Celso Orsini

Characteristics of fine and coarse particles of natural and urban aerosols of Brazil

Abstract Fine and coarse particles have been sampled from 1982 to 1985 in one natural forest seacoast site (Jureia) and five urban-industrial cities (Vitoria, Salvador, Porto Alegre, Sao Paulo and Belo Horizonte). The time variations of concentrations in air and the relative elemental compositions of fine and coarse particle fractions, sampled by Nuclepore stacked filter units (SFU), have been determined gravimetrically and by PIXE analysis, respectively. Enrichment factors and correlation coefficients of the trace elements measured led to unambiguous characterization of soil dust and sea salt, both major aerosol sources that emit coarse particles, and soil dust is also a significant source of fine particles.

RELATION BETWEEN AEROSOL SOURCES AND METEOROLOGICAL PARAMETERS FOR INHALABLE ATMOSPHERIC PARTICLES IN SAO-PAULO CITY, BRAZIL.

Abstract Stacked filter units were used to collect atmospheric particles in separate coarse and fine fractions at the Sao Paulo University Campus during the winter of 1989. The samples were analysed by particle-induced X-ray emission (PIXE) and the data were subjected to an absolute principal component analysis (APCA). Five sources were identified for the fine particles: industrial emissions, which accounted for 13% of the fine mass; emissions from residual oil and diesel, explaining 41%; resuspended soil dust, with 28%; and emissions of Cu and of Mg, together with 18%. For the coarse particles, four sources were identified: soil dust, accounting for 59% of the coarse mass; industrial emissions, with 19%; oil burning, with 8%; and sea salt aerosol, with 14% of the coarse mass. A data set with various meteorological parameters was also subjected to APCA, and a correlation analysis was performed between the meteorological “absolute principal component scores” (APCS) and the APCS from the fine and coarse particle data sets. The soil dust sources for the fine and coarse aerosol were highly correlated with each other and were anticorrelated with the sea breeze component. The industrial components in the fine and coarse size fractions were also highly positively correlated. Furthermore, the industrial component was related with the northeasterly wind direction and, to a lesser extent, with the sea breeze component.

PIXE and receptor models applied to remote aerosol source apportionment in Brazil

Abstract We have used PIXE together with receptor modeling for quantiative source apportionment of remote aerosols. The Sao Paulo PIXE system uses an 8 MeV alpha particle beam, and it can detect elements with Z > 11 with a detection limit between 1 and 30 ng m 3 . We sampled atmospheric aerosols in six remote sites: Amazon Basin. Atlantic Forest, Arembepe, Fernando de Noronha, Firminopolis and Itaherai. Sampling was performed using stacked filter units, cascade impactors with 6 or 10 stages and linear streakers. Source apportionment of the coarse mode aerosols was done using three receptor models: chemical mass balance, principal factor analysis and stepwise multiple regression analysis. Three sources of aerosols were quantiatively distinguished: marine aerosols, soil dust and aerosols released by plants. The emission of aerosols by vegetation is very clear for all sampling sites and with all three receptor models. In the Amazon Basin and Atlantic Forest the aerosol released by plants is the major source and accounts for 60–80% of airborne concentrations.

Preliminary data on atmospheric aerosol of the Amazon Basin

Abstract Preliminary distributions of the trace-elements Al, Si, P, S, Cl, K, Ca, Ti, V and Fe in the atmospheric aerosol of the Amazon Basin have been determined through samples collected from 23 August to 2 September 1980, at a remote place located in the Amazon Forest, about 30km NE of the city of Manaus. In all, 10 complete cascade impactor samples of 6-stage, Battelle model, have been exposed but only 8 with success, thus generating 48 samples. From these, 33 samples have been successfully analyzed by the PIXE method (Particle Induced X-Ray Emission), using particle beam of the Pelletron Accelerator of the University of Sao Paulo, and the results revealed that the trace-elements S and K have a noticeable predominance, mainly in the fine particle size range relative to the others. The high correlation factor found between the fine particle concentrations of S and K (0.96) support the assumption of their common airborne transport on the same particulates, originated from the gas-to-particle conversion of gases exuded by the trees of the forest, their only existing sources. Coarse airborne particles, of a probable soil origin, are also present but in unusally small amounts, as it was revealed by the Al, Si, Ca, Ti and Fe size distribution curves.

The São Paulo PIXE system and its use on a national monitoring air quality program

Abstract The SP-PIXE has been installed on the 8 MV pelletron accelerator of the IFUSP. Its characteristics and calibration procedure are presented below. The SP-PIXE is being used as the reference analytical method on a national monitoring air quality program performed by the Group for Air Pollution Studies (GEPA) of the IFUSP. As part of the program eight sites around Brazil, seven having critical air pollution problems and one with a natural atmosphere, have been selected for sampling. Each aerosol is sampled continuously for one year using stacked filter units and, for period of 3 to 5 days, by ten stage cascade impactors. Partial results from the two first sites investigated are presented here by element size distribution curves and histograms of relative concentrations.

A PIXE system for air pollution studies in south america

Abstract As an outgrowth of the PIXE-SP Project an experimental sut-up for particle-induced X-ray emission analysis will soon be available at Sao Paulo. The irradiation chamber, nearly completed, will provide sequential measurements of several (up to 70) sample-targets. The major focus of the Project will be on the investigation of atmospheric aerosols using this set-up. Preliminary results on urban Sao Paulo aerosol characteristics are given for the first time both as an illustration of the analytical potential of the PIXE method and as an example of aerosol data analysis. Samples collected in Sao Paulo were subjected to PIXE analysis at Tallahassee, using the Florida State University facilities. The results for S, Pb and Br are presented.

