Comprehensive laboratory measurements ofbiomass-burning emissions:2. First intercomparison of open-path FTIR,PTR-MS, and GC-MS//FID//ECDT. J. Christian,1B. Kleiss,2R. J. Yokelson,1R. Holzinger,2,3P. J. Crutzen,2,4W. M. Hao,5T. Shirai,6,7and D. R. Blake6Received 17 June 2003; revised 14 October 2003; accepted 7 November 2003; published 28 January 2004.[1] Oxygenated volatile organic compounds (OVOC) can dominate atmospheric organicchemistry, but they are difficult to measure reliably at low levels in complex mixtures.Several techniques that have been used to speciate nonmethane organic compounds(NMOC) including OVOC were codeployed/intercompared in well-mixed smoke generatedby 47 fires in the U.S. Department of Agriculture Forest Service Fire Sciences CombustionFacility. The agreement between proton transfer reaction mass spectrometry (PTR-MS)and open-path Fourier transform infrared spectroscopy (OP-FTIR) was excellent formethanol (PT/FT = 1.04 ± 0.118) and good on average for phenol (0.843 ± 0.845) and acetol(0.81). The sum of OP-FTIR mixing ratios for acetic acid and glycolaldehyde agreed(within experimental uncertainty) with the PTR-MS mixing ratios for protonated mass 61(PT/FT = 1.17 ± 0.34), and the sum of OP-FTIR mixing ratios for furan and isoprene agreedwith the PTR-MS mixing ratios for protonated mass 69 (PT/FT = 0.783 ± 0.465). Thesum of OP-FTIR mixing ratios for acetone and methylvinylether accounted for most of thePTR-MS protonated mass 59 signal (PT/FT = 1.29 ± 0.81), suggesting that one of thesecompounds was underestimated by OP-FTIR or that it failed to detect other compounds thatcould contribute at mass 59. Canister grab sampling followed by gas chromatography (GC)with mass spectrometry (MS), flame ionization detection (FID), and electron capturedetection (ECD) analysis by two different groups agreed well with OP-FTIR for ethylene,acetylene, and propylene. However, these propylene levels were below those observed byPTR-MS (PT/FT = 2.33 ± 0.89). Good average agreement between PTR-MS and GC wasobtained for benzene and toluene. At mixing ratios above a few parts per billion the OP-FTIR had advantages for measuring sticky compounds (e.g., ammonia and formic acid) orcompounds with low proton affinity (e.g., hydrogen cyanide and formaldehyde). Even atthese levels, only the PTR-MS measured acetonitrile and acetaldehyde. Below a few ppbvonly the PTR-MS measured a variety of OVOC, but the possibility of fragmentation,interference, and sampling losses must be considered.INDEX TERMS: 0315 AtmosphericComposition and Structure: Biosphere/atmosphere interactions; 0322 Atmospheric Composition and Structure:Constituent sources and sinks; 0365 Atmospheric Composition and Structure: Troposphere—composition andchemistry; 0394 Atmospheric Composition and Structure: Instruments and techniques; KEYWORDS: instrumentintercomparision, biomass burning, oxygenated organic compoundsCitation: Christian, T. J., B. Kleiss, R. J. Yokelson, R. Holzinger, P. J. Crutzen, W. M. Hao, T. Shirai, and D. R. Blake (2004),Comprehensive laboratory measurements of biomass-burning emissions: 2. First intercomparison of open-pathFTIR, PTR-MS, and GC-MS/FID/ECD, J. Geophys. Res., 109, D02311, doi:10.1029/2003JD003874.JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 109, D02311, doi:10.1029/2003JD003874, 20041Department of Chemistry, University of Montana, Missoula, Montana,USA.2Atmospheric Chemistry Department, Max Planck Institute forChemistry, Mainz, Germany.3Now at Department of Environmental Science, Policy and Manage-ment, University of California, Berkeley, California, USA.Copyright 2004 by the American Geophysical Union.0148-0227/04/2003JD003874$09.00D023114Also at Scripps Institution of Oceanography, University of California,San Diego, La Jolla, California, USA.5Fire Sciences Laboratory, USDA Forest Service, Missoula, Montana,USA.6Department of Chemistry, University of California, Irvine, California,USA.7Now at Earth Observation Research Center, Japan AerospaceExploration Agency, Tokyo, Japan.1of121. Introduction[2] It is widely accepted that detailed models are neededto quantify the complex chemistry of the atmosphere. It isalso true that the output of these models is sensitive to tracegas species present at very low levels. A dramatic exampleof this is O1D, a key species in atmospheric photochemicalmodels that is present at levels so low that it may never bemeasured in the troposphere [Albritton et al., 1990]. OH andHO2(collectively known as HOx) are two other species ofprime importance (largely derived from O1D) that havebeen measured reliably only recently [Mount and Williams,1997; Brune et al., 1998].[3] The organic chemistry of the atmosphere is of greatinterest for many reasons including its major influence onO3(and thus O1D), HOx, and the oxidizing effi ciency(power) of t he atmosp here [ Singh et al., 1995, 2001;McKeen et al., 1997; Mason et al., 2001]. Until recently,it was widely assumed that the organic chemistry of theatmosphere was well understood, mainly because hydro-carbons (which were thought to be the main organic con-stituents of the atmosphere) are routinely speciated at partsper billion or trillion (ppb, 109, ppt, 1012) levels by avariety of GC-based techniques [e.g., Blake et al., 1996].The hydrocarbon/O3observations were reasonably consis-tent with atmospheric models, especially for smog chambers[Carter et al., 1979] or urban airsheds [McRae and Seinfeld,1983]. However, in the 1990s, it became clear that OVOCaccount for most of the NMOC from biomass burning[Griffith et al., 1991; Yokelson et al., 1996, 1997, 1999,2003a; Worden et al., 1997; Holzinger et al., 1999] and alarge fraction of the biogenic emissions from plants [Ko¨niget al., 1995; Kirstine et al., 1998; Schade and Goldstein,2001]. Together these sources are est imated to producemore total trace gases and more VOC than the main globaltrace gas source: fossil fuel burning [Schimel et al., 1995;Andreae and Merlet, 2001; Guenther et al., 1995]. Inaddition, the nonmethane hydrocarbons (NMHC) emittedby fossil fuel burning quickly become OVOC in theatmosphere, mainly by reaction with OH. The above obser-vations probably help explain why recent campaigns tocharacterize the background troposphere found that OVOCwere up to five times more abundant than NMHC [Singh etal., 2001], more reactive than the NMHC [Singh et al.,1995] and consequently more important. Further, currentphotochemical models (at various local-global