Environ. Sci. Technol. 1996, 30, 810-816
Concentrations of Volatile Organic Compounds in the Passenger Compartments of Automobiles NICHOLAS J. LAWRYK AND CLIFFORD P. WEISEL* Exposure Measurement and Assessment Division, Environmental and Occupational Health Sciences Institute, 681 Frelinghuysen Road, P.O. Box 1179, Piscataway, New Jersey 08855-1179
In-vehicle concentrations of selected gasoline-derived volatile organic compounds (VOCs) and formaldehyde were examined on 113 commutes through suburban New Jersey and 33 New Jersey/New York commutes. Overall median concentrations were lowest in a typical suburban commute, slightly higher on the New Jersey Turnpike, and highest in transit through the Lincoln Tunnel. Median in-vehicle concentrations of benzene, ethylbenzene, m&p-xylene, and o-xylene were 14, 6.8, 36, and 15 µg/m3, respectively. One vehicle, with a carbureted engine, developed malfunctions that caused gasoline emissions within the engine compartment during driving, resulting in the gasoline-derived VOC concentrations in this vehicle being much higher than in the properly maintained fuel-injected vehicle, particularly for the low ventilation extreme. The highest in-vehicle benzene concentration measured during these malfunctions was 45.2 µg/m3. The air concentration in the vehicle driven in tandem was a factor of 25 less (1.8 µg/ m3).
Introduction Over the past several decades, the reduction of ambient concentrations of pollutants from automobiles has been accomplished by the use of various emission controls. These efforts have been successful for most vehicles in the domestic fleet. Currently, the majority of automobile pollution is generated by high emitting vehicles that constitute 10-30% of the fleet (1). These high-emitting vehicles may have malfunctions in systems involving emission control, fuel distribution, and exhaust. Such excessive emissions from these components under the hood may result in elevated in-vehicle concentrations of gasolinederived compounds. Recent studies (2-5) found that in-vehicle concentrations of gasoline-derived VOCs were up to 8 times higher than the corresponding ambient air at nearby monitoring sites. Highest in-vehicle VOC concentrations were found * Corresponding author fax: (908)445-0116.
810
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 3, 1996
FIGURE 1. Maps of suburban (a) and urban (b) commutes.
during high traffic density periods in the morning and evening rush hours (4). Two studies used randomly selected vehicles without regard for vehicle maintenance (2, 3). In one study, the vehicles were well-characterized and maintained, and little difference in VOC concentrations was reported between vehicle and prevailing interior ventilation conditions during a 2-month period in the summer (4). The present study used two well-characterized vehicles that were maintained for personal use. The long-term observations of these vehicles, which extended over a 19month period, allowed for a better estimate of real-life invehicle exposure scenarios. Furthermore, continued sampling during any malfunctions in the pollutant controls, exhaust system, or fuel distribution systems can demonstrate how in-vehicle VOC concentrations vary with respect to those of the roadway air.
