://www.epa.gov/iris/subst/0436.htm
111 U.S. EPA. (1998). Toxicological Review of Naphthalene. Environmental Protection Agency, Integrated Risk
Information System (IRIS), Research and Development, National Center for Environmental Assessment,
Washington, DC http://www.epa.gov/iris/subst/0436.htm.
112 Health Effects Institute Panel on the Health Effects of Traffic-Related Air Pollution. (2010). Traffic-related air
pollution: a critical review of the literature on emissions, exposure, and health effects. HEI Special Report 17.
[Online at http://www.healtheffects.org]
113 Bureau of Labor Statistics. (2009). American Time Use Survey. [Online at http://www.bls.gov/tus]
114 Bureau of Transportation Statistics. (2003). Highlights of the 2001 National Household Travel Survey. Report
BTS03-05. [Online at http://www.bts.gov]
115 Baja, E.S.; Schwartz, J.D.; Wellenius, G.A.; Coull, B.A.; Zanobetti, A.; Vokonas, P.S.; Suh, H.H. (2010).
Traffic-related air pollution and QT interval: modification by diabetes, obesity, and oxidate stress gene
polymorphisms in the Normative Aging Study. Environ Health Perspect 118: 840-846. doi:10.1289/ehp.0901396
116 Zanobettia, A.; Stone, P.H.; Speizer, F.E.; Schwarz, J.D.; Coull, B.A.; Suh, H.H.; Nearing, B.D.; Mittleman,
M.A.; Verrier, R.L.; Gold, D.R. (2009). T-wave alterans, air pollution and traffic in high-risk subjects. Am J
Cardiol 104: 665-670. doi:10.1016/j.amjcard.2009.04.046
117 Brook, R.D.; Rajagopalan, S.; Pope, C.A.; Brook, J.R.; Bhatnagar, A.; Diez-Rouz, A.V.; Holguin, F.; Hong, Y.;
Luepker, R.V.; et la. (2010). Paniculate matter air pollution and cardiovascular disease: an update to the scientific
statement from the American Heart Association. Circulation 121: 2331-2378. doi:10.1161/CIR.Ob013e3181dbecel
118 Bastain, T.M.; Gilliland, F.D.; Li, Y.; Saxon, A.; Diaz-Sanchez, D. (2003) Intraindividual reproducibility of nazal
allergic responses to diesel exhaust particles indicates a susceptible phenotype. Clinical Immunol 109: 130-136.
119 Gilliland, F.D.; Li, Y.; Diaz-Sanchez, D. (2004) Effect of glutathione-S-transferase Ml and PI genotypes on
xenobiotic enhancement of allergic responses: randomized, placebo-controlled crossover study. Lancet 363:119-
125.
120 Svartengren, M., Strand, V.; Bylin, G. Ja'rup, L.; Pershagen, G. (2000) Short-term exposure to air pollution in a
road tunnel enhances the asthmatic response to allergen. Eur Respir J 15: 716-724.
121 Vrijheid, M.; Martinez, D.; Manzanares, S.; Dadvand, P.; Schembari, A.; Rankin, F.; Nieuwenhuijsen, M. (2011).
Ambient air pollution and risk of congenital anomalies: a systematic review and meta-analysis. Environ Health
Perspect 119: 598-606. doi: 10.1289/ehp. 1002946
122 Raaschou-Nielson, O.; Reynolds, P. (2006). Air pollution and childhood cancer: a review of the epidemiological
literature. IntJ Cancer 118: 2920-2929. doi: 10.1002/ijc.21787 [Online at http://dx.doi.org]
123 National Research Council, (1993). Protecting Visibility in National Parks and Wilderness Areas. National
Academy of Sciences Committee on Haze in National Parks and Wilderness Areas. National Academy Press,
Washington, DC. This book can be viewed on the National Academy Press Website at
http://www.nap.edu/books/0309048443/html/
124 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009.
6-37
-------
125 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009. pg 9-19 through 9-23.
126 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009.
127 U.S. EPA. (2005). Review of the National Ambient Air Quality Standard for Paniculate Matter: Policy
Assessment of Scientific and Technical Information, OAQPS Staff Paper. EPA-452/R-05-005.
128 U.S. EPA. (2008). Integrated Science Assessment for Oxides of Nitrogen and Sulfur- Ecological Criteria (Final).
U.S. EPA, WashingtonD.C., EPA/600/R-08/082F.
129 U.S. EPA. (2008). Integrated Science Assessment for Oxides of Nitrogen and Sulfur- Ecological Criteria (Final).
U.S. EPA, WashingtonD.C., EPA/600/R-08/082F.
130 Environmental Protection Agency (2003). Response Of Surface Water Chemistry to the Clean Air Act
Amendments of 1990. National Health and Environmental Effects Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency. Research Triangle Park, NC. EPA 620/R-03/001.
131 Fenn, M.E. and Blubaugh, TJ. (2005). Winter Deposition of Nitrogen and Sulfur in the Eastern Columbia River
Gorge National Scenic Area, USDA Forest Service.
132 Galloway, J. N.; Cowling, E. B. (2002). Reactive nitrogen and the world: 200 years of change. Ambio 31: 64-71.
133Bricker, Suzanne B., et al., (1999). National Estuarine Eutrophication Assessment, Effects of Nutrient
Enrichment in the Nation's Estuaries, National Ocean Service, National Oceanic and Atmospheric Administration,
September, 1999.
134 Smith, W.H. (1991). "Air pollution and Forest Damage." Chemical Engineering News, 69(45): 30-43.
135 Gawel, J.E.; Ahner, B.A.; Friedland, A.J.; and Morel, F.M.M. (1996). "Role for heavy metals in forest decline
indicated by phytochelatin measurements." Nature, 381: 64-65.
136 Cotrufo, M.F.; DeSanto, A.V.; Alfani, A.; et al. (1995). "Effects of urban heavy metal pollution on organic matter
decomposition in Quercus ilix L. woods." Environmental Pollution, 89: 81-87.
137Niklinska, M.; Laskowski, R.; Maryanski, M. (1998). "Effect of heavy metals and storage time on two types of
forest litter: basal respiration rate and exchangeable metals." Ecotoxicological Environmental Safety, 41: 8-18.
138 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009. Section 9.4.5.2.
139 Mason, R.P. and Sullivan, K.A. (1997). "Mercury in Lake Michigan." Environmental Science & Technology, 31:
942-947. (from Delta Report "Atmospheric deposition of toxics to the Great Lakes").
140 Landis, M.S. and Keeler, G. J. (2002). "Atmospheric mercury deposition to Lake Michigan during the Lake
Michigan Mass Balance Study." Environmental Science & Technology, 21: 4518-24.
141 U.S. EPA. (2000). EPA453/R-00-005, "Deposition of Air Pollutants to the Great Waters: Third Report to
Congress," Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
142 National Science and Technology Council (NSTC) (1999). "The Role of Monitoring Networks in the
Management of the Nation's Air Quality."
143 Callender, E. and Rice, K.C. (2000). "The Urban Environmental Gradient: Anthropogenic Influences on the
Spatial and Temporal Distributions of Lead and Zinc in Sediments." Environmental Science & Technology, 34: 232-
238.
144 Rice, K.C. (1999). "Trace Element Concentrations in Streambed Sediment Across the Conterminous United
States." Environmental Science & Technology, 33: 2499-2504.
145 Ely, JC; Neal, CR; Kulpa, CF; et al. (2001). "Implications of Platinum-Group Element Accumulation along U.S.
Roads from Catalytic-Converter Attrition." Environ. Sci. Technol. 35: 3816-3822.
6-38
-------
146U.S. EPA. (1998). EPA454/R-98-014, "Locating and Estimating Air Emissions from Sources of Polycyclic
Organic Matter," Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
147 U.S. EPA. (1998). EPA454/R-98-014, "Locating and Estimating Air Emissions from Sources of Polycyclic
Organic Matter," Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
148 Simcik, M.F.; Eisenreich, S.J.; Golden, K.A.; et al. (1996). "Atmospheric Loading of Polycyclic Aromatic
Hydrocarbons to Lake Michigan as Recorded in the Sediments." Environmental Science and Technology, 30: 3039-
3046.
149 Simcik, M.F.; Eisenreich, S.J.; and Lioy, P.J. (1999). "Source apportionment and source/sink relationship of
PAHs in the coastal atmosphere of Chicago and Lake Michigan." Atmospheric Environment, 33: 5071-5079.
150 Arzayus, K.M.; Dickhut, R.M.; and Canuel, E.A. (2001). "Fate of Atmospherically Deposited Polycyclic
Aromatic Hydrocarbons (PAHs) in Chesapeake Bay." Environmental Science & Technology, 35, 2178-2183.
151 Park, J.S.; Wade, T.L.; and Sweet, S. (2001). "Atmospheric distribution of polycyclic aromatic hydrocarbons and
deposition to Galveston Bay, Texas, USA." Atmospheric Environment, 35: 3241-3249.
152Poor, N.; Tremblay, R.; Kay, H.; et al. (2002). "Atmospheric concentrations and dry deposition rates of
polycyclic aromatic hydrocarbons (PAHs) for Tampa Bay, Florida, USA." Atmospheric Environment 38: 6005-
6015.
153 Arzayus, K.M.; Dickhut, R.M.; and Canuel, E.A. (2001). "Fate of Atmospherically Deposited Polycyclic
Aromatic Hydrocarbons (PAHs) in Chesapeake Bay." Environmental Science & Technology, 35, 2178-2183.
154 U.S. EPA. (2000). EPA453/R-00-005, "Deposition of Air Pollutants to the Great Waters: Third Report to
Congress," Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
155 Van Metre, P.C.; Mahler, B.I; and Furlong, E.T. (2000). "Urban Sprawl Leaves its PAH Signature."
Environmental Science & Technology, 34: 4064-4070.
156 Cousins, I.T.; Beck, A.J.; and Jones, K.C. (1999). "A review of the processes involved in the exchange of semi-
volatile organic compounds across the air-soil interface." The Science of the Total Environment, 228: 5-24.
157 Tuhackova, J. et al. (2001). Hydrocarbon deposition and soil microflora as affected by highway traffic.
Environmental Pollution, 113: 255-262.
158U.S. EPA. (1999). The Benefits and Costs of the Clean Air Act, 1990-2010. Prepared for U.S. Congress by U.S.
EPA, Office of Air and Radiation, Office of Policy Analysis and Review, Washington, DC, November; EPA report
no. EPA410-R-99-001.
159 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
160 Winner, W.E., and CJ. Atkinson. (1986). "Absorption of air pollution by plants, and consequences for growth."
Trends in Ecology and Evolution 7:15-18.
161 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
162Tingey, D.T., and Taylor, G.E. (1982). Variation in plant response to ozone: a conceptual model of physiological
events. InM.H. Unsworth & D.P. Omrod (Eds.), Effects of Gaseous Air Pollution in Agriculture and Horticulture.
(pp. 113-138). London, UK: Butterworth Scientific.
163 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
164 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
165 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
6-39
-------
166U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
167 Ollinger, S.V., Aber, J.D., Reich, P.B. (1997). Simulating ozone effects on forest productivity: interactions
between leaf canopy and stand level processes. Ecological Applications, 7, 1237-1251.
168 Winner, W.E. (1994). Mechanistic analysis of plant responses to air pollution. Ecological Applications, 4(4), 651-
661.
169 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
170 Chappelka, A.H., Samuelson, LJ. (1998). Ambient ozone effects on forest trees of the eastern United States: a
review. New Phytologist, 139,91-108.
171 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
172 Fox, S., Mickler, R. A. (Eds.). (1996). Impact of Air Pollutants on Southern Pine Forests, Ecological Studies.
(Vol. 118, 513 pp.) New York: Springer-Verlag.
173 De Steiguer, I, Pye, I, Love, C. (1990). Air Pollution Damage to U.S. Forests. Journal of Forestry, 88(8), 17-
22.
174 Pye, J.M. (1988). Impact of ozone on the growth and yield of trees: A review. Journal of Environmental Quality,
77,347-360.
175 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
176 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
177 McBride, J.R., Miller, P.R., Laven, R.D. (1985). Effects of oxidant air pollutants on forest succession in the
mixed conifer forest type of southern California. In: Air Pollutants Effects On Forest Ecosystems, Symposium
Proceedings, St. P, 1985, p. 157-167.
178 Miller, P.R., O.C. Taylor, R.G. Wilhour. (1982). Oxidant air pollution effects on a western coniferous forest
ecosystem. Corvallis, OR: U.S. Environmental Protection Agency, Environmental Research Laboratory (EPA600-D-
82-276).
179 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
180 Grulke, N.E. (2003). The physiological basis of ozone injury assessment attributes in Sierran conifers. In A.
Bytnerowicz, M. J. Arbaugh, & R. Alonso (Eds.), Ozone air pollution in the Sierra Nevada: Distribution and effects
on forests, (pp. 55-81). New York, NY: Elsevier Science, Ltd.
181 U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. EPA,
Washington, DC, EPA/600/R-05/004aF-cF, 2006.
182Kopp, R. J., Vaughn, W. J., Hazilla, M., Carson, R. (1985). Implications of environmental policy for U.S.
agriculture: the case of ambient ozone standards. Journal of Environmental Management, 20, 321-331.
183 Adams, R. M., Hamilton, S. A., McCarl, B. A. (1986). The benefits of pollution control: the case of ozone and
U.S. agriculture. American Journal of Agricultural Economics, 34, 3-19.
184 Adams, R. M., Glyer, J. D., Johnson, S. L., McCarl, B. A. (1989). A reassessment of the economic effects of
ozone on U.S. agriculture. Journal of the Air Pollution Control Association, 39, 960-968.
185 Abt Associates, Inc. (1995). Urban ornamental plants: sensitivity to ozone and potential economic losses. U.S.
EPA, Office of Air Quality Planning and Standards, Research Triangle Park. Under contract to RADIAN
Corporation, contract no. 68-D3-0033, WA no. 6. pp. 9-10.
6-40
-------
186 White, D., Kimerling, A.J., Overton, W.S. (1992). Cartographic and geometric component of a global sampling
design for environmental monitoring. Cartography and Geographic Information Systems, 19, 5-22.
187 Smith, G., Coulston, I, Jepsen, E., Prichard, T. (2003). A national ozone biomonitoring program—results from
field surveys of ozone sensitive plants in Northeastern forests (1994-2000). Environmental Monitoring and
Assessment, 87, 271-291.
188 Coulston, J.W., Riitters, K.H., Smith, G.C. (2004). A preliminary assessment of the Montreal process indicators
of air pollution for the United States. Environmental Monitoring and Assessment, 95, 57-74.
189U.S. EPA. (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants. EPA/600/R-05/004aF-
cF.
190U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants. EPA/600/R-05/004aF-cF.
191 Smith, G., Coulston, I, Jepsen, E., Prichard, T. (2003). A national ozone biomonitoring program—results from
field surveys of ozone sensitive plants in Northeastern forests (1994-2000). Environmental Monitoring and
Assessment, 87, 271-291.
192 U.S. EPA. (1991). Effects of organic chemicals in the atmosphere on terrestrial plants. EPA/600/3-91/001.
193 Cape JN, ID Leith, J Binnie, J Content, M Donkin, M Skewes, DN Price AR Brown, AD Sharpe. (2003). Effects
of VOCs on herbaceous plants in an open-top chamber experiment. Environ. Pollut. 124:341-343.
194 Cape JN, ID Leith, J Binnie, J Content, M Donkin, M Skewes, DN Price AR Brown, AD Sharpe. (2003). Effects
of VOCs on herbaceous plants in an open-top chamber experiment. Environ. Pollut. 124:341-343.
195 Viskari E-L. (2000). Epicuticular wax of Norway spruce needles as indicator of traffic pollutant deposition.
Water, Air, and Soil Pollut. 121:327-337.
196 Ugrekhelidze D, F Korte, G Kvesitadze. (1997). Uptake and transformation of benzene and toluene by plant
leaves. Ecotox. Environ. Safety 37:24-29.
197 Kammerbauer H, H Selinger, R Rommelt, A Ziegler-Jons, D Knoppik, B Hock. (1987). Toxic components of
motor vehicle emissions for the spruce Picea abies. Environ. Pollut. 48:235-243.
6-41
-------
Chapter 7 Impacts of the Proposed Rule on Emissions and Air
Quality
This chapter presents the overall emissions and air quality impacts of the proposed Tier 3
standards. Section 7.1 describes national average impacts on criteria and toxic emissions
resulting from the Tier 3 proposal. Section 7.2 describes the air quality effects of the proposed
emission reductions. Because the air quality analysis requires emission inventories with greater
geographical resolution than the national average inventories, the emission inventories described
in the two sections were developed separately, as described in each portion of this chapter.
Section 7.3 discusses the impact of the proposed program on greenhouse gas emissions.
7.1 Criteria and Toxic Pollutant Emission Impacts
7.1.1 Overview
This section presents the projected national emission impacts of the Tier 3 proposal on
criteria and toxic air pollutants for selected calendar years, and the methodology used to estimate
these reductions. The proposed fuel and vehicle standards will directly reduce emissions of
nitrogen oxides (NOx), including nitrogen dioxide (NO2), volatile organic compounds (VOC),
carbon monoxide (CO), particulate matter (PM^.s), air toxics, and sulfur dioxide (802). The
implementation of lower sulfur gasoline will reduce criteria and air toxic emissions from the
existing gasoline fleet, and cause some reductions in SO2 from the nonroad gasoline sector. The
largest reductions come immediately following the implementation of the fuel standard, as a
significant share of overall emissions are produced by Tier 2 and older vehicles. To reflect these
early reductions, we are presenting emission reductions in calendar year 2017, near the
beginning of the proposed fuel program.
The proposed vehicle standards will incur reductions as the cleaner cars and trucks begin
to enter the fleet in model year 2017. The magnitude of reduction will grow as the contribution
of these vehicles to fleet emissions becomes more prominent - to reflect this, we are also
presenting emission reductions through calendar year 2030, when 2017 and later model year cars
and trucks comprise 80 percent of the light-duty fleet, and travel 90 percent of vehicle miles
travelled (VMT). 2030 is a standard out-year for evaluation; it is used for air quality modeling in
this proposal as well as recent EPA rules. However, the full impact of the vehicle program
would be realized after 2030. For this reason, we are also presenting emissions reductions in
calendar year 2050, when the fleet will have fully turned over to the proposed vehicle standards.
As explained in Section 7.2, air quality modeling was done only for 2017 and 2030.
Emission impacts presented in this section are estimated on an annual basis, for all 50
U.S. states plus the District of Columbia, Puerto Rico and the U.S. Virgin Islands. The
reductions from onroad sources were estimated using an updated version of EPA's MOVES
model, as described in detail in Section 7.1.3; and the NONROAD model for offroad sources.
Reductions were estimated compared to a reference case that assumed continuation of the Tier 2
vehicle program indefinitely and an average gasoline sulfur level of 30 ppm (10 ppm in
California).
7-1
-------
The emission inventory methodology applied to generate these national estimates does
differ from the methodology used to generate the finely resolved emission inventories needed for
the air quality modeling, leading to some differences in absolute estimates of tons reduced
between the two analyses. These differences are discussed in Section 7.2.
7.1.2 Scenarios Modeled
We analyzed emission impacts of the Tier 3 vehicle emissions and fuel standards by
comparing projected emissions for future years without the Tier 3 rule (reference scenario) to
projected emissions for future years with the Tier 3 standards in place (control scenario). Table
7-1 below presents an overview of the reference and control scenarios for calendar years 2017
and 2030. Both scenarios reflect the renewable fuel volumes mandated in the Energy Policy Act
of 2005 (EPAct) and Energy Independence and Security Act of 2007 (EISA). We thus refer to
this renewable fuel level as "post-EPAct/EISA". A key update in the analysis of renewable fuel
volumes from that performed for the RFS2 final rule is the inclusion of gasoline containing 15
percent ethanol by volume (El 5) in the fuel pool, per the approval in 2011 of a waiver which
allows El5 to be used in model year 2001 and later cars and trucks. We estimated the market
shares for El5 and gasoline containing 10 percent ethanol by volume (E10) in the market in 2017
and 2030, as discussed in Section 7.1.3.2.1.
7-2
-------
Table 7-1 Overview of Reference and Control Scenarios
Reference Scenario
Control Scenario
2017
Renewable Fuels: RFS2 program51
21.6 B gallons renewable fuels
(24 B ethanol-equivalent gallons):
17.8 B gallons ethanol: E10b, E15C
Fuel Sulfur Level:
30 ppm (10 ppm California)
Fleet:
100 percent Tier 2 and older vehicles
Renewable Fuels: RFS2 program51
21.6 B gallons renewable fuels
(24 B ethanol-equivalent gallons):
17.8 B gallons ethanol: E10b, E15C
Fuel Sulfur Level:
10 ppm
Fleet:d
93 percent Tier 2 and older vehicles
7 percent Tier 3 vehicles
2030
Renewable Fuels: RFS2 program51
30.5 B gallons renewable fuels
(36 B ethanol-equivalent gallons):
22.2 B gallons ethanol: E15°
Fuel Sulfur Level:
30 ppm (10 ppm California)
Fleet:
100 percent Tier 2 and older vehicles
Renewable Fuels: RFS2 program51
30.5 B gallons renewable fuels
(36 B ethanol-equivalent gallons):
22.2 B gallons ethanol: E15°
Fuel Sulfur Level:
10 ppm
Fleet:d
20 percent Tier 2 and older vehicles
80 percent Tier 3 vehicles
a RFS2 primary mid-ethanol scenario, 75 FR 14670 (March 26, 2010)
b Gasoline containing 10 percent ethanol by volume
0 Gasoline containing 15 percent ethanol by volume
d Fraction of the vehicle population
Our reference scenarios assumed an average fuel sulfur level of 30 ppm in accordance
with the Tier 2 gasoline sulfur standards. Under the proposed Tier 3 program, federal gasoline
would contain no more than 10 ppm sulfur on an annual average basis by January 1, 2017
(Section V of the preamble), and we therefore assumed a nationwide fuel sulfur level of 10 ppm
for both future year control cases. A more detailed description of our fuel inputs and
assumptions for this analysis can be found in Section 7.1.3.2.
