Airborne carbonyls from motor vehicle emissions in two highway tunnels.

Carbonyls (aldehydes and ketones) continue to receive scientific and regulatory attention as toxic air contaminants, mutagens, and carcinogens. Vehicle emissions are a major source of carbonyls in outdoor air, but information about the nature and magnitude of carbonyl emissions by motor vehicles is limited. The objective of this study was to identify speciated carbonyls emitted by motor vehicles under real-world, on-road conditions and to calculate on-road carbonyl emission factors. We collected air samples at the inlet and outlet of two highway tunnels, the Caldecott Tunnel near San Francisco and the Tuscarora Mountain Tunnel in Pennsylvania. At the Caldecott Tunnel, the fleet consisted almost entirely of light-duty (LD) vehicles that used California phase 2 reformulated gasoline. Vehicle count, speed and other parameters relevant to carbonyl emissions were nearly the same from one assessment to the next. At the Tuscarora Mountain Tunnel, the fleet included LD vehicles and heavy-duty (HD) diesel trucks. This part of the study was designed to capture differences in percentage of LD and HD vehicles from one assessment to the next. Air downstream of KI oxidant scrubbers was sampled on silica gel cartridges coated with 2,4-dinitrophenylhydrazine (DNPH). Carbonyls were identified as their DNPH derivatives by liquid chromatography (LC) with detection by diode-array, UV-visible spectroscopy and by atmospheric pressure negative-ion chemical ionization mass spectrometry (MS). About 100 carbonyls were identified. For about 30 of these carbonyls, concentrations were measured at the inlet and outlet of both tunnels. This information was used to calculate on-road carbonyl emission factors for LD vehicles (Caldecott Tunnel) and for the overall fleet (Tuscarora Mountain Tunnel). At the Tuscarora Mountain Tunnel, data for the fleet were used to calculate carbonyl emission factors for LD vehicles and for HD diesel trucks, the majority of which were weight class 7-8 trucks. Carbonyl emission factors at the Caldecott Tunnel were calculated as milligrams of emissions per liter of fuel consumed. Those at the Tuscarora Mountain Tunnel were calculated as milligrams of emissions per distance traveled and then converted to milligrams per liter using the fuel economy reported by Gertler et al (2000) for this tunnel (14.75 km/L for LD vehicles and 3.15 km/L for HD vehicles). At the Caldecott Tunnel, the LD vehicles emission factor was 68.4 mg/L for total measured carbonyls; the ten most abundant carbonyls were, in decreasing order, formaldehyde, acetaldehyde, benzaldehyde, acetone, m-tolualdehyde, p-tolualdehyde, methacrolein, o-tolualdehyde, 2,5-dimethylbenzaldehyde, and crotonaldehyde. At the Tuscarora Mountain Tunnel, the LD emission factor was 94.9 mg/L for total measured carbonyls; the ten most abundant carbonyls were formaldehyde, acetone, acetaldehyde, heptanal, crotonaldehyde, 2-butanone, propanal, acrolein, methacrolein, and benzaldehyde. The weight class HD 7-8 vehicle emission factor at the Tuscarora Mountain Tunnel was 82.1 mg/L for total measured carbonyls; the ten most abundant carbonyls were formaldehyde, acetaldehyde, acetone, crotonaldehyde, m-tolualdehyde, 2-pentanone, a C5 saturated aliphatic carbonyl, 2-butanone, benzaldehyde, and methacrolein. The most abundant carbonyl was formaldehyde, which accounted for 45.4% (Caldecott, LD vehicles), 40.1% (Tuscarora Mountain, LD vehicles), and 25.8% (Tuscarora Mountain, HD vehicles) of total measured carbonyl emissions. The three most abundant carbonyls, formaldehyde, acetaldehyde, and acetone, together accounted for 63.0% (Caldecott, LD vehicles), 76.5% (Tuscarora Mountain, LD vehicles), and 50.5% (Tuscarora Mountain, HD vehicles) of total carbonyl emissions. At the Tuscarora Mountain Tunnel, HD vehicles emitted more unsaturated carbonyls, aromatic carbonyls, and dicarbonyls (as a percentage of total carbonyl emissions) than did LD vehicles. For LD vehicles, less acetone and more aromatic carbonyls (as a percentage of total carbonyl emissions) were emitted at the Caldecott Tunnel than at the Tuscarora Mountain Tunnel. The highway tunnel studies described in the main body of the report also offered an opportunity to examine the role of the sampling substrate, a critical aspect of the carbonyl sampling protocol. The results are described in Appendix A. Co-located samples, one collected using a DNPH-coated silica gel cartridge and the other using a DNPH-coated C18 cartridge, were collected downstream of KI oxidant scrubbers at the inlet and outlet of the Caldecott Tunnel. Statistical comparisons of the concentrations measured for about 30 carbonyls indicated good agreement between silica gel cartridges and C18 cartridges for about 25 carbonyls, including formaldehyde and acetaldehyde. Concentrations of acetone and 2-butanone measured using C18 cartridges were lower than those measured using silica gel cartridges.