Materials and Methods Study Design. A series of experiments were conducted to determine the concentrations of gasoline-derived VOCs in the passenger compartments of two automobiles on two different commuting routes. The commuting routes were characteristic of driving patterns for the New Jersey/New York metropolitan area (Figure 1). One was a local highway commute in the central New Jersey area, and the other was a commute from suburban New Jersey into New York City. Ventilation conditions in both vehicles were recorded. Low ventilation conditions involved closed windows with the outside vents off and interior fan off. High ventilation conditions were obtained by opening the front windows 30 cm, opening the outside vents, and turning on the interior fan. The two ventilation conditions offered extremes in air
0013-936X/96/0930-0810$12.00/0
1996 American Chemical Society
exchange rates between the roadway air and the vehicle interior (6). The two automobiles used were a 1988 Chevrolet Celebrity and a 1987 Plymouth Horizon. The Plymouth Horizon was a 4-door hatchback, with a carbureted 4-cylinder engine; the Chevrolet Celebrity was a 4-door sedan with a fuel-injected 6-cylinder engine. At the time of the study, these two automobiles represented domestic vehicles in widespread use, according to Ward’s 19851988 surveys of production automobiles and commonly requested options. The Chevrolet was a low mileage automobile (with under 30 000 mi), and the Plymouth represented a high mileage vehicle (with more than 60 000 mi). Body integrity tests for both vehicles showed no leakage around the door, lift back seams, firewalls, or through the passenger compartment floors when they were parked, based on smoke releases under the cars and near the seams and visually observing for penetration into the passenger compartment. This does not mean that the passenger compartment is impervious to outside air however. There is an inherent leakiness in all vehicles, even when all pathways are apparently closed to the outside air. A previous study found that even when the windows were closed tightly, the air turnover rate in an automobile was over 9 times greater when the vehicle was traveling at 32 kph than when it was parked, with the largest rates occurring during periods of acceleration (6). Therefore, air concentrations of gasoline-derived VOCs in the passenger compartment should be similar to those of the roadway air, although malfunctions that result in increased emissions under the hood could, in turn, result in increased concentrations of these compounds in the passenger compartment. The fuel, engine, and exhaust systems on the Chevrolet Celebrity did not malfunction during the experiments. However, the Plymouth Horizon required a new carburetor just before the study began and developed several fuel distribution problems during the study, including a malfunctioning carburetor, a perforated fuel supply line under the carburetor, and a defective electric choke. Four commutes on the suburban loop were completed while these malfunctions were present. Commuting Studies. The suburban loop was a 23-km drive along a 6-lane state highway in East Brunswick, NJ (Figure 1). It included traffic lights, a landfill access road frequently used by heavy trucks, residential areas, and shopping malls. Travel time for a single round trip was typically 30 min during peak traffic times. About half of the time was spent in stop-and-go traffic. Commutes along this route began between 5:00 and 5:30 PM, and the two round trips were completed by 7:00 PM. Average vehicle speeds per trip ranged from 28 to 56 kph, with a median speed of 43 kph. Two trips were usually done in one sampling day, with one VOC sample collected for each trip. On most sampling days, one trip was done with high interior ventilation, and the other with low interior ventilation as described above. The New York City commuting route was a 72-km drive from New Brunswick to New York City via the 12-lane New Jersey Turnpike, Route 495 East, and the Lincoln Tunnel (Figure 1). The one-way metropolitan commute began between 7:00 and 7:30 AM and was completed by 9:00 AM. Two sample sets were collected during this commute. The first was collected from the point of departure in New Brunswick to the turnpike toll booth for the tunnel (turnpike