Fine and coarse elemental components in the urban aerosol of São Paulo, Brazil

Abstract Using cascade impactor sampling and PIXE analysis we have measured particle size distributions for approximately 15 elements in the Sao Paulo urban atmosphere. These elements, in our case, may be classified, according to their occurrence in fine or coarse aerosol log-normal modes, into three groups: (a) soil dust reference elements (coarse particle mode): Ti, Si and Ca; (b) anthropogenic fine particle mode: Zn, Br and Pb; and (c) mixed bimodal elements: S, K, V, Cr, Mn, Ni, Cu and Fe. All of the soil dust reference elements show consistently the log-normal parameters MMAD ≈ 5.5 μm and σg ≈ 3.2 (mass median aerodynamic diameter and geometric standard deviation, respectively). Enrichment factor calculations for Ti, Si, Ca and K in the coarse particle fraction (>2 μmad), relative to Fe and the “standard crustal aerosol” values of Lawson and Winchester, show that Ti and Si are mainly soil derived while Ca and K may have significant industrial components in this particle fraction (i.e. coarse mode). The fine mode parameters for the other elements show variations with element suggesting different air pollution sources (such as motor vehicles, resuspended dust, refuse burning, industrial activities, etc.) and/or different chemical pathways, which presumably could be identifiable. For example, the modal parameters for group (b) are: Zn, MMAD=0.9 μm σg=2.2; Br, MMAD=0.5 μm, σg=4.0; Pb, MMAD=0.6 μm, σg=3.0; thus suggesting a common source (automotive) for Br and Pb, unrelated to the source of Zn (possibly refuse burning).

Receptor modeling for inhalable atmospheric particles in Sao Paulo, Brazil

Abstract Aerosol samples were collected with stacked filter units (SFUs) at the Sao Paulo University Campus during the winter of 1989. The fine and coarse filters of the SFU samples were analyzed by particle-induced X-ray emission (PIXE) and the data were subjected to an absolute principal component analysis (APCA) in order to identify the major aerosol sources and to apportion the particulate mass to these sources. Five sources were identified for the fine particles: industrial emissions, which accounted for 13% of the fine mass; emissions from residual oil and diesel, explaining 41%; resuspended soil dust, with 28%; and emissions of Cu and of Mg, with together 18% (these elements could not be attributed to any specific source). For the coarse particles, four sources were identified: soil dust, accounting for 59% of the coarse mass; industrial emissions, with 19%; oil burning, with 8%; and sea-salt aerosol, with 14% of the coarse mass.

PIXE ANALYSIS FOR AIR POLLUTION SOURCE APPORTIONMENT IN URBAN AREAS OF BRAZIL

PIXE analysis has been applied to an air quality monitoring and assessment program in south-eastern Brazil. Fine and coarse particle size fractions were sampled selectively, using Nuclepore in stacked filter units, once a week for one year in one background and five urban sites. The procedures used for sampling and analysis have been developed and refined in close conjunction with the application to aerosol studies, and 138 pairs of filters have been analyzed by PIXE. Once a national source signatures library for particulate matter has been created, the data are suitable for receptor modeling and source apportionment for anthropogenic and natural particulate constituents by chemical mass balance models.

Variation of Elemental Composition with Particle Size in the St. Louis Aerosol

Detailed particle size distribution of elemental constituents have been determined in the St. Louis aerosol by sampling with cascade impactors for particles >0.25 μm aerodynamic diameter and by analyzing using proton-induced X-ray emission, PIXE, for S, Cl, K, Ca, Ti, Mn, Fe, Zn, Br, and Pb. Two rooftop sites were selected for intensive sampling in industrial and residential locations in St. Louis during February 1974, every 12 h for a 7 d period which included clear, rainy, and snowy weather regimes. The data could be compared with the results of a similar study of the St. Louis aerosol during August 1973 when the weather was generally warm and without rain. Both data sets could also be compared with results of studies of Tallahassee, situated in coastal north Florida, and the north Atlantic marine atmosphere of Bermuda. Although variation in elemental concentrations and their size distributions were observed which could be associated with local weather conditions during the February sampling period, the data set was sufficiently large for meaningful average characteristics to be computed. As examples, two elements of pollution importance, sulfur and iron, show the following comparison between St. Louis and the Florida and Atlantic locations: Sulfur in all data sets was most abundant in the smallest particle size classes. Winter and summer concentration in St. Louis air were the same within a factor of two but generally exceeded north Florida concentrations. Bermuda concentrations were intermediate between St. Louis and north Florida, except for largest particles where sea spray derived material was present. The results reflect the importance of gas phase conversion to particulate forms of sulfur, e.g. sulfate, in all three locations. Iron concentrations in St. Louis generally exceeded those in north Florida for smallest particles and to a lesser extent for larger sizes. Concentrations in Bermuda were about 10-fold lower than in St. Louis for all particle sizes. The results reflect the terrestrial dust origin of iron and also suggest that iron is partially pollution-derived in the St. Louis atmosphere, especially in smaller size ranges. (Less)