JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 104, NO. D17, PAGES 21 685-21 693, SEPTEMBER 20, 1999, Soot and sulfate aerosol particles in the remote marine troposphere Mihély Pésfai,! James R. Anderson, and Peter R. Buseck Deparment of Geology and Chemistry, Arizona State Univesity, Tempe Herman Sievering Global Change and Enviroamental Quality Program and Center for Environmental Sciences, University of Colorado at Denver Abstract. Sulfate aerosol particles containing soot aggregates were obsorved inthe marine troposphere in both hemispheres under condition that ranged from extremely clean to heavily polluted. Even in clean air above the remote Southern Ocean during the First Aerosol (Characterization Experiment (ACE 1), depending on the sample, between 10 and 45% of sulfate particles contained soot inclusions. We identified aircraft emissions and biomass burning asthe most likely major sources of soot. internally mixed soot and sulfate appear to comprise a globally significant fraction of aerosols in the troposphere. Anthropogenic combustion aerosols can thus potentially change the radiative climate effects of sulfate aerosols and may have an impact an cloud properties even in the remote troposphere. 1. Introduction Sulfate ad carbonaccous particles are the two most important anthropogenic aerosol constituents influencing climate [Charleon aly 1992; Chuang ot a, 1997; Haywood and Ramaswamy, 1998! Schult et al. 1997], Riomass huming. [Crutzen and Andreae, 1990; Hobbs et al, 1997] and fosi uel combustion [Penner, 1995) produce both werosol types. Sulfates axe aso fixed fou gases emited by marin soures [Charlson ef al 1987] and oceur across the globe, The concenration of black carbon (BC) is variable: 800 ng/m* in continental rural regions [Penner ot al, 1995}, 40) ngim? in the Arctic (Bodhaine, 1995). and 20 ng/m? in remote oceanic regions [Clarke, 1989; Raw and ‘ha, 1993) however, it sees Wo be ubiquitous in atmosphere aerosols [Cooke al, 197 Liousse eal, 1993]. “The terms "BC" and "soot" are often used interchangeably. As sssed by Penner and Novakow 199], soot refers to primary ‘omiuston-generafed carbonaceous aerosol, whereas BC Is used to emphasize the llgheabsorbing propery vf cabonaccous pattces, Sine i his study we use microscope methods nd the bbserved carbonaceous parisles have morphological features, characters of combustion aerosol, we use the tem soot instead of RC Whereas sullates and thet climae-cooling effects have been studied extensively during the past decade, the globe importance ‘Of soot aciosls is just being recognized (Cooke und Wilson, 1996; Liousse et al, 1996; Haywood and Ramaswamy, 1998; Sehulte al, 1997]. Soot is generally the major lightabsorbing ermal species [Cipot of al, 1995) can also contnibute 10 cloud condensation nuclei (CCN), thereby changing the raditve Now at Bepanment of Eanh and Environmental Sciences, University ‘of Vesapém, Veszprém, Hungary Copyright 1999 by the American Geophyical Union. Paper number 199950900208. ‘0148-0227/997199950900208809.00 properties and microphysics of clouds [Hudson er al, 1991; Kayfnan and Praser, 997]. ‘The combined climate effects of soot and sulfate aerosols are difficult to estimate. There are insufficient data onthe ‘concentrations and vertical distribution of soot inthe atmosphere [aywood and Ramaswamy, 1998}; liule is also known about ‘what fraction, if any, of the soot and sulfate particles are agarcgated with one another (internally mixed) in remote regions. Sulfatlsoot aggregates have been observed in urban and polled continental and marine atmospheres [Katrinak at al, 1997; Okada, 1985; Mészéros and Mészdros, 1989; Parungo et al. 1992, 1994] and some investigations huve shown correlation between sulfate and DC concentrations in remote marine regions [O'Dowd et al, 1993; Van Dingenen ot a, 1995; Clarke etal, 1996). Knowledge of the extent of internal mixing is important for understanding the mechanisms of sulfate formation and soot rewoval and for assessing the climate effets of both types of particles. For example, if the sulfates contain strongly absorbing inclusions such as soot, then their direct climate cooling effect is moderated [Chek et a, 1995]; on the other hand, the originally hydrophobic soot particles can become active CCN if they are ‘coated with hygroscopic sulfites. ‘We used transmission electron microscopy (TEM) to study soot asrozol particles in order to identify their possible sources and to determine whether individual sulfte particles contain soot inclusions in (1) the polluted marine houndary layer (MBL), (2) ‘the relatively clean free troposphere (FT) above the Atlantic Oceun, and (3) he remote, undisturbed MBL and FT above the Souther Ocean. Our study provides the frst direct evidence that 1 significant fraction of sulfate particles occurs internally mixed ‘with soot even in the essentially npolluted trapnsphere 2, Experiment Results obtained ftom three sets of aerosol. samples. are discussed in this paper. We collected