FIGURE 2. Sampling train for roadway air collection from the automobile hood.
phase). Typical sampling time was 45 min, with the overall sample representing pollutant concentrations in the automobile passenger compartment while driving on the turnpike. Traffic generally moved at highway speeds (average speed per trip ranged from 63 to 84 kph, with a median of 72 kph) on this portion of the route, although traffic density was very high in sections. The second sample was collected from just before the toll booth across 495 East, through the Lincoln Tunnel, and into New York City. The Lincoln Tunnel phase averaged 31 min and was characterized by very high traffic density, with the average vehicle speed per trip ranging from 11 to 29 kph (median speed 16 kph). Sampling and Analysis. All in-vehicle VOC samples were collected by suspending adsorbent tubes containing either Tenax (500 mg) or a layered adsorbent (100 mg of Tenax GC, 100 mg of Carboxen 569, and 100 mg of Carbosieve SIII) at a height 50 cm above the passenger seat. Adsorbent tubes were connected to constant flow pumps, which drew air at approximately 60 cm3/min for commuting air samples. The pumps were calibrated before and after sample collection. The average of these two rates was used as the sample flow rate in all volume calculations. Departures from the initial flow rate larger than 10% excluded the associated samples from analysis and inclusion in the database. No samples were voided during the study due to excessive flow rate changes in the sampling pumps. Average volumes of air collected during the commuting studies were 1.5 L. Formaldehyde samples were collected on dinitrophenylhydrazine (DNPH) silica cartridges (7). Air was drawn through the cartridge with a direct current vacuum pump powered by the cigarette lighter jack. The flow rate was set at 1500 cm3/min. During a commute, VOC samples of the ambient roadway air were also collected through a 0.25-in. outside diameter copper tube secured to the hood of the vehicle with a large fishing magnet (Figure 2). The inlet of the tube was turned downward to prevent precipitation from entering the tube. The inlet was set at least 15 cm above the surface of the hood to avoid excessive turbulence associated with streamlining. The tube entered the vehicle interior through the passenger window. For the low ventilation conditions, the window was sealed with duct tape. The VOCs collected on the cartridges were thermally desorbed (Perkin Elmer ATD 400) and completely transferred to a gas chromatograph (GC) equipped with a capillary column. The GC was connected to a mass selective
VOL. 30, NO. 3, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
811
Results and Discussion
TABLE 1
Sampling Precision. The precision of the sampling and analytical techniques was tested by collecting 28 duplicate samples. The mean relative standard deviations (RSD) for nearly all compounds were less than 0.20 (Table 1), with the exceptions of the two lightest alkanes, 3-methylpentane (RSD ) 0.23) and n-hexane (RSD ) 0.48), which were more difficult to quantify than the compounds of greater mass and bulk. Also, samples of n-hexane began later in the study than the other compounds, and only nine duplicate samples were measured. The mean VOC concentrations measured on the suburban loop for both ventilation conditions are given in Table 2. With the exception of benzene, all compound concentrations were higher for the low ventilation condition. Concentrations of the xylenes and toluene were significantly elevated (p < 0.05) during the low ventilation condition. The concentrations of substituted aromatics in the interior of the vehicles were about the same after a 30-min idle as those measured on the suburban loop, while all other compound concentrations after idling were considerably lower than all commuting studies. On some daily commutes, compound concentrations varied between vehicles driven in tandem. Comparison of these concentrations between the two vehicles (Table 3) showed significantly higher concentrations (p < 0.05) in car 2 (Plymouth Horizon) for all substituted aromatics and n-hexane. The two suburban loop drives were done in a 90-min time period, with each loop using a different ventilation condition. Therefore, paired samples for the two ventilation conditions were collected under similar meteorological conditions, although slight differences in traffic density may have occurred. Concentration Differences by Route Driven. The mean in-vehicle VOC concentrations measured in the Lincoln Tunnel are greater than those measured on the suburban loop and the Turnpike, with the differences being significant (p < 0.05) for all compounds except n-hexane (Table 2). VOC concentrations measured on the Turnpike were similar to or slightly higher than those measured on the suburban loop. Formaldehyde concentrations for all three routes