We assumed a continuation of the existing Tier 2 standards for model years 2017 and
later in modeling emissions for our reference scenario. Our Tier 3 control scenario modeled the
suite of exhaust and evaporative emission standards proposed for light-duty vehicles (LDVs),
light duty trucks (LDTs: 1-4), medium passenger vehicles (MDPVs) and large pick-ups and vans
(Class 2b and 3 trucks) described in Section IV of the preamble, including:
• Fleet average Federal Test Procedure (FTP) NMOG+NOx standards of 30
mg/mi for LDVs, LDTs and MDPVs, phasing in from MYs 2017 to 2025
7-3
-------
• Fleet average Supplemental Federal Test Procedure (SFTP) NMOG+NOx
standards of 50 mg/mi for LDVs, LDTs and MDPVs, phasing in from MYs
2017 to 2025
• Per-vehicle FTP PM standard of 3 mg/mi for LDVs, LDTs and MDPVs,
phasing in from MYs 2017 to 2022
• Per-vehicle US06-only PM standard of 10 mg/mi for LDVs through LDT2s
and 25 mg/mi for LDT3s and LDT4s, phasing in for MYs 2017 to 2022
• New standards for Class 2b and 3 trucks phasing in by MY 2022 including
NMOG+NOx declining fleet average, more stringent PM standards, and a
regulatory useful life of 150,000 miles
• More stringent evaporative emission standards for diurnal plus hot soak
emissions, a new canister bleed test and emissions standard, a new
requirement to measure emissions using 9 RVP E15 certification test fuel, and
new requirements addressing evaporative leaks on in-use vehicles.
The Tier 3 standards are expected to reduce onroad criteria and toxic emissions, and to a
much smaller extent, nonroad SO2 emissions, but will not affect upstream, refueling or portable
fuel container criteria or toxic emissions. The methodology for estimating emission impacts and
the results for onroad and nonroad emissions are described in Section 7.1.3 and Section 7.1.4,
respectively.
Implementation of the proposed Tier 3 standards is aligned with the model year 2017-
2025 Light-Duty GHG standards1 to achieve significant criteria pollutant and GHG emissions
reductions while providing regulatory certainty and compliance efficiency to the auto and oil
industries. The LD GHG standards were still in a preliminary state of development (pre-
proposal) at the time we developed our assumptions for the Tier 3 emissions, air quality, and cost
analyses presented here, so we were not able to reflect them in these analyses. However, our
analyses for the final Tier 3 rule will include the final LD GHG standards in both the reference
and control scenarios, and will thus account for their impacts on the future vehicle fleet and
future fuel consumption. We do not anticipate that this change will substantially modify our
results or conclusions. Improved fuel efficiency does not have a direct impact on criteria
pollutant emissions, because non-GHG emissions standards for light-duty vehicles are expressed
as grams per mile driven; thus, non-GHG emissions are a function of the emission control
systems rather than the amount of fuel burned. As a result, we do not believe that changes to
vehicle efficiency as a result of the LD GHG standards will affect the emissions benefits of the
Tier 3 rule, but we will include the LD GHG requirement in our analysis for the final rule. The
analysis described here accounts for the following national onroad rules:
• Tier 2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control
Requirements (65 FR 6698, February 10, 2000)
• Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur
Control Requirements (66 FR 5002, January 18, 2001)
7-4
-------
• Mobile Source Air Toxics Rule (72 FR 8428, February 26, 2007)
• Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard
Program (75 FR 14670, March 26, 2010)
• Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average
Fuel Economy Standards for 2012-2016 (75 FR 25324, May 7, 2010)
The analysis also accounts for many other national rules and standards. In addition, the
modeling accounts for state and local rules including local fuel standards,
Inspection/Maintenance programs, Stage II refueling controls, the National Low Emission
Vehicle Program (NLEV), and the Ozone Transport Commission (OTC) LEV Program. See the
Tier 3 emissions modeling TSD for more detail.
Decisions on the emissions and other elements used in the air quality modeling were
made early in the analytical process for this proposal. For this reason, the Tier 3 emission
control scenario used in the air quality modeling includes emission reductions from Tier 3 across
the nation, assuming no reductions associated with California's LEV III program (as opposed to
including California's LEV III program and its associated emission reductions in the baseline
scenario). We will conduct new air quality modeling for the final rule that will include emission
reductions from California's LEV III program, both in California and in states that by that time
have adopted the LEV III program, in the baseline scenario.A
7.1.3 Onroad Emissions
7.1.3.1 Methodol ogy Overvi ew
EPA's official model for use in estimating mobile source emissions is known as the
Motor Vehicle Emission Simulator (MOVES), with the most recent version approved for use in
State Implementation Plan (SIP) and transportation conformity analyses being MOVES2010b.2
A version of MOVES2010a, updated specifically for this analysis, was used to estimate
emissions of criteria and air toxic emissions from on-road gasoline and diesel vehicles for the
entire U.S. for the reference and control scenarios described in Section 7.1.2 above, for calendar
years 2017 through 2030, and 2050.B
The MOVES model updates affecting reference and control case emissions were
extensive, and are documented in a separate memorandum to the Tier 3 docket.3 Updates made
to MOVES2010a for this analysis primarily incorporated major new research in three areas. The
A The California Air Resources Board (CARB) approved the LEV III program in January 2012. Several states have
adopted the LEV III program under Section 177 of the Clean Air Act. These states include Connecticut, Maryland,
Maine, Massachusetts, New Jersey, New York, Pennsylvania, Rhode Island, Washington, and Vermont.
T3
The MOVES updates are reflected in a version of the MOVES model code (April 14, 2011 Version a) and a
concurrently updated version of the MOVES default database (May 12, 2011), available in the Tier 3 docket As
these updates are still draft, this code and/or database are not approved for official use in SIP and conformity
analyses.
7-5
-------
first involves fuel effects on exhaust emissions from Tier-2 vehicles. The second involves
improvements in estimation of evaporative emissions from all vehicles, including Tier-2
vehicles. The third involves accounting for the effects of fuel sulfur level on exhaust emissions.
The effects of changes in fuel properties on exhaust emissions of Tier 2 vehicles, which
comprise the majority of the fleet by 2017, were assessed through the results of the EPAct Phase-
3 Program. Specific fuel properties addressed include ethanol level, aromatics, distillation
properties, and volatility (Reid Vapor Pressure, or RVP). Methods used to account for the
effects of these properties in inventory modeling are described in a separate memorandum to the
docket.4 Improvements in estimating and projecting evaporative emissions are described in this
document (see 7.1.3.3.7). Finally, because the updates to fuel sulfur effects are critical for
estimating reductions from the Tier 3 program, they are also presented in detail in Section
7.1.3.4.
Other than sulfur, these changes had more bearing on updating reference case emissions
than on the projected reductions from the Tier 3 standards. One key exception to this, however,
were updates made to PM2.5 fuel effects that result in an increase in emissions with increases in
aromatics and T90 (reduction in E300). As discussed in Section 7.1.3.2.2, this update had an
impact on control case emissions, as changes in aromatics and E300 were projected by refinery
modeling as sulfur levels were reduced from 30 ppm to 10 ppm. However, other recent studies
suggest that the PM emission increases observed are likely linked to the impacts of heavier
hydrocarbons, and in particular heavy aromatics.5'6'7 More study is needed to understand these
relationships, and whether the changes in refinery operation to make up for lost octane, should it
occur, would be likely to impact heavy aromatics or not.
In addition to fuel effects, we also improved emission estimates in other areas. The
sulfate, sulfur dioxide, organic carbon and elemental carbon emission rates for 2007-and-later
heavy-duty diesel vehicles were updated to include information from a recent study that
examined the composition of particulate emissions from advanced diesel engines.8 HC, CO and
NOx start and running emission rates for light heavy-duty gasoline vehicles were updated to fix
an error in these rates for 2007-and-later emissions, and we repaired errors in the MOVES2010a
*~1 Q
emission rates for NH3, NO and NO2. '
The MOVES version used for this analysis also includes an added capability to model
many hazardous air pollutants. And other changes were made to the MOVES2010a model to
facilitate the large number of parallel runs that needed to be done to complete the Tier 3 air
quality modeling inventories. These changes are also detailed in the docket memo addressing
MOVES updates.
In addition to the model updates needed to incorporate new research, a set of custom
inputs were developed to allow MOVES to model the reference and control scenarios. Some of
these inputs were required to reflect regional variations in fuels for both the reference and control
c The changes to the NO and NO2 rates did not impact the total NOX emissions, but facilitated the output of separate
results for nitric oxide and nitrogen dioxide.
7-6
-------
scenarios, as discussed in detail in Section 7.1.3.2. Other inputs were required to model the
vehicle program for exhaust and evaporative emissions, discussed in Section 7.1.3.3.
The national emission inventories presented in this section were developed with a simpler
and quicker method than we used for air quality modeling, because the abbreviated approach
enabled analysis for calendar years in addition to 2017 and 2030, and makes the analysis easier
for stakeholders and other commenters to replicate. National emission inventories were
developed using the pre-aggregation feature of MOVES. For criteria pollutants, the default pre-
aggregation level of 'hour' was selected, which condenses the county-level temperatures into a
single national average temperature by hour of the day.0 While the model and many of the
inputs are identical for the emission inventory modeling performed for the air quality analysis,
the pre-aggregation approach is coarser than the approach described in Section 7.2 used to
develop the gridded/hourly emission inventories needed for air quality modeling. In addition to
this difference in temperature resolution (national average vs. gridded hourly temperature), the
national emission inventory analysis also used information contained in the MOVES2010a
default database for vehicle miles travelled, fleet age distributions, activity inputs (e.g., speeds),
temperatures, emission standards, and inspection/maintenance programs. As discussed in
Section 7.2, this contrasts with the air quality modeling inventory methodology, which used data
supplied by state and regional modelers for many of these inputs, employed hourly
meteorological data, and accounts for Section 177 states adoption of California LEV standards in
the modeling baseline.
To assure that adequate temperature resolution was incorporated into the national
emission inventory processes, MOVES was run separately for January and July, and annual
emissions were extrapolated (for all pollutants except PM) by scaling up each month by a factor
of 5.88. For PM, to offset the disproportionate effect of the colder temperature January results, a
scaling factor of 7.5 was applied to July and 4.3 to January; these factors were determined based
on analysis of annual PM emissions during modeling for the RFS2 rule.10 The updated MOVES
version, and all inputs and outputs that produced the results presented in Section 7.1.5 of this
Chapter are contained in the Tier 3 rulemaking docket.
7.1.3.2 Fuel Inputs
Estimating national emission inventories required translation of the reference and control
fuel scenarios presented in Table 7-1 into a discrete set of fuels (defined by RVP, sulfur content,
ethanol level, aromatics content, olefin content, T50 and T90), and the market share of these
fuels, by month and county. These data were converted into "fuel supply" database tables used
by MOVES to estimate emission inventories. Even for the national emission inventories
calculated at a pre-aggregated level, these county-level fuel supply tables are retained to develop
composite emissions that reflect the market share of the entire set of fuels that define the U.S.
fuel pool. The crux of estimating emission impacts for the Tier 3 fuel program was the
For air toxics, pre-aggregation level of 'month' was used instead to improve the model run time. Monthly pre-
aggregation averages the temperatures of all selected days and hours into a national average hourly temperature for
the month.
7-7
-------
development of fuel supply database tables that reflected the difference between the reference
and control scenarios, discussed in the following sections.
7.1.3.2.1 Reference Scenario
The reference scenario was developed to reflect updated assumptions regarding the
ethanol blends used to meet the EISA-mandated volumes. For simplicity, we assumed the same
biofuel volumes analyzed in the RFS2 final rule, referred to in that rule as the primary "mid-
ethanol" case.11 The RFS2 analysis also considered a "low-ethanol" and a "high-ethanol"
scenario; the difference being an increase in cellulosic ethanol and a decrease in cellulosic diesel
going from the low-ethanol to the high-ethanol scenario. We believe that the RFS2 renewable
fuel volumes still bracket the realm of realistic potential fuel scenarios assuming the EISA
mandate is met. However, given the practical limitations of conducting air quality modeling for
all three scenarios, we focused our Tier 3 modeling on the primary mid-ethanol case.
We relied on biofuel volumes taken directly from the RFS2 mid-ethanol case, which
made projections for every year between 2010 and 2022. We assumed that renewable fuel
production/consumption would remain constant after EISA reaches full implementation in 2022.
Accordingly, for 2030, we relied on 2022 biofuel volumes from the RFS2 primary mid-ethanol
case.
In the RFS2 final rule, we assumed that all ethanol would be consumed as either E10 or
E85. Since then, EPA issued a waiver permitting 15 volume percent ethanol blends (E15) to be
used in 2001 and newer model year light-duty vehicles.E While E15 has only limited
commercial availability currently, EPA believes it may compete favorably with E85 in the
marketplace and could become a major gasoline blend in the future to meet the mandated RFS2
volumes. Accordingly, the reference scenario assumes an increasing utilization of E15 over time
as infrastructure ramps up and owners become more aware of their vehicle's ethanol usage
potential.
To estimate future El 5 fuel consumption, we started by estimating the fraction of
vehicles capable of legally refueling on El 5 (or higher ethanol blends) in the future. To do this,
we relied on many vehicle assumptions made under the RFS2 final rule.F For flexible fuel
vehicles (FFVs), we started with EPA certification data and assumed that the "Detroit 3" (GM,
Ford and Chrysler) would follow through with their voluntary commitment to produce 50
percent FFVs by 2012.° We also assumed that the Detroit 3 would continue to comprise
approximately 45 percent of total light-duty vehicle sales. In addition, we relied on total light-
duty vehicle sales projections from EIA's Annual Energy Outlook (AEO). Factoring in EPA's
projected vehicle scrappage rates, vehicle miles traveled (VMT) by model year (MY), and fuel
economy assumptions based on EPA's 2012-2016 Light-Duty Greenhouse Gas Rule,H we
E On Nov. 4 2010, EPA issued a partial waiver for MY2007 and newer light-duty vehicles (75 FR 68094). On
January 26, 2011, EPA extended the waiver to MY 2001-2006 light-duty vehicles (76 FR 4662).
F Refer to Section 1.7 of the RFS2 RIA for more on our FFV/vehicle assumptions.
G We assumed that non-domestic automakers would produce around 2 percent FFVs in current and future years.
H 75 FR 25324 (May 7, 2010).
7-8
-------
estimate that by 2017, 13.6 percent of gasoline demand would come from FFVs, 68.7 percent
would come from 2001 and later model year non-FFVs, and 17.7 percent would come from
legacy vehicles, nonroad, motorcycles and other small engines not approved for E15. By 2030,
the contributions would be 22.0, 66.8, and 11.2 percent, respectively. A graph of the projected
gasoline fuel fractions over time is provided in Figure 7-1.
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030
Year
I Gas from Nonroad, Motorcycles, etc.
Gas from Legacy Non-FFVs • Gas from 2001+ Non-FFVs • Gas from FFVs
Figure 7-1 Gasoline Fuel Fraction by Vehicle Type
The next step was to estimate how often E15-capable vehicles would fill up on E15. E15
utilization is a function of how accessible El 5 is to consumers (i.e., number of stations offering
it) and how often E15-capable vehicles choose to fill up on it (based on retail pricing and other
factors). We assumed that El5 utilization for 2001 and later model year vehicles would ramp up
to 50 percent by 2017, increasing to almost 100 percent utilization by 2030. We assumed that
new nonroad equipment (and other small engines not covered by the waiver) would be designed
for and/or certified on El5 in the future. As such, we assumed that El5 utilization in nonroad
equipment would ramp up from zero percent in 2017 to almost 100 percent by 2030. This
resulted in a very small amount of E85 needing to be used in 2017 and 2030 to reach the RFS2
primary mid-ethanol volumes. FFV owners were assumed to utilize E15 at the same rate as 2001
and later model year conventional vehicles when not refueling on E85. When not refueling on
E15 (or E85), all E15-capable vehicles were assumed to refuel on E10. As a simplifying
assumption, we assumed EO usage was negligible. Figure 7-2 shows the El 5 and E85 utilization
assumptions and Figure 7-3 shows the resulting fuel volumes by year. Total fuel volumes were
computed based on motor gasoline energy projections provided in the Annual Energy Outlook
(AEO) 2011 early release.
7-9
-------
2001+ MY Vehicles
Nonroad, etc.
* " 98.9% /
' /
* /
y /
/
c
.2 60%
ro
3 50%
~a>
00 40%
2
I
30%
50.0%
Year
Figure 7-2 Post-EPAct/EISA Fuel Utilization Assumptions/Results
7-10
-------
160,000 -
140,000
120,000 -
100,000 -
in
c
o
ra
15 80,000
O
i
60,000
40,000
20,000
0
2010 2011 2012 2015 201-1 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030
Year
• Total E10 • Total E15 Total E85
Figure 7-3 Assumed Post-EPAct/EISA Gasoline Fuel Volumes by Calendar Year
As shown above, our post-EPAct/EISA ethanol blend assumptions (i.e., ramp up of E15)
are capable of meeting the RFS2 primary mid-ethanol scenario with little to no E85 needed -
especially in our 2017 and 2030 modeling years. Therefore, for emissions and air quality
modeling, we made the simplifying assumption that E85 use would be negligible. While E85
used in FFVs could be part of the future biofuel picture, we believe that given EPA's waiver
decision and the ability for El5 to compete in today's gasoline marketplace, El5 is a viable
ethanol blend for meeting the RFS2 volumes.
These projected fuel scenarios do not reflect the impact of the LD GHG standards, as the
rule was still in a preliminary state at the time we needed to finalize our assumptions for these
Tier 3 analyses. The LD GHG standards will reduce gasoline demand in 2017 and beyond which
has two impacts on our analysis. First, the reduced fuel consumption associated with the LD
GHG standards will likely result in somewhat lower costs for sulfur control, because less
gasoline will need to be desulfurized. Second, if ethanol volumes are constant but gasoline
demand is reduced due to the LD GHG standards, perhaps as much as five percent of the fuel
pool could be E85 or other higher-level ethanol blends by 2030. However, we do not expect this
to significantly impact criteria pollutant reference or control case emissions because recent data
on NMOG and NOx emissions do not show significant differences when run on E85,12'13 and the
7-11
-------
percentage of the fleet impacted is small. There could be a small impact on sulfur related costs
and emission reductions due to an effect of E85 on overall fuel pool sulfur concentrations,
depending on what could be assumed for future E85 (or other higher level ethanol blend) sulfur
levels. An increase in E85 volumes could also lead to changes in toxic emissions, with increases
in acetaldehyde emissions from ethanol combustion, but decreases in most other toxics due to
dilution. However, EPA air quality modeling suggests that changes in ambient levels of
acetaldehyde are likely to be much smaller than changes in direct emissions.14'15 This is because
emissions of acetaldehyde precursors, particularly alkenes, are lower in E85 emissions. For the
Tier 3 final rule's analysis, we will be including the impact of the LD GHG standards, and we
will revisit our assumed future fuel scenarios accordingly, including higher-level ethanol blends
as appropriate.
For this analysis we assumed that El5 would first appear in significant volumes in
reformulated gasoline (RFG) areas. All RFG is subject to a de facto 7-psi RVP standard. And
since neither E10 nor E15 can take advantage of the 1-psi waiver in RFG areas, the two fuels are
essentially the same with respect to vapor pressure limits. In conventional gasoline (CG) areas,
E15 is held to a 9-psi standard whereas E10 is allowed to be 10-psi in-use (through the use of a
1-psi waiver). As a result, in order for E15 to enter into CG areas, refiners would need to remove
butanes/pentanes to make an 8-psi blendstock to account for the 1-psi increase associated with
ethanol blending. Conversely, refiners could keep their CG blendstocks around 9-psi to blend up
10-psi E10 which is eligible for the 1-psi waiver and thus is considered 9-psi for compliance
purposes. Due to the additional actions needed to market El 5 during the summer in CG areas, it
makes sense that more El5 would be used in RFG areas, at least in the near term. Accordingly,
we assumed that in 2017, 75 percent of RFG would be E15 and the balance of E15 would be
used in CG (resulting in 25 percent E15 utilization and 75 percent E10 utilization in those areas).
By 2030, since we modeled a much larger volume of ethanol, El 5 was assumed to be used
virtually everywhere.
Converting these assumptions into post-EPAct/EISA fuel properties on a county-by-
county basis required an assessment of fuel properties prior to the EPAct/EISA requirements for
2017 and 2030, termed "pre-EPAct/EISA". We developed these starting with 2005 fuel
volumes as the baseline for our analysis of renewable fuel impacts because these volumes were
in use immediately before EPAct was enacted, creating volume requirements for 2006, and
because they provide consistency with the base case used for air quality modeling described in
Section 7.2. In 2005, gasoline contained over 4 billion gallons of ethanol. In 2005, gasoline also
contained approximately 2 billion gallons of MTBE.
For translation into the fuel supply inputs needed by MOVES, the MOVES2010a default
fuel supply for 2005 (based on the National Emission Inventory16) was updated to serve as a
better basis for the other modeling cases. These updates are detailed below.
• In 2005, California fuel was required to have an ethanol level of 5.7 percent for all fuel.
The MOVES2010a default fuel supply listed California counties with a 57 percent market
share of E10 and a 43 percent market share of EO. The fuel supplies for these counties
were refined to indicate E5.7 at 100 percent market share. Also, Alameda County,
California, had anomalous fuel property data. The fuel properties for this county were
replaced with an average of the fuel properties of the neighboring counties.