aerosol particles during the 21,68521,686 1992 Allanic Stratocumulus Transition Experimen/Marine ‘Actosol and Gus Exchange (ASTEX/MAGE) experiment sbove the North Alantic Ocean nea the Azores Islands on June 19 and 20, days on which apolited air mas of European origin artived atthe sampling site [Hubert eta, 1996]. The results fom the ASTEX/MAGE experiment are use here a a efrence to which other TEM observations can be compared. Also included are results from two sazples Ua were culled in he North Anti FT during the Second Aerosol Characterization Experiment (ACE 2) Much of our discussion focuses on data obtained fiom 2 Set of Samples that was collected during the Necemer 1095 ‘ACE I campaign, conducted above the Souther Ocean, neat Tasmania [Dates ea, 1998) “The ASTEX/MAGE samples were collected on the bow tower of the research vessel Malcolm Baldrige, 18-m above sea level Particles were colected directly on TEM grids with carbone coated Formvar fms. The grids were placed onthe fourth stage of a Casella cascade impactor te Rourth stage of the impactor callets parila with nominal diamcters inthe 0.2-to 0.7-um range. Sampling conditions and the analyses of third sage impactor samples are given by Paso eta. (1995) Cur ACE 2 samples were collected at a 256U-m alttude on lua, Teerfe, Canary Islands. Cu TEM grids with Roey carbon films wore placed on the lower stage ofa two-stage filter stack. The upper cut size is 1. ym, and particles downto 0.08 jum were collected. Results fm twa simples ate discused here; one was coileted trom clean marine air (on July 20, 1997) wheres the other was fom an air mas that coulsned mineral dust fom Africa (on July 25, 1997). hring the ACE 1 experiment we sampled aerosols with a ‘one-stage impactor on the National Center for Atmospheric Resear (NCAR) C-130alrerat and obralned TEM samples in clear air at altitudes ranging from 30 to S400 a during Lagrangian experiment "B" on Deceuct 7, 8, ad 9 (ghts 24, 25, and 26). The community aerosol inlet (CAN) was used itis described hy Rates eral. [1998]. According to Hucbert et al 11998), the CAI allowed te efficient sampling of aerosol pavtclessialler than 3.5 ym. The impactor had @ nominal lower ut sizeof 0.2 pm; howeves, we also observed s00t paricles smaller than .1 jm In this Lagrangian experiment, an air mass was tagged by balloons and tracked so tt the same aie mess could be sampled sepeatelly during each of three research fights. Because the sir flew circles with diameters uf 60 km relative 0 one ofthe balloons, the chance of our actualy sampling the exhaust of out ‘own airplane was negligible. Contamination was detetable by spike-ike increases in NO and aerosol concentrations when We Jncenionalycrosed the exhaust plume. Most of the ACE 1 samples are representative of a clean marine scrosol; the particular air parcel for the Lagrangian ‘experiment was chosen because it had ot pased over ld for several days prior to ariving at the sampling site. Median concentrations of Ka, a passive tracer of continental si, were alighly lower inthe baseline sector at Cape Grim in December 1995 than during the preseing 4 years (1990 wo 1994), whereas elemental carbon conccatrations were slightly greater (9.3 ng ur? compared with 7.3 ng m® during the reference period) Uainswarth eta, 1998) Concentrations of fine aerosol (2 um diameter) non-see-sal sulfate were lower than normal throoghout December, with average values around | nmol m® (Hainsworth etal, 1998) For TEM analysis, we weed a JEOL 2000°X election microscope operated” st a 2004V accelerating voltage POSFAI ET AL.: SOOT AND SULFATE AFROSOL Compositions were determined by energy-dispersive X-ray spectrometry (KIS) using an attached uitrathin-window detector that allowed the detection of C and heavier elements. For X rays mitted by elements ighter han Na, dhe absorption within the specimen is large, and the overlapping peaks make intensity ‘measurements dificult. Since the partileedizoussed inthis study contained mainly §, N, O, and C, we obtained qualitative results ‘from EDS spectra. The crystal structures of ammonium sulfate puiticles were confirmed using selected-area electron diffraction (SAED). High-resolution images were obtained using a 400-KV, JEOL 40008 electron microscope. 3. Results 3.L. SulfatesSoot in the Polluted North ‘Adluntic MBL (ASTEX/MAGE) In the North Atlantic MBL samples fom ASTEX/MAGE, the submicron fraction consists almost entirely of ammonium sulfate particles. A combined ‘TEM-atomic force microscopy (AFM) sudy [Pésfat et al, 1998} showed that the particles become Matteued ou the collection surface and eit appascat diameters (ae aoen on the TEM grid; Figure 1a) are about double those of the original spherical, airborne particles. These sulfates typically contain soot (Figure Ta). Soot aggregates are either compact oF ‘chain-like associations of spherules (Figure 1b) with diameters of 10 to 30 um, The splicrules within the aggregates ae typically attached to one another and coated by a carbon film (Figure 1b); their structures are revealed by high-resolution TEM (HRTEM) ‘images that show curved, disordered graphitic layers, as is ‘common for combustion soot [Lahaye and Prado, 1981}. Where spacings can be measured between the rather chaotically aanged layers, larger values are obtained than the layer spacing (0.335 nm) in graphite (Figure 1d). In addition to C, the soot aggregates typically contain S; some also contain minor Na, K, Si, of Ca. Spherules of fly ash are present (Figure Ie) and are amorphous, mainly consisting of SiO. with variable Na, Mg. K, Ca, Fe, Ma, and Zn, The association of soot with silica fly ash is consistent ‘9 coal-combustion origin. [Ramsden and ‘Shibaoka, 1982; Sheridan, 1989}, Sont-tfee sulfate particles from ASTEX/MAGE tend to be smaller than soot-bearing ones; about 50% ofthe smallest (0.1- to 0.2-yun diameter) and as much as 90% of the larger (~1 um) sulfite particles contain soot. There is almost no "ice soot in dhe samples; all soot aggregates are internally mixed with sulfates ‘These results are in agreement with aerosol volatility studiee (Clarke et al, 1996, 1997] that indicate an internal mixture af sulfate with @ nonvolatile species within the pollution plume that vas sampled during ASTEX/MAGE. The sovUullate particles were probably advected from the continental atmosphere and aged in the MBL; the number concentration of sulfate aerocal ‘could have heen determined hy the snot and fly-ash particles that provided condensation sites for sulfate. The good agreement ‘etweeu vur resuls und those of Clarke er a. (1996, 19971 serves 1s independent proof that the aggregation of soot with sulfate, as s2en by TEM, is not an artifuet but represents truc atmospheric distributions ofthe two species relative to each other. 3.2. Sulfate/Soot Above the Southern Ocean (ACE 1) Soot/sulfate ageregntes occur in all samples from ACE I that Were collected fom the C-130 during Lagrangian experiment “BY; collections were at altitudes ftom 30 10 5400 m above the Southem Ocean, Depending on the sample, 11 w 46% of sulfitePOSPAIET AL. SOOT AND SULFATE AEROSOL Figure 1. TEM images of werosol paticles collected in the polluted MBL of the North Atlantic Ovcan, near the ‘Azores Islands. (a) Ammonium sulfite particles containing soot (marked by the small arous) and fly-ash spheres (marked by the bold arrow in the lower-right comer). (b) A typical chain-like soot agerezate; the utrows point to a carbon film that connects individual spherules within the aggregate. (c) Fly-ash spheres, consisting of amorphous silica. (4) High-resolution image showing di ‘araphitic layers within a soot aggregate. Ihe background of ‘the image Isa support film of amorphous cazbuu outo which the patcles were collected. The inst in the upper-left comer isa copy ofthe boxed area; the graphitic ayers are highlighted inthe inset 21,68721,688 particles contained soot (Figure 2). The ratio of soot-beating to sootrice sulfates scemed to increase slightly with the sizeof the host sulfite particles. Air mass trajectories for most samples did not pase overland for atleast several days prior to arriving atthe sampling site (see map insert in Figure 2), suggesting that the observed sooUsulfale aggregation is representative for an ‘unpoltuted marine aerosol [No significant differences in aggregation exist among the sets of particles collected during the flights of Lagrangian experiment B (Figure 2), indicating that particle aging did not play a major forming sooUsulfate aggregates over the 33 hours of the 0 10 20 30 40 % of eulfate particles that contain soot 50 POSPAILT AL: SOOT AND SULTATE AEROSOL expetimeul, This observation peiiod is long enough t0 monitor changes in the acrosol; the aging of sca-salt particles as they ‘converted into culfates was observed during the came experiment [Buseck and Péafai, 1999), The aggregation of soot with sulfate likely occured hefore the air mass arrived at the first sampling site, The differences between the ratios of soot-contaning sulfates in the samples collected at an allude of 5.4 kin may result fin the fact thatthe exact same air mass may not have been sampled; as indicated by the air mass bock-trjestories and relatively high CN counts [Hudson ot a, 1998}, the sample from the last flight (26) may have been’slghily influenced hy 0-90. san 440190 490 san x80 ACEt Flight #26 Bi Fright #25 BB Flight #24 Flight #23 ACE2 | ‘clean’ conditions “dusty” conditions Figure 2, Peroenuge of soot-contaning sulfates relative to all (00.1 yam) sulfate panicles. The samples were collected at several altitudes above the Souther Ocean, during a Lagrangian experiment of ACE 1. In this experiment, the same tagged air mass was followed and sampled in three research lights (Rights 24, 25, nd 26; ‘ight 23 is pre-Layrangian fight) over 33-hour period. Two North Antic samples (ACE 2) ae also included in the dingram,; thece were collected at an altitude of 2.4 km on Izana, Canary Islands. Between 100 and 200 particles were shied in each sample (fora