Number of Observations and RSD for Duplicate In-Vehicle VOC Samples compound
no. of observations
RSD
benzene ethylbenzene n-hexane isooctane 3-methylpentane m&p-xylene o-xylene 1,2,4-trimethylbenzene toluene
28 28 9 28 28 28 28 24 28
0.15 0.12 0.48 0.17 0.23 0.11 0.19 0.13 0.13
detector (MSD) (Hewlett Packard 5890/5970B). The software was capable of reconstructing mass spectra and ion chromatograms. Chromatogram search routines were written to reconstruct the ion chromatograms of the mass of interest for each compound inside a time window of 1 min around the expected retention time. The ion chromatograms were integrated and the results stored in a computer file. Quantification was based on multipoint external standard curves constructed from the single ion chromatograms. The resulting reports were examined manually for errors due to shifts in retention time or peak identification. Peak identification was based on retention time and the ratios of ion fragments to a primary ion fragment used for quantification. Calibration curves were verified for all sample process days to be within (20% of the concentration present in the external standard (8, 9). Mass spectrometer drift was tested at the beginning of each analytical day with an MSD tune check based on procedures outlined in EPA Method 625 and Eichelberger et al. (10). Analysis of the formaldehyde samples was done on a gradient high-performance liquid chromatography (HPLC) system with an ultraviolet (UV) detector at 360 nm (7). The analytical system used an automated injection valve, a stainless steel 4-µm Nova-Pak reversed-phase C18 column (3.9 mm diameter, 150 mm length), and a recorder/ integrator. TABLE 2
Mean In-Vehicle VOC Concentrations (µg/m3) ( Standard Deviation by Route for Both Vehicles suburban loop compound
n-hexane isooctane 3-methylpentane benzene toluene
m&p-xylene o-xylene ethylbenzene 1,2,4-trimethylbenzene formaldehyde
812
9
low ventilation
high ventilation
Turnpike
Lincoln Tunnel
6.5 ( 13.8 (N ) 18) 13.0 ( 19.9 (N ) 51) 9.6 ( 15.1 (N ) 49) 13.1 ( 8.8 (N ) 52) 60.2 ( 49.2 (N ) 52) 34.6 ( 30.1 (N ) 52) 14.5 ( 15.7 (N ) 52) 11.5 ( 18.8 (N ) 52) 18.1 ( 16.2 (N ) 36) 0.3 ( 0.2 (N ) 11)
5.5 ( 4.5 (N ) 12) 10.3 ( 11.8 (N ) 43) 6.7 ( 5.3 (N ) 42) 13.8 ( 7.8 (N ) 43) 40.4 ( 26.6 (N ) 43) 22.5 ( 19.7 (N ) 43) 10.1 ( 10.7 (N ) 43) 8.5 ( 11.2 (N ) 43) 12.9 ( 12.4 (N ) 31) 0.3 ( 0.2 (N ) 10)
9.8 ( 14.5 (N ) 20) 21.9 ( 48.0 (N ) 32) 15.5 ( 33.1 (N ) 32) 16.2 ( 19.5 (N ) 32) 71.0 ( 90.4 (N ) 32) 31.4 ( 30.0 (N ) 32) 12.5 ( 13.9 (N ) 32) 8.8 ( 10.8 (N ) 32) 13.9 ( 15.5 (N ) 26) 0.1 ( 0.1 (N ) 6)
9.9 ( 7.2 (N ) 19) 29.2 ( 37.3 (N ) 32) 26.0 ( 81.1 (N ) 32) 26.4 ( 27.1 (N ) 32) 101 ( 85 (N ) 32 52.9 ( 29.8 (N ) 32) 20.7 ( 13.1 (N ) 32) 14.3 ( 10.2 (N ) 32) 23.2 ( 17.9 (N ) 26) 0.3 ( 0.2 (N ) 3)
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 3, 1996
TABLE 3
Comparison of Mean In-Vehicle VOC Concentrations (µg/m3) ( Standard Deviation for Car 1 (Chevrolet Celebrity) and Car 2 (Plymouth Horizon) on the Suburban Loopa compound
n-hexane isooctane 3-methylpentane benzene toluene
m&p-xylene o-xylene ethylbenzene 1,2,4-trimethylbenzene formaldehyde
car 1
car 2
3.4 ( 3.0 (N ) 14) 9.2 ( 7.3 (N ) 44) 6.4 ( 5.4 (N ) 43) 12.9 ( 7.0 (N ) 45) 35.0 ( 19.0 (N ) 45) 19.1 ( 9.7 (N ) 45) 7.8 ( 4.4 (N ) 16) 6.3 ( 7.1 (N ) 45) 9.9 ( 5.5 (N ) 32) 0.4 ( 0.2 (N ) 12)
8.5 ( 14.5 (N ) 16) 14.1 ( 21.7 (N ) 50) 10.0 ( 15.2 (N ) 48) 13.9 ( 9.4 (N ) 50) 65.8 ( 50.4 (N ) 50) 38.2 ( 32.9 (N ) 50) 16.8 ( 17.6 (N ) 16) 13.6 ( 20.2 (N ) 50) 21.0 ( 18.2 (N ) 35) 0.3 ( 0.2 (N ) 9)
a All concentration values include both ventilation conditions. Bold underscored concentrations for a given route indicate a significant difference between cars at p < 0.05.
