summer and winter days. The emissions speciation of 81 hydrocarbons is designed for use with the Canadian AURAMS model, and will hereafter be referred to as the AURAMS81.
Since the previous year’s report, biogenic VOC emissions have been added, using the CEPS CBEIS emissions processor. The RO2’s produced by biogenics dominate in terms of the net effect across the grid (see Figure 1), but in urban centres, the anthropogenically emitted RO2
precursors may have an equal (Atlanta and south-eastern US cities) or up to six times greater impact (Northern US and Canadian cities) than the biogenic RO2’s. The implication of this finding is that VOC controls will be less effective in southern regions in which the biogenic emissions are high, and will be more effective in northern urban regions than outside of the urban areas.
The potential for the formation of organic particulate matter from each of the VOCs was also examined. This was done in two stages. VOCs with chemical properties allowing partitioning to the condensed phase directly after emissions were used in conjunction with the absorption partitioning formulae of Pankow (1994a,b) to determine directly partitioned particle mass.
The 20 VOC classes in the AURAMS81 speciation scheme considered to be potential condensable VOCs are listed in Table 1. Many of these classes were selected on the basis of observational evidence showing their presence in atmospheric aerosols. Others (methyl, ethyl, and propyl alcohol, di-and tri-alcohols, esters, and petroleum distillation spirits) were included on the basis of high aqueous solubility and boiling points higher than the typical range of ambient temperatures, characteristics that might allow the given compounds to exist in the liquid or aqueous phase under ambient conditions.
Table 1.
Species or Species Class Water
Alkanes (C20 to C26) Alkanes (C27 to C35) Alkanes (C36 to C43)
High-C-Number Alkyl Phthalates Low-Vapour-Pressure PAHs Phenol
Cresols C6+ Aldehydes Aromatic Aldehydes C6+ Ketones
Formic Acid Acetic Acid C4+ Diacids Methyl Alcohol Ethyl Alcohol Propanol C4+ Alcohols Di- & Tri-Alcohols Low-Reactivity Esters High-Reactivity Esters Petroleum Distillation Spirits Furandiones
The “direct partitioning” calculation was followed by a calculation for the amount of organic aerosol mass produced by oxidation of the more volatile emitted compounds using the parameterization of Odum et al. (1996). The AURAMS81 VOC classes considered for VOC oxidation/SOA formation and the sources for αi and Ki data are listed in Table 2. Note that three of these classes − Phenol, Cresol, and Aromatic Aldehydes − are also listed in Table 1. It can be seen from Table 2 that 9 of these 19 species classes have laboratory-based αi and Ki values available; the remaining 10 have been observed to produce organic aerosols in chamber experiments, but ORGM-dependent yields are not available at the current time. For the latter species, αi and Ki were assigned from the 9 species for which data are available – the values for these species are therefore approximations, and subject to future revision.
Table 2.
VOC Species or Class Alpha-Pinene
Beta-Pinene D-Limonene D-3-Carene End C9-19 Alkenes Toluene
Mono-Alkyl Aromatics Di-Alkyl Aromatics Tri-Alkyl Aromatics Alkene Aromatics Cresol
C9-10 Alkanes C11-19 Alkanes End C6-8 Alkenes Internal C6-8 Alkenes Internal C9-19 Alkenes Napthalene
Aromatic Aldehydes Phenol
The results of these calculations over the model grid showed that the effect of the high-molecular mass, directly partitioning species may be as important or more important towards organic particle formation than the formation of aerosols through secondary oxidation.
Figure 2 shows the net grid condensed-phase VOC mass; the main part of the mass is taken up by high molecular mass “alkane” groups (each of these are actually an amalgam of many hydrocarbons of similar structure and mass) and anthropogenic diacids. Figure 3 shows how the composition may vary between individual cities. For large urban centres (New York City, Toronto), the SOA portion of the aerosol mass is equal to the directly condensed aerosol mass.
Main conclusions
The most significant conclusion on the work to date is that a significant proportion of the organic aerosol mass may originate in VOCs which were not resolved in previous VOC emissions speciations. These VOCs make up a small fraction of the total emitted hydrocarbon mass, and have a small impact on reactivity, but may outweigh the SOA fraction of organic aerosols. Emissions inventories which lack this level of speciation will seriously underestimate the total organic aerosol mass.
Other conclusions stemming from this work:
• Biogenically emitted species were found to have a lower impact than anthropogenic species (noting that the calculations here included local effects only; no transport and hence interaction between biogenic and anthropogenic species was considered).
• The relative contribution of different VOCs towards particle production varies substantially between different urban centres.
• RO2 production from anthropogenic emissions was shown to be approximately six times higher than biogenic RO2 production in northern cities, but approximately equal to biogenic RO2 production in southern cities.
• Biogenic emissions are dominated by isoprene, high-reactivity esters, and alpha-pinene on a mass basis, and by isoprene, high-reactivity terpenes, and internal C6-8 alkenes on a reactivity basis. Anthropogenic emissions were dominated by C4-5 n- and iso-alkanes, C6-8 n- and iso-alkanes, and toluene on a mass basis, and by internal C4-5 alkenes, internal C6-8 alkenes, ethene, and high-reactivity amines and amides on a reactivity basis. The latter group of compounds is usually not resolved in regional-air-quality-model VOC speciations, while the current work suggests that it may be the fourth greatest contributor to RO2 production. Its inclusion in future reaction mechanisms is therefore recommended.
Aim for the coming year
Two papers have been submitted for review with the Journal of Geophysical Research on the work. The final stage of the work will be to install the new emissions speciation into the AURAMS regional air quality model, allowing these effects to be examined in the context of transported, size and composition resolved aerosols. The final stage of the current study will be an examination of the effects of reactivity lumping on ozone and particulate matter formation through 3D model simulations.
Acknowledgements
This work has been supported by the Air Quality Research Branch of Environment Canada.
References
Middleton, P., W.R. Stockwell and W.P.L. Carter; Aggregation and analysis of volatile organic compound emissions for regional modelling, Atmos. Environ. 24 (1990) 1107-1123.
Odum, J.R., T. Hoffman, F. Bowman, D. Collins, R.C. Flagan and J.H. Seinfeld; Gas/particle partitioning and secondary organic aerosol yields, Environ. Sci. Technol. 30 (1996) 2580-2585.
Pankow, J.F.; An absorption model of gas/aerosol partitioning involved in the formation of secondary organic aerosol, Atmos. Environ. 28 (1994b) 189-193.
Pankow, J.F.; An absorption model of gas/particle partitioning of organic compounds in the atmosphere, Atmos.
Environ. 28 (1994a) 185-188.