total of aver 2300 particles), randomly chasen fenm several regions on ‘each TEM grid. The estimated eur in the indicated percentages is +8, 3% since some small soot particles may hhave escaped attention because oftheir low contrast in the TEM, underestimating the number of soot containing sulfates is more likely than overestimating it ‘The map inthe upper-right comer shows three types of generalize, ‘36-hour air was back tajectories for dhe Southern Oveau samples; dhe coresponding numbers are marked on the ‘columns that represent the ACE 1 samples in the diagram.POSPAIET AL: SOOT AND SULFATE AEROSOL, Ausraian sources. Refatvely high concentrations of soot bearing sulfates occur in the to 0.9-km samples; these samples were collected from a"buler aye” tht separated the MBL from the FT [Russell et al, 1998]. In general, the faction of soot- hearing slfates seems to increase with lite, atlas for those sulfate particles tht are lager than 0.1 um, those tht were efcienly collected on ou grids ‘Typieal soot-bearing ammonium sollte patles fom the Souther Ossan wopoophers are shown in Figures 3,36, and 3d Some sulfites are surrounded by halos of smaller particles (Figure 32); such halos are artifacts resulting from interactions hetueen the hydrated particle and the collection substrate and presumably indicate the acidic nature of the original suite [Sheridan eral 1993; Pésfat er al, 1994). However, ekecron Aifiaction patems confirm that the particles are crystalline fmmonium sulfate when they are in the TEM; cither during storage or upon exposure tothe vacuum ofthe TEM, the assumed acidic sulfates presumably converted into ammonium sulfite The sulfite can be easily sublimed with the electron beam, leaving behind a S-bearing residue and the soot inclusions (Firure 30. Some soot aggregates are only tens, whereas others, ae ulieds of wanomcirs across. In general, the s00t ggrcgatesffom the Souther Occan consist of 10- to 20-nm spheres that do not seem to be atached by a common film (Figue 36) and are smaller and have more ordered structures than those in the ASTEX/MAGE samples (compare Figure 1b. High-magnitication images (Figure 3d) show that i the typical Soot paricls from the Southern Ooean several wavy but roughly conventic waphie layer are uistingubhuble win each spherul, forming onionlike structures (Figure 3a). Soot occurs onthe TEM grids either in the centers or at the cages of the sulfates (Figures 3a, 3e, and 34). The typically Hight loadings of parties on the grids male it highly unikaly that soot cours atthe edges of sulfates because af Scavenging of soot by hydrated sulfates a they spend on the supporting film. We believe that soot was forced to the peripheries of the panicles when de sulites eystalized from the orignal aerosol dropes, Thus the position of sot within the sulfate, as observed on the ‘TEM grid, docs not provide much useful infomation about the likely spat relationship of the two aggregated species in tho original srbome particles. The air mass back trajectories indicate that combustion Sources on the Ansan landniass only could have contributed 10 the aerosol in a few samples. Most of the soot above the Souther Ocean could originate fom elther biomass buming or fiom ship and sical caissions. Soot fom biomass burning invariably contains K [Andreae, 1983; Artaxo eta, 1994; Turn ct al, 1997; Cacher etal, 1991). About a quarter of analyzed fot particle thow detectable K (the detection limit for K i éstmated to be about 2 at 8), but most do ot contain elements heavier than O. About half of the soot particles contained sianlfeant K inthe two samples collected at 09 km altitude, In ast 9 the vier Southern Osean samples. Since these Samples were collected from a "bulfer layer" that served as. reservoir and mixing 2one for MBL and FT sir [Rusell tal, 1998}, itis likely that distinct aerosol layer was sampled; it may Inve contnned particles from South African biomass hiring oth ship [Corbet and Fiachbeck, 1997] and aircraft traffic {Gardner et al, 1997 could contbute to atmospheric sot the studied region. Most ships bum low-grade residual ol! that onlin contaminant suc a yas ad metals [Corbet cand Fischbeck, 1997), Since we did not observe metal peaks in the 21,689 EDS specira or Ay-ush associated with the soot particles, we conclude that most of the Southem Ocean soot originated from siroraft emissions, with a smaller fraction possibly coming from South African (and perhaps Australian) biomass burning. ‘The ‘morphologies and sizes of the smaller soot aggregates resemble those of aircraf-emined panicles [Blate and Karo, 1995, Pueschel etal 1992, P. J. Sheridan, personal communication, 1997], and the growing ratio of internally mixed soovulfte particles with altitude (Figure 2) is also consistent with areraft as the likely source of most soot. 33. Sulfate/Soot inthe North Atlantic PT (ACE 2) In the ACE 2 (Izana) samples (Figure 2, at 24 km altitude). lhe raius of soot-bearing wo soot-tiee sulfates are similar tw Use in the Southem