Development of molecular marker source profiles for emissions from on-road gasoline and diesel vehicle fleets.

ABSTRACT

As part of the Gasoline/Diesel PM Split Study, relatively large fleets of gasoline vehicle[s.sup.53] and diesel vehicle[s.sup.34] were tested on a chassis dynamometer to develop chemical source profiles for source attribution of atmospheric particulate matter in California's South Coast Air Basin. Gasoline vehicles were tested in cold-start and warm-start conditions, and diesel vehicles were tested through several driving cycles. Tailpipe emissions of particulate matter were analyzed for organic tracer compounds, including hopanes, steranes, and polycyclic aromatic hydrocarbons. Large intervehicle variation was seen in emission rate and composition, and results were averaged to examine the impacts of vehicle ages, weight classes, and driving cycles on the variation. Average profiles, weighted by mass emission rate, had much lower uncertainty than that associated with intervehicle variation. Mass emission rates and elemental carbon/organic carbon (EC/OC) ratios for gasoline vehicle age classes were influenced most by use of cold-start or warm-start driving cycle (factor of 2-7). Individual smoker vehicles had a large range of mass and EC/OC (factors of 40 and 625, respectively). Gasoline vehicle age averages, data on vehicle ages and miles traveled in the area, and several assumptions about smoker contributions were used to create emissions profiles representative of on-road vehicle fleets in the Los Angeles area in 2001. In the representative gasoline fleet profiles, variation was further reduced, with cold-start or warm-start and the representation of smoker vehicles making a difference of approximately a factor of two in mass emission rate and EC/OC. Diesel vehicle profiles were created on the basis of vehicle age, weight class, and driving cycle. Mass emission rate and EC/OC for diesel averages were influenced by vehicle age (factor of 2-5), weight class (factor of 2-7), and driving cycle (factor of 10-20). Absolute and relative emissions of molecular marker compounds showed levels of variation similar to those of mass and EC/OC.

INTRODUCTION

Profiles of organic compounds in particulate matter emissions from various sources are becoming more widely used for apportionment modeling to investigate the sources contributing to ambient particulate matter levels. (1) Several studies have been performed to characterize tailpipe emissions of gasoline-powered and diesel-powered vehicles, (2-8) many with the goal of developing representative profiles for on-road vehicle tailpipe emissions. Most of these studies have investigated a small number of vehicles but have seen variation in emission rate and composition related to vehicle age, fuel type, driving cycle, engine temperature, and engine repair. (9-12) With so many factors contributing to differences in emissions, it is necessary to understand the relationships between emissions of tested vehicles and the emissions of an actual on-road fleet of vehicles in an area.

This study was conducted with a relatively large number of gasoline and diesel vehicles, in several age groups and weight classes, to represent average fleet emissions and to assess the uncertainty in average vehicle emissions profiles. With emissions measurements for a larger number of vehicles, average profiles were created that are representative of on-road gasoline and diesel vehicle fleets in the Los Angeles (LA) area in 2001. Although the intervehicle and intertest variability in both the total mass emission rate of particulate matter and the composition of particulate matter are very large, weighted averaging of these measurements to reflect the known composition of vehicle fleets in the area provides a sense of the actual variability in emissions from the fleet. The uncertainty in these profiles and their sensitivity to changes in the fleet composition were explored by varying the composition of the constructed fleet profiles in terms of vehicle age distribution, the fraction of the fleet that was high-emitting smokers, vehicle weight distribution, and tested vehicle driving cycle.

The focus in these measurements was on organic compounds previously used as tracers for vehicle exhaust in particulate matter apportionment studies, (1) including hopanes, steranes, and polycyclic aromatic hydrocarbons (PAHs).

METHODS

Gasoline Vehicle Testing

A total of 53 gasoline-powered spark-ignition (SI) vehicles were recruited by the California Bureau of Automotive Repair and the South Coast Air Quality Management District (SCAQMD) for testing. The tested vehicles fit into several age groups (1975-1984, 1985-1994, and 1995-2001) plus five recruited high-emitting visible smokers, and one light-duty diesel vehicle (discussed here with diesel vehicles). After testing, two additional vehicles with emissions of elemental carbon (EC) of more than 50 mg x [mi.sup.-1] were also classified as smokers because their EC emission rates were higher than most tested recruited smokers, and an order of magnitude higher than any other tested nonsmoker. The tested gasoline and smoker vehicles and their ages are summarized in Table 1a.