FIGURE 4. Simultaneous measurements of selected gasoline-derived VOCs in roadway air and the passenger side breathing zone during suburban loop drives for car 1 (Chevrolet Celebrity).
FIGURE 3. Simultaneous measurements of selected gasoline-derived VOCs in roadway air and the passenger side breathing zone during suburban loop drives for the two principal vehicles.
were very low and did not show any significant trend with route driven. The mean air concentrations of all compounds except formaldehyde were greater in the interior of car 2 (Plymouth
Horizon) than in car 1 (Chevrolet Celebrity) on the suburban loop (Table 3). Formaldehyde was slightly (but not significantly at p < 0.05) higher in car 1 than car 2. The pattern of higher VOC concentrations in car 2 than car 1 also prevailed across the two phases of the urban route. In-Vehicle and Roadway Air Concentrations. The mean concentrations of all substituted aromatic compounds were higher inside the two vehicles combined than in the roadway air (Figure 3). However, when each vehicle was compared to the roadway concentration individually, the in-vehicle and corresponding roadway VOC concentrations for car 1 (the 1988 Chevrolet Celebrity) did not differ significantly at p < 0.05 (Figure 4), while the concentrations of the substituted aromatics were significantly higher (p < 0.05) in the interior of car 2 (the 1987 Plymouth Horizon) (Figure 5). In-Vehicle Concentration Differences between Vehicles. Under the high in-vehicle ventilation condition (Figure 6), higher concentrations of most compounds were recorded in car 2 than in car 1, but these differences were not statistically significant (p < 0.05). Under the low ventilation condition, all substituted aromatics and nhexane were significantly elevated (p < 0.05) in car 2 (Figure 7). Malfunctions in the fuel system of car 2, differences in the way fuel is introduced into the engine between the carburetor of car 2 and the fuel injection system of car 1, and/or differences in wear and tear between the cars may have contributed to these concentration differences. On some days, high roadway air VOC concentrations caused uncharacteristically high concentrations in both cars, as seen in sample 28 taken on March 4, 1992 (Figure
VOL. 30, NO. 3, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
813
FIGURE 5. Simultaneous measurements of selected gasoline-derived VOCs in roadway air and the passenger side breathing zone during suburban loop drives for car 2 (Plymouth Horizon).
FIGURE 6. Selected gasoline-derived VOCs measured in the passenger side breathing zone during suburban loop drives under high ventilation conditions in the two principal cars.
8). There were also days where malfunctions in car 2 resulted in VOC concentrations that greatly exceeded those of the car 1 and the roadway air. One such malfunction occurred on November 5, 1991 (sample 18, Figure 8), when the choke in car 2 malfunctioned. The butterfly valve could not completely open as the engine warmed up, resulting in a richer air/fuel mixture than normal. Some of the uncombusted mixture backed up through the butterfly valve, exited the air cleaner, and entered the engine compartment. As the vehicle was driven, the vapors entered the passenger compartment. Another concentration spike was noted on December 4, 1991 (sample 20, Figure 8), when a leak at the carburetor of car 2 caused fuel to drip onto the manifold, allowing vaporized fuel to enter the passenger compartment. A considerable difference in the VOC ratios was seen for these two malfunctions in car 2. Isooctane concentrations were much higher during the December malfunction (fuel dripping onto the manifold) than in early November (faulty choke), while the benzene concentrations were equally high in both cases. The data was reanalyzed for the low ventilation condition without days when problems were occurring in car 2. Again, the concentrations of all VOCs except benzene and formaldehyde remained higher in car 2 than in car 1, with significant differences (p < 0.05) for the substituted aromatics. Therefore, the in-vehicle concentration difference between cars appears to be a function of inherent differences in the fuel distribution systems of the two cars. On the two days when car 2 was experiencing the malfunctions, the VOC concentration difference between cars was considerably larger between car 2 and car 1, which