Ocean samples, and there is no. significant difference between ratios in "lean" and "dusty" samples ‘Wherear sulfate particles with diameters <0.1 ym were not clficintly collected in the ACE 1 and ASTEX/MAGE samples, the ACE 2 samples contain a continuous size range of ammonium sulfate trom ~0.04 um to ~2 jm. As mentioned inthe Experiment section, the ACE 2 particles were collected om grids coated with holey carbon. It is likely that the different collection method resulted in the deposition of 0.1 jm particles. Even though the size distributions of sulfates are slightly different in the ACE 1 and ACE.2 samples, and the number af analyzed ACE 2 samples {s limited, he relative numbers of soot-bearing sulfate particles above the Southem Ocean and North Atlantic ae similar, These results suxuest al souUsulfate aguregates occur extensively in the remote FT 4. Discussion and Conclusions ‘The hygroscopic propeitcs of soot could be coutrolled by its composition (Lamme! and Novakov, 1995]. Organic surface ‘oups can enhance water adsorption [Chugeal otal, 1996}, and trace metals such as Fe increase the catalytic activity of soot [Grete et al, 1993]. Almost all soot particles in the ASTEXIMAGE samples occur within sulfaies: the nucleating ability of sot may be enbanced by is Sand the metal contents of the associated fly-ash particles. In the Southern Ocean samples, K bearing soot particles are typically included in sulfates; on the other hand, a large number of K-fre, presumably aircraft omitted soot particles from the Southem Qeean occur without being aggregated with sulfates. ‘The low contrast of these individual soot particles in the TEM hinders their counting. so we could not ‘lablis a reliable relationship between the numbers of free and sulfterincluded soot particles in the different samples. The presenee of these fee soot particles indicates that they are poor nuclei, consistent with their chemical purty. These observations sare supported by studies that identify most coal combustion LGrgit et af, 1993] and biomase-hurming soot particles active ‘CCN, and particles produced by jet fuel combustion as poor CCN [Htudson eta, 19911 ‘Numerous studies show that contrails behind cruising ascraf ‘contain both sulfate and soot particles [Schumann et al, 1996; Petzold otal, 1997; Karcher and Fahey, 1997; Yu and Turco, 1997] A significant number of soot particles can be activated in ‘plimes that form finm the products of high-S fuel combustion [Andronache and Chametdes, 1997: Brown et al. 1997): this activation is thought to happen primarily through dhe ‘condensation of Hy$0, on the surface of soot [Rogaski et al 1997; Yue and Turco, 1998]. Some sulfate particles in a dispersing21,690 POSFAIET AL.: SOOT AND SULFATE AEROSOL. Figure 3. Aerosol particles collected above the Southern Ocean, south of Tasmania, (a) Ammonium sulfate particles with soot inclusions (marked by arrows). The Inset displays two different zone-axis SAED patterns that were obtained from a similar particle; spacings are consistent with the structure of ammonium sulfats and arc indexed accordingly. (6) A typical, indi ual soot particle that is not enclased by sulfate (c) Residues and soot Inclusions (arrowed) that remained afer three ammonium sulfwe particles were intentionally sublimed with the cleetron beam. (4) An ammonium sulfate particle witha small sot lusion (arrowed). Bubbles that resulted from electron-beam damage are visible nthe particles in Figures 5a and 5d. The high-resolution mage shows the onion Tike suuctue of spherules aircraft contrail_ will thus contain soot inclusions. If our interpretation is correct that most soot above the Southem Ocean originated in aircraft contrals, then the concentrations of eoot ae ‘well as sulfate particles inthe troposphere are ciretly influenced by alrrat emissions at higher altitudes. the arrowed soot inclusion Heterogeneous processes on the surfiwes of sovt particles uring atmospheric transport could further increase the concentration of sulfate aerosols. Even though there ems to be ria consencs about the atmospheric importance of SO, oxidation ‘on soot (Nowakov e al, 1974; Chang ef al, 1981; Mamane andPOSFAIET AL. SOOT AND SULFATE AFROSOI Goutlicb, 1989}, it is possible thatthe formation of sulfate is facilitated by soot particles that serve as heteragrnenie nuclei Since some of the soot particles can acquire a hygroscopic coating soon after emission, either by aircraft or by busning biounas, these ean become CCN and provide favorable sites for 80; oxidation in the aerosol water or in cloud droplets. It remains to be determined how important this mechanism is in rodcing the snat-hearing sulfate particles in remote marine regions Regardless of the processes that fom sootsulfate ageregates, the climate effects of such particles are quite diferent from those ‘of an extemal mixture of the same aerosol species. The direct radiative foreing (DRE) by sulfate is signiicantly reduced i they are intemally mixed with soot (Haywood et al. 