Individual aerosol particles from the remote marine atmosphere were investigated by scanning electron microscopy and electron microprobe analysis. A large fraction of the silicate mineral component of the aerosol was found to be internally mixed with sea-salt aerosol particles. This observation explains the unexpected similarity in the size distributions of silicates and sea salt that has been observed in remote marine aerosols. Reentrainment of dust particles previously deposited onto the sea surface and collision between aerosol particles can be excluded as possible source mechanisms for these internally mixed aerosols. The internal mixing could be produced by processes within clouds, including droplet coalescence. Cloud processes may also be responsible for the observed enrichment of excess (nonsea-salt) sulfate on sea-salt particles.

As a part of the NASA Global Tropospheric Experiment (GTE), aerosols were sampled in the tropical rain forest of the Amazon Basin during the Amazon Boundary Layer Experiment (ABLE 2B) in April and May 1987, in the wet season, when no forest burning occurs. Fine (dp < 2.0 μm) and coarse (2.0 < dp < 15 μm) aerosol fractions were collected using stacked filter units, at three sites under the forest canopy and at three levels of a tower inside the jungle. Particle‐induced X ray emission (PIXE) was used to measure concentrations of 22 elements (Na, Mg, Al, Si, P, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, Br, Rb, Sr, Zr, and Pb). Morphological and trace element measurements of individual particles were carried out by automated electron probe X ray microanalysis. Gravimetric analysis was performed to obtain the fine and coarse aerosol mass concentration. Absolute factor analysis was used to interpret the large data set of the trace element concentrations and to obtain elemental source profiles. Hierarchical cluster analysis was used to derive groups of individual particles. The concentrations of soil dust related elements (Al, Si, Ti, Fe, Mn) were 5 times larger in the wet season compared to the 1985 ABLE 2A dry season experiment. Biogenic aerosol related elements in the fine fraction showed lower concentrations in the wet season. Fine aerosol mass concentration averaged only 2.1±0.7 μg m−3, while the average coarse mass concentration was 6.1±1.8 μg m −3. Sulphur concentrations averaged 76±14 ng m −3 in the fine fraction and 37±9 ng m −3 in the coarse fraction. Biogenic aerosol‐related elements were dominant under the forest canopy, while soil dust dominated at the top of the forest canopy. Only two factors explained about 90% of the data variability for the fine and coarse aerosol fractions. These were soil dust (represented mainly by Al, Si, Ti, Mn, and Fe) and biogenic aerosol (represented by K, P, Cl, S, Zn, and the aerosol mass concentration). Source profiles showed a homogeneous aerosol distribution with similar elemental compositions at the different sampling sites. Enrichment factor calculations revealed a soil dust elemental profile similar to the average bulk soil composition, and a biogenic component similar to the plant bulk elemental composition. Total aerosol mass source apportionment showed that biogenic particles account for 55–95% of the airborne concentrations. The analysis of individual aerosol particles showed that the biogenic particles consist of leaf fragments, pollen grains, fungi, algae, and other types of particles. Several groups of particles with K, Cl, P, S, and Ca as minor elements could easily be identified as biogenic particles on the basis of their morphology. Considering the vast area of tropical rain forests and the concentrations measured in this work, it is possible that biogenic particles can play an important role in the global aerosol budget and in the global biogeochemical cycles of various elements.

Institute for Nuclear Sciences, Rijksuniversiteit Gent, Ghent, Bel•tiumAerosols were sampled in the Amazon Basin, as part of the Global Tropospheric Experiment (GTE), during the Amazon Boundary Layer Experiment (ABLE 2A) in July-August 1985. Fine-and coarseparticle fractions were analyzed for 22 elements by particle-induced X ray emission. Gravimetric mass, black carbon, sulfate, and nitrate concentrations were also determined. Morphological and trace element measurements of individual particles were carried out by automated electron probe X ray microanalysis. Various receptor models, including multivariate methods and a chemical mass balance model, were employed in the interpretation of the bulk trace element concentrations. Three factors explained over 85% of the variability of fine-and coarse-mode variables. On the basis of the elemental composition of the factors, two could be identified as plant related, and the third was a soil dust component. Of the coarse-mode aerosol mass concentration (of 7.6 + 1.6 /•g/m3), 62% could be attributed to aerosols released by the vegetation and 11% to soil dust. In the fine mode, soil dust accounted for less than 10% of the measured mass concentration (of 6.8 + 3.9 /•g/m3). The variables related to the plant component were K, P, S, Ca, Mg, C1, Rb, and the gravimetric mass. The elemental profile of the plant component resembled the bulk plant composition. By single-particle analysis coupled with hierarchical cluster analysis, six to nine different biogenic-related particle groups could be identified in the fine-and coarse-aerosol modes. Almost all particle types consisted predominantly of carbonaceous material, with trace amounts of K, S, Ca, P, C1, and Na. Only one group, comprising less than 11% of the total number of particles, consisted of soil dust-related aerosol.

No abstract

The Armstrong–Buseck correction for absorption effects in electron probe x‐ray microanalysis of particles considers seven specific particle shapes, and for these geometries exact correction equations are used. This procedure implies that the analyst has to associate the particle to be analysed with a certain particle type; an arbitrary relative thickness is sometimes assumed. A theoretical study was made of this absorption correction as a function of the particle composition, type and thickness for micrometre‐sized particles. It appears that a correct choice of the particle type is critical. However, when the analytical results are normalized to 100%, the differences between the models are much less pronounced, and it is justified to assume a spherical model in all cases.