UNIVERSIDADE ESTADUAL PAULISTA INSTITUTO DE GEOCIÊNCAS E CIÊNCAS EXATAS CAMPUS DE RIO CLARO Isotopic and Elemental Determination of Lead in Particulate Matter in the Cities of Goiânia (GO) and Rio Claro (SP) using ICP-MS technique Hendryk Gemeiner Orientador: Prof. Dr. Amauri Antonio Menegário Co-Orientador: Prof. Dr. Didier Gastman Dissertação de Mestrado apresentada ao Instituto de Geociências e Ciências Exatas do Campus de Rio Claro, da Universidade Estadual Paulista - “Júlio de Mesquita Filho” como parte dos requisitos para obtenção do título de Mestre em Geociências e Meio Ambiente. Rio Claro - SP Setembro de 2016 Gemeiner, Hendryk Isotopic and elemental determination of lead in particulate matter in the cities of Goiânia (GO) and Rio Claro (SP) using ICP-MS technique / Hendryk Gemeiner. - Rio Claro, 2016 80 f. : il., figs., gráfs., tabs., fots., mapas Dissertação (mestrado) - Universidade Estadual Paulista, Instituto de Geociências e Ciências Exatas Orientador: Amauri Antonio Menegário Coorientador: Didier Gastmans 1. Geochemistry. 2. Lead (Pb). 3. Lead isotope ratio. 4. Particulate matter. 5. ICP-MS. I. Título. 551.9 G322i Ficha Catalográfica elaborada pela STATI - Biblioteca da UNESP Campus de Rio Claro/SP UNIVERSIDADE ESTADUAL PAULISTA INSTITUTO DE GEOCIÊNCAS E CIÊNCAS EXATAS CAMPUS DE RIO CLARO Isotopic and Elemental Determination of Lead in Particulate Matter in the Cities of Goiânia (GO) and Rio Claro (SP) using ICP-MS technique Hendryk Gemeiner Dissertação de Mestrado apresentada ao Instituto de Geociências e Ciências Exatas do Campus de Rio Claro, da Universidade Estadual Paulista - “Júlio de Mesquita Filho” como parte dos requisitos para obtenção do título de Mestre em Geociências e Meio Ambiente. Comissão Examinadora Prof. Dr. Amauri Antonio Menegario Orientador – CEA/UNESP - Rio Claro/SP Profa. Dra. Veridiana Teixeira de Souza Martins IG – USP – São Paulo/SP Dra. Juliana Aparecida Galhardi CEA/UNESP – Rio Claro/SP Conceito: Aprovado Rio Claro - SP 13 de Setembro de 2016 1 Rio Claro, 2016 Isotopic and Elemental Determination of Lead in Particulate Matter in the Cities of Goiânia (GO) and Rio Claro (SP) using ICP-MS technique List of figures Figure 1: Photograph of the low volume sampler STACKER (left) and the integrated filter holder device with Teflon filter (right) for particulate matter sampling ................ 22 Figure 2: Gravimetric measurement of Teflon filter before sampling (left), filter with sampled particulate matter before digestion in front of microwave tube (right) ............ 23 Figure 3: Location of the sampling point in Goiania (Source: COSTA e SILVA, 2015) ........................................................................................................................................ 27 Figure 4: Location of the sampling point in Rio Claro (Source: Lorenzeto de Abreu, 2014) ............................................................................................................................... 30 Figure 5: Theoretically calculated and experimentally obtained relative standard deviations of 206Pb/207Pb and 208Pb/207Pb isotope ratios for different concentrations at 5 min and 10 min of acquisition time ................................................................................ 33 Figure 6: Normalized 208Pb/207Pb isotope ratio plotted as a function of the value applied for dead time correction of the 'raw' results obtained for solutions with a Pb concentration ranging from 2 to 10 ppb ......................................................................... 35 Figure 7: Analytical curve for the quantification of Pb via ICP-MS ............................ 36 Figure 8: Pb concentrations (Conc.) and respective standard deviations (s) in ng ml-1 in sample and blank dilutions of Goiânia samples from April and May 2014 (BL Ac = Acid blank, BL F = Filter blank) ............................................................................................. 37 Figure 9: Pb concentrations (Conc.) and respective standard deviations (s) in ng ml-1 in sample and blank dilutions of Goiânia samples from August 2014 (BL Ac = Acid blank, BL F= Filter Blank) ........................................................................................................ 37 Figure 10: Pb concentration in ng m-3 for Goiânia samples from April and May 2014 (X = mean value; s = standard deviation; RSD% = relative standard deviation; n = quantity) ........................................................................................................................................ 38 Figure 11: Pb concentration in ng m-3 for Goiânia samples from August 2014 (X = mean value; s = standard deviation; RSD% = relative standard deviation; n = quantity) ........................................................................................................................................ 