1997). Each percent inrease in the soot/sulfate mass ratio is estimated to Teal w # 0.034 War" “top of atmosphere" positive foreng for internal mixtures [Chjlek eta, 1995]. In the Southern Ocean samples, the eooUulfate mass ratio in internally mixed particles is highly variable; we estimate that it ranges from less than 0.01 to about (0.29 (as in particles X and Y., respectively, in Figure 3). The DRF of soot is very seusitive to the albedo of the underlying. surface: the higher the surface reflectivity, the more effective is the absorption by soot [Haywood and Ramaswamy, 1998} ‘Therefore soot above ice or snow and shove clouds has a much higher mass efficiency in tems ofits DRF effect than soot above the oceans. Since high-altitude soot has a yreater probability of being above clouds than low-altitude soot, aireraemitted soot could have a greater effect on climate than its relatively insignificant mass would suggest. ‘Soot particles can he effective as CCN (Chylek and Halter, 1992; Kawfinan and Fraser. 1997], especially when they are Anemnally mixed with hygroscopic sulfates. For example, we suggest that heterogeneous reactions forming sulfate on soot particles may be responsible forthe observed larger cloud droplet ‘concentrations in ship tracks [Rade of al, Y9X9}. Soot aerosols have complex effets on cloud properties, Smoke particles trom biomass burning can increase cloud reflevime by reducing the droplet size [Kaufman und Fraser, 1997]; on the other hand, the high absorption by soot-contaning cloud droplets enhances cloud absorption [Chilek and Hallet, 1992) ‘From our observations, we infer that both the indiect and irect radiative forcing by sulfae aerosols are afexied by thei soot inclusions in the remote marine atmosphere. However, before tle effects of intemal mixing can be taken into account in ‘lobal climate models, it will be necessary to establish reliahle numbers for the fraction of sulfates that enntains soot in the label background aerasol and the average soousulfate mass ratio ‘i these particles. Although TEM is an indispensable method ‘when particle species and their mixing properties are to be determined, itis 50 labor intensive that it cannot provide the statistical depth that ie needed for such data. In the future, We plan to combine TEM ohservations with other types of aerosol ‘measurements in order to obtain more quantitative data about soov/sulfate mixing, Acknowledgments We thank Steve Howell for providing the mass back tajectorics. Steve Howell and David Wifi helped collet the [ACE I samples, and Hal Mang. Tom Snowden, and others ftom the Univesity of Miami helped collet the ACE 2 samples. This resarch was supported by grants trom the Atmosphenc Chemisty Division of the NSF (ATM 9408704 end ATM.97I3101), and. minor support was ‘wovided by the Hungarian Science Foundation (OTKA‘F23393) 21,691 Electron miroscopy was performed at Arizona State University inthe Center for Hish-Resolaton Elesson Miiosapy. This research i ‘Sotthaton to te inemaiona Cio Americ Chemisty OGAQ) Core Project of the Innainal Gcospher-Dosphere Programme (UGB snd spat ofthe IGAC Acoso Charceization Experiment (ace References ‘Andreae, M.O, Soot cafbon and excess fine potassium: Long-range ‘eansprt of combustion derived aurzols, Slonce, 220, 11481151, 1983. Andronache, C., and W.L, Chameides, Interactions betwen sulfur and ‘soot emissions fom sicraft and et role In conrall formation, J. Geophys Rex, 102,21 443-2148, 1997 ‘vax, Py F. Ger, MA. Yamasoe, and J.V. Manns, Fine mode ‘aezosol composition at three longtemm atmospheric monitoring sites in ue Amuaon Basin J. Geopiys. Ret, 9, 2,857-22,868, 199. Rates, T'S, BJ Huebert, jL- Gras, FB. Geifiths, and P'A. Burke, The ntcwaional Global Aunospheti’ Chemisty (IGAC) Projects First, Aernsol Characterization Experiment (ACE I) Overview, J. Geophys Fes, 103, 16297-16318, 1998 Blake, Find Kato, Latina dsebtion of black chon sot in the upper wopasplete an lower snp, J. Geopiys. Res, 100, ‘7198-7202, 1998 Dodiaine, BA, Aciosol ebsoption measurements at Baum, Mauna ‘oa, and the south pole, Geaphye Rex, 100, 8967-8975, 1905, Brown, RC., RC, Miake-Lye, MIR. Anderson, and CLE. Kelb, Aircraft Ulf emissions andthe formation of Visible conta. Genny, Rex Lett, 34, 385-388 1997. ‘Buseck’ PAR, and M Psi Airbome minerals and related aerosol Particles: Eifets on climate and the envionment, Proc. Nath dead. Sci US. A. 96, 3372-3379, 1999, CCochiee, HJ. Duert, M-P. Brémond, V. Yoboug, J-P. Lacaux, A. ‘Gaudehe, and J. Baudet, Biomass bummg aerosols im a savanna region ofthe Ivory Coast, in Global Biomass Burning, edited by 1S. Levine, pp. 174-184, MIT Pres, Cambridge, Mass. 1991 (Chang, §.G,,R. Too and T, Novak, Th immporance of sot partes ‘and nigous acd in oxidizing SU: in atmosphere aqueous droplets, Atmos. Environ, 15, 1287-1292, 1081 (Chalson, RI, 1. Lovelock, MO, Andreae, and S.G, Warten, Oceanic ‘hytoplanion, stmosphoric sulphur, cioud albedo and’ climate, Nature, 326, 635-61, 1987 Chars, R., SE. Schwartz, JM. Hales, RD. Ces, A, Coakley