Vehicle driving simulation was conducted on a chassis dynamometer (Clayton Model CTE-50-0). Particulate matter samples were collected for both cold-start and warm-start engine conditions over the same driving cycles. After an engine-off period of at least 12 hr, vehicles were operated with the Unified Driving Cycle. (13) The cold-start samples were taken over the first two phases of the Unified Driving Cycle, for a total sample time of 1436 sec and an average speed of 24.6 mph. After a 600-sec engine-off period, warm-start samples were collected over identical cycles.

Diesel Vehicle Testing

Thirty-three diesel-powered compression-ignition (CI) vehicles were recruited by West Virginia University (WVU) and identified by model year and weight class in testing, as summarized in Table 2a. All heavy-heavy duty vehicles were tested at a test weight of 46,000 lb and all other vehicles were tested with a test weight of 70% of the gross vehicle weight rating (maximum weight the vehicle is designed to bear). The tests were conducted on two dynamometers owned and operated by West Virginia University. The WVU Heavy-Duty Vehicle Chassis Dynamometer, with two rollers, was used for vehicles over 22,000 lb, and the single-roller WVU Medium-Duty Vehicle Chassis Dynamometer was used for smaller vehicles. The vehicles were tested with three driving cycles, including an engine idle cycle, the city/suburban heavy vehicle route (CSHVR) for low speeds, and the highway cycle for higher speeds. Additionally, subsets of vehicles were operated with a cold-start CSHVR, a CSHVR with the engine brake enabled, a CSHVR with federal fuel (instead of California reformulated fuel), and a heavy-duty urban dynamometer driving schedule (UDDS). (14) Two of the vehicles were buses, and were operated on a CSHVR, a vehicle idle, and on the Manhattan cycle, which approximates slow speeds in stop-and-go traffic. One vehicle was a passenger car, and was operated through cold-start and warm-start Unified Driving cycles like the gasoline-powered passenger cars described above.

Sample Collection and Analysis

Vehicle exhaust was sampled using a dilution-sampling tunnel. Dilution air was charcoal treated and HEPA filtered to remove gaseous and particulate contaminants. Exhaust and dilution air were well mixed in the tunnel. For gasoline vehicle tests, a probe downstream drew diluted exhaust through stainless steel tubing to fine particulate matter ([PM.sub.2.5]) cyclones and filter holders (University Research Glassware). For diesel exhaust tests, the primary dilution tunnel was a full-scale dilution tunnel operated with HEPA-filtered dilution air. A sample probe drew diluted exhaust into a secondary dilution chamber, and particulate matter samples were collected with cyclones and filter holders attached to the secondary dilution chamber.

[PM.sub.2.5] samples were collected on pre-baked quartz fiber filters (Pall) for organic compound speciation and EC and organic carbon (OC), and on preweighed Teflon membrane filters (Pall) for mass determination. EC and OC were analyzed with the National Institute for Occupational Safety and Health 5040 method, chosen to be parallel to the ACE-Asia study (15) and other recent climate change studies. Mass was determined gravimetrically from conditioned filters with a microbalance (Mettler-Toledo).

Organic compounds were quantified with solvent extraction and gas chromatography (GC)-mass spectroscopy (GCMS). Details of sample preparation (16) and GCMS analysis (17) have been described elsewhere. Before extraction, filters were spiked with internal standards across a range of molecular weights and volatility to enable quantification of a wide range of compounds. Two sequential extractions were performed with a Soxhlet apparatus (one with dichloromethane and one with methanol) and the two extracts were combined. Extracts were rotary evaporated to approximately 10 mL and further evaporated with nitrogen down to 0.25 mL. The GC column was 30 m long with a diameter of 0.25 mm and a film thickness of 0.25 mm. The injector and GCMS interface temperatures were held at 300 [degrees]C. The initial oven temperature was held at 65 [degrees]C for 10 min, then ramped at 10 [degrees]C per minute to 300 [degrees]C and held for 26.5 min. The carrier gas was helium, flowing at 1 mL x [min.sup.-1].