38 Figure 12: Boxplot diagram for Pb concentrations of samples from rainy and dry season 2014 from Goiânia .............................................................................................. 40 Figure 13: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 2 Rio Claro, 2016 m AGL) on episodic dates with elevated Pb concentrations in rainy season arriving at sampling point in Goiânia............................................................................................... 42 Figure 14: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with low Pb concentrations in rainy season arriving at sampling point in Goiânia............................................................................................... 43 Figure 15: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with elevated Pb concentrations in dry season arriving at sampling point in Goiânia............................................................................................... 44 Figure 16: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with low Pb concentrations in dry season arriving at sampling point in Goiânia............................................................................................... 45 Figure 17: Pb–Pb diagram according to the Pb isotope ratios for samples taken in Goiânia ........................................................................................................................... 48 Figure 18: Pb–Pb diagram according to the Pb isotope ratios for samples taken in Goiânia compared with Pb isotope ratios from recent studies from Brazil .................... 49 Figure 19: 208Pb/206Pb-207Pb/206Pb diagram according to the Pb isotope ratios for samples taken in Goiânia ................................................................................................ 50 Figure 20: 208Pb/206Pb-207Pb/206Pb diagram according to the Pb isotope ratios for samples taken in Goiânia compared with Pb isotope ratios from recent studies from Brazil .............................................................................................................................. 51 Figure 21: Comparison of the values of 206Pb/207Pb isotope ratio here reported with values of recent studies from Brazil ............................................................................... 52 Figure 22: Pb concentrations (Conc.) and respective standard deviations (s) in ng ml-1 in sample and blank dilutions of Rio Claro samples from July 2015 (BL Ac = Acid blank, BL F = Filter blank) ............................................................................................. 53 Figure 23: Pb concentrations (Conc.) and respective standard deviations (s) in ng ml-1 in sample and blank dilutions of Rio Claro samples from December 2015 (BL Ac = Acid blank, BL F = Filter blank) .................................................................................... 54 Figure 24: Pb concentrations (Conc.) and respective standard deviations (s) in ng ml-1 in sample and blank dilutions of Rio Claro samples from February and March 2016 (BL Ac = Acid blank, BL F = Filter blank) ........................................................................... 54 Figure 25: Pb concentration in ng m-3 for Rio Claro samples from July 2015 (X = mean value; SD = standard deviation; RSD% = relative standard deviation; n = quantity) .... 55 Figure 26: Pb concentration in ng m-3 for Rio Claro samples from December 2015 (X = mean value; s = standard deviation; RSD% = relative standard deviation; n = quantity) ........................................................................................................................................ 55 3 Rio Claro, 2016 Figure 27: Pb concentration in ng m-3 for Rio Claro samples from February and March 2016 (X = mean value; s = standard deviation; RSD% = relative standard deviation; n = quantity) .......................................................................................................................... 56 Figure 28: Boxplot diagram for Pb concentrations of rainy (July 2015) and dry season (December 2015, February/March 2016) samples from Rio Claro ................................ 58 Figure 29: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with elevated Pb concentrations in dry season arriving at sampling point in Rio Claro ........................................................................................... 61 Figure 30: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with low Pb concentrations in dry season arriving at sampling point in Rio Claro ........................................................................................... 62 Figure 31: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with elevated Pb concentrations in rainy season arriving at sampling point in Rio Claro ........................................................................................... 63 Figure 32: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with low Pb concentrations in rainy season arriving at sampling point in Rio Claro ........................................................................................... 