Jt, “LE. Hansen, and DJ. Hofmann, Climate fering by anthropogenic aerosols, Sion, 255, 423-43, 1992. (Chuang, CC, LE, Peer, KE. Taylor, AS. Gros, and J. Walton, "An assesment of the radiative effete of antropogenic sulfate, J Geophys. Rex, 102, 3761-3718, 197. CChugtal, AR., ME. Brooks, and DM. Smith, Hydration of black carbon, “J Geopins. Rex, 101, 13305-19514, 1956 (Chile Band J. Hale, Enhanced absorption of soit radiation by ‘oud droplets cntaning soot patices in thet sufce, Q. J. Metearal Soc, 178, 167-172, 1987 Chile, P, O. Videe, D. Ngo, RG. Pinch, and J.D. Klet, Efe of ‘black carbon of the optical peopenies and’ climate forcing of sft scrsoly J Geophys, es, 100, 16325-16332, 1995. Clarke, A, Aerosol light shsoption hy sot in em environments, “Aeroel Sc, Technol, 10, O17, 1989, Clarke, ADD, IN Porer, FP. Valeo, and P. Pilewskie, Vertical piles, srosl microphysics, and optical closure during the Allantc Stratocumulus Transition Experiment: Measured and modeled cola ‘optical properties, Geophse Rez, 107, 4443-4453, 1996, Clarke, AD, T- Uehara, and IN. Pore, Aimospbenc uct! and reated ‘erro fields over the Alante: Clean subsiding air and continental pollion during ASTEX, J. Geaphs. kes. 102, 29.281-23.292 1397 Cooke, W.F. and LN. Wilson, A global black carbon acoso model. J ‘Geophys Rex, 101, 19,398-19,409, 1996 (Cooke, WE. SU, Jennies. an 1G, Spain, Black carbon measurements ‘st Mace Head, 1989°1996, J. Geophys. Ree, 102, 25,330-28346, sy. Corbet, 1, and P Fisehbeck, Emissions from ship, Selnce, 278, #23- 1824, 199721,692 CCruzen, P., and M0. Andreae, Blomass burning inthe wopis: Impact ‘on aimospherie chemistry and biogeochemical eels, Selene, 250, 1669-1678, 1980. Gardner, RM, et al, The ANCATIEC lobal inventory of NO, emissions ftom aera, Amos, Environ, 31, 1791-176, 1991. corgi TV. Hodnik, and ML Bigjak, Agucour S(QV) sxiation, TI, Calaljtc effect of wot partic, dos. Environ, Part 4, 27, 1409 1416, 1098, Haluswou, AHLW., AL. Dick, and 11, Gras, Climatic context ofthe ist Aerosol Characterization Experiment (ACE 1): A meteorological a level overview J. Geops. Res, 103, 16319-16340, 1998 Heywood, JM, and V. Ramaswamy, Global sensiity studies ofthe ‘uct radiative forcing eet du to amdropogenic sulfate and black ‘arbon aerosols, Geopiys. Rex, 103, 6043-6058, 1998 Haywood, JM, D.L. Roberts, A. Slingo, JM. Edwards, and KP Shin, ‘General circulation modal calculations of the dies radiative forcing by anropogeni sulfate and fossl-uel soot aerosol, J. Clam, 10, 1862-1877, 1997 Hobbs, PY. }.S. Rel, RA, Kotchennuber, RJ. Ferek, and R. Weiss, Diret radiative forcing by smoke from biomass buming, Science, 273, 17761778, 1997, Hudson, 1G, J.’Hallot, and CF. Rogers, Field and laboratory measurements of cloud forming properties of combustion aerosols, J Gooplys. Ras, 96, 10817-10889, 1991 Hudson, 1G," Xie, and 5.5. Yum, Verical distributions of cloud ‘condencation muse spectra over the summertime Souther Osean, J Geophys. Res. 103, 10609-16824. 5. Husben, BJ, A. Pazcany, and B. Blomquist, The ASTEXMAGE experiment. J Geophos. Kes, 101, 4819-4529, 1996 Husbert, BJ, 6. Howell, L Zhuang, 1A. Teath, MR. Litchy, DJ. Wyle, JL. Kreider Moss, 8. Coppeas, and SE Pei, itr and impacior measurements of anions and cations during the Psst Aerosol (Characterization Experiment (ACE 1). Geopiys Res. 103, 16493- 16,509, 1998 cher, Band D.W. Fahey, The role of sufer emission in volatile particle formation in jet aircraft exhaust plames, Geophys. Res. Let, 24, 389-302, 1997 Katrnak, KA, LR. Anderson, and PR. Buseck, Structural variations in individual arbonaceous particles frm an urban seo, Envir. Set Technol, 26, 1967-1976, 1992. Knafinan, Vi, and RS. Fer, The effet of smoke paris on conde ‘and climate acing, Science, 277, 1636-1639, 1997 Lahaye, and Gra, Morphology and internal statue of sot and ‘aubon’ blacks, in Purdeulate "Carbon. Formation During Combustion, edited by DIC. Siegla and GW. Smith, pp 38-55, Plenum, New Yotk, 1981 Lammel, Gand T. Novakov, Water nucleation propertis of carton ‘black and diesel soot particles, mas. Evron, 29, 813-823, 1995, Luss, C, C. Cachir, and'S.G. Jemnings, Optical and’ thermal Teasuremenis of lack carbon aerosol” coment in diffent vironments: Variation of the specie atlenantion crossetion, Sma (0), amos. Environ, 27, 1203-1211, 1993. Liogeee,C, LE. Pemer, C. Chuang, 34. Walton, H, Eddleman, and H. Cacier” A. global treedimensional model study of carbonaceous soso, J) Geophys. Res, 101, 1911-19432, 1986. Mamane,Y., and J. Gone, The study of heterogencous reactions of ‘aronacooue patisles with 2ulfr and mirogen exis using a sngle- Datcle approach, J. Aerosol Sct, 20, 575884, 1989 ‘Maatros, Ay and E, Méczars, Sulfate formation on elemental carbon Dares, derasol Set Technol, 10, 337-342, 1989 Novakos, Ty SG. Chang, and’ AB, Harker, Sulfaes as polation arcuate: Caayle formation on carbon (Sot) paces, Sctence, 186, 289-261, 1974 Dowd, CD, MH, Sm, and $.G. Jemings, Submicron panicle radon, and toot carbon characeritice over the northeast Atlantic, J Geophys Res. 98, 1137-1189, 1993. Okada, K, Numbersize distibuton and formation process of ‘submicrometer Sulfte-contalning parcles inthe urban aosphere of Nagoya, Atmos. Environ, 19, 718-757, 1988. Parungo. FB. Kopeewiez C. Nagamoro, R. Schnell, P. Sheridan, C. “Zhu, and J. Har, Aerosol patele inthe Kuvat oil fie plumes Their morphology, size distribution, chemical compostion, anspor, and poten! effec on climate. Gaophye. Ren, 97, 15867-15882, 992, POSFALET AL.; SOOT AND SULFATE AEROSOL Parungo. F.C. Nagamoto. MY, Zhou, AD.A. Hansen, and J. Hams “Alianiranspert of aerosol back earbon from China tothe ocean, ‘Atmos. Environ. 