Environ. Sci. Technol. 2007, 41, 4059-4064
Effects of Fuel Ethanol Content and Volatility on Regulated and Unregulated Exhaust Emissions for the Latest Technology Gasoline Vehicles T H O M A S D . D U R B I N , * ,† J. WAYNE MILLER,† THEODORE YOUNGLOVE,‡ T A O H U A I , †,§ A N D K A T H A L E N A C O C K E R † College of Engineering, Center for Environmental Research and Technology, University of California, Riverside, California 92521, and Statistical Consulting Collaboratory, Department of Statistics, University of California, Riverside, California 92521
Oxygenate content and fuel volatility (distillation) variables are important parameters affecting vehicle exhaust emissions, and data on their effects on the latest technology vehicles are quite limited. For this study, 12 Californiacertified LEV to SULEV vehicles were tested on a matrix of 12 fuels with varying levels of ethanol concentration (0, 5.7, and 10 vol %), T50 (195, 215, and 235 °F), and T90 (295, 330, and 355 °F). There were statistically significant interactions between ethanol and T90 for NMHC, ethanol, and T50 for CO and ethanol and T50 for NOx. NMHC emissions increased with increasing ethanol content at the midpoint and high level of T90 but were unaffected at the low T90 level. CO emissions decreased as the ethanol content increased from the low to the midpoint level for all levels of T50, but between the 5.7 and 10% ethanol levels, CO showed only an increase for the high level of T50. NOx emissions increased with ethanol content for some conditions. Non-methane organic gases (NMOG) and toxic emissions were examined for only a subset of fuels with the highest T90 level, with NMOG, acetaldehyde, benzene, and 1-,3-butadiene all found to increase with increasing ethanol content.
Introduction As vehicle technologies and fuel formulations continue to change to meet more stringent environmental standards, it is necessary to understand the effects of fuel properties on vehicle emissions. Over the years, the effect of various fuel properties on vehicle emissions has been the subject of numerous studies and programs. Data from these earlier programs were used in the development of regulations for fuel properties. For example, these data were incorporated into the Environmental Protection Agency’s (EPA) Complex * Corresponding author. Tel.: (951) 781-5791; fax: (951) 781-5790; e-mail:
[email protected]. † College of Engineering, Center for Environmental Research and Technology. ‡ Statistical Consulting Collaboratory, Department of Statistics. § Present address: California Environmental Protection Agency, Air Resources Board, Research Division, 1001 I St., Sacramento, CA 95812. 10.1021/es061776o CCC: $37.00 Published on Web 05/04/2007
2007 American Chemical Society
Model for federal reformulated gasoline (RFG) and the California Air Resources Board’s (CARB) Predictive Model for cleaner burning gasoline (CBG). Although the database on the emissions impacts of fuel properties is large, data on the effects of fuel properties on the latest technology vehicless-those certified to California’s Low-Emission Vehicle (LEV), Ultralow-Emission Vehicle (ULEV), and SuperUltralow-Emission Vehicle (SULEV) standardss-are quite limited. In the future, these vehicle technologies will account for an increasing share of the emissions of the in-use vehicle fleet. Today, regulatory agencies continue to consider changes in gasoline composition in response to environmental regulations and other initiatives like the use of renewable fuels. Two fuel parameters that are considered to be important in determining vehicle emissions are oxygenate content and volatility. Both properties are included in the models used by the EPA and CARB. Many states have banned methyl t-butyl ether (MTBE), leading to greater use of ethanol (EtOH). Further, the Renewable Fuel Standard (RFS) adopted as part of the federal Energy Policy Act of 2005 requires significant and increasing volumes of renewables to be blended into the transportation fuel pool between 2006 and 2012, much of which is likely to be ethanol. The effects of fuel volatility and ethanol/oxygenates on emissions have been investigated extensively in past studies (1-21). These studies have shown some general trends for the effects of these properties on emissions. Reducing T50 and T90 generally reduces exhaust hydrocarbon emissions (2, 3, 5, 8). Adding ethanol and other oxygenates generally reduces total hydrocarbon (THC) and carbon monoxide (CO) emissions (1, 2, 7-9, 11, 12). Oxides of nitrogen (NOx) have increased with oxygenates in some studies but not for all test fleets (1, 2, 7-9, 11, 12). While these studies provide important information, there is scarce data on how these fuel parameters affect emissions in vehicles with advanced emission control technology. Some of the more recent studies also include some contradictory data, including a recent study in which slightly higher NOx emissions were found for a fuel with no oxygenates in comparison with the oxygenated fuels (1). The effects of ethanol and fuel volatility on the detailed speciation of the exhaust have been studied in a number of previous programs. Early Auto/Oil Air Quality Improvement Research Program (AQIRP) studies provided some of the most comprehensive databases for fuel effects on speciated exhaust emissions and toxics (2, 4, 7, 13). The EPA also conducted several extensive studies to better understand the impacts of the introduction of reformulated gasoline (8, 9). A number of other smaller studies has also evaluated fuel effects on toxics in the exhaust, including those conducted by or in conjunction with the EPA (14-17), studies by Environment Canada (18, 19), and studies by the American Petroleum Institute (20). The goal of this study was to expand the database of information available on the effects of gasoline ethanol content and volatility parameters on exhaust emissions. This study included a comprehensive set of test fuels with varying ethanol content and mid- and back-end volatility, and a test fleet of vehicles with the newest emission control technology. The information obtained from this study will help in understanding and predicting the implications of changes in fuel properties for regulatory or other reasons. The results of this study are summarized in the following paper and discussed in greater detail in reference 22. VOL. 41, NO. 11, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4059
TABLE 1. Description of Test Vehicles MY 1 2 3 4 5 6 7 8 9 10 11 12 a
2002 2003 2003 2003 2003 2003 2002 2003 2001 2003 2003 2003
OEM Ford Chevrolet Ford Dodge Ford Chevrolet Toyota Buick VW Ford Chevrolet Honda
model Taurus Cavalier F-150 Caravan Explorer Trailblazer Camry LeSabre Jetta Windstar Silverado Accord
California certification LEV LEV LEV LEV LEV LEV ULEV ULEV ULEV ULEV ULEV SULEV
type
engine size (L)
mileage
engine family
PCa
3.0 2.2 4.6 3.3 4.0 4.2 2.4 3.8 2.0 3.8 5.3 2.4
19 414 28 728 13 856 18 342 16 445 13 141 14 731 10 364 28 761 20 523 10 298 12 432
1FMXV03.0VF4 1GMXV02.2025 3FMXT05.4PFB 3CRXT03.32DR 3FMXT04.02FB 3GMXT04.2185 1TYXV02.4JJA 3GMXV03.8044 1VWXV02.0223 3FMXT03.82HA 3GMXT05.3176 3HNXV02.4KCP
PC LDT LDT LDT LDT PC PC PC LDT LDT PC
PC ) passenger car and LDT ) light-duty truck. Vehicles equipped with catalysts aged to 100 000 miles for testing.