64 Figure 33: Pb–Pb diagram according to the Pb isotope ratios for samples taken in Rio Claro ............................................................................................................................... 67 Figure 34: Pb–Pb diagram according to the Pb isotope ratios for samples taken in Rio Claro compared with Pb isotope ratios from recent studies from Brazil........................ 68 Figure 35: 208Pb/206Pb-207Pb/206Pb diagram according to the Pb isotope ratios for samples taken in Goiânia ................................................................................................ 69 Figure 36: 208Pb/206Pb-207Pb/206Pb diagram according to the Pb isotope ratios for samples taken in Rio Claro compared with Pb isotope ratios from recent studies from Brazil .............................................................................................................................. 70 Figure 37: Comparison of the values of 206Pb/207Pb isotope ratio here reported with values of recent studies from Brazil ............................................................................... 71 4 Rio Claro, 2016 List of tables Table 1: Standard values for PM10; PM2.5; TSP and Pb established by CETESB (Source: CETESB, 2013) .............................................................................................................. 13 Table 2: Criteria of acute episodes of air pollution for PM10 and PM2.5 established by CETESB (Source: CETESB, 2013) ............................................................................... 14 Table 3: Air quality indexes for PM10 and PM2.5 established by CETESB (Source: CETESB, 2013) .............................................................................................................. 14 Table 4: Standard values for PM10; PM2.5 and Pb established by the European Union (Source: Directive 2008/50/EC, 2008) ........................................................................... 15 Table 5: Standard values for PM10; PM2.5 and Pb established by the United States Environmental Protection (Source: EPA, 2015) ............................................................. 16 Table 6: Parameter settings of Speedwave®four by BERGHOF for the digestion of the particulate matter samples .............................................................................................. 22 Table 7: Parameter settings of Speedwave®four by BERGHOF for the cleaning of the microwave tubes ............................................................................................................. 23 Table 8: ICP-MS instrumental parameters..................................................................... 25 Table 9: Data acquisition parameters ............................................................................. 25 Table 10: Descriptive statistics of rainy season (April 2014) and dry season (August 2014) samples from Goiânia........................................................................................... 39 Table 11: Corrected lead isotope ratios of Goiânia samples from rainy season (April/May 2014) with their relative standard deviations (X = mean value; s = standard deviation; RSD% = relative standard deviation) ............................................................ 46 Table 12: Corrected lead isotope ratios of Goiânia samples from dry season (August 2014) with their relative standard deviations (X = mean value; s = standard deviation; RSD% = relative standard deviation) ............................................................................. 47 Table 13: Descriptive statistics of rainy season (July 2015) and dry season (December 2015, February/March 2016) samples from Rio Claro................................................... 57 Table 14: Corrected lead isotope ratios of Rio Claro samples from dry season (July 2015) with their relative standard deviations (X = mean value; s = standard deviation; RSD% = relative standard deviation) ............................................................................. 65 Table 15: Corrected lead isotope ratios of Rio Claro samples from rainy season (December 2015 and February/ March 2016) with their relative standard deviations (X = mean value; s = standard deviation; RSD% = relative standard deviation) ................ 