7-12
-------
• Counties in the southeast Michigan fuel program area erroneously had been assigned
higher sulfur and benzene levels than the rest of the state of Michigan. Sulfur and
benzene for these counties were corrected to an average of all other Michigan counties.
Some Michigan counties also had very low levels of MTBE (<0.3 percent), despite an
MTBE ban in Michigan in 2005. These fuels have been corrected to a 0 percent MTBE
conventional fuel to properly reflect the absence of MTBE in Michigan.
• Many counties in the 2005 default fuel supply had fuel RVP levels that were drastically
inconsistent with the mandated level for RVP in that county for a given summer month.
Summer RVP levels for all counties were corrected to their mandated level (including a
0.25psi compliance margin) based on knowledge of fuel programs in place for 2005.
Winter RVP levels remain unchanged from the 2005 default fuel supply.
• The corrections above were applied to the months of January (winter fuel properties) and
July (summer fuel properties) for the 2005 base case. The fuel properties for these two
months were then duplicated to other months to complete the 2005 fuel supply for all
months. The corrected January fuel supply was duplicated to February, March, April,
October, November and December; the corrected July fuel supply was duplicated to May,
June, August and September. Although this duplication eliminates fuel property changes
during the shoulder seasons in April and September, these intermediate month fuel
properties were used only when modeling refueling for the air quality inventories; we
expect that this simplification will not significantly impact the modeling run results.
Diesel fuel properties were not adjusted in the 2005 fuel supply.
The 2017 pre-EPACT/EISA base was developed using the corrected 2005 fuel supply
discussed above as a foundation. From this fuel supply, benzene and sulfur levels for all counties
were further corrected to properly reflect the introduction of the Control of Hazardous Air
Pollutants from Mobile Sources (MSAT2) (2007) rule and the Tier 2 Motor Vehicle Emissions
Standards and Gasoline Sulfur Control Requirements (1999). Benzene corrections were made by
PADD following results from the MSAT2 analysis of downstream benzene levels. No other fuel
property changes were found to change significantly with a change in benzene levels. Benzene
levels by PADD follow in Table 7-2 below:
Table 7-2- MSAT2 Downstream Fuel Benzene Levels
PADD
1
2
3
4
5
CA
CG
0.61
0.63
0.63
0.86
0.65
N/A
RFC
0.54
0.60
0.54
N/A
0.61
0.62
Sulfur corrections were made to all counties based on the default sulfur level found in the
2005 fuel supply. Counties with a sulfur level higher than 30 ppm were reduced to 30 ppm to
reflect the gasoline sulfur standards of the Tier 2 rule (counties subjected to lower fuel sulfur
7-13
-------
standards, such as in California, were not changed). Refinery modeling showed that there is a
significant effect on aromatics level when sulfur is reduced. Corrections to the aromatics level
based on refinery modeling for counties with reduced sulfur level were made as follows:
1) A "high sulfur" aromatics level was determined using the following equation:
/ 1 \
high sulfur aromatics level = I * 0,479 + 24,60
\s-ulfur level /
Equation 7-1 Aromatics Level from Initial Sulfur Concentration
2) A "low sulfur" aromatics level was determined using the same equation, substituting
30ppm for the initial sulfur level of the county
3) An aromatics delta was calculated by subtracting the "low sulfur" aromatics level from
the "high sulfur" aromatics level
4) This aromatics delta was applied as a correction for sulfur reduction to the original
aromatics level for the county as appearing in the 2005 fuel supply
Diesel fuel sulfur levels were also adjusted to 30ppm to reflect low sulfur diesel levels.
There were no other changes from the 2005 base case to the 2017 base case. The fuel properties
for the 2030 base case were identical to the 2017 base case for every county; the only
modification made to create the 2030 pre-EPAct/EISA base case is a change in the year of the
fuel supply.
The post-EPAct/EISA cases used as a reference for Tier 3 were developed using the pre-
EP Act/EISA cases for 2017 and 2030 as a foundation, and were then created by inserting the
ethanol market share assumptions for each county discussed at the beginning of this section.
Fuel properties for all counties were then adjusted based on relationships found in refinery
modeling for this additional ethanol. The process used for adjusting ethanol market shares and
fuel properties for every county is described below:
1) Counties with multiple fuels were aggregated to one set of temporary fuel properties
using the market shares of these fuels. Counties with only one fuel remain unchanged in
this step.
2) Two new fuels, a 10 percent ethanol blend and a 15 percent ethanol blend, were
created from the temporary fuel properties by raising the ethanol level from the
temporary aggregate to the appropriate new ethanol level.
3) For each new fuel, other fuel properties were adjusted based on refinery modeling to
reflect changes due to increased ethanol use. These fuel property changes (on a percent
ethanol basis) are shown in Table 7-3 below:
7-14
-------
Table 7-3- Reference Case Fuel Property Changes
Fuel Property"
E200
E300
Aromatics
Olefins
Summer
E10
0.54834
0.13946
-0.40887
-0.13788
E15
0.76481
0.09549
-0.46684
-0.15915
Winter
E10
0.73693
0.00634
-0.50555
-0.25198
E15
1.04067
0.39788
-0.49956
-0.15296
Note:
a These fuel property changes are listed as a per-ethanol percent change.
4) Market shares for the E10 and E15 fuels were adjusted to reflect RFS2 level ethanol
use. As explained above, in RFG counties, E10 was added to the fuel supply at a 25
percent market share, El5 was added to the fuel supply at a 75 percent market share. For
counties not using RFG, E10 was added to the fuel supply at a 75 percent market share,
and E15 was added at a 25 percent market share. In 2030, the E10 market share is set at 1
percent and El 5 market share set at 99 percent, regardless of fuel program.
5) RVP levels for the new E10 and El5 fuels were corrected to reflect El5 fuel not
receiving a Ipsi waiver for maximum RVP level. As discussed for the base case fuel
supplies in the previous section, RVP levels were adjusted to within a 0.25 psi
compliance margin depending on county fuel programs.
The result of this effort was two alternate fuel supply databases tables for use in MOVES,
reflecting the reference case fuels in 2017 and 2030 - these tables were used for the development
of national emissions inventory as well as air quality modeling. New fuel formulations were also
required for MOVES, and were created for each of the new E10 and E15 fuels created for the
2017 and 2030 reference case.
7.1.3.2.2 Control Scenario
The Tier 3 control fuel scenarios for the years 2017 and 2030 used the fuel supplies
constructed for the 2017 and 2030 reference cases described in the previous section as a
foundation. To develop the control scenario fuel supplies, we modified the reference fuel
supplies to reflect the sulfur program proposed in the Tier 3 control case by reducing sulfur from
30 ppm to 10 ppm for all gasoline.1 Associated fuel properties determined by refinery modeling
were also adjusted to reflect the implications of sulfur reductions on other fuel properties, such as
an increase in aromatics and decrease in olefins and distillation properties, as shown in Table 7-4
1 An inconsistency in our approaches for estimating California fuel sulfur levels led us to model an increase in
sulfur in California. More specifically, for the reference case, we used MOVES default fuels for California. These
were developed based on a mix of survey data and other sources, which resulted in a range of sulfur levels from 8-
19 ppm. However, the control case assumed 10 ppm sulfur throughout California. As a result, although there
should have been no change in California fuel properties due to Tier 3, some areas had small modeled increases or
decreases.
7-15
-------
below. These changes were made to every county with fuel exceeding a sulfur level of lOppm.
No changes were made to the diesel fuel supply for the control scenario.
Table 7-4 Tier 3 Control Case Sulfur Fuel Property Changes
Fuel Property
E200 (%)
E300 (%)
Aromatics (%)
Olefms (%)
Summer
-0.78
-0.75
0.63
-0.82
Winter
-1.27
-0.68
0.48
-1.12
The result of this effort was two additional alternate fuel supply databases tables for use
in MOVES, reflecting the control case fuels in 2017 and 2030; these tables were used for the
development of national emissions inventory as well as air quality modeling. For the national
emission inventories the 2017 fuel supply was applied to 2018 through 2021 as well, to
approximate the fuel supply prior to full RFS2 implementation; the 2030 fuel supplies were
applied to 2022 through 2029, reflecting full RFS2 implementation. New fuel formulations were
also required for MOVES, and were created for the low sulfur versions of the E10 and E15 fuels
created for the 2017 and 2030 reference case.
Upon further analysis, we believe that the increase in aromatics and reduction in E300
shown in Table 7-4 is unlikely. The process of hydrotreating in the fluid catalytic cracking
(FCC) of gasoline to reduce sulfur tends to saturate olefins formed by the FCC unit and which
are present in FCC gasoline, thus reducing its octane. Because refiners have historically been
pressed to make as much octane as possible to supply market needs, various technologies and
catalyst formulations have been developed for hydrotreating FCC gasoline to avoid and/or
minimize this loss of octane. Nevertheless, our analysis estimated that desulfurizing gasoline to
an average of 10 ppm would result in about a half number loss in the octane of FCC gasoline,
and we also conservatively assumed that this octane loss would need to be compensated for with
other changes to gasoline, such as increased reformate (an aromatic rich stream), isomerate and
alkylate. These assumptions are conservative and not consistent with more recent trends. For
example, there is less demand for octane due to the dramatic rise in the use of ethanol (which has
very high octane), and the declining demand for gasoline. It is not clear that the loss in octane
resulting from gasoline desulfurization would need to be compensated for. In fact, there is
considerable "give away" of octane in the gasoline pool today as ethanol is still often splash
blended on top of finished gasoline instead of blending it with a sub-octane gasoline blendstock.
Nevertheless, for our Tier 3 analysis, we conservatively assumed that the full half number loss in
the octane of FCC gasoline would have to be compensated for by the refinery and not made up
by blending in more ethanol. In particular, in running the refinery model to make up for the lost
octane, we conservatively constrained the refinery model such that the only options for
additional octane were changes within the refinery. Additional ethanol was not a modeled
option, despite the fact that ethanol will be providing additional octane to the gasoline pool as
ethanol use continues to rise through 2017 and beyond.
Furthermore, the increase in aromatics and decrease in E300 may simply be a function of
additional assumptions in the refinery modeling. We did not model any revamps in FCC feed
hydrotreaters. For some refineries, revamping FCC pretreaters is expected to realize the
7-16
-------
decrease in FCC gasoline sulfur levels without any decrease in FCC gasoline octane levels.
Furthermore, when refiners are faced with tighter sulfur standards, one strategy for compliance
would be to undercut the heaviest portion of the FCC naphtha, which is also the portion highest
in sulfur, into either jet fuel or diesel fuel. By doing so, refiners would not only reduce the
hydrotreating severity of their FCC posttreaters (reducing olefin saturation and some of the
octane loss), but they would be increasing E300 (lightening up the gasoline pool). One check of
our refinery modeling output, which shows increasing aromatics and decreasing E300, would be
to compare the output of our refinery modeling with that of other studies. Our refinery modeling
estimated a 0.48 volume percent increase in the aromatic content and 0.68 percent reduction in
the E300 of gasoline in the winter, and a 0.63 volume percent increase in the aromatic content
and 0.75 percent reduction in the E300 of gasoline in the summer. Recent modeling performed
by Mathpro for the International Council on Clean Transportation (ICCT) showed a 0.1 volume
percent increase in aromatic content and a 0.2 volume percent decrease in E300 in the summer,
and a 0.3 volume percent decrease in aromatics and 0.8 percent volume percent increase in E300
in the winter.17 Overall, the Mathpro refinery modeling showed an annual average decrease in
aromatics and increase in E300. Recent modeling by Baker & O'Brien for API (Sensitivity Case
3) showed about half the increase in aromatics that our analysis showed, however the API study
also included the octane impacts of lower RVP, so we could not determine whether modeling by
API would show an increase in aromatics or not if it solely modeled sulfur control.18 API's
addendum to its original study which added an additional control case that solely modeled a 10
ppm gasoline sulfur standard (no change in RVP), showed a 0.2 volume percent decrease in
aromatics during the summertime (no wintertime fuel quality data was presented, nor was any
distillation data presented either summer or winter). Furthermore, after we completed this
analysis for the NPRM, we discovered that the LP refinery cost model that we licensed to use
required some improvements to correctly characterize the qualities of the light and heavy
naphtha streams from the reformer to improve its estimation of E300 and aromatics content.
Thus, unlike our modeling results shown in Table 7-4, which show a meaningful impact on
aromatics and E300, we believe, consistent with the Mathpro and Baker and O'Brien refinery
modeling studies, that there will be little to no change. Note that these improvements are not
expected to have any impact on the cost estimates made by the refinery model.
The air quality analysis included the changes to aromatics and E300 shown in Table 7-4.
However, because of the concerns above, we did not reflect these changes in the national
emission inventories presented in Section 7.1.5.
7.1.3.3 Vehi cl e Program Inputs
Modeling the controls introduced by the Tier 3 vehicle program required the development
of another set of alternate MOVES database tables to reflect each aspect of the proposed Tier 3
program. These database tables included:
• Gaseous exhaust emissions rates (HC/CO/NOx) for light duty cars, trucks, and
light-heavy trucks (gas and diesel) to reflect the proposed Tier 3 FTP and US06
standards and their phase-in.
7-17
-------
• Elemental carbon (EC) and organic carbon (OC) exhaust emission rates for light
duty cars, trucks, and light-heavy trucks (gas and diesel) to reflect the proposed
Tier 3 FTP and US06 PM standards and phase-in.
• Evaporative HC permeation emission rates to reflect the proposed diurnal test
standard, certification fuel, and phase-in.
• Leak prevalence rates for tank vapor venting and liquid leaks to reflect proposed
requirements for evaporative leak detection.
The development of these alternative emission rates is discussed below by pollutant, fuel
and vehicle regulatory class.
7.1.3.3.1 Gasoline LD HC/CO/NOX Exhaust
Gaseous emission rates in MOVES are contained in a database table
(EmissionRateByAge) that expresses emission rates as mass per time, distinguished by emission
process (start and running), fuel type (gas and diesel), vehicle regulatory class (LDV, LDT, Light
HD, etc.), model year, age, and operating mode (power/speed for running, vehicle soak time for
start). Developing these rates on Tier 3 vehicles required accounting for expected changes in
each of these dimensions.
The development of Tier 3 emission rates followed the same procedures used to develop
National LEV (NLEV, covering model years 2001-2003) and Tier 2 rates (covering model years
2004 and later) in the default MOVES database, as described in the documentation for
development of light-duty exhaust emission rates for MOVES2010 (known as the "MOVES
Light-Duty report").19 However, specific modifications were made to represent the introduction
of Tier 3 standards, summarized below. Where no modifications to methods were made, we will
refer the reader to the appropriate section of the MOVES2010 report. In particular, see Section
1.3.4.
MOVES emission rates are estimated by standard level, model year, age, and vehicle
regulatory class. There are separate rates for areas with Inspection/Maintenance programs (I/M)
and those without. Developing the rates involves six steps, listed below.
1. Project average Federal Test Procedure (FTP) results by standard level and vehicle
regulatory class. As in the development of the default MOVES2010 database outlined in the
MOVES Light Duty Report, we made use of data measured on the FTP cycle in the course of
EPA's In-use Verification Program (IUVP) to project emissions under the proposed standards.
For Tier 3, we developed estimates of FTP results for Tier 3 vehicles based on IUVP data from
Tier 2 Bin 2 and 3 vehicles, including composite results, "cold-start" emissions" (Bagl minus
Bag3) and "hot-running" emissions (Bag 2 FTP and US06).
2. Develop phase-in assumptions for model years (MY) 2017 - 2031, by standard level,
vehicle class and model year, including phase-in assumptions representing the introduction of
Tier 3 standards.
7-18
-------
3. Merge FTP results and Phase-in assumptions. For running emissions, calculate
weighted ratios of emissions in each model year relative to those for cars (LDV) in MY2000,
which represent Tier 1 LDV (as discussed in the MOVES Light Duty report, default MOVES
rates were projected forward based on model year 2000 data from state I/M data, in conjunction
with IUVP data for later model years).
4. Estimate Emissions by Operating Mode. Calculate emissions by operating mode in
each model year by multiplying the MY2000 emission rates by the weighted ratio for each model
year. We assume that the emissions control at high power (outside ranges of speed and
acceleration covered by Bag 2 of the FTP) is not as effective as at lower power (within the range
of speed and acceleration covered by Bag 2).
5. Apply Deterioration to estimate emissions for three additional age groups (4-5, 6-7
and 8-9). We assume that Tier 3 vehicles will deteriorate similarly to other vehicles, when
viewed in logarithmic terms, but we modified deterioration to represent a useful life of 150,000
miles, as opposed to a useful life of 120K miles, assumed for Tier 2 and NLEV vehicles. This is
the outcome of applying In-linear deterioration to the rates developed in steps 1-4. For the
remaining three groups (10-14, 15-19 and 20+), emissions are assumed to stabilize as described
in the MOVES2010 report.
6. Estimate non-I/Mreference rates. The rates in steps 1-6 represent rates under a
reference inspection/maintenance (I/M) program. Corresponding non-I/M rates are calculated by
applying the ratios applied to the Tier 1 and pre-Tier 1 rates.
Each of these six steps is described in more detail below. Addition information is
available in a separate memo available in the docket.20
7.1.3.3.1.1 Average FTP Results (Step 1) (Standard)
Our projected emissions for Tier 3 vehicles are driven by the proposed NMOG+NOx
standard, set at 30 mg/mi. However, because MOVES projects NOx and THC emissions
separately, we apportioned the aggregate standard into NMOG and NOx components, which we
will refer to as the "effective standards" for each pollutant. For purposes of apportionment, we
assumed that NMOG control would pose a greater technical challenge than NOx control.
Accordingly, we assumed "effective standards" for NMOG and NOx would be 20 mg/mi and 10
mg/mi, respectively. To implement this assumption, we further assumed that for NOx, vehicles
would be effectively brought into Bin 2, and that for NMOG, vehicles would be brought to a
level between Bin 2 and Bin 3, but closer to Bin 2.
In addition, MOVES models start and running processes separately. It is therefore
necessary to translate the composite standard into start and running components. One
component represents a "cold start" on the FTP cycle, represented as "Bagl minus Bag3"
emissions. A second component represents "hot-running" emissions, represented by the hot-
running phase of the FTP (Bag 2). A third component represents emissions on the US06 cycle,
representing emissions at high speed and power.
7-19
-------
Estimated FTP and US06 emissions levels for hydrocarbons (NMOG and NMHC) are
shown in Table 7-5, for several Tier 2 Bins and for Tier 3. Values for all standards except Tier 3
are identical to those used to develop rates in the default database. The values for Tier 3 are
calculated as a weighted average of those for Bins 2 and 3, using Equation 7-2.
T3 = 0.775 -B2 + 0.225 -B3
Equation 7-2
Table 7-5 Hydrocarbons (HC): Useful-Life FTP Standards and Associated Cold-Start and
Hot-Running Results on the FTP and US06 Cycles. Values for the FTP and US06 represent
NMOG and NMHC, respectively.
Bin
8
5
4
3
2
Tier 3C
Useful-life Standard
(mg/mi)
125
90
70
55
10
20
FTP Composite51
(mg/mi)
41.3
35.5
24.8
21.5
5.6
9.2
FTP Cold Starta
(mg)
591
534
383
329
87
142
FTP hot Running51
(Bag 2)
(mg/mi)
3.56
2.63
2.28
1.74
0.42
0.7
US06b
(mg/mi)
35.8
35.8
35.8
35.8
2.6
10.0
a Values represent "non-methane organic gases" (NMOG).
b Values represent "non-methane hydrocarbons" (NMHC).
0 Values for Tier 3 calculated using Equation 7-2.
Under a general assumption that CO standards are not forcing, but that CO emissions
tend to track NMOG emissions, corresponding values for CO were calculated in the same
manner, and are presented in Table 7-6.
Table 7-6 CO: Useful-Life FTP Standards and Associated Cold-Start and Hot-Running
Results on the FTP and US06 Cycles.
Bin
8
5
4
O
2
Tier 3a
Useful-life Standard
(mg/mi)
4,200
4,200
4,200
2,100
2,100
2,100
FTP Composite
(mg/mi)
861
606
537
463
235
286
Cold Start
(mg)
6,680
5,510
5,500
3,470
1,620
2,040
FTP hot Running
(Bag 2)
(mg/mi)
451
238
201
119
70
81
US06
(mg/mi)
2,895
2,895
2,895
2,895
948
1,390
a Values for Tier 3 calculated using Equation 7-2.
Corresponding results for NOx are presented in Table 7-7. In contrast to HC and CO, the
values for Tier 2 Bin 2 were adopted for Tier 3, as the FTP composite of 5.5 mg/mi suggests that
Bin 2 vehicles can meet the "effective standard" of 10 mg/mi with a reasonable compliance
margin.
7-20
-------
Table 7-7 NOX: Useful-Life FTP Standards and Associated Cold-Start and Hot-Running
Results on the FTP and US06 Cycles.
Bin
8
5
4
3
2
Tier3
Useful-life Standard
(mg/mi)
200
70
40
30
20
10
FTP Composite
(mg/mi)
64.2
21.2
8.7
5.7
5.5
5.5
Cold Start
(mg)
418
165
90
71
67
67
FTP hot Running
(Bag 2)
(mg/mi)
35.1
8.2
4.7
3.8
0.4
0.4
US06
(mg/mi)
61.3
45.9
30.6
30.6
18.4
18.4
7.1.3.3.1.2 Develop Phase-In Assumptions (Step 2)
We designed phase-in assumptions so as to project compliance with the Tier 3 fleet
average NMOG+NOx requirements. The requirements are illustrated in Figure 7-4. The phase-
in begins in model year 2017 and ends in model year 2025. Note the sharp drop in emissions at
the outset of the Tier 3 phase-in, also that the truck standards (LDT2,3,4) are slightly higher than
the lighter vehicles' (LDV-T1). After 2017, the reduction in the fleet average is linear. The fleet
averages for cars and trucks no longer differ at the completion of the phase-in.