was not experiencing any such difficulties, under the low ventilation condition (Table 4). It is also seen, especially on December 4, that the concentration differences between cars was greater for the alkanes than for the aromatics. In-Vehicle Concentrations vs Other Studies. Several studies have measured in-vehicle VOC concentrations for automobile commuters. The concentrations of selected in-vehicle VOCs measured in this and the earlier studies are listed in Table 5. The mean concentration of benzene in this study on the suburban loop was 21% lower than the Boston study, essentially the same as downtown Raleigh, and 35% higher than the Raleigh highway loop concentration. The urban commute concentrations exceeded the levels measured in both cities. In-vehicle concentrations in the Los Angeles study were well in excess of this and those of the earlier studies. The concentrations of the other aromatics followed a similar pattern. These differences may be regional, as fuel composition, traffic patterns, and local meteorology influence roadway air concentrations. For the studies in Boston and Los Angeles, these factors are compounded by the lack of information on the repair history, type of fuel distribution system, or body integrity of the cars used. The role of self-contamination as opposed to roadway air in the observations is therefore uncertain. The samples collected in the Los Angeles study were measured just before the 1988 California Air Resources Board (CARB) Program was initiated. Presumably, the concentrations would now be lower in California after the regulations have been in place for several years and closer to the values measured for the NY/NJ region measured in this study.
814
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 3, 1996
TABLE 4
Comparison of In-Vehicle VOC Concentrations (µg/m3) for Car 1 (Chevrolet Celebrity) and Car 2 (Plymouth Horizon) on the Suburban Loop when Car 2 Was Malfunctioninga 11/05/91b
12/04/91c
compound
car 1
car 2
car 1
car 2
3-methylpentane n-hexane benzene isooctane toluene ethylbenzene m&p-xylene o-xylene 1,2,4-trimethylbenzene
11.9 [12.4] no data 17.8 [16.4] 19.1 [27.9] 44.8 [49.1] 6.3 [7.6] 21.4 [26.6] 8.3 [9.9] 14.9 [14.6]
14.6 [44.5] no data 19.1 [44.6] 20.0 [49.0] 57.8 [131] 8.3 [22.7] 29.7 [74.3] 11.3 [29.2] 8.9 [29.9]
1.3 [1.8] 0.5 [1.6] 2.2 [3.6] 2.5 [4.3] 27.0 [20.7] 1.1 [1.6] 4.7 [7.9] 1.4 [2.5] 3.6 [2.3]
2.7 [91.1] 2.3 [61.4] 5.8 [45.2] 4.5 [128] 49.0 [255] 2.5 [71.3] 10.8 [185] 3.4 [100] 3.2 [4.6]
a Values outside the brackets are the high ventilation condition observations; those enclosed in brackets correspond to the low ventilation concentrations. b Faulty electric choke in car 2. c Leak under carburetor in car 2.
TABLE 5
Mean In-Vehicle VOC Concentrations (µg/m3) for the Present and Previous Studies Raleigh Raleigh
Los Angeles
New Jersey/New York
compound
Boston 1989 urban4
1988 urban2
1988 highway2
1987 urban3
1988 urban3
1991-2 suburbana
1991-2 urbana
benzene toluene ethylbenzene m&p-xylene o-xylene formaldehyde
17.0 33.3 5.8 20.9 7.3 5.1
13.6 45.7 11.6 39.3 14.8 b
9.9 34.5 6.7 23.1 8.6 b
31.2 107 b 127 b 13.7
50.4 158 b 154 b 16.5
13.4 51.2 10.1 29.2 12.5 0.4
20.6 82.9 11.1 40.5 16.0 0.2
a
This study.
b
Compound not measured.
FIGURE 7. Selected gasoline-derived VOCs measured in the passenger side breathing zone during suburban loop drives under low ventilation conditions in the two principal cars.
