Asphalt-related emissions are a major missing nontraditional source of secondary organic aerosol precursors

Abstract

Asphalt-based materials are abundant and a major nontraditional source of reactive organic compounds in urban areas, but their emissions are essentially absent from inventories. At typical temperature and solar conditions simulating different life cycle stages (i.e., storage, paving, and use), common road and roofing asphalts produced complex mixtures of organic compounds, including hazardous pollutants. Chemically speciated emission factors using high-resolution mass spectrometry reveal considerable oxygen and reduced sulfur content and the predominance of aromatic (~30%) and intermediate/semivolatile organic compounds (~85%), which together produce high overall secondary organic aerosol (SOA) yields. Emissions rose markedly with moderate solar exposure (e.g., 300% for road asphalt) with greater SOA yields and sustained SOA production. On urban scales, annual estimates of asphalt-related SOA precursor emissions exceed those from motor vehicles and substantially increase existing estimates from noncombustion sources. Yet, their emissions and impacts will be concentrated during the hottest, sunniest periods with greater photochemical activity and SOA production.

Fig. 1. Asphalt’s life cycle and temperature-dependent emissions.

(A) Different stages in asphalt’s life cycle with potential to emit reactive organic gases into the atmosphere. (B) Temperature dependence of total gas-phase emissions from asphalt, ranging from in-use (40° to 60°C) and storage (80° to 140°C) to paving and overheating (120° to 200°C) temperature conditions (filled circles). The corresponding potential SOA that could be produced is also shown (hollow squares). The orange, green, and blue curves show the IVOC, SVOC, and aromatic fractions of the total emissions, respectively. The error bars indicate SD of the emission and SOA production factors, and the error bands indicate the SD in the volatility fractions. (C) Variation in the emission factors of select hazardous PAHs and total n-alkanes (C₁₀-C₃₂) with applied temperature.

Fig. 2. Detailed chemical composition of hydrocarbons and functionalized organic compound emissions.

Laboratory test results of the chemical composition of gas-phase complex mixture emissions from PG 64-22 road asphalt at typical in-use (60°C) and paving (140°C) temperatures. The emission factors and volatility distributions are shown in (A) and (D) for the hydrocarbon (C_xH_y) emissions, (B) and (E) for sulfur-containing compound (C_xH_yS) emissions, and (C) and (F) for oxygen-containing compound (C_xH_yO) emissions. The legend indicates distribution of molecular structures within each carbon number bin ranging from n-alkanes to PAHs. The trace signals on PAHs with formulas C_xH_2x−20-C_xH_2x−28 (and corresponding heteroatom-containing formulas) are lumped together and labeled in black.

Fig. 3. Changes in temperature-related emissions over prolonged heating and large emission enhancements with solar exposure.

(A and B) Variation in total gas-phase emission factors over time from commonly used PG 64-22 road asphalt as observed in laboratory experiments simulating (A) in-use temperatures (60°C) and (B) paving temperatures (140°C). Individual markers for the C_xH_yS and C_xH_yO fitted decay curves in (B) are shown in fig. S7, including for 60°C where heteroatom-containing emissions were minor and did not exhibit a clear decaying trend. Additional details including SDs on equations in (A) and (B) can also be found in fig. S7. (C and D) Emission enhancements due to solar exposure shown with total gas-phase emission factors (solid circles) and SOA production factors (hollow squares) over time for (C) PG 64-22 road asphalt under solar exposure at 60°C and (D) commonly used liquid roofing asphalt under solar exposure at 75°C. The orange, green, and blue curves in both panels show the IVOC, SVOC, and aromatic fractions of the total emissions, respectively, and are shown as a function of discrete time points on the x axes. The error bars and error bands indicate SD in emission factors and SOA production factors, respectively. Emissions from liquid roofing asphalt exclude any potential contributions from Stoddard solvent off-gassing.

Fig. 4. Other asphalt-containing materials have similar emissions and solar enhancements.

(A) Summary of emission factors and SOA yields for different highly used asphalt-based materials compared to gasoline and diesel emissions. The red circles and red-bordered squares indicate emission factors and SOA yields, respectively, determined in the absence of solar exposure, while the blue circles and blue-bordered squares show the corresponding values measured with the material exposed to solar radiation. Variation in chemical speciation of hydrocarbon (C_xH_y) emissions observed during laboratory tests without (left) and with (right) sunlight are shown in (B) and (C) for asphalt shingles at 75°C, (D) and (E) for asphalt-based sealant at 75°C, and (F) and (G) for commonly used liquid roofing asphalt at 75°C. Hydrocarbon emissions made up more than 80% of total emissions in most cases.

Fig. 5. Consistent results from other asphalt samples and ambient measurements.

(A and B) Chemical speciation of complex gas-phase hydrocarbon emissions from ulterior (i.e., minor) road asphalt collected from Pittsburgh, PA, during laboratory tests at (A) 40°C and (B) 60°C. In both cases, hydrocarbons made up over 95% of the total emissions. (C and D) Confirmatory ambient measurements following road asphalt application demonstrate similar emissions (shown as a function of carbon number) for (C) primary roadway asphalt immediately following application in New Haven, CT (detailed composition in fig. S9) and (D) over 3 days of measurements at an ulterior roadway in Pittsburgh, PA, both compared to laboratory experiments. To evaluate continued vertical fluxes (i.e., emissions) over 3 days, the main plot in (D) shows the aromatic vertical concentration differentials at 8 cm versus 2 m (i.e., Conc. at 8 cm − Conc. at 2 m) via simultaneous adsorbent tube collection at the 2 heights. Single-ring aromatics and PAHs are shown here to remove biogenic interferences, but supporting vertical gradient data via GC-TOF and GC-EI-MS for alkanes, aromatics, and PAHs can be found in figs. S9 and S10. The inset in (D) shows the sum of vertical concentration differential for aromatic compounds. The field results from the 3 days show good agreement with laboratory data, including SVOC enhancements with solar exposure.