0.180
LDV/LDT1
LDT234
Model Year
Figure 7-4 NMOG+NOX FTP Fleet Average Requirements during Phase-In of the Tier 3
Exhaust Emissions Standards for Light-Duty Vehicles.
7.1.3.3.1.3 Merge Cycle Results and Phase-In Assumptions (Step 3)
The goal of this step is to calculate weighted averages of the FTP (cold-start and hot-
running) results for all standards in each model year, with the emissions results weighted by
applicable phase-in fractions. We do this step for each vehicle class separately, then weight the
four truck classes together using a set of fractions also derived from the weighted sales estimates.
7-21
-------
Start and running emissions in each model year are simply calculated as weighted
averages of the emissions estimates and the phase-in fractions. The resulting weighted start
estimates are used directly to represent cold-start emissions for young vehicles in each model
year (ages 0-3). For running emissions, however, the averages are not used directly; rather, each
is expressed as a ratio to the corresponding Tier 1 value.
7.1.3.3.1.4 Estimate Emissions by Operating Mode (Step 4)
To project emissions for the 2016-and-later vehicles, we divided the operating modes for
running exhaust into two groups. These groups represent the ranges of speed and power covered
by the hot-running phase (Bag 2) of the FTP standards (< -20 kW/Mg), and the ranges covered
by the SFTP standards (primarily the US06 cycle). For convenience, we refer to these two
regions as "the hot-running FTP region" and "US06 region," respectively (See Figure 7-5). Data
measured on the SC03 cycle did not play a role in emission rate development.
To estimate emissions by operating mode, the approach was to multiply the emission
rates for MY 2000, representing Tier 1, by a specific ratio for each model year from 2016 to
2025, to represent emissions for that model year.
To estimate rates for the US06 modes, we followed a procedure similar to that for the
"FTP" modes, but using the "US06" columns in Table 7-5 through Table 7-7. For HC and CO,
we used Equation 7-2, as before. For NOx, we applied the Bin-2 values. Figure 7-6 and Figure
7-7 show application of the ratios to the FTP and US06 operating modes in model years 2010,
2017, and 2025, representing fully phased-in Tier 2 standards, an interim year during the Tier 3
phase-in, and the fully phased-in Tier 3 standards, respectively. Figure 7-3 displays the
information on linear scale, highlighting the differences at the higher operating modes, while
Figure 7-7 shows the same information on a logarithmic scale, illustrating the patterns for the
lower operating modes.
7-22
-------
30 +
1? 27-30
c
£ 24-27
1 27-24
1 78-27
£ 75-78
u
fc 72-75
u
Q. (}_•/?
CO y 'Z
-c 6"9
a) 3-6
' * 0-3
<0
Speed Class
7-25 25-50
16 30
29
28
27
15 25
14 24
13 23
12 22
11 21
(mph)
50 +
40
39 .^
^
38
37
35
«—
33
^
The "US06 Region"
High Power
The "hot-running FTP Region"
Low to Moderate
Power
(also includes braking (0)
and idle (1)
Figure 7-5 Operating modes for running Exhaust Emissions, divided broadly into "hot-
running FTP" and "US06" regions.
7-23
-------
s
I
g
'>
(A
E
LU
-5
10
15
20
25 30 35 40
Vehicle Specific Power (kW/Mg)
O>
I
K
c
.0
'
LU
5 10 15 20 25 30
Vehicle Specific Power (kW/Mg)
35
40
ra
c
o
E
LLI
-5
5 10 15 20 25 30
Vehicle Specific Power (kW/Mg)
35
40
Figure 7-6 Projected Emission Rates for Cars in Operating modes 21-30, vs. VSP, in
ageGroup 0-3 years, for three model years, for (a) CO, (b) THC and (c) NOX (LINEAR
SCALE).
7-24
-------
10,000
£ 1,000
I
(A
(A
E
LU
100 !-
0 10 20 30
Vehicle Specific Power (kW/Mg)
40
O)
I
oi
c
,0
'35
M
I
LU
100.00
10.00
1.00
0.10
0.01
0 10 20
Vehicle Specific Power (kW/Mg)
30
40
0 5 10 15 20 25
Vehicle Specific Power (kW/Mg)
30
35
40
Figure 7-7 Projected Emission Rates for Cars in Operating modes 21-30, vs. VSP, in
ageGroup 0-3 years, for Four Model Years, for (a) CO, (b) THC and (c) NOX
(LOGARITHMIC SCALE).
7-25
-------
7.1.3.3.1.5 Apply Deterioration (Step 5)
Based on our extensive emissions analysis during MOVES2010 development, we assume
that deterioration for different technologies is best represented by a multiplicative model, in
which different technologies, represented by successive model-year groups, show similar
deterioration in relative terms but markedly different deterioration in absolute terms. We
implemented this approach by translating emissions for the 0-3 age group, as calculated above,
into natural logarithms and applying uniform logarithmic age trends to all model-year groups.
We derived logarithmic deterioration slopes for Tier 1 vehicles (MY 1996-98) and applied them
to Tier 2 vehicles. In this process we applied the same logarithmic slope to each operating mode,
which is an extension of the multiplicative deterioration assumption. For 2017 and later model
year vehicles, the deterioration assumptions were modified to represent the extension of the full
useful life (FUL), which is increased from 120,000 mi to 150,000 mi. However, we did not
extrapolate the deterioration trend beyond the 8-9 year age group, as we know that emissions
tend to stabilize beyond this age, while the In-linear emissions model would project an
increasingly steep and unrealistic exponential emissions trend. For the 10-14, 15-19 and 20+ age
groups, the "stabilization of emissions with age" was estimated as for MOVES2010 (MOVES
Light Duty report, section 1.3.3.7).
7.1.3.3.1.6 Estimate Non-I/M References (Step 6)
Completion of the preceding steps provided a set of rates representing I/M reference rates
for MY 2016-2025. As a final step, we estimated non-I/M reference rates by applying the same
ratios used in MOVES2010 (section 1.3.3.6).
7.1.3.3.1.7 Start Emissions
The values for "Cold Start" shown in Tables 8- 4 through 6 above were used to represent
cold-start emissions for the various standard levels. These are designated as opModeID=108 in
the emissionRateByAge table; emission rates for starts following shorter soak periods were
developed by applying standard soak curves (found in the MOVES Light Duty report) to the
updated cold start rates. Deterioration was applied to start emissions, using the same approach as
used for developing MOVES2010 base rates discussed in the MOVES Light Duty report. Start
deterioration is expressed relative to deterioration for running emissions.
7.1.3.3.1.8 Final Estimates of Composite FTP and US06
In producing emission inventory estimates, MOVES combines emission rates with
activity patterns derived from surveys of in-use vehicles. These emissions do not necessarily
correlate directly with the test procedures used for compliance; for example, in-use activity
shows that more miles are driven per start event than assumed on the FTP. Likewise, the US06
cycle is focused on compliance, and represents a relatively small portion of in-use driving.
However, to give a relative sense of the changes projected by the proposed Tier 3 standards,
emissions can be constructed for FTP composite and US06 from MOVES emission rates for the
Tier 2 (labeled as model year 2010) and Tier 3 (labeled as model year 2025) cases. These are
shown in Figures 7-5 through 7-8 below. Note that the Tier 3 rates shown below are for the
MOVES base fuel of 30 ppm. In modeling the control scenarios on 10 ppm, these emission rates
were further lowered by the sulfur reductions outlined in Section 7.1.3.4.1.
7-26
-------
0.12
0.00
0
10 15
Age (Years)
20
•2010
•2025
25
Figure 7-8 FTP Composite NOX emissions for reference (2010) and Tier 3 (2025)
constructed from MOVES rates
0.16
0.00
0
•2010
•2025
10 15
Age (Years)
20
25
Figure 7-6 FTP Composite THC emissions constructed from MOVES rates
7-27
-------
0.18
0.16
0.00
•2010
•2025
10 15
Age (Years)
20
25
Figure 7-7 US06 NOx emissions constructed from MOVES rates
0.06
0.00
•2010
•2025
10 15
Age (Years)
20
25
Figure 7-8 US06 THC emissions constructed from MOVES rates
7-28
-------
7.1.3.3.2 Diesel LD HC/CO/NOX Exhaust
Emission rates representing light-duty diesel vehicles under Tier 3 standards were
calculated identically to those representing gasoline vehicles, with the exception that the
"effective standards" were set differently. Again, diesel vehicles are projected to meet the same
NMOG+NOx standard as gasoline vehicles (30 mg/mi). However, for diesel vehicles, we
assumed that light-duty vehicles would meet Bin-2 standards following completion of the phase
in. Accordingly, the "effective standards" for NMOG and NOx were set at 10 and 20 mg/mi,
respectively. As mentioned, all remaining steps were conducted as described in 7.1.3.3.1 above.
As a result of the different effective standards, however, the ratios and other numeric results
specific to diesel vehicles vary slightly from their counterparts for gasoline vehicles.
7.1.3.3.3 Gasoline MD HC/CO/NOX
The proposed Tier 3 program will affect not just light duty vehicles (below 8,500 pounds
GVW), but also chassis certified vehicles between 8,500 and 14,000 pounds. These vehicles are
referred to here as medium duty vehicles, but are also commonly known as Class 2b and 3 heavy
trucks. In MOVES, these vehicles are designated regulatory class 41. Regulatory class 41 also
captures certain other vehicles, namely engine certified trucks and medium duty passenger
vehicles, which are not regulated underneath the proposed medium duty standards. As described
in the reference case medium duty updates to MOVES,21 we assumed that during this timeframe
engine certified vehicles and medium-duty passenger vehicles (MDPV) are five percent and
fifteen percent of the regulatory class respectively.
Table 7-8 Population Percentage
Category
MDPV
Class 2B
Class 3
Engine Certified
Percent
of Reg Class 41
15%
60%
20%
5%
The Class 2b and Class 3 vehicle program was modeled to begin in model year 2017 and
fully phase in during the 2022 model year (Figure 7-9). This yields an aggregate standard of
0.178 g/mile NMOG+NOx for Class 2b vehicles and 0.247 gram/mile for Class 3 vehicles in
2022.
7-29
-------
2Band 3 Phase-in
550
500
450
400
350
300
250
200
150
100
Class 2B
Class 3
2017 2018 2019 2020 2021 2022
MY
Figure 7-9 Class 2B and 3 Standard Phase-in
We combined the proposed Tier 3 phase-in for Class 2b and 3 vehicles with the existing
emission standards MDPV and engine certified vehicles that comprise MOVES regulatory class
41. For this analysis, we assumed that MDPVs met the Tier 3 SULEV 30 standard, and that
engine certified vehicles would perform at 1.2x their standard on the chassis FTP.22 Calculated
using Table 7-8, the weighted average of the Class 2b, Class 3, MDPV, and engine standards is
0.181.J To account for the real world performance of these vehicles, we related this average back
to Tier 2 light duty vehicles, for which we have a significant amount of data from EPA's in-use
verification program. Using the same tool that was used for developing the Tier 2 and Tier 3
light duty emission rates, we modeled MOVES Regulatory Class 41 in 2022 as 90 percent Bin 5
and 10 percent Bin 8 vehicles.K For the phase-in years of 2017-2021, we calculated new
MOVES rates as a weighted average of the Tier 3 rates and the existing MOVES rates for
regulatory class 41 such that the appropriate weighted composite was calculated each year Table
7-9.
1 Due to a calculation error, an NMOG+ NOX standard of 0.177 was used for Regulatory Class 41. As a result, a
slightly different phase-in of the NMOG+NOX and CO rates was used in the proposal modeling. This error
overstates the NMOG+ NOX reductions for this vehicle class by approximately 2.5% in 2022.
K By basing the data on light duty vehicles, it is possible that we are misstating the emission profile of these larger
vehicles, but as emissions decrease, it is also possible that the emission profile for the larger vehicles will more
closely resemble that of light duty vehicles.
7-30
-------
Table 7-9 Phase-in of MD Tier 3 Rates
Model Year
2017
2018
2019
2020
2021
2022
Tier 3 Rate
34%
49%
62%
75%
87%
100%
MOVES DB Rates
66%
51%
38%
25%
13%
0%
Composite Standard
(g/mile)
0.33
0.29
0.26
0.24
0.21
0.18
The CO standards for MD vehicles are less stringent than those for Tier 2 Bin 5 and Bin 8
vehicles. For Bin 5 and Bin 8 vehicles, the CO standard is 4.2 g/mile. For engine certified
vehicles, the standard is approximately 17.3 grams per mile (14.4 grams per bhp multiplied by
1.2),23 and for the Tier 3 MD vehicles, the standard ranges from 4.2 to 7.3 g/mile. Using the
same weighted averages as before, we calculated an aggregate CO standard of 4.4 grams/mile,
which is 5.5 percent higher than the Tier 2 Bin 5/8 standards. To compensate for the lower CO
emissions in the Tier 2 vehicles that were used to develop the Tier 3 MD emission rates, we
multiplied the running CO rates by 1.1 and the start CO rates by 1.05.
Like in the light duty vehicles, deterioration modeled to represent a 150,000 mile useful
life. The same methodology was used for light duty and medium duty vehicles.
7.1.3.3.4 Diesel MD HC/CO/NOX
For medium duty diesel vehicles, the emission rates currently in MOVES are
significantly below the proposed Tier 3 HC and CO standards. When MOVES is used to
generate a simulated FTP estimate for NMHC, the model calculates a rate of approximately 0.05
grams per mile, while the simulated FTP estimate for CO is less than 1 gram/mile.
Consequently, we assumed no HC and CO emission benefits from Tier 3 standards on MD diesel
vehicles.
By contrast, we estimate that the Tier 3 NOx standard will produce a reduction in diesel
Class 2b and Class 3 NOx emissions. Because data on current NOx emissions are limited, as
there is little in-use data on MY 2010 and 2011 vehicles which use selective catalytic reduction
as a NOx control strategy, we used a proportional approach to estimate the Tier 3 effect,
reducing NOx in proportion to the change in the emission standard. Because emission standards
tend to impact start and running emissions differently, we applied a greater portion of the
reduction to running emissions and a smaller reduction to start emissions. These reductions were
phased-in over the same schedule as for gasoline vehicles, as detailed in Table 7-10. Also, to
account for the change in "useful life", we duplicated the Tier 3 age 0-3 NOx rates to the 4-5
year age-group.
7-31
-------
Table 7-10 Phase-in of MD Diesel Tier 3 NOX Rates
Model Year
2017
2018
2019
2020
2021
2022
Tier 3 Phase In
20%
38%
54%
69%
85%
100%
Reduction in NOx
Running Emission
Rate
12%
23%
33%
42%
52%
61%
Reduction in NOx
Start Emission Rate
5%
9%
12%
16%
19%
23%
7.1.3.3.5 Gasoline PM
2.5
Tier 3 will reduce direct particulate matter (PlV^.s) emissions from light-duty vehicles
through a full-useful life (FUL) PM standard on the FTP of 3 mg/mi. Additionally, the Tier 3
standards will include more stringent SFTP PM standards for light-duty vehicles, with a 10
mg/mi standard on the US06 for light-duty passenger vehicles, and 20 mg/mi on the US06 for
light-duty trucks. These standards are targeting several processes that contribute to particulate
matter in light-duty gasoline vehicles: cold starts, high-power operation, and deterioration of
engine and emission control technology over the life of the vehicle.
To achieve the FUL PM standards without sacrificing fuel economy, it is anticipated that
manufacturers may try a variety of strategies including reducing lubrication oil consumption over
the life of the vehicle.L For our analysis, we projected that reductions in lube oil consumption
would reduce the organic carbon (OC) fraction of PM by 30 percent for both cold start and
running emissions. These reductions are based on an analysis of the Kansas City Light-duty
Vehicle Emissions Study, which collected PM emissions from a randomly recruited sample of
vehicles in the Kansas City area in 2005. Using chemical tracers found in both the lube oil and
the particulate samples, the lube oil contribution to PM was estimated to be 25 percent of the PM
emissions, primarily from the OC fraction of PM. The 25 percent reduction is estimated by
weighting the sample to represent the distribution of vehicle ages and vehicle types in the Kansas
City Metropolitan Area in 2004.
From this analysis, a 30 percent OC reduction is projected in Tier 3 gasoline vehicles due
to decreased lube oil consumption, which represents approximately a 25 percent PM reduction
for gasoline vehicles during cold start and running emission processes. The modeling
assumptions and overview of the Kansas City analysis are located in the Supporting Technical
L Since high oil consumption leads directly to elevated PM emissions, it is unlikely that manufacturers would be
able to meet the proposed Tier 3 PM standards without addressing oil consumption over the full useful life of the
vehicle.
7-32
-------
Document: Estimated Reductions in P articulate Matter Emissions from Light-duty and Medium-
duty Gasoline Vehicles through Implementation of Tier 3 Regulations.24
For the modeling runs described here, the OC reductions are anticipated to be phased-in
over a four year period, with the full-implementation occurring in 2020. The OC reductions are
7.5 (2017), 15 (2018), 22.5 (2019), and 30 percent (2020). The percent reductions are applied to
cold-start and running emission processes, to passenger car and truck sources, and to PM
emissions as the vehicles age.
7.1.3.3.6 Diesel PM
The Tier 3 controls were modeled as having no impact on light duty diesel PM emissions.
7.1.3.3.7 Gasoline Evaporative Emissions
The proposed Tier 3 evaporative program, requiring lower emissions on the diurnal test
procedure on El 5 certification fuel and strengthened in-use detection of vapor leaks, is projected
to cause a significant reduction in evaporative hydrocarbon emissions. For this analysis, tighter
diurnal standards in conjunction with El 5 certification fuel were attributed to reductions in
evaporative permeation emissions, since the current certification standards are aimed at not
allowing any vented vapor emissions during the test. The new requirements for in-use leak
detection were modeled as resulting in a reduced prevalence (frequency rate) of fuel system
vapor and liquid leaks.M
7.1.3.3.7.1 Permeation Improvements
Permeation emissions include fuel vapors that escape from a vehicle through micro pores
in the various fuel system components and materials. Tier 3 will reduce the allowable emissions
from this process. Light duty vehicles will see a reduction from 0.50 g/test to 0.300 g/test,
approximately a 40 percent reduction.
The Tier 2 permeation rate in MOVES is 0.0102 g/hour on EO fuel. Analysis of the
impact of ethanol on permeation emissions conducted as part of the RFS2 final rule, and
included in MOVES2010, suggests that the use of El 5 as Tier 3 certification test fuel will
effectively double permeation emissions over the test procedure. Therefore, the combination of
lowering the vehicle standard and certifying on a fuel with higher propensity to permeate must
be accounted for in Tier 3 permeation rates.
The Tier 3 rate is developed by estimating permeation emissions over one day of diurnal
activity (65F-105F) on an ethanol-containing fuel using various reductions in the base rate. The
total permeation emissions for the day should equal about 75 percent of the standard (~0.225g)
as the other 25 percent of the standard can be attributed to the Hot Soak portion of the
certification test. The result is a Tier 3 permeation rate of 0.0026g/hour (a 75 percent reduction
from the Tier 2 rate).
M One of the updates to MOVES for this analysis was to enable direct input of the leak prevalence rates.
7-33
-------
Table 7-11 Tier 3 Permeation Rates
Model Year
Tier 2
2016 2017
2018 2019
2020 2021
2022
Tier 3 Phase-
in
0%
40%
60%
80%
100%
Permeation
(g/hr)
0.0102
0.0072
0.0056
0.0041
0.0026
7.1.3.3.7.2 Reduced frequency of vapor leaks
EPA, in conjunction with the state of Colorado and the Coordinating Research Council,
undertook multiple research programs to help quantify the prevalence of evaporative system
leaks in the real world, and the emissions they cause. 5'26 The proposed evaporative leak
provisions grew from this work, and informed the emission inventory contribution of evaporative
leaks, and the level of reductions possible from an in-use program focused on reducing the
incidence of these leaks. To establish the reference case, the frequency of evaporative system
leaks were estimated from the prevalence of high evaporative emission vehicles in the Colorado
field study. Because the Colorado study was not able to distinguish leaks from other high
evaporative emission sources in the broader population, we analyzed data from CRC's E-77
program and found that insufficient canister purge is also contributing to high evaporative
emissions, as this causes the canisters to become oversaturated, venting vapor into the
atmosphere.
Because the Colorado data suggested high evaporative emissions on newer vehicles as
well as old, we assumed that the leak prevalence rates from the Colorado study included a degree
of emissions from insufficient purge. Because insufficient purge would not be a function of age,
we assume most (all but 1 percent) of the Colorado prevalence rates at age 0-3 are due to
insufficient purge, which is assumed constant with age. The increase in high evaporative
emissions incidence with age was then attributed to an increasing prevalence of leaks. This is
illustrated in the chart to the left in Figure 7-10.
To model the control scenario, because the proposed leak requirement would only
address leaks and not insufficient purge, only the leak portion of high evaporative emissions
were reduced. Colorado data on a subset of vehicles that were selected for more detailed
observation and testing suggested that 70 percent of the evaporative leaks detected were due to
durability of the evaporative and/or fuel system - e.g. problems like corroded fuel lines, filler
neck, cracked hoses etc. that could be addressed with improved durability (the other 30 percent
were due to issues beyond the manufacturers control, such as improper maintenance or missing
gas caps). We estimated that the leak program would lead to a 70 percent reduction in the
occurrence of vapor leaks. This is reflected in the chart on the right in Figure 7-10; note that the
emissions attributed to insufficient purge were not changed from the reference case.
