Nature of unresolved complex mixture in size-distributed emissions from residential wood combustion as measured by thermal desorption-gas chromatography-mass spectrometry

Unresolved complex mixture (UCM) is an analytical artifact of gas chromatographs of combustion source-related fine aerosol extracts. In this study the UCM is examined in size-resolved fine aerosol emissions from residential wood combustion. The aerosols are sorted by size in an electrical low-pressure impactor (ELPI) and subsequently analyzed by thermal desorption/ gas chromatography/ mass spectrometry (TD/GC/MS). A semiquantitative system for predicting the branched alkane, cycloalkane, alkylbenzene, C-3-, C-4-, C-5-alkylbenzene, methylnaphthalene, C-3-, C-4-, C-5-alkylnaphthalene, methylphenanthrene C-2-, C-3-alkylphenanthrene, and dibenzothiophene concentrations in the UCM is introduced. Analysis by TD/GS/MS detects UCM on each ELPI stage for all six combustion tests. The UCM baseline among the different fuel types is variable. In particular, the UCM of Pseudotsuga sp. is enriched in later-eluting compounds of lower volatility. A high level of reproducibility is achieved in determining UCM areas. UCM fractions (UCM ion area/total extracted ion chromatograph area) by individual ELPI stage return a mean relative standard deviation of 19.1% over the entire combustion test set, indicating a highly consistent UCM fraction across the ELPI size boundaries. Among the molecular ions investigated, branched alkane (m/z 57) and dibenzothiophene (m/z 212 and 226) constituents are most abundant in UCM emissions from RWC, collectively accounting for 64-95% of the targeted chemical species. The total UCM emissions span 446-756 mg/kg of dry biomass burned and correspond to an upper limit of 7.1% of the PM2.5 mass. The UCM emissions are primarily accumulation mode (0.1 mum less than or equal to aerodynamic diameter (d(a)) less than or equal to 1 mum), with a geometric mean diameter (d(g)) range of 120.3-518.4 nm. UCM in PM2.5 is chemically asymmetric (shifted to finer d(a)), typically clustering at d(a) less than or equal to 1 mum. Measurable shifts in d(g) and changes in distribution widths (sigma(g)) on an intratest basis suggest that the particle density may be a function of size within PM1. The potential effects these results have on regulatory affairs, human health studies, and the state of the analytical science covering organics in PM2.5 are discussed.