28, 3251-3280, 199%. Penner, JE, Carbonacoous sarozalsialuencing atmosphere radiation: ‘Black and organic carbon in Aerosol Forcing of Climate edited by RJ. Chargon and J. Heintanberg, pp. 91109, John Wiley, New “York, 195 Penner, 1B, and T, Novakov, Carbonaccous particles inthe amoephee: 'Ahistovial perspective to the Fh Intemational Conference on Caibonaccous Pats in the Atamspc, J: Geopiys. Res, 10, 19,373-10.378, 1996 Penner, , H. Eddleman, and T, Novakov, Towards the development of 1 global inventory fr black carbon emissions, dimes, Environ, 27, 1277-1285, 1993, Potold, A, otal, Nearfild measurements on conta properties from els eth diferent sulfur content, J. Geoplps. Ree, 102, 20,867~ 29,880, 197, Pésfai, M, LR. Anderson, PR. Buseck, TW. Shatuck, and NW. “Tindale” Consents of a remote Pacife marine aerosol: A TEM sudy, donee Environ, 28, 177-1756, 1094 Postal, MR Andetson, P.R. Buseck, and H. Sleverng, Composttonal ‘aration of seasalt mode aerosol particle fom the North Atlantic, 2-Geophs. Res, 100, 23083-23074, 1995. osha, M, H. Xu, ZR. Anderson, and PR. Busock, Wet and dy sass of stmospheric aerosol "patclés: A combined AFM-TEM study, Geophys. Res Let, 25,1907 1910, 1998 Pueschel, REDE Rlake, KG Soetings, ATA Hancen, S Verma, ‘nd K. Kato, Black carbon (sot aerosol in the lowe suatasphee and ‘pperrapocphee, Cape Rex Tat, 19, 1689-1662, 1092 Radke, LF, JA. Coakley 1, and M.D. King, Diet and venot sensing ‘eerutione ofthe efce of shipe on ed Siena, 246, 11a 1149, 1988 Rameden, AR, and M. Shibaoks, Characterization and analysis of individual fy-ash panicles from coa-fred power stations by a combination of optical microscopy, electron microscopy and quantitate electron microprobe analss. Amos. Environ. 16, 2191- 2206, 1982. Ran, JA, and MAK Khali, Anthropogenic conthutions 10 the Taubonaccuus conic of acisols ore the Pacific Ovean, ames Environ Part 4, 27, 1287-1303, 1993 [Rogaski, CA.. DIM. Golden, and 1-K. Willams, Keacive uptake and hydration experiments on amorphous carbon treated with NO., 50, Os, HNOs, and H28Us, Gopi, Kes Lett, 24, 381-384, 1997 Russel, LM, DH. Lenschow, KK Lauren, PB. Krum, 67. Siem, ‘AK. Bandy, D.C, Thomion, and TS. Bates. Bidirectional mixing in ‘an ACE I marine boundary layer overiin bye second turbulent lye, 4 Geopins. Kes, 105, 16411-1632, 1998 Sebuly 1, J Teihtr, and WIP. Cooke, ie of black carbon and sulfite ‘aerosols on the global radiation budget J. Geaphox. Res, 02 30,107-30,117, 197, Schumana, U, J. Stom, R. Busen, R. Baumana, K. Gierens, M. raulstunk, FP. Schoder, and J Sting, In sta observations of Particles in jet sirrat exhausts and conti for diffrent sulfur ontaining Fels. Geophns. Res. 10 6853-8869, 1996. Shevidan, PL, Charotrization of sie apregated pails collected over ‘Alaska and the Canadian high Acuc, AGASPAI flights 208206, ‘Amos. Environ, 38,2371 2386, 1989. Sheridan, PT, RC” Schnell, 11) Kahl, 1F. Boatman, nd DM. Garvey, ‘Mivtvanaysis of Un aero elleced over soulncctal New Mexi uring the ALIVE field experiment, May-December, 1989, mas Environ, Part 4, 27, 11691183, 1983. Tum, 6Q, DIM, Jenkins, JC. Chow, LC. Pitchet, D. Camptell, (Call and S.A. Wain, Elemental characterization of particulate mater emitted from biomass buming- Wind tunel derived source profiles for herhacens and woad ful, Genple Rex, 102, 2683. 3699, 1997. ‘Van Dingenen, RF. Rass, and NR. Jensen, Evidence for anthropogenic ‘impact on’ number concentration and” suite content of cloud: procected aerosol particles over the North Atlantic, J: Ceop. es, 100. 21097-2067, 1995 ‘Yu,F, and RP. Teo, The oe of fons in the formation and evolution of articles i acraft plumes, Geophys es Lett, 24, 1927-1930, 1997,POSFAIET AL.: SOOT AND SULFATE AEROSOL, 21,693 ‘Yo, F, and RP. Turco, Contal formation and impacts on scrosol Univenity of Vesrpém, Vesreém, POB 158, 1.8201 Hungary ‘properties in aircraft plumes: Effects of fel sulfur content, Geophys (postamialmos vin hu) es. Let, 25, 313316 199, TL Sicvering, Global Changs and Cavronmental Quality Program tnd Center for Environmental Setenees, University of Colorado at Denver, Denver, CO 80217, TR Anderson and P_ R Buseck, Departments of Geology and Chemisty, Arizona Stat University, Tempe, AZ 5267, (essed October 15,1998; revised March 18, 1999; IM. Pista. Department of Earth and Environmental Sciences, accepted March 25, 199.)