FIGURE 1. CRC E-67 fuel cube design.
Experimental Procedures The experimental procedures are briefly summarized in the following section, with more extensive description in reference 22. Test Vehicles. The test fleet consisted of 12 vehicles. The fleet sizing was based on experimental design statistics used in previous large AQIRP fuel studies (23). The vehicles included present day technologies with California lowemission vehicle (LEV), ultralow emission vehicle (ULEV), and super-ultralow-emission vehicle (SULEV) certification. The test fleet was evenly split between passenger cars and light-duty trucks. The vehicles were fitted with catalysts that had been bench-aged to the equivalent of 100 000 miles for testing. The catalysts were aged for 75 h using the Rapid Aging Test-A (RAT-A) protocol at Johnson Matthey Testing in Taylor, MI (24). The specific details of the vehicles used in this project are listed in Table 1. Fuels. Twelve fuels were prepared and provided for testing for this project. These 12 fuels were designed to encompass three levels of ethanol content (0, 5.7, and 10%), three levels of T50 (195, 215, and 235 °F), and three levels of T90 (295, 330, and 355 °F). The values for ethanol represent typical ethanol concentrations found in California (5.7%) and the rest of the U.S. (10%). Previous studies had shown that in addition to the main effects of ethanol, curvature effects for T50 and T90 and interactions between ethanol and T50 and T90 might be important, so these variables were included in the experimental design. The fuel design matrix is shown graphically in Figure 1. The fuels were blended from refinery streams with the general properties targeted to be constant for all fuels to reduce or eliminate any potential confounding effect of these properties with the design parameters. A summary of the design and actual values of the target fuel properties 4060
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 11, 2007
and a full analysis for each fuel is provided in the Supporting Information. Test Protocol. Each vehicle was tested over two lightduty Federal Test Procedure (FTP)s, with a third test performed if the difference in the composite FTP emissions exceeded the following: HC 33%, NOx 29%, and CO 70% (provided that the absolute difference in the measurements was greater than 5 mg/mi). The fuel test order within each vehicle was fully randomized. All fuels were tested once before starting the second test of each fuel. This resulted in the fuels being tested in two blocks. This randomized approach allows the use of more robust statistical methods during the data analysis phase. A multiple drain and fill procedure with onroad conditioning was used between tests on different fuels to minimize or eliminate carryover effects between fuels. Vehicle Emissions Measurements. All tests were conducted in CE-CERT’s Vehicle Emissions Research Laboratory (VERL) equipped with a Burke E. Porter 48 in. single-roll electric dynamometer. Standard bag measurements were obtained for total hydrocarbons (THC), non-methane hydrocarbons (NMHC), carbon monoxide (CO), nitrogen oxides (NOx), and carbon dioxide (CO2) with a Pierburg AMA-4000 bench. Full NMOG speciation measurements were made on fuels D, E, K, and L for all vehicles. These fuels represent the right face of the fuel cube in Figure 1, with T90 ) 355 °F for each fuel. The California Air Resources Board definition of nonmethane organic gases (NMOG) was used in this study
NMOG ) NMHC(FID) +
∑alcohols + ∑carbonyls
The NMOG speciation measurements were performed in accordance with protocols developed previously as part of the AQIRP, including the validation criteria (25). The basic methodologies include Tedlar bag sampling with gas chromatography with a flame ionization detector (GC/FID) for the C1-C12 hydrocarbons and sampling on cartridges coated with dinitrophenyl-hydrazine (DNPH) followed by highperformance liquid chromatography (HPLC) analysis for the carbonyls. Some additional procedures were incorporated to enhance the detection levels for the NMOG species, including doubling the size of the sample loop from 5 to 10 mL, changing the FID makeup gas from helium to nitrogen, which roughly doubles the FID sensitivity, and adding a liquid nitrogen cold trap to improve the signal-to-noise level. The combined enhancements provided approximately a factor of 3-4 improvement on the minimum detection limits (MDLs) as compared to the earlier AQIRP methods. MDLs were approximately 6 and 7 ppbc1, respectively, for benzene and 1,3-butadiene. MDLs for formaldehyde and acetaldehyde were