7-34
-------
0 J 45 «_7 8_
-------
loading. Numerous studies have shown the direct impact of fuel sulfur levels above 30 ppm on
emissions; these formed the basis of the Tier 2 rulemaking, which considered the impact of
sulfur in terms of immediate impact, and irreversible impact due to permanent catalyst damage.29
With the advent of the Tier 2 sulfur standards, new research has focused on the emission
reduction potential of lowering sulfur levels below 30 ppm, particularly on Tier 2 technology
vehicles, under the hypothesis that increased reliance on the catalytic converter would result in a
higher sensitivity to sulfur accumulation. A study conducted by EPA and the auto industry on
nine Tier 2 vehicles in support of the Mobile Source Air Toxics (MS AT) rule, found significant
reductions in NOx, CO and total HC when the vehicles were tested on low sulfur fuel, relative to
32 ppm fuel.30 In particular, the study found a nearly 50 percent increase in NOx when sulfur
was increased from 6 ppm to 32 ppm. Another recent study by Umicore showed reductions of
41 percent for NOx and 17 percent for HC on a PZEV operating on fuel with 33 ppm and 3 ppm
sulfur.31 Both of these studies conducted testing on high and low sulfur after running the test
vehicles through test cycles meant to clean the catalyst from the effects of prior sulfur exposure.
Both of these studies showed the emission reduction potential of lower sulfur fuel on Tier
2 and later technology vehicles over the FTP cycle. However, assessing the potential for
reduction on the in-use fleet requires understanding how sulfur exposure over time impacts
emissions, and the state of sulfur loading for the typical vehicle in field. In response, EPA
conducted a new study to assess the emission reductions expected from the in-use Tier 2 fleet
with a reduction in fuel sulfur level from current levels. It was designed to take into
consideration what was known from prior studies on sulfur build-up in catalysts over time and
the effect of periodic regeneration events that may result from higher speed and load operation
over the course of day-to-day driving.
The study sample described in this analysis consisted of 81 cars and light trucks recruited
from owners in southeast Michigan, covering model years 2007-9 with approximately 20,000-
40,000 odometer miles. The makes and models targeted for recruitment were chosen to be
representative of high sales vehicles covering a range of types and sizes. Test fuels were two
non-ethanol gasolines with properties typical of certification fuel, one at a sulfur level of 5 ppm
and the other at 28 ppm. A nominal concentration of approximately 25 ppm was targeted for the
high level to be representative of retail fuel available to the public in the vehicle recruiting area.
All emissions data was collected using the FTP cycle at a nominal temperature of 75°F.
Using the 28 ppm test fuel, emissions data were collected from vehicles in their as-
received state, and then following a high-speed/load "clean-out" procedure consisting of two
back-to-back US06 cycles intended to reduce sulfur loading in the catalyst. A statistical analysis
of this data showed highly significant reductions in several pollutants including NOx and
hydrocarbons (Table 7-11), suggesting that reversible sulfur loading exists in the in-use fleet and
has a measurable effect on aftertreatment performance; for example, Bag 2 NOx emissions
dropped 32 percent between the pre- and post-cleanout tests on 28 ppm fuel.
7-36
-------
Table 7-11 Average Clean-out Effect on In-use Emissions using 28 ppm Test Fuel
Bagl
Bag 2
Bag3
FTP Composite
Bag 1 - Bag 3
NOX
(p-value)
—
31.9%
(0.0009)
38.3%
(O.OOOl)
11.4%
(O.OOOl)
—
THC
(p-value)
—
16.5%
(0.0024)
21.4%
(O.OOOl)
4.1%
(0.0187)
—
CO
(p-value)
4.7%
(0.0737)
—
19.5%
(0.0011)
7.6%
(0.0008)
4.2%
(0.0714)
NMHC
(p-value)
—
17.8%
(0.0181)
27.8%
(O.OOOl)
3.0%
(0.0751)
—
CH4
(p-value)
—
15.3%
(0.0015)
12.0%
(O.OOOl)
6.9%
(0.0003)
—
PM
(p-value)
15.4%
(< 0.0001)
—
24.5%
(O.OOOl)
13.7%
(O.OOOl)
—
Where no reduction estimate is provided, the clean-out effect is not significant at a = 0.10.
To assess the impact of lower sulfur fuel on in-use emissions, a representative subset of
vehicles was kept to conduct testing on 28 ppm and 5 ppm fuel with accumulated mileage. A
first step in this portion of the study was to assess differences in the effectiveness of the clean-
out procedure when done using different fuel sulfur levels. Table 7-12 presents a comparison of
emissions immediately following (<50 miles) the clean-out procedures at the low vs. high sulfur
level. These results show significant emission reductions for the 5 ppm fuel relative to the 28
ppm fuel immediately after this clean-out; for example, Bag 2 NOx emissions were 47 percent
lower on the 5 ppm fuel vs. the 28 ppm fuel. This indicates that the catalyst is not fully
desulfurized, even after a clean out procedure, as long as there is sulfur in the fuel.
Table 7-12 Reduction in emissions from 28 ppm to 5 ppm immediately following clean-out
Bagl
Bag 2
Bag3
FTP
Composite
Bag 1 - Bag 3
NOX
(p-value)
5.9%
(0.0896)
47.3%
(0.0010)
51.2%
(O.OOOl)
17.7%
(0.0001)
_?
THC
(p-value)
5.4%
(0.0118)
40.2%
(O.OOOl)
35.0%
(O.OOOl)
11.2%
(O.OOOl)
_?
CO
(p-value)
7.3%
(0.0023)
_?
10.1%
(0.0988)
8.3%
(0.0003)
5.8%
(0.0412)
NMHC
(p-value)
4.6%
(0.0465)
34.4%
(0.0041)
45.0%
(O.OOOl)
8.8%
(0.0003)
_?
CH4
(p-value)
11.1%
(O.OOOl)
53.6%
(O.OOOl)
25.4%
(O.OOOl)
21.4%
(O.OOOl)
_?
PM?
—
—
—
—
—
1 Sulfur level not significant at a = 0.10.
To assess the overall in-use reduction between high and low sulfur fuel, a mixed model
analysis of all data as a function of fuel sulfur level and miles driven after cleanout was
performed. This analysis found highly significant reductions for several pollutants, as shown in
Table 7-13; reductions for Bag 2 NOx were particularly high, estimated at 59 percent between
28ppm and 5ppm overall For some pollutants, such as Bag 2 NOx, the model fitting did not find
7-37
-------
a significant miles-by-sulfur interaction, suggesting the relative differences were not dependent
on miles driven after clean-out. Other results, such as Bag 1 hydrocarbons, did show a
significant miles-by-sulfur interaction. In this case, determination of a sulfur level effect for the
in-use fleet required estimation of a miles-equivalent level of sulfur loading, which can be
gleaned from the cleanout results obtained from the baseline testing on the vehicles as-received.
Table 7-13: In-use emission reductions from 28 to 5 ppm sulfur
Bagl
Bag 2
Bag3
FTP
Composite
Bagl -Bag
3
NOX (p-
value)
10.7%
(0.0033)
59.2%
(< 0.0001)
62.1%
(< 0.0001)
23.0%T
(0.0180)
_ t
THC
(p-value)
8.5%T
(0.0382)
48.8%
(< 0.0001)
40.2%
(< 0.0001)
13.0%T
(0.0027)
5.2%
(0.0063)
CO
(p-
value)
7.5%T
(0.0552)
_t
20.1%
(< 0.0001)
11.9%T
(0.0378)
4.3%
(0.0689)
NMHC
(p-value)
7.5%
(< 0.0001)
44.8%T
(0.0260)
49.9%
(< 0.0001)
10.6%T
(0.0032)
5.1%
(0.0107)
CH4
(p-value)
13.9%T
(< 0.0001)
49.9%
(< 0.0001)
29.2%
(< 0.0001)
25.8%T
(< 0.0001)
4.6%
(0.0514)
NOX+NMO
G
(p-value)
N/A
N/A
N/A
17.3%
(0.0140)
N/A
PM
*
—
—
—
—
"
T Model with significant sulfur and mileage interaction term. * Sulfur level not significant at a = 0.10. For THC Bag
1 and CH4 Bag 1, because the effect of clean-out was not statistically significant, the reduction estimates are based
on the estimates of least squares means.
Major findings from this study include:
• Reversible sulfur loading is occurring in the in-use fleet of Tier 2 vehicles and has
a measureable effect on emissions of NOx, hydrocarbons, and other pollutants of
interest.
• The effectiveness of high speed/load procedures in restoring catalyst efficiency is
limited when operating on higher sulfur fuel.
• Reducing fuel sulfur levels from 28 to 5 ppm is expected to achieve significant
reductions in emissions of NOx, hydrocarbons, and other pollutants of interest in
the in-use fleet.
The overall reductions found in this study are in agreement with other low sulfur studies
conducted on Tier 2 vehicles. The magnitude of NOx and HC reductions found in this study
when switching from 28 ppm to 5 ppm fuel are consistent with those found in MSAT and
Umicore studies mentioned above.
7.1.3.4.2 Implementation in MOVES
The results shown in Table 7-13 were applied in MOVES for model year 2004 and later
gasoline vehicles (the nominal start of the Tier 2 phase-in) to estimate sulfur effects below 30
ppm. The sulfur fuel effect applies multiplicatively to other gasoline fuel effects in MOVES,
and applies only for sulfur levels below 30ppm. For sulfur levels above 30 ppm, and for all pre-
7-38
-------
2004 model year vehicles, the sulfur effect originally used in MOVES remains in place. Equation
7-3 shows the generic form of the new sulfur correction; Table 7-14 shows the specific values for
the sulfur coefficients by pollutant, process, and vehicle type.
sulfur effect = [l.O - Coeffsulfur * (30 - sulfurLevel}]
Equation 7-3 Low Sulfur Effect
Table 7-14 Low Sulfur Coefficients by Vehicle Type, Process and Pollutant
Vehicle Type
Motorcycle
Passenger Car,
Passenger Truck &
Light Commercial Truck
All other Vehicle Types
THC
Starts
0
0.002237
0
Running
0
0.020336
0.015488
CO
Starts
0
0.001866
0
Running
0
0
0.009436
NOX
Starts
0
0
0
Running
0
0.024459
0.027266
PM
Starts
0
0
0
Running
0
0
0
These equations were then used to fill in the database table that houses fuel effect
equations in the MOVES database ("GeneralFuelRatioExpression"). This table allows the
MOVES model to compute fuel effects based on the fuel properties of any fuel contained in the
fuel supply and fuel formulation database tables.
7.1.4 Nonroad Emissions
The nonroad sector includes a wide-range of mobile emission sources ranging from
locomotives and construction equipment to hand-held lawn tools. In the nonroad sector, the only
emissions that are directly affected by the proposed Tier 3 regulation are the emissions from
gasoline-powered equipment such as lawn-mowers, recreational boats and all-terrain vehicles.
Their SO2 emissions are reduced with the proposed decrease in gasoline sulfur levels. As with
onroad, reference and control case emissions were generated using the fuel supply inputs
reflecting the post-EISA/EPAct renewable fuel volumes.
Gasoline and land-based diesel nonroad emissions were estimated using EPA's
NONROAD2008 model, as run by the EPA's consolidated modeling system known as the
National Mobile Inventory Model (NMEVI).32 The fuels in the NMEVI database,
NCD2010201 Tier 3, were developed from the reference and control fuels used for onroad
vehicles, as described in Section 7.1.3. In 2005, onroad and nonroad gasoline formulations are
assumed to be identical. In 2017, E10 and E15 ethanol blends are available in every county, but
nonroad equipment is assumed to use E10 only. In 2030, we assume almost all nonroad
equipment can use either E10 or El 5, so onroad and nonroad gasoline formulations are again
assumed to be identical. For all years, the reference case included the higher sulfur reference
gasoline and the control case met the proposed sulfur limits.
7-39
-------
Since aircraft, locomotive and commercial marine emission sources do not burn gasoline,
their emission factors are unaffected by the changes in gasoline fuels that were developed for this
rule. Hence, their emissions are the same for both the reference and control cases. The
emissions from these sources used for this rule are the same as they were for the Heavy-Duty
Greenhouse Gas Rule (2011)33. The procedures for calculating emissions from locomotives and
C1/C2 commercial marine were developed for the Locomotive Marine Rule (2008) and are
detailed in the RIA for that rule.34 The procedures used for calculating C3 commercial marine
emissions are those developed in the recent C3 Rule (2010).
35
7.1.5 Criteria and Toxic Emission Impact Results
The proposed Tier 3 rule will reduce NOx (including NCh), VOC, CO, and 862 from all
gasoline-powered on road vehicles immediately upon implementation of lower sulfur fuel, and
will further reduce these emissions as well as PM2.5 from cars, light trucks and light heavy-duty
trucks (gas and diesel) as tighter emission standards from these vehicles phase in. There also
will be reductions in 862 emissions from the nonroad gasoline fleet as a result of sulfur
standards. The reductions are summarized in this section for each pollutant.
NOx reductions are shown in Table 7-15 for Calendar Years 2017 through 2030, and
2050 when the fleet will have turned over completed to Tier 3 standards. We project significant
reductions immediately upon implementation of the program, growing to a nearly 30 percent
reduction in onroad emissions by 2030, and nearly 40 percent reduction in onroad emissions with
full fleet turnover by 2050.
Table 7-15 Tier 3 NOx Reductions by Calendar Year (Annual U.S. Short Tons)
Year
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2050
Onroad
mobile
reference
3,452,314
3,179,942
2,942,238
2,732,380
2,557,573
2,427,305
2,304,619
2,196,463
2,115,326
2,040,468
1,989,071
1,942,891
1,913,717
1,890,403
2,410,373
Onroad
mobile
with
control
3,167,934
2,890,474
2,646,150
2,429,517
2,241,209
2,091,971
1,950,687
1,823,335
1,717,083
1,617,809
1,540,270
1,469,181
1,413,482
1,365,613
1,507,204
Reduction
284,381
289,469
296,088
302,863
316,364
335,335
353,932
373,128
398,244
422,659
448,801
473,710
500,234
524,790
903,170
Percent
reduction
in onroad
8.2%
9.1%
10.1%
11.1%
12.4%
13.8%
15.4%
17.0%
18.8%
20.7%
22.6%
24.4%
26.1%
27.8%
37.5%
7-40
-------
Table 7-16 shows the reduction in NOx emissions, in annual short tons, projected in
calendar years 2017 and 2030. The reductions are split into those attributable to the introduction
of low sulfur fuel in the pre-Tier 3 fleet (defined for this analysis as model years prior to 2017);
and reductions attributable to vehicle standards enabled by low sulfur fuel (model year 2017 and
later). As shown, in 2017 over 90 percent of the program reductions are coming from lower
sulfur gasoline on the fleet already on the road. By 2030, nearly 90 percent of the reduction is
coming from 2017 and later model year vehicles, with remaining reduction coming from lower
sulfur fuel on pre-Tier 3 vehicles.N We project that about one tenth of the reductions from the
Tier 3 fleet in 2030 are attributable to the proposed Tier 3 heavy-duty emission standards.
Table 7-16 Projected NOx Reductions from Tier 3 Program (Annual U.S. Short Tons)
Total reduction
Reduction from pre-Tier 3
fleet due to sulfur standard
Reduction from Tier 3 fleet
due to vehicle and sulfur
standards
2017
284,381
264,653
19, 728
2030
524,790
66,286
458,504
VOC reductions are shown in Table 7-17 for Calendar Years 2017 through 2030, and
2050 when the fleet will have turned over completed to Tier 3 standards. We project reductions
of over 40,000 tons (3 percent of the onroad fleet emissions) immediately upon implementation
of the program, growing to a 23 percent reduction in onroad emissions by 2030, and 36 percent
reduction in onroad emissions with full fleet turnover by 2050.
N This is an approximate breakdown, as there will be some NOX emission reduction from heavy-duty gasoline
vehicles greater than 14,000 pounds beyond the 2017 model year that are counted in the "Tier 3 fleet" here
7-41
-------
Table 7-17 Tier 3 VOC Reductions by Calendar Year (Annual U.S. Short Tons)
CY
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2050
Onroad
mobile
reference
1,614,384
1,506,001
1,411,787
1,333,867
1,266,250
1,211,710
1,163,986
1,123,476
1,084,012
1,045,808
1,025,657
1,004,922
988,529
977,067
1,151,908
Onroad
mobile with
control
1,569,603
1,456,342
1,356,874
1,272,497
1,195,112
1,128,740
1,067,208
1,012,446
955,680
899,298
858,927
818,008
781,665
751,040
735,548
Reduction
44,782
49,659
54,913
61,370
71,138
82,970
96,778
111,029
128,332
146,509
166,730
186,913
206,863
226,028
416,361
Percent
reduction in
onroad
2.8%
3.3%
3.9%
4.6%
5.6%
6.8%
8.3%
9.9%
11.8%
14.0%
16.3%
18.6%
20.9%
23.1%
36.1%
Table 7-18 shows the VOC reductions in 2017 and 2030 split into those attributable to
the pre-Tier 3 fleet, and the Tier 3 fleet. The Tier 3 fleet reductions are further subdivided into
the contribution of the proposed exhaust and evaporative standards. In 2017 over 80 percent of
the program reductions are coming from lower sulfur gasoline on the fleet already on the road.
By 2030, over 90 percent of the reduction is coming from 2017 and later model year vehicles,
with remaining reduction coming from lower sulfur fuel on pre-Tier 3 vehicles. The evaporative
standards account for about one quarter of the Tier 3 fleet reductions in 2030.
Table 7-18 Projected VOC Reductions from Tier 3 Program (Annual U.S. Short Tons)
Total reduction
Reduction from pre-Tier 3
fleet due to sulfur standard
Reduction from Tier 3 fleet
due to vehicle and sulfur
standards
Exhaust
Evaporative
2017
44,782
39,561
5,222
41, 433
3,349
2030
226,028
13,739
212,289
168, 264
57,764
CO reductions are shown in Table 7-19 for Calendar Years 2017 through 2030, and 2050
when the fleet will have turned over completed to Tier 3 standards. We project significant
reductions immediately upon implementation of the program, growing to a 30 percent reduction
in onroad emissions by 2030, and 46 percent reduction in onroad emissions with full fleet
turnover by 2050.
7-42
-------
Table 7-19 Tier 3 CO Reductions by Calendar Year (Annual U.S. Short Tons)
CY
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2050
Onroad
mobile
reference
20,915,593
20,297,089
19,782,476
19,374,712
19,122,796
18,830,157
18,736,417
18,647,649
18,647,177
18,621,982
18,699,029
18,745,355
18,850,303
18,951,626
24,839,365
Onroad
mobile with
control
20,168,910
19,454,190
18,830,656
18,295,010
17,747,325
17,119,383
16,610,543
16,095,479
15,584,593
15,054,696
14,558,766
14,047,003
13,594,884
13,186,263
13,339,791
Reduction
746,683
842,899
951,820
1,079,702
1,375,471
1,710,775
2,125,874
2,552,170
3,062,585
3,567,286
4,140,263
4,698,352
5,255,420
5,765,362
11,499,574
Percent
reduction
in onroad
3.6%
4.2%
4.8%
5.6%
7.2%
9.1%
11.3%
13.7%
16.4%
19.2%
22.1%
25.1%
27.9%
30.4%
46.3%
Table 7-20 shows the reductions for CO, broken down by pre- and post-Tier 3 in the
manner described for NOx and VOC above. The immediate reductions in the onroad fleet from
sulfur control comprise about 80 percent of total reductions in 2017. By 2030, the proposed
vehicle standard, enabled by low sulfur fuel, are accounting for 98 percent of program
reductions. Of the Tier 3 vehicle standard reductions in 2030, we estimate that about 5 percent
are contributed by the proposed heavy-duty tailpipe standards.
Table 7-20 CO Reductions from Tier 3 Program (Annual U.S. Short Tons)
Total reduction
Reduction from pre-Tier 3
fleet due to sulfur standard
Reduction from Tier 3 fleet
due to vehicle and sulfur
standards
2017
746,683
608,502
138,181
2030
5,765,362
139,074
5,626,288
Direct PM2.5 impacts are shown in Table 7-21 for calendar years 2017 through 2030, and
2050 when the fleet will have turned over completed to Tier 3 standards. For direct PM, the
impact shown is solely from the proposed tailpipe standards. Thus, unlike other pollutants,
reductions do not become significant until the fleet has turned over to cleaner vehicles. By 2030,
we project a reduction of about 7,500 tons annually, which represents approximately 10 percent
of the onroad direct PIVb.s inventory. However, since the PM standards are mainly focused on
improving engine durability through the end of a vehicle's useful life, the relative reduction in
onroad emissions is projected to grow to 17 percent with full fleet turnover in 2050.
7-43
-------
As discussed in Section 7.2.1.1, the control scenario emissions inventory prepared for air
quality modeling included the impact of an increase in aromatics as sulfur is reduced from 30
ppm to 10 ppm fuel. While this assumption had a minor effect on control case emissions for
several pollutants, the effect was more visible for direct PM2.5 emissions, as it resulted in a
projected increase in emissions (roughly 700 tons nationwide) in 2017. As discussed in Section
7.1.3.2.2, this emissions increase results from a series of conservative assumptions and
uncertainties related to fuel parameters in 2017, and is not expected to occur in reality. As a
result, we have not included the PM emissions increase from sulfur reduction in the emission
inventory impacts shown in Table 7-21.
Table 7-21 Tier 3 PM2.5 Reductions by Calendar Year (Annual U.S. Short Tons)
CY
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2050
Onroad mobile
reference
115,098
107,295
100,885
95,192
90,480
85,144
81,859
78,955
76,935
75,204
73,880
72,504
71,990
71,554
92,895
Onroad mobile
with control
114,977
106,932
100,162
93,992
88,767
82,943
79,081
75,561
72,890
70,502
68,496
66,431
65,210
64,096
77,279
Reduction
121
362
724
1,200
1,713
2,201
2,778
3,393
4,045
4,702
5,384
6,073
6,780
7,458
15,616
Percent
reduction in
onroad
0.1%
0.3%
0.7%
1.3%
1.9%
2.6%
3.4%
4.3%
5.3%
6.3%
7.3%
8.4%
9.4%
10.4%
16.8%
Emissions of air toxics also would be reduced by the proposed sulfur, exhaust and
evaporative standards. Air toxics are generally a subset of compounds making up VOC, so the
reduction trends tend to track the VOC reductions presented above. Table 7-22 presents
reductions for certain gaseous air toxics and polycyclic aromatic hydrocarbons (PAHs)°,
reflecting reductions of a few percent in 2017, and 20 to 40 percent of onroad emissions,
depending on the individual pollutant, in 2030.