Formaldehyde concentrations measured in this study were more than 1 order of magnitude lower than those in
FIGURE 8. Simultaneous measurements of benzene and isooctane concentrations in the passenger side breathing zones of the two principal cars and roadway air on the suburban loop.
the Boston and Los Angeles studies. The reason behind this difference is uncertain.
VOL. 30, NO. 3, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
815
Future Technology Directions. Carbureted engine production is decreasing in the light-duty motor vehicle fleet. As this trend continues, problems of carburetor and choke related self-contamination will diminish. Since roadway air is the primary VOC source in automobiles with fuel injection and properly functioning carburetors, regulations such as enhanced inspection and maintenance procedures can also be expected to reduce in-vehicle VOC concentrations for these vehicles. The increasing proportion of fuel-injected vehicles in combination with enhanced inspection and maintenance programs will minimize in-vehicle concentrations of gasoline-derived VOCs from the associated engine compartment sources. However, millions of carbureted automobiles and other malfunctions causing excessive emissions under the hood remain in the fleet. Occupants of these vehicles will be exposed to elevated concentrations of gasoline-derived VOCs. Although current light-duty motor vehicle pollution regulations involve overall vehicle emissions and are important for meeting ambient air quality standards, they do not directly address concentrations of the in-vehicle microenvironment. Automobile interior air should be considered in future exposure and risk analyses prior to making any decisions on the merits of new fuels and pollution control technologies for light-duty motor vehicles. Author-Supplied Registry Numbers: Benzene, 71-432; ethylbenzene, 100-41-4; formaldehyde, 50-00-0; nhexane, 110-54-3; 3-methylpentane, 96-14-0; isooctane, 54084-1; tetrachloroethene, 127-18-4; toluene, 108-88-3; 1,2,4-
816
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 3, 1996
trimethylbenzene, 108-67-8; m-xylene, 108-38-3; o-xylene, 95-47-6; p-xylene, 106-42-3.
Literature Cited (1) U.S. EPA. Fed. Regist. 1992, 57, 52950-52952. (2) Chan, C.-C.; Spengler, J. D.; O ¨ zkaynak, H.; Lefkopoulou, M. J. Air Waste Manage. Assoc. 1991, 41, 1594-1600. (3) In-Vehicle Characterization Study in the South Coast Air Basin; South Coast Air Quality Management District: Los Angeles, 1989. (4) Chan, C.-C.; O ¨ zkaynak, H.; Spengler, J. D.; Sheldon, L. Environ. Sci. Technol. 1991, 25, 964-972. (5) Weisel, C. P.; Lawryk, N. J.; Lioy, P. J. J. Exp. Anal. Environ. Epidemiol. 1992, 2, 79-96. (6) Ott, W.; Switzer, P.; Willits, N. Carbon Monoxide Exposures Inside an Automobile Traveling on an Urban Arterial Highway; 84th Meeting of the Air & Waste Management Association, Vancouver, BC; Air & Waste Management Association: Pittsburgh, 1991; Abstract 91-138.1. (7) Tejada, S. B. DNPH-Coated Silica Cartridges for Sampling Carbonyl Compounds in Air and Analysis by High Performance Liquid Chromatography; U.S. Environmental Protection Agency: Research Triangle Park, NC, 1986; Publication L-09. (8) Hampton, C. V.; Pierson, W. R.; Harvey, T. M.; Updegrove, W. S.; Marano, R. S. Environ. Sci. Technol. 1982, 16, 287-298. (9) Krost, K. J.; Pellizzari, E. D.; Walburn, S. G.; Hubbard, S. A. Anal. Chem. 1982, 54, 810-817. (10) Eichelberger, J. W.; Harris, L. E.; Budde, W. L. Anal. Chem. 1975, 47, 995-1000.
Received for review March 31, 1995. Revised manuscript received October 23, 1995. Accepted October 25, 1995.X ES950225N X
Abstract published in Advance ACS Abstracts, January 1, 1996.