Table 7-22 Reductions for Certain Individual Compounds (Annual U.S. Short Tons)
2017
Percent reduction 2030 Reduction Percent reduction
0 PAHs represents the sum of the following 15 PAH compounds: acenaphthene, acenaphthalene, anthracene,
benz(a)anthracene, benzo(a)pyrene, benzo(b)fluoranthene, benze(g,h,i)perylene, benzo(k)fluoranthene, chrysene,
dibenzo(a,h)anthracene, fluoranthene, fluorine, indeno(l,2,3,cd)pyrene, phenanthrene, and pyrene. These PAHs are
included inEPA's national emissions inventory (NEI).
7-44
-------
Acetaldehyde
Formaldehyde
Acrolein
1,3 -Butadiene
Benzene
Naphthalene
Ethanol
2,2,4-
Trimethylpentane
Ethyl Benzene
Hexane
Propi onal dehy de
Styrene
Toluene
Xylene
PAHs
Reduction
762
727
23
322
1,625
96
2,684
840
724
857
29
98
3,504
2,856
7
in onroad
3%
3%
1%
5%
4%
2%
2%
2%
3%
3%
2%
4%
3%
3%
1%
4,414
2,707
184
1,087
8,581
420
27,821
5,616
3,204
3,525
102
755
16,965
14,238
67
in onroad
26%
12%
15%
37%
36%
17%
24%
18%
22%
30%
20%
38%
22%
22%
17%
The totals shown in Table 7-23 represent the sum of all toxic species in Table 7-22. As
shown, in 2030 the overall on-road inventory of total toxics would be reduced by over 20
percent, with nearly one third of the vehicle program reductions coming from the proposed
evaporative standards.
Table 7-23 Reductions in Total Mobile Source Air Toxics (Annual U.S. Short Tons)
Total reduction
Reduction from pre-Tier 3 fleet due to
sulfur standard
Reduction from Tier 3 fleet due to
vehicle and sulfur standards
Exhaust
Evaporative
2017
15,156
13,184
1,972
13, 748
1,408
2030
89,685
5,022
84,663
64,144
25,541
SC>2 emissions from mobile sources are a direct function of sulfur in the fuel, and
reducing sulfur in gasoline would result in immediate reductions in 862 from the on and off-road
fleet. The reductions, shown in Table 7-24, represent a roughly 50 percent reduction in onroad
SC>2 emissions. The breakdown of the relative contribution of on-road vehicles and off-road
equipment is shown; the contribution of off-road sources is a function of off-road gasoline
consumption accounting for approximately 5 percent of overall gasoline use.36
7-45
-------
Table 7-24 Projected SOi Reductions from Tier 3 Program (Annual U.S. Short Tons)
Total reduction
Reduction from onroad
vehicles due to sulfur
standard
Reduction from off-road
equipment due to sulfur
standard
Percent reduction in
onroad SOi emissions
2017
16,261
15,494
767
51%
2030
17,267
16,370
897
51%
7.2 Criteria and Toxic Pollutant Air Quality Impacts
7.2.1 Emission Inventories for Air Quality Modeling
To estimate the benefits of the proposed Tier 3 rule, we performed air quality modeling
for the years 2017 and 2030. As noted in Section 7.1, emission inventories for air quality
modeling were required for the entire U.S. by 12 km grid cell and hour of the day for each day of
the year, requiring a methodology with much greater detail than the national emission inventories
presented above. While most of the modeling tools and inputs used for estimating national
emission inventories were also used in developing inputs for air quality modeling, the application
of these tools (particularly MOVES) to produce the gridded / hourly emissions was quite
different, and in essence a separate analysis. As explained in Section 7.2.1.1, the different
analyses generated different onroad inventory totals, but the relative reduction from reference to
control scenarios was consistent. The summary of the methodology for each sector is contained
in the following sections; for brevity, details of the process for developing air-quality ready
emission inventories are available in a separate technical support document.37
7.2.1.1 Onroad Emissions
For the onroad vehicle emissions inputs to our air quality modeling, we used an emission
inventory approach that provided more temporal and geographical resolution than the approach
used for the national inventories described above. While modeling at this level is time-
consuming and resource intensive, this detail is needed when generating inputs to air quality
models because it allows us much more precision in accounting for local ambient temperatures
and local fuel properties in our air quality modeling. For this purpose, we used county-specific
inputs and tools that integrated the MOVES model of onroad emissions with the Sparse Matrix
Operator Kernel Emissions tool (SMOKE) emission inventory model to take advantage of the
gridded hourly temperature information used in air quality modeling.
In particular, we used an automated process to run MOVES to produce emission factors
by temperature and speed for the fleet mix, fuels, and I/M program for more than 100
"representing counties", to which every other county could be mapped. The emission factors
then were multiplied by activity at the grid-cell-hour level to produce gridded hourly emissions
7-46
-------
for the entire continental U.S. These emissions were input into the Community Multiscale Air
Quality Modeling System (CMAQ). We summarize this approach in the sections below.
We used the same approach to model both the reference and the control cases, except, for
the Tier 3 control case, we used the Tier 3 emission rates and fuels developed for the national
inventories and described in Section 7.1.3.
Because of differences in methodology, particularly the treatment of vehicle speed
distributions and the non-linear temperature effects in MOVES, the more detailed approach used
for the air quality inventory produced different emission estimates than those described in the
national inventory section above. The two sets of results are compared in Table 7-25 below.
Table 7-25 Comparison of
Inventories
Calendar Year 2030 Onroad Emission National Inventories and
Used for Air Quality Modeling [U.S. Short tons]
Pollutant
NOX
VOC
CO
PM2.5
Benzene
Ethanol
Acrolein
1,3-Butadiene
Formaldehyde
Acetaldehyde
S02
Reference
National
Inventory
1,890,403
977,067
18,951,626
71,554
23,654
107,912
1,223
2,915
21,967
16,757
31,983
Air Quality
Inventory
1,846,571
911,513
17,021,674
88,516
22,221
116,762
863
2,932
14,810
13,926
30,526
Difference
AQ vs. NI
-2%
-7%
-10%
24%
-6%
8%
-29%
1%
-33%
-17%
-5%
Control
National
Inventory
1,365,613
751,040
13,186,263
64,917
15,073
80,091
1,039
1,828
19,260
12,343
15,613
Air Quality
Inventory
1,371,925
699,592
11,984,061
83,842
14,352
89,574
699
1,955
12,270
9,946
15,068
Difference
AQ vs. NI
0%
-7%
-9%
29%
-5%
12%
-33%
7%
-36%
-19%
-3%
The differences between the national inventories and air quality inventories reflect the
non-linear response to the more detailed handling of temperature and other local variables such
as speed in the air quality inventory; this is pronounced in pollutants with strong temperature
sensitivities in MOVES, such as PM2 5, where the finer temperature resolution in the air quality
approach produced significantly higher emissions than the aggregate national inventory
approach.
Because the reference and control case emissions rates were the same for the national
inventory and air quality inventory runs, the percent reductions due to the proposed Tier 3 rule
are very similar, as shown in Table 7-26. The exception is PM2 5, where the air quality inventory
shows a slight increase in emissions in 2017. As discussed in Section 7.1.3.2.2, this increases
resulted from a series of conservative assumptions and uncertainties related to fuel parameters in
2017 which we do not expect to occur in reality.
7-47
-------
Table 7-26 Comparison of Emission Reductions from Reference to Control Case in
"National" and "Air Quality" Onroad Inventories
Pollutant
NOX
VOC
CO
PM2.5
Benzene
Ethanol
Acrolein
1,3 -Butadiene
Formaldehyde
Acetaldehyde
S02
2017
National
Inventory
Reduction
-8%
-3%
-4%
-0.1%
-4%
-3%
-1%
-5%
-2%
-3%
-51%
Air Quality
Inventory
Reduction
-7%
-3%
-3%
1%
-4%
-2%
-1%
-5%
-3%
-3%
-50%
2030
National Inventory
Reduction
-28%
-25%
-30%
-10%
-36%
-26%
-15%
-37%
-12%
-26%
-51%
Air Quality
Inventory
Reduction
-26%
-23%
-30%
-5%
-35%
-23%
-19%
-33%
-17%
-29%
-51%
The following sections summarize the analysis done to generate the air quality
inventories.
7.2.1.1.1 Representing Counties
Air quality modeling requires emission inventories for nearly all of the more than 3,000
counties in the United States. Although EPA compiles county-specific databases for all counties
in the nation, actual county-specific data is rare. Instead, much of our "county" data is based on
state-wide estimates or national defaults. For this proposal, rather than explicitly model every
county in the nation, we have done detailed modeling for some counties and less detailed
estimates for the other counties.
In this approach, we group counties that have similar properties that would result in
similar emission rates. We explicitly model only one county in the group (the "representing"
county) to determine emission rates. These rates are then used in combination with county
specific activity and meteorology data, to generate inventories for all of the counties in the
group. This approach dramatically reduces the number of modeling runs required to generate
inventories and still takes into account differences between counties.
More specifically, the representing counties are chosen so they can be used to compute
g/mi factors that will be representative across the group of counties. To assure this, the counties
are grouped based on fuel parameters, emission standards, I/M programs and altitude. However,
representative counties are not meant to represent VMT. VMT is estimated for every
Continental U.S. county. As explained in Section 7.2.1.1.3, the SMOKE model calculates
emissions by multiplying the county-specific VMT by the county-group specific g/mi emission
rates produced in the MOVES run. The characteristics used to group the counties are
summarized in Table 7-27 below.
7-48
-------
Table 7-27 Characteristics for Representing County Groupings
County Grouping Characteristic
PADD
Fuel Parameters
Emission Standards
Inspection/Maintenance Programs
Altitude
Description
Petroleum Administration for Defense Districts (PADDs).
PADD 1 is divided into three sub-PADD groupings and
each sub-group is treated as a separate PADD (la, Ib and
Ic). Each state belongs to a PADD and all counties in any
state are within the same PADD.
Average gasoline fuel properties for January and July
2005, including RVP, sulfur level, ethanol fraction and
percent benzene
Some states have adopted California highway vehicle
emission standards or plan to adopt them. Since
implementation of the standards varies, each state with
California standards is treated separately.
Counties were grouped within a state according to whether
or not they had an I/M program. All I/M programs within
a state were considered as a single program, even though
each county may be administered separately and have a
different program design.
Counties were categorized as high or low altitude based
on the criteria set forth by EPA certification procedures
(4,000 feet above sea level).
The result is a set of 106 county groups with similar fuel, emission standards, altitude and
I/M programs. For each group, the county with the highest VMT was chosen as the representing
county. Of these, only 103 were needed to model the 48 states included in the air quality
analysis inventory.
For each county group, SMOKE-MOVES generated a set of rates that varied by vehicle
type, speed and temperature, thus we did not need to consider the fleet mix, speed or temperature
range in our grouping characteristics. This greatly increases the number of counties that can be
in each grouping, and reduces the number of MOVES runs required.
More detail on the process for selecting representative counties and a list of all of the
3,322 counties in the nation and the counties selected to represent is provided in the emission
inventory technical support document.
38
7.2.7.7.2 SMOKE-MOVES
The official EPA highway vehicle emissions model (MOVES) was updated as described
in Section 7.1.3 for national emission inventory development, but in order to take advantage of
the gridded hourly temperature information used in air quality modeling, MOVES and SMOKE
have been integrated into an inventory generation system called SMOKE-MOVES.39 MOVES
can be run in "inventory mode" to calculate the mass of pollutant emissions, as was done for the
national inventories, or in "emission rate" mode, in which it calculates emissions in grams per
7-49
-------
mile (for running emissions) or grams per vehicle (for start and evaporative emissions). For our
air quality runs, we used the rates approach. This creates a set of "lookup tables" with emission
rates by temperature, speed, pollutant, and vehicle class (Source Classification Code (SCC)).
SMOKE then transforms these rates into emission inventories for the air quality modeling by
multiplying these emission factors by activity specific to each grid cell hour,40
The SMOKE-MOVES process generates MOVES run specification files to produce the
emission rate lookup tables (in MOVES there are three per run to cover all emission processes:
Rate Per Distance, Rate per Vehicle, and Rate per Profile) covering the range of temperatures
needed, across each combination of fuel and I/M program in the nation. For a given scenario,
this resulted in over 16,000 run specification files. A series of post-processing scripts were
developed to take the raw MOVES emission rate table results and translate it into the emission
rates tables needed by SMOKE to produce mass emissions by 12 km grid and hour of the day,
for an entire year. Note, an update to these post-processing scripts was made between the
reference case and control case runs, which inadvertently introduced a small inconsistency in
emission rates between reference and control for about one-third of the counties. Specifically,
there were 1,218 counties, out of 3,109 total counties, which were impacted. The result was that
for some counties, control case emission rates were a few percent higher than they should have
been; fixing this error would increase the magnitude of reduction in the air quality analysis.41
For expediency, MOVES lookup tables were generated for July and January to get the
full range of temperatures needed for an entire year's worth of meteorology data. This efficiency
step introduces uncertainty because it does not account for fuel "shoulder" seasons in the fall and
spring, where the actual fuel pool is a blend of winter and summer fuel. This is mainly an issue
for fuel RVP, which is not changing between the reference and control scenarios.
7.2.1.1.1 Inputs to MOVES
The county-level fuel-property inputs for the air quality runs were the same as for the
national inventories described in Section 7.1.3. However, for the air quality runs, we were able
to use grid-level temperatures. We also needed county-specific information on vehicle
populations, age distributions, and inspection-maintenance programs for each of the representing
counties. The source data for each of these inputs is described below.
7.2.1.1.1.1 Temperature and Humidity
Ambient temperature can have a large impact on emissions. Cold temperatures are
associated with high start emissions for many pollutants. High temperatures are associated with
greater running emissions due to the higher engine load of air conditioning. High temperatures
also are associated with higher evaporative emissions. And, of course, the interaction between
emissions and ambient temperatures is an important consideration in air quality modeling. Thus
accurately accounting for ambient temperatures was important for our air quality modeling work.
The gridded meteorological input data for the entire year of 2005 were derived from
simulations of the Pennsylvania State University / National Center for Atmospheric Research
Mesoscale Model.42 This model, commonly referred to as MM5, is a limited-area,
nonhydrostatic, terrain-following system that solves for the full set of physical and
7-50
-------
thermodynamic equations which govern atmospheric motions. A description of how this tool
was used to determine temperatures is found in the documentation for the recent Heavy-Duty
Greenhouse Gas rule.43
SMOKE-MOVES uses the MM5 temperatures and the county groups described in
Section 7.2.1.1.1 to generate a list of all the possible temperatures and temperature profiles that
are needed in the lookup tables. (Temperature profiles are vectors of 24 temperatures that
describe how temperatures change over a day. They are needed to estimate vapor venting
emissions.) To do this, SMOKE-MOVES determines the minimum and maximum temperatures
in the county group for January and for July, and the minimum and maximum temperatures for
each hour of the day. It then generates a list of all possible temperatures between these limits,
using a five degree Fahrenheit interval. The model also uses these temperatures (using a 10
degree interval) to develops a collection of possible temperature profiles.
SMOKE-MOVES then runs MOVES for each of the listed temperatures and temperature
profiles, generating emission rate look-up tables that cover the desired temperature ranges.
Finally, the original grid cell temperatures are used to find the appropriate emission rate for each
cell. The 2005 temperatures were used for all scenarios.
The treatment of humidity is simpler. SMOKE-MOVES calculates an average day-time
(6am to 6pm) relative humidity for the county group for July and for January. The appropriate
(July or January) humidity is used for all runs of the county group.
7.2.1.1.1.2 Vehicle Population Inputs
Vehicle population data is a required input for MOVES when modeling on a county
basis. Using the technical guidance provided to states by EPA, a contractor generated
appropriate estimates for vehicle populations for use in the MOVES databases using the county
specific VMT and national average ratios of vehicle populations versus vehicle VMT from the
MOVES application. This method is described in Section 3.3 of the document, "Technical
Guidance on the Use of MOVES2010 for Emission Inventory Preparation in State
Implementation Plans and Transportation Conformity" (EPA-420-B-10-023, April 2010), which
is available on the EPA web site at: http://www.epa.gov/otaq/models/moves/index.htm
7.2.1.1.1.3 Other Local Inputs
In addition to temperature, vehicle population and fuels, we also needed inputs such as
age distribution and Inspection Maintenance program descriptions for each of the representing
counties. These inputs are required for the model to run at the county level and provided an
opportunity to assure that the model was properly accounting for the most recent available local
data. These county inputs were derived from the inputs used for the National Emissions
Inventory (NEI). This inventory covers the 50 United States (U.S.), Washington DC, Puerto
Rico and U.S. Virgin Islands. The NEI was created by the U.S. Environmental Protection
Agency's (EPA's) Emission Inventory Group (EIG) in Research Triangle Park, North Carolina,
in cooperation with the Office of Transportation and Air Quality in Ann Arbor, Michigan. The
inputs for the NEI are stored in the National Mobile Inventory Model (NMIM) county database
7-51
-------
(NCD). Details of how the NCD was developed are documented for the NEI.44 These inputs
were then converted to a format consistent with MOVES.
7.2.1.1.2 Parallel Processing on the "Cloud"
Providing the level of detail desired for the air quality modeling required an enormous
amount of data. Even with the "representing county" approach, we ran MOVES over 83,000
times to support Tier3 rule making. Processing just one of the Tier 3 scenarios required 16,604
runs, which if run serially would take over 200 days. Early on, we recognized that this would be
infeasible, even with the fastest computers available to us. Therefore, we developed a Linux-
based environment at Amazon Web Services that enabled us to process the Tier 3 base case in
less than 48 hours. We split the 16,604 runs into 206 batches (103 representative counties, for
January and July). We then ran 206 Linux processor instances in parallel, with each instance
processing 80 - 120 individual MOVES runs.
7.2.1.1.3 VMT, Population, and Speed
In addition to the lookup tables, SMOKE requires county VMT, population, and average
speed by road type to calculate the necessary emissions for air quality modeling.
VMT by county and Source Classification Code (SCC) was developed using
MOVES2010a and the National County Database. MOVES2010a has the EPA's most recent
projections of VMT growth at the national level, based on estimates from the Annual Energy
Outlook45. The National County Database (NCD20101201)46 has our most recent estimates of
2005 VMT and our best estimates of allocation of VMT from national to the county level.
Accordingly, for the 2005 base year, our estimates of VMT by county and SCC were taken
directly from the NCD. For the 2017 and 2030 inventories, we ran MOVES2010a with default
inputs to generate total national VMT by SCC. But, because MOVES uses a static (1999)
default allocation of VMT to county, we did not use MOVES for these allocations. Instead, the
2017 county VMT was created by interpolating between the NCD VMT values for 2015 and
those for 2020 and computing the NCD fraction for each county, then multiplying these fractions
by the MOVES VMT. The 2030 county allocation was computed similarly, using the NCD
VMT for 2030. The VMT was also adjusted to account for increased onroad transportation of
ethanol fuels and the resulting increase in travel by large tanker trucks. For both the reference
and control scenarios, impacts of this activity on emissions from tank trucks (Class 8) are
accounted by adjusting VMT used in SMOKE-MOVES.47 The VMT adjustments were derived
from the Oak Ridge National Laboratory analysis of ethanol transport, scaled to account for the
ethanol volumes. Nationwide impacts of this adjustment on VMT are small - less than 0.03
percent of total diesel truck VMT in 2017, and less than 0.05 percent in 2030.
Vehicle populations by county and SCC were developed similarly to the VMT, using
MOVES to generate national totals for each year and using the NCD to allocate to county.
However, the NCD does not include population estimates, so we used MOVES to generate the
2005 national population and we assumed that, for each calendar year (2005, 2017 and 2030) and
for each SCC, the allocation of national vehicle population to county was proportional to the
allocation of VMT (summed across roadtypes).
7-52
-------
The average speeds provided to SMOKE for each county were derived from the default
national average speed distributions found in the default MOVES2010a database
AvgSpeedDistribution table. These average speeds are the average speeds developed for the
,10
previous EPA highway vehicle emission factor model, MOBILE6. In MOVES, there is a
distribution of average speeds for each hour of the day for each road type. The average speeds in
these distributions were used to calculate an overall average speed for each hour of the day.
These hourly average speeds were weighted together using the default national average hourly
vehicle miles traveled (VMT) distribution found in the MOVES default database
HourlyVMTFraction table, to calculate an average speed for each road type. This average speed
by road type was provided to SMOKE for each county.
7.2.1.2 Nonroad Emissions
The "primary" nonroad emissions used in air quality modeling are identical to those used
for national inventories as presented in Section 7.1.4 above. The NMEVI model was run to
generate county-month inventories by SCC, which were processed to gridded-hourly emissions
by SMOKE. For more details on SMOKE processing of nonroad emissions, see the emissions
modeling technical support document.
49
Table 7-28 Comparison of Calendar Year 2030 Nonroad Emission National Inventories
and Inventories Used for Air Quality Modeling [U.S. Short tons]
Pollutant
NOX
VOC
CO
PM2.5
Benzene
Acrolein
1,3-Butadiene
Formaldehyde
Acetaldehyde
SO2
Reference
National
Inventory
765,026
1,209,534
12,921,772
68,308
23,246
571
1,772
13,580
7,899
3,154
Air Quality
Inventory
765,026
1,209,534
12,921,772
68,308
23,390
570
1,774
13,522
7,888
3,154
Difference
0.0%
0.0%
0.0%
0.0%
0.6%
-0.2%
0.1%
-0.4%
-0.1%
0.0%
Control
National
Inventory
765,026
1,209,452
12,921,772
68,308
23,245
571
1,772
13,580
7,899
2,257
Air Quality
Inventory
765,026
1,209,452
12,921,772
68,308
23,389
570
1,774
13,522
7,888
2,257
Difference
0.0%
0.0%
0.0%
0.0%
0.6%
-0.2%
0.1%
-0.4%
-0.1%
0.0%
7.2.1.:
Refueling
7.2.1.3.1 Methodology
This section describes how the emission inventories for refueling from on-road vehicles
in calendar years 2017 and 2030 for Tier 3 reference and control cases were generated for air
quality modeling. The refueling inventory includes emissions from spillage loss and
displacement vapor loss. The displacement emissions vary from scenario to scenario depending
on the RVP of the modeled fuels. For this analysis, the refueling emissions were estimated using
the revised version of EPA's Motor Vehicle Emissions Simulator (MOVES2010a) at the county
7-53
-------
level for all twelve months. We used the same fuel supply inputs as we used for the onroad
inventories described in Section 7.1.3.2.
As for onroad emissions, described above, we used a "representing county" approach to
reduce MOVES runtime. Additional information on the use of representing counties to model
refueling emissions is available in a technical support document.
50
7.2.1.3.2 Emission Inventory Results
The annual refueling emission inventories for air quality modeling for the lower 48 states
are shown in Table 7-29, along with the percent changes between the two scenarios. The Tier 3
proposal to eliminate the RVP waiver for El0 fuels reduces refueling emissions in 2017, but has
no impact in 2030, where we model El5 as the only gasoline available for onroad vehicles in
both the reference and control cases.
Table 7-29 Refueling Emissions for Tier 3 reference and control (U.S. Short Tons)
voc
Benzene
Ethanol
2017
Reference
75,397
158
8,582
Control
74,235
160
8,426
Percent
Reduction
-1.5%
1.6%
-1.8%
2030
Reference
56,405
90
7,143
Control
56,402
90
7,142
Percent
Reduction
0.0%
0.0%
0.0%
7.2.1.4 Portable Fuel Container and Upstream Emissions
The Tier 3 proposed rule has no impact on portable fuel container (PFC) emissions. The
proposed standards are also not expected to impact upstream emissions associated with fuel
transport/distribution. For fuel production, the results of our refinery permitting analysis
described in Section V.B. of the preamble and Chapter 4 of the RIA project minor emissions
increases at some refineries due to the reductions in fuel sulfur content that would be required by
the proposed Tier 3 standards. We did not include these emission impacts in our modeling
because the projected increases are small and may be even less than projected if refineries apply
emissions controls to reduce emissions increases.
Although there is no modeled impact of the proposed Tier 3 standards on upstream or
PFC emissions a significant number of modifications were made to the 2005v.4.2 platform
inventory to account for impacts of renewable fuel requirements under EISA in the reference
case air quality inventory. These modifications are described in detail in a memorandum to the
docket.51 Modifications to point and nonpoint inventories include adjustments to agricultural
emissions, increases in emissions associated with production of corn ethanol, cellulosic ethanol,
cellulosic diesel, and biodiesel, decreases in petroleum refinery emissions to account for gasoline
displacement, and increased vapor loss emissions from transport of ethanol and gasoline/ethanol
fuel blends. Modifications to mobile source inventories include increases in combustion
emissions from water, rail and truck transport of biofuels. PFC emissions were adjusted to
account for impacts of RVP changes associated with use of gasoline/ethanol blends.
7-54
-------
7.2.1.5 Hydrocarbon Speciation Profiles and SMOKE
We used the Community Multi-scale Air Quality (CMAQ) model, described in detail in
the following section, to conduct air quality modeling for this analysis. The SMOKE tool is used
to process emission inventories for air quality modeling.1" Specifically, SMOKE converts our air
quality emissions inventories into CMAQ-ready inputs by transforming the emission inventories
based on the temporal allocation, chemical speciation, and spatial allocation requirements of
CMAQ. In processing our Tier 3 emissions inventories for CMAQ, SMOKE uses hydrocarbon
speciation profiles to break total hydrocarbons down into individual constituent compounds and
create the needed chemical speciation inputs required for CMAQ. Given the complexity of the
atmospheric chemistry, the hydrocarbon speciation can have an important influence on the air
quality modeling results. We recently created a number of new hydrocarbon speciation profiles
for vehicle exhaust and evaporative emissions and headspace vapor. These profiles were created
using data from the EPAct test program,52 CRC's E-77 series of evaporative emissions programs
described in Section 7.2 above,53'54 and recent measurements of speciated gasoline headspace
vapors collected by EPA's Office of Research and Development (ORD).55 Mobile source
hydrocarbon speciation profiles used in this analysis are listed in a memo to the docket.56
7.2.2 Air Quality Modeling Methodology
Air quality models use mathematical and numerical techniques to simulate the physical
and chemical processes that affect air pollutants as they disperse and react in the atmosphere.
Based on inputs of meteorological data and source information, these models are designed to
characterize primary pollutants that are emitted directly into the atmosphere and secondary
pollutants that are formed as a result of complex chemical reactions within the atmosphere.
Photochemical air quality models have become widely recognized and routinely utilized tools for
regulatory analysis by assessing the effectiveness of control strategies. These models are applied
at multiple spatial scales - local, regional, national, and global. This section provides detailed
information on the photochemical model used for our air quality analysis (the Community Multi-
scale Air Quality (CMAQ) model), atmospheric reactions and the role of chemical mechanisms
in modeling, and model uncertainties and limitations. Further discussion of the modeling
methodology is included in the Air Quality Modeling Technical Support Document (AQM TSD)
found in the docket for this rule. Results of the air quality modeling are presented in Section
7.2.4.
7.2.2.1 Modeling Methodology
A national-scale air quality modeling analysis was performed to estimate future year 8-
hour ozone concentrations, annual PM2.5 concentrations, 24-hour PM2.5 concentrations, annual
NO2 concentrations, air toxics concentrations, visibility levels and nitrogen and sulfur deposition
levels for 2017 and 2030. The 2005-based CMAQ modeling platform was used as the basis for
the air quality modeling for this proposed rule. This platform represents a structured system of
connected modeling-related tools and data that provide a consistent and transparent basis for
' For more information, please see the website for SMOKE: http://www.smoke-model.org/index.cfm.
7-55
-------
assessing the air quality response to projected changes in emissions. The base year of data used
to construct this platform includes emissions and meteorology for 2005. The platform was
developed by the U.S. EPA's Office of Air Quality Planning and Standards in collaboration with
the Office of Research and Development and is intended to support a variety of regulatory and
research model applications and analyses.
The CMAQ modeling system is a non-proprietary, publicly available, peer-reviewed,
state-of-the-science, three-dimensional, grid-based Eulerian air quality model designed to
estimate the formation and fate of oxidant precursors, primary and secondary PM concentrations,
acid deposition, and air toxics, over regional and urban spatial scales for given input sets of
meteorological conditions and emissions.57'58'59 The CMAQ model version 4.7 was most
recently peer-reviewed in February of 2009 for the U.S. EPA.Q The CMAQ model is a well-
known and well-respected tool and has been used in numerous national and international
applications.60'61'62 This 2005 multi-pollutant modeling platform used the most recent CMAQ
code available at the time of air quality modeling (CMAQ version 4.7.1R) with a minor internal
change made by the U.S. EPA CMAQ model developers intended to speed model runtimes when
only a small subset of toxics species are of interest.
CMAQ includes many science modules that simulate the emission, production, decay,
deposition and transport of organic and inorganic gas-phase and particle-phase pollutants in the
atmosphere. We used CMAQ v4.7.1 which reflects updates to version 4.7 to improve the
underlying science. These include aqueous chemistry mass conservation improvements,
improved vertical convective mixing and lowered CB05 mechanism unit yields for acrolein from
1,3-butadiene tracer reactions which were updated to be consistent with laboratory
measurements. Section 7.2.3 of this draft RIA discusses the chemical mechanism and SOA
formation.
7.2.2.2 Model Domain and Configuration
The CMAQ modeling domain encompasses all of the lower 48 States and portions of
Canada and Mexico. The modeling domain is made up of a large continental U.S. 36 kilometer
(km) grid and two 12 km grids (an Eastern U.S. and a Western U.S. domain), as shown in Figure
7-11. The modeling domain contains 14 vertical layers with the top of the modeling domain at
about 16,200 meters, or 100 millibars (mb) of atmospheric pressure.
Q Allen, D., Burns, D., Chock, D., Kumar, N., Lamb, B., Moran, M. (February 2009). Report on the Peer Review of
the Atmospheric Modeling and Analysis Division, NERL/ORD/EPA. U.S. EPA, Research Triangle Park, NC.,
http://www.epa.gov/amad/peer/2009_AMAD_PeerReviewReport.pdf.
R CMAQ version 4.7 was released on December, 2008. It is available from the Community Modeling and Analysis
System (CMAS) as well as previous peer-review reports at: http://www.cmascenter.org.
7-56
-------
Figure 7-11 Map of the CMAQ Modeling Domain
7.2.2.3 Model Inputs
The key inputs to the CMAQ model include emissions from anthropogenic and biogenic
sources, meteorological data, and initial and boundary conditions. The CMAQ meteorological
input files were derived from simulations of the Pennsylvania State University/National Center
for Atmospheric Research Mesoscale Model63 for the entire year of 2005 over model domains
that are slightly larger than those shown in Figure 7-11. This model, commonly referred to as
MM5, is a limited-area, nonhydrostatic, terrain-following system that solves for the full set of
physical and thermodynamic equations which govern atmospheric motions.64 The meteorology
for the national 36 km grid and the two 12 km grids were developed by EPA and are described in
more detail within the AQM TSD. The meteorological outputs from MM5 were processed to
create model-ready inputs for CMAQ using the Meteorology-Chemistry Interface Processor
(MCIP) version 3.4. Outputs include: horizontal wind components (i.e., speed and direction),
temperature, moisture, vertical diffusion rates, and rainfall rates for each grid cell in each vertical
layer.
65
The lateral boundary and initial species concentrations are provided by a three-
dimensional global atmospheric chemistry model, the GEOS-CHEM model.66 The global
GEOS-CHEM model simulates atmospheric chemical and physical processes driven by
assimilated meteorological observations from the NASA's Goddard Earth Observing System
(GEOS). This model was run for 2005 with a grid resolution of 2 degree x 2.5 degree (latitude-
longitude) and 20 vertical layers. The predictions were used to provide one-way dynamic
boundary conditions at three-hour intervals and an initial concentration field for the 36 km
CMAQ simulations. The future base conditions from the 36 km coarse grid modeling were used
as the initial/boundary state for all subsequent 12 km finer grid modeling.
7-57
-------
The emissions inputs used for the 2005 base year and each of the future year base cases
and control scenarios analyzed for this rule are summarized in Section 7.2.1 of this draft RIA.
7.2.2.4 CMAQ Evaluation
An operational model performance evaluation for ozone, PM2.5 and its related speciated
components (e.g., sulfate, nitrate, elemental carbon, organic carbon, etc.), nitrate and sulfate
deposition, and specific air toxics (formaldehyde, acetaldehyde, benzene, 1,3-butadiene, and
acrolein) was conducted using 2005 state/local monitoring data in order to estimate the ability of
the CMAQ modeling system to replicate base year concentrations. Model performance statistics
were calculated for observed/predicted pairs of daily/monthly/seasonal/annual concentrations.
Statistics were generated for the following geographic groupings: domain wide, Eastern vs.
Western (divided along the 100th meridian), and each Regional Planning Organization (RPO)
Q
region. The "acceptability" of model performance was judged by comparing our results to those
found in recent regional PM2.s model applications for other, non-EPA studies.1 Overall, the
performance for the 2005 modeling platform is within the range or close to that of these other
applications. The performance of the CMAQ modeling was evaluated over a 2005 base case.
The model was able to reproduce historical concentrations of ozone and PM2.5 over land with
low bias and error results. Model predictions of annual formaldehyde, acetaldehyde and benzene
showed relatively small bias and error results when compared to observations. The model
yielded larger bias and error results for 1,3 butadiene and acrolein based on limited monitoring
sites. A more detailed summary of the 2005 CMAQ model performance evaluation is available
within the AQM TSD found in the docket of this rule.
7.2.2.5 Model Simulation Scenarios
As part of our analysis for this rulemaking, the CMAQ modeling system was used to
calculate 8-hour ozone concentrations, daily and annual PM2.5 concentrations, annual NO2
concentrations, annual and seasonal (summer and winter) air toxics concentrations, visibility
levels and annual nitrogen and sulfur deposition total levels for each of the following emissions
scenarios:
- 2005 base year
- 2017 Tier 3 reference case
- 2017 Tier 3 control case
- 2030 Tier 3 reference case
s Regional Planning Organization regions include: Mid-Atlantic/Northeast Visibility Union (MANE-VU), Midwest
Regional Planning Organization - Lake Michigan Air Directors Consortium (MWRPO-LADCO), Visibility
Improvement State and Tribal Association of the Southeast (VISTAS), Central States Regional Air Partnership
(CENRAP), and Western Regional Air Partnership (WRAP).
T These other modeling studies represent a wide range of modeling analyses which cover various models, model
configurations, domains, years and/or episodes, chemical mechanisms, and aerosol modules.
7-58
-------
- 2030 Tier 3 control case
The emission inventories used in the air quality and benefits modeling are different from
the proposed rule inventories due to the considerable length of time required to conduct the
modeling. As noted above, emission inventories for air quality modeling were required for the
entire U.S. by 12 km grid cell and hour of the day for each day of the year, requiring a
methodology of much greater detail than the national emission inventories presented in Section
7.1. While most of the modeling tools and inputs used for estimating national emission
inventories were also used in developing inputs for air quality modeling as well, the application
of these tools (particularly MOVES) to produce the gridded / hourly emissions was quite
different, and in essence a separate analysis. As explained in Section 7.2.1.1, the different
analyses generated different onroad inventory totals, but the reduction from reference to control
scenarios was consistent. The emission inventories used for air quality modeling are discussed
in Section 7.2.1 of this draft RIA. The emissions modeling TSD, found in the docket for this rule
(EPA-HQ-OAR-2011-0135), contains a detailed discussion of the emissions inputs used in our
air quality modeling.
We use the predictions from the model in a relative sense by combining the 2005 base-
year predictions with predictions from each future-year scenario and applying these modeled
ratios to ambient air quality observations to estimate 8-hour ozone concentrations, daily and
annual PM2.5 concentrations, annual NC>2 concentrations and visibility impairment for each of the
2017 and 2030 scenarios. The ambient air quality observations are average conditions, on a site-
by-site basis, for a period centered around the model base year (i.e., 2003-2007).
The projected daily and annual PM2.5 design values were calculated using the Speciated
Modeled Attainment Test (SMAT) approach. The SMAT uses a Federal Reference Method
(FRM) mass construction methodology that results in reduced nitrates (relative to the amount
measured by routine speciation networks), higher mass associated with sulfates (reflecting water
included in FRM measurements), and a measure of organic carbonaceous mass that is derived
from the difference between measured PM2.5 and its non-carbon components. This
characterization of PM2.5 mass also reflects crustal material and other minor constituents. The
resulting characterization provides a complete mass balance. It does not have any unknown
mass that is sometimes presented as the difference between measured PM2.5 mass and the
characterized chemical components derived from routine speciation measurements. However,
the assumption that all mass difference is organic carbon has not been validated in many areas of
the U.S. The SMAT methodology uses the following PM2.s species components: sulfates,
nitrates, ammonium, organic carbon mass, elemental carbon, crustal, water, and blank mass (a
fixed value of 0.5 |ig/m ). More complete details of the SMAT procedures can be found in the
report "Procedures for Estimating Future PM2.5 Values for the CAIR Final Rule by Application
of the (Revised) Speciated Modeled Attainment Test (SMAT)".67 For this latest analysis, several
datasets and techniques were updated. These changes are fully described within the technical
support document for the Final Transport Rule AQM TSD.68 The projected 8-hour ozone design
values were calculated using the approach identified in EPA's guidance on air quality modeling
attainment demonstrations. 9
Additionally, we conducted an analysis to compare the absolute and percent differences
between the future year reference and control cases for annual and seasonal ethanol,
7-59
-------
formaldehyde, acetaldehyde, benzene, 1,3-butadiene, and acrolein, as well as annual nitrate and
sulfate deposition. These data were not compared in a relative sense due to the limited
observational data available.
7.2.3 Chemical Mechanisms in Modeling
This rule presents inventories for NOx, VOC, CO, PM2.s, 862, NH3, and seven air
toxics: benzene, 1,3-butadiene, formaldehyde, acetaldehyde, ethanol, naphthalene and acrolein.
The air toxics are explicit model species in the CMAQv4.7 model with carbon bond 5 (CB05)
mechanisms.70 Emissions of all the pollutants included in the rule inventories, except ethanol,
were generated using the Motor Vehicle Emissions Simulator (MOVES) VOC emissions and
toxic-to-VOC ratios calculated using EPAct data.71 Ethanol emissions for air quality modeling
were based on speciation of VOC using different ethanol profiles (EO, E10 and E85) (see Section
7.2.1.5 for more information). In addition to direct emissions, photochemical processes
mechanisms are responsible for formation of some of these compounds in the atmosphere from
precursor emissions. For some pollutants such as PM, formaldehyde, and acetaldehyde, many
photochemical processes are involved. CMAQ therefore also requires inventories for a large
number of other air toxics and precursor pollutants. Methods used to develop the air quality
inventories can be found in Section 7.2.1.
In the CB05 mechanism, the chemistry of thousands of different VOCs in the
atmosphere are represented by a much smaller number of model species which characterize the
general behavior of a subset of chemical bond types; this condensation is necessary to allow the
use of complex photochemistry in a fully 3-D air quality model.72
Complete combustion of ethanol in fuel produces carbon dioxide (CO2) and water (F^O).
Incomplete combustion results in the production of other air pollutants, such as acetaldehyde and
other aldehydes, and the release of unburned ethanol. Ethanol is also present in evaporative
emissions. In the atmosphere, ethanol from unburned fuel and evaporative emissions can
undergo photodegradation to form aldehydes (acetaldehyde and formaldehyde) and peroxyacetyl
nitrate (PAN), and also plays a role in ground-level ozone formation. Mechanisms for these
reactions are included in CMAQ. Additionally, alkenes and other hydrocarbons are considered
because any increase in acetyl peroxy radicals due to ethanol increases might be counterbalanced
by a decrease in radicals resulting from decreases in other hydrocarbons.
CMAQ includes 63 inorganic reactions to account for the cycling of all relevant oxidized
nitrogen species and cycling of radicals, including the termination of NO2 and formation of nitric
acid (HNO3) without PAN formation.11
NO2 + -OH + M^HNO3+M k = 1.19 x 10'11 cm3moleculeV 73
The CB05 mechanism also includes more than 90 organic reactions that include alternate
pathways for the formation of acetyl peroxy radical, such as by reaction of ethene and other
r All rate coefficients are listed at 298 K and, if applicable, 1 bar of air.
7-60
-------
alkenes, alkanes, and aromatics. Alternate reactions of acetyl peroxy radical, such as oxidation
of NO to form NO2, which again leads to ozone formation, are also included.
Atmospheric reactions and chemical mechanisms involving several key formation
pathways are discussed in more detail in the following sections.
7.2.3.1 Acetaldehyde
Acetaldehyde is the main photodegradation product of ethanol, as well as other precursor
hydrocarbons. Acetaldehyde is also a product of fuel combustion. In the atmosphere,
acetaldehyde can react with the OH radical and O2 to form the acetyl peroxy radical
[CH3C(O)OO-].V When NOx is present in the atmosphere this radical species can then further
react with nitric oxide (NO), to produce formaldehyde (HCHO), or with nitrogen dioxide (NO2),
to produce PAN [CH3C(O)OONO2]. An overview of these reactions and the corresponding
reaction rates are provided below.w
CH3CHO + -OH -> CH3C-O + H2O k = 1.5 x 10'11 cn^molecule'V1 74
CH3C-O + O2 + M -> CH3C(O)OO + M
CH3C(O)OO +NO^ CH3C(O)O- + NO2 k = 2.0 x 10'11 cn^molecule'V1 75
CH3C(O)O- -> -CH3 + CO2
•CH3 + O2 + M -> CH3OO- + M
CH3OO- + NO -> CH3O- + NO2
CH3O- + O2 -> HCHO + HO2
CH3C(O)OO + NO2 + M -> CH3C(O)OONO2 + M k = 1.0 x 10'11 cn^molecule'V1 76
Acetaldehyde can react with the NO3 radical, ground state oxygen atom (O3P) and
chlorine, although these reactions are much slower. Acetaldehyde can also photolyze (hv),
which predominantly produces -CH3 (which reacts as shown above to form CH3OO-) and HCO
(which rapidly forms HO2 and CO):
CH3CHO + hv +2 O2 -> CH3OO- +HO2 + CO A, = 240-3 80 nm 77
As mentioned above, CH3OO- can react in the atmosphere to produce formaldehyde
(HCHO). Formaldehyde is also a product of hydrocarbon combustion. In the atmosphere, the
most important reactions of formaldehyde are photolysis and reaction with the OH, with
v Acetaldehyde is not the only source of acetyl peroxy radicals in the atmosphere. For example, dicarbonyl
compounds (methylglyoxal, biacetyl, and others) also form acetyl radicals, which can further react to form
peroxyacetyl nitrate (PAN).
w All rate coefficients are listed at 298 K and, if applicable, 1 bar of air.
7-61
-------
atmospheric lifetimes of approximately 3 hours and 13 hours, respectively.78 Formaldehyde can
also react with NC>3 radical, ground state oxygen atom (C^P) and chlorine, although these
reactions are much slower. Formaldehyde is removed mainly by photolysis whereas the higher
aldehydes, those with two or more carbons such as acetaldehyde, react predominantly with OH
radicals. The photolysis of formaldehyde is an important source of new hydroperoxy radicals
(HO2), which can lead to ozone formation and regenerate OH radicals.
HCHO + hv + 2 O2 -> 2 HO2 + CO A, = 240-360 nm 79
HO2 + NO -> NO2+ OH
Photolysis of HCHO can also proceed by a competing pathway which makes only stable
products: H2 and CO.
CB05 mechanisms for acetaldehyde formation warrant a detailed discussion given the
increase in vehicle and engine exhaust emissions for this pollutant and ethanol, which can form
acetaldehyde in the air. Acetaldehyde is represented explicitly in the CB05 chemical
mechanism80'81 by the ALD2 model species, which can be both formed from other VOCs and can
decay via reactions with oxidants and radicals. The reaction rates for acetaldehyde, as well as for
the inorganic reactions that produce and cycle radicals, and the representative reactions of other
VOCs have all been updated to be consistent with recommendations in the literature.82
The decay reactions of acetaldehyde are fewer in number and can be characterized well
because they are explicit representations. In CB05, acetaldehyde can photolyze in the presence
of sunlight or react with molecular oxygen (O3(P)), hydroxyl radical (OH), or nitrate radicals.
The reaction rates are based on expert recommendations,83 and the photolysis rate is from
IUPAC recommendations.
In CMAQ v4.7, the acetaldehyde that is formed from photochemical reactions is tracked
separately from that which is due to direct emission and transport of direct emissions. In CB05,
there are 25 different reactions that form acetaldehyde in molar yields ranging from 0.02 (ozone
reacting with lumped products from isoprene oxidation) to 2.0 (cross reaction of acylperoxy
radicals, CXOs). The specific parent VOCs that contribute the most to acetaldehyde
concentrations vary spatially and temporally depending on characteristics of the ambient air, but
alkenes in particular are found to play a large role. The IOLE model species, which represents
internal carbon-carbon double bonds, has high emissions and relatively high yields of
acetaldehyde. The OLE model species, representing terminal carbon double bonds, also plays a
role because it has high emissions although lower acetaldehyde yields. Production from
peroxyproprional nitrate and other peroxyacylnitrates (PANX) and aldehydes with 3 or more
carbon atoms can in some instances increase acetaldehyde but because they also are a sink of
radicals, their effect is smaller. Thus, the amount of acetaldehyde (and formaldehyde as well)
formed in the ambient air as well as emitted in the exhaust (the latter being accounted for in
emission inventories) is affected by changes in these precursor compounds due to the addition of
ethanol to fuels (e.g., decreases in alkenes would cause some decrease of acetaldehyde, and to a
larger extent, formaldehyde).
7-62
-------
The reaction of ethanol (CHaCH^OH) with OH is slower than some other important
reactions but can be an important source of acetaldehyde if the emissions are large. Based on
kinetic data for molecular reactions, the only important chemical loss process for ethanol (and
QA
other alcohols) is reaction with the hydroxyl radical ('OH). This reaction produces
acetaldehyde (CH3CHO) with a 90 percent yield.85 The lifetime of ethanol in the atmosphere can
be calculated from the rate coefficient, k, and due to reaction with the OH radical, occurs on the
order of a day in polluted urban areas or several days in unpolluted areas.x
In CB05, reaction of one molecule of ethanol yields 0.90 molecules of acetaldehyde. It
assumes the majority of the reaction occurs through H-atom abstraction of the more weakly-
bonded methylene group, which reacts with oxygen to form acetaldehyde and hydroperoxy
radical (HO2), and the remainder of the reaction occurs at the -CHS and -OH groups, creating
formaldehyde (HCHO), oxidizing NO to NO2 (represented by model species XO2) and creating
glycoaldehyde, which is represented as ALDX:
CH3CHOH + OH -> HO2 + 0.90 CH3CHO + 0.05 ALDX + 0.10 HCHO + 0.10 XO2
7.2.3.2 S econdary Organi c Aero sol s ( S O A)
Secondary organic aerosol (SOA) chemistry research described below has led to
implementation of new pathways for secondary organic aerosol (SOA) in CMAQ 4.7, based on
recommendations of Edney et al. and the recent work of Carlton et al.86'8? In previous versions
of CMAQ, all SOA was semivolatile and resulted from the oxidation of compounds emitted
entirely in the gas-phase. In CMAQ v4.7, parameters in existing pathways were revised and new
formation mechanisms were added. Some of the new pathways, such as low-NOx oxidation of
aromatics and particle-phase oligomerization, result in nonvolatile SOA.
Organic aerosol can be classified as either primary or secondary depending on whether it
is emitted into the atmosphere as a particle (primary organic aerosol, POA) or formed in the
atmosphere (SOA). SOA precursors include volatile organic compounds (VOCs) as well as low-
volatility compounds that can react to form even lower volatility compounds. Current research
suggests SOA contributes significantly to ambient organic aerosol (OA) concentrations, and in
Southeast and Midwest States may make up more than 50 percent (although the contribution
varies from area to area) of the organic fraction of PM2.s during the summer (but less in the
QO OQ
winter). ' A wide range of laboratory studies conducted over the past twenty years show that
anthropogenic aromatic hydrocarbons and long-chained alkanes, along with biogenic isoprene,
monoterpenes, and sesquiterpenes, contribute to SOA formation.90'91'92'93'94 Modeling studies, as
well as carbon isotope measurements, indicate that a significant fraction of SOA results from the
oxidation of biogenic hydrocarbons.95'96 Based on parameters derived from laboratory chamber
experiments, SOA chemical mechanisms have been developed and integrated into air quality
models such as the CMAQ model and have been used to predict OA concentrations.97
Over the past 10 years, ambient OA concentrations have been routinely measured in the
U.S. and some of these data have been used to determine, by employing source/receptor
All rate coefficients are listed at 298 K and, if applicable, 1 bar of air.
7-63
-------
methods, the contributions of the major OA sources, including biomass burning and vehicular
gasoline and diesel exhaust. Since mobile sources are a significant source of VOC emissions,
QR
currently accounting for almost 40 percent of anthropogenic VOC, mobile sources are also an
important source of SO A, particularly in populated areas.
Toluene is an important contributor to anthropogenic SOA. Mobile sources are the most
significant contributor to ambient toluene concentrations as shown by analyses done for the 2005
National Air Toxics Assessment (NATA)99 and the Mobile Source Air Toxics (MSAT) Rule.100
The 2005 NATA indicates that onroad and nonroad mobile sources accounted for almost 60
percent (1.46 |ig/m3) of the total average nationwide ambient concentration of toluene (2.48
|ig/m3), when the contribution of the estimated "background" is apportioned among source
sectors.
The amount of toluene in gasoline influences the amount of toluene emitted in vehicle
exhaust and evaporative emissions, although, like benzene, some toluene is formed in the
combustion process. In turn, levels of toluene and other aromatics in gasoline are potentially
influenced by the amount of ethanol blended into the fuel. Due to the high octane quality of
ethanol, it greatly reduces the need for and levels of other high-octane components such as
aromatics including toluene (which is the major aromatic compound in gasoline). Since toluene
contributes to SOA and the toluene level of gasoline is decreasing, it is important to assess the
effect of these reductions on ambient PM.
In addition to toluene, other mobile-source hydrocarbons such as benzene, xylene, and
alkanes form SOA. Similar to toluene, the SOA produced by benzene and xylene from low-NOx
pathways is expected to be less volatile and be produced in higher yields than SOA from high-
NOx conditions.101 Alkanes form SOA with higher yields resulting from the oxidation of longer
1 09
chain as well as cyclic alkanes.
It is unlikely that ethanol would form directly from SOA or affect SOA formation
indirectly through changes in the radical populations from increasing ethanol exhaust.
Nevertheless, scientists at the U.S. EPA's Office of Research and Development recently directed
experiments to investigate ethanol's SOA forming potential.103 The experiments were conducted
under conditions where peroxy radical reactions would dominate over reaction with NO (i.e.,
irradiations performed in the absence of NOx and OH produced from the photolysis of hydrogen
peroxide). This was the most likely scenario under which SOA formation could occur, since a
highly oxygenated C4 organic would be potentially made. As expected, no SOA was produced.
From these experiments, the upper limit for the aerosol yield would have been less than 0.01
percent based on scanning mobility particle sizer (SMPS) data. Given the expected negative
result based on these initial smog chamber experiments, these data were not published.
In general, measurements of organic aerosol represent the sum of POA and SOA and the
fraction of aerosol that is secondary in nature can only be estimated. One of the most widely
applied method of estimating total ambient SOA concentrations is the EC tracer method using
ambient data which estimates the OC/EC ratio in primary source emissions.104'105 SOA
concentrations have also been estimated using OM (organic mass) to OC (organic carbon) ratios,
which can indicate that SOA formation has occurred, or by subtracting the source/receptor-based
total primary organic aerosol (POA) from the measured OC concentration.106 Aerosol mass
7-64
-------
spectrometer (AMS) measurements along with positive matrix factorization (PMF) can also be
used to identify surrogates for POA and SOA in ambient as well as chamber experiments. Such
methods, however, may not be quantitatively accurate and provide no information on the
contribution of individual biogenic and anthropogenic SOA sources, which is critical information
needed to assess the impact of specific sources and the associated health risk. These methods
assume that OM containing additional mass from oxidation of OC comes about largely (or
solely) from SOA formation. In particular, the contributions of anthropogenic SOA sources,
including those of aromatic precursors, are required to determine exposures and risks associated
with replacing fossil fuels with biofuels.
Upon release into the atmosphere, numerous VOC compounds can react with free
radicals in the atmosphere to form SOA. While this has been investigated in the laboratory, there
is relatively little information available on the specific chemical composition of SOA compounds
themselves from specific VOC precursors. This absence of compositional data from the
precursors has largely prevented the identification of aromatically-derived SOA in ambient
samples which, in turn, has prevented observation-based measurements of the aromatic and other
SOA contributions to ambient PM levels.
As a first step in determining the ambient SOA concentrations, EPA has developed a
tracer-based method to estimate such concentrations.107'108 The method is based on using mass
fractions of SOA tracer compounds, measured in smog chamber-generated SOA samples, to
convert ambient concentrations of SOA tracer compounds to ambient SOA concentrations. This
method consists of irradiating the SOA precursor of interest in a smog chamber in the presence
of NOx, collecting the SOA produced on filters, and then analyzing the samples for highly polar
compounds using advanced analytical chemistry methods. Employing this method, candidate
tracers have been identified for several VOC compounds which are emitted in significant
quantities and known to produce SOA in the atmosphere. Some of these SOA-forming
compounds include toluene, a variety of monoterpenes, isoprene, and p-caryophyllene, the latter
three of which are emitted by vegetation and are more significant sources of SOA than toluene.
Smog chamber work can also be used to investigate SOA chemical formation
mechanisms.109'110'111'112
Although these concentrations are only estimates, due to the assumption that the mass
fractions of the smog chamber SOA samples using these tracers are equal to those in the ambient
atmosphere, there are presently no other means available for estimating the SOA concentrations
originating from individual SOA precursors. Among the tracer compounds observed in ambient
PM2.5 samples are two tracer compounds that have been identified in smog chamber aromatic
SOA samples.113 To date, these aromatic tracer compounds have been identified, in the
laboratory, for toluene and w-xylene SOA. Additional work is underway by the EPA to
determine whether these tracers are also formed by benzene and other alkylbenzenes (including
o-xylene, />-xylene, 1,2,4-trimethylbenzene, and ethylbenzene).
One caveat regarding this work is that a large number of VOCs emitted into the
atmosphere, which have the potential to form SOA, have not yet been studied in this way. It is
possible that these unstudied compounds produce SOA species which are being used as tracers
for other VOCs. This means that the present work could overestimate the amount of SOA
formed in the atmosphere by the VOCs studied to date. This approach may also estimate entire
7-65
-------
hydrocarbon classes (e.g., all methylsubstituted-monoaromatics or all monoterpenes) and not
individual precursor hydrocarbons. Thus the tracers could be broadly representative and not
indicative of individual precursors. This is still unknown. Also, anthropogenic precursors play a
role in formation of atmospheric radicals and aerosol acidity, and these factors influence SOA
formation from biogenic hydrocarbons. This anthropogenic and biogenic interaction, important
to EPA and others, needs further study. The issue of SOA formation from aromatic precursors is
an important one to which EPA and others are paying significant attention.
The aromatic tracer compounds and their mass fractions have also been used to estimate
monthly ambient aromatic SOA concentrations from March 2004 to February 2005 in five U.S.
Midwestern cities.114 The annual tracer-based SOA concentration estimates were 0.15, 0.18,
0.13, 0.15, and 0.19 ug carbon/m3 for Bondville, IL, East St. Louis, IL, Northbrook, IL,
Cincinnati, OH and Detroit, MI, respectively, with the highest concentrations occurring in the
summer. On average, the aromatic SOA concentrations made up 17 percent of the total SOA
concentration. Thus, this work suggests that we are finding ambient PM levels on an annual
basis of about 0.15 ug/m3 associated with present toluene levels in the ambient air in these
Midwest cities. Based on preliminary analysis of recent laboratory experiments, it appears the
toluene tracer could also be formed during photooxidation of some of the xylenes.11
Over the past decade a variety of modeling studies have been conducted to predict
ambient SOA levels. While early studies focused on the contribution of biogenic monoterpenes,
additional precursors, such as sesquiterpenes, isoprene, benzene, toluene, and xylene, have been
implemented in atmospheric models such as GEOS-Chem, PMCAMx, and CMAQ.116'117'118'119'
120,121,122 studies jiave incjicated that ambient OC levels may be underestimated by current model
parameterizations.123 While the treatment of new precursors has likely reduced the
model/measurement bias, underestimates can persist.124 In general, modeling studies focus on
comparing the sum of the POA and SOA concentrations with ambient OC or estimated OA
concentrations. Without a method to attribute measured OC to different sources or precursors,
identifying causes of the underestimates in modeled OC via model/measurement comparisons
can be challenging. Oxidation of low-volatility organic compounds as well as particle-phase
reactions resulting from acidity have been explored as potential missing sources of OC in
models.125'126
7.2.3.3 Ozone
As mentioned above, the addition of ethanol to fuels has been shown to contribute to
PAN formation and this is one way for it to contribute therefore to ground-level ozone formation
downwind of NOx sources. PAN is a reservoir and carrier of NOx and is the product of acetyl
radicals reacting with NO2 in the atmosphere. One source of PAN is the photooxidation of
acetaldehyde (Section 7.2.3.1), but many VOCs have the potential for forming acetyl radicals
and therefore PAN or a PAN-type compound.Y PAN can undergo thermal decomposition with a
lifetime of approximately 1 hour at 298K or 148 days at 250K. z
Y Many aromatic hydrocarbons, particularly those present in high percentages in gasoline (toluene, m-, o-, p-xylene,
and 1,3,5-, 1,2,4-trimethylbenzene), form methylglyoxal and biacetyl, which are also strong generators of acetyl
7-66
-------
CH3C(O)OONO2 + M -> CH3C(O)OO- + NO2 + M k = 3.3 x 10'4 s'1 127
The reaction above shows how NO2 is released in the thermal decomposition of PAN,
along with a peroxy radical which can oxidize NO to NC>2 as previously shown in Section
7.2.3.1. NC>2 can also be formed in photodegradation reactions where NO is converted to NO2
(see OH radical reaction of acetaldehyde in Section 7.2.3.1). In both cases, NO2 further
photolyzes to produce ozone (Os).
NO2 + hv -> NO + O(3P) A, = 300-800 nm 128
O(3P) + O2 + M -> O3 + M
The temperature sensitivity of PAN allows it to be stable enough at low temperatures to
be transported long distances before decomposing to release NO2. NO2 can then participate in
1 9Q
ozone formation in regions remote from the original NOx source. A discussion of CB05
mechanisms for ozone formation can be found in Yarwood et al. (2005). 13°
Another important way that ethanol fuels contribute to ozone formation is by increasing
the formation of new radicals through increases in formaldehyde and acetaldehyde. As shown in
Section 7.2.3.1, the photolysis of both aldehydes results in two molecules of either hydroperoxy
radical or methylperoxy radical, both of which oxidize NO to NO2 leading to ozone formation.
7.2.3.4 Uncertainties Associated with Chemical Mechanisms
A key source of uncertainty with respect to the air quality modeling results is the
photochemical mechanisms in CMAQ 4.7.1. Pollutants such as ozone, PM, acetaldehyde,
formaldehyde, acrolein, and 1,3-butadiene can be formed secondarily through atmospheric
chemical processes. Since secondarily formed pollutants can result from many different reaction
pathways, there are uncertainties associated with each pathway. Simplifications of chemistry
must be made in order to handle reactions of thousands of chemicals in the atmosphere.
Mechanisms for formation of ozone, PM, acetaldehyde and peroxyacetyl nitrate (PAN) are
discussed in Section 7.2.3.
For PM, there are a number of uncertainties associated with SO A formation that should
be addressed explicitly. As mentioned in Section 7.2.3, a large number of VOCs emitted into the
atmosphere, which have the potential to form SO A, have not yet been studied in detail. In
addition, the amount of ambient SOA that comes from benzene is uncertain. Simplifications to
the SOA treatment in CMAQ have also been made in order to preserve computational efficiency.
These simplifications are described in release notes for CMAQ 4.7 on the Community Modeling
and Analysis System (CMAS) website.131
radicals (Smith, D.F., T.E. Kleindienst, C.D. Mclver (1999) Primary product distribution from the reaction of OH
with m-, p-xylene and 1,2,4- and 1,3,5-Trimethylbenzene. J. Atmos. Chem., 34: 339- 364.).
z All rate coefficients are listed at 298 K and, if applicable, 1 bar of air.
7-67
-------
7.2.4 Impacts of the Proposed Rule on Air Quality
Air quality modeling performed for this proposed rule estimates the changes in ambient
concentrations of PM2.5, ozone and NC>2, as well as changes in ambient concentrations of ethanol
and the following air toxics: acetaldehyde, acrolein, benzene, 1,3-butadiene, and formaldehyde.
The air quality modeling results also include changes in deposition of nitrogen and sulfur and
changes in visibility levels due to this proposed rule.
This section describes current ambient levels of the modeled pollutants and presents the
projected future ambient levels resulting from the proposed rule.
7.2.4.1 Ozone
As described in Section 6.2.1 of this draft RIA, ozone causes adverse health effects, and
the EPA has set national ambient air quality standards (NAAQS) to protect against those health
effects. In this section, we present information on current and model-projected future ozone
levels.
7.2.4.1.1 Current Levels of Ozone
Figure 7-12 shows a snapshot of measured ozone concentrations in 2010. The highest
ozone concentrations were located in California.
7-68
-------
Concentration Range (ppm)
• 0.025-0.059(81 Sites)
O 0.060 - 0.075 (835 Sites)
O 0.076 - 0.095 (279 Sites)
• 0.096-0.120 (18 Sites)
Puerto Rico
Alaska
Figure 7-12 Ozone Concentrations (fourth highest daily maximum 8-hour concentration) in
ppm for 2010AA
The primary and secondary NAAQS for ozone are 8-hour standards set at 0.075 ppm. The most
recent revision to the ozone standards was in 2008; the previous 8-hour ozone standards, set in
1997, had been set at 0.08 ppm. In 2004, the U.S. EPA designated nonattainment areas for the
1997 8-hour ozone NAAQS (69 FR 23858, April 30, 2004). B As of December 14, 2012, there
were 41 8-hour ozone nonattainment areas for the 1997 ozone NAAQS, composed of 221 full or
partial counties, with a total population of over 129 million. Nonattainment areas for the 1997 8-
hour ozone NAAQS are pictured in Figure 7-13. Nonattainment designations for the 2008 ozone
standards were finalized on April 30, 2012 and May 31, 2012.132 These designations include 46
areas, composed of 227 full or partial counties, with a population of over 123 million.
Nonattainment areas for the 2008 ozone NAAQS are pictured in Figure 7-14. As of December
/-i/-i
14, 2012, over 138 million people are living in ozone nonattainment areas.
^ From U.S. EPA, 2011. Our Nation's Air: Status and Trends through 2010. EPA-454/R-12-001. February 2012.
Available at: http://www.epa.gov/airtrends/2011/.
BB A nonattainment area is defined in the Clean Air Act (CAA) as an area that is violating an ambient standard or is
contributing to a nearby area that is violating the standard.
cc The 138 million total is calculated by summing, without double counting, the 1997 and 2008 ozone
nonattainment populations contained in the Summary Nonattainment Area Population Exposure report
7-69
-------
8-Hour Ozone Nonattainment Areas (1997 Standard)
12-2012
Stir Ozone Classifications
^^| Extreme
I | Severe 17
I I'Sevete 15
Nonattainment areas are indicated by color-
When only a portion of a county is shown in color.
it indicates that only that part of the county is within
a nonattainment area boundary
I I Moderate
I | Marginal
The St Louis. MO-IL 8-hr Ozone (1997 Standard) multi-state nonattainment area has a
state that has been redesignated but it is not considered a maintenance area until all states
in the area are redesignated. The counties for this area are displayed as nonattainment areas
Figure 7-13 1997 8-hour Ozone Nonattainment Areas
(http://www.epa.gov/oar/oaqps/greenbk/popexp.html). If there is a population associated with both the 1997 and
2008 nonattainment areas, and they are not the same, then the larger of the two populations is included in the sum.
7-70
-------
8-Hour Ozone Nonattainment Areas (2008 Standard)
7/2012
Nona |