66 5 Rio Claro, 2016 Abstract The toxic metal lead (Pb) can be harmful to human health in various manners, but is also considered as a distinguished tracer of environmental pollution, since the relative abundance of its four stable isotopes with the atomic masses of 204, 206, 207 and 208 varies with the emission source. This study is focused on the lead concentrations and isotope ratios in the particulate matter of the Brazilian cities of Goiânia (GO) and Rio Claro (SP), in order to determine the main Pb pollution sources. Particulate matter samples were collected on clean Teflon filters during the rainy and dry season between 2014 and 2016 on the campus of the State University of São Paulo (UNESP) in Rio Claro city and in the centre of Goiânia city near main roads with a high traffic volume. The Pb concentrations as well as the 206Pb/207Pb and 208Pb/207Pb stable isotope ratios of the particulate matter samples were analysed by Inductively-Coupled Plasma Mass Spectrometry. To apply this analytical technique successfully, it was necessary to optimize parameters in case of acquisition time, detector dead time and mass discrimination, which affect the measurement accuracy and precision. Results showed that lead concentrations in Goiânia were different between rainy and dry season. In Goiânia, Pb concentrations showed higher values in dry season than in rainy season, while Pb concentrations were more similar in both sampling periods in Rio Claro. Back trajectories were analysed with the HYSPLIT model to investigate associations between Pb concentration levels and the direction of incoming air masses. However, the comparison of the obtained 206Pb/207Pb and 208Pb/207Pb isotope ratios data with data of potential Pb sources from previous studies indicated that gasoline may be considered as main Pb sources in the particulate matter of Goiânia and Rio Claro. Pb isotope ratios in Goiania were slightly different between dry and rainy season, while in Rio Claro, 206Pb/207Pb isotope ratios showed markedly higher values in dry season than in rainy season. Hence, Pb in Rio Claro in dry season also seems to be influenced by industrial emissions. These assumptions were supported by the calculation of 208Pb/206Pb-207Pb/206Pb diagrams and the contribution factor of Pb coming from gasoline by applying binary mixing equations. Key words: Lead, Lead isotope ratio, Particulate Matter, ICP-MS 6 Rio Claro, 2016 Resumo O metal tóxico de chumbo (Pb) pode ser prejudicial para a saúde humana em várias maneiras, mas também pode ser utilizado como um traçador da poluição ambiental, porque a abundância relativa dos seus quatro isótopos estáveis de massas 204, 206, 207 e 208 conhecidos varia de acordo com a fonte de emissão. Este estudo é focado nas concentrações de chumbo e nas razões isótopicas de material particulado das cidades brasileiras de Rio Claro (SP) e Goiânia (GO), a fim de determinar as principais fontes de poluição por Pb. Amostras de material particulado foram recolhidas em filtros de teflon limpos, durante a estações chuvosa e seca entre anos de 2014 e 2016 na UNESP – Campus de Rio Claro e no centro de Goiânia nas proximidades das estradas principais com um grande volume de tráfego. As concentrações de Pb e as razões isótopicas estáveis de 206Pb/207Pb e 208Pb/207Pb das amostras de material particulado foram analisadas por ICP-MS. Para aplicar esta técnica, foi necessário otimizar os parâmetros como o tempo de aquisição, tempo morto de detector e discriminação de massa, que afetam a exatidão da medida e precisão. Em Goiânia, as concentracões de Pb exibiram valores mais elevados na estação seca do que na estação chuvosa, enquanto as concentracões de Pb foram similares em ambas as campanhas de amostragem em Rio Claro. Trajetórias de volta do modelo HYSPLIT foram analisadas, a fim de investigar- se associações entre os níveis de Pb e as direções das massas de ar. Entretanto, a comparação entre os valores das razões isótopicas 206Pb/207Pb e 208Pb/207Pb e os dados das fontes potenciais de Pb a partir de estudos prévios indicaram que a gasolina pode ser considerada como a principal fonte de Pb para o material particulado em Goiânia e Rio Claro. As razões isótopicas de Pb em Goiânia foram levemente distintas entre os períodos seco e chuvoso, enquanto em Rio Claro as razões 206Pb/207Pb exibiram de forma marcante valores mais elevados na estação seca. Portanto, o Pb em Rio Claro na estacão seca tambem aparenta ser influenciado por emissões industriais. Tais considerações são corroboradas pelo cálculo do fator de contribuição de Pb advindo da gasolina ao aplicar-se equações de mistura binária. Palavras-chave: Razão isotópica de chumbo, Material Particulado, ICP-MS 7 Rio Claro, 2016 Contents 1 Introduction and Objectives ................................................................................... 8 1.1 Objectives ....................................................................................................... 9 2 Literature Review .................................................................................................. 9 2.1 Particulate Matter............................................................................................ 9 2.2 Properties of Lead ........................................................................................ 10 2.3 Legislation and Standard Values for Particulate Matter and Lead ................. 12 2.4 Related Studies ............................................................................................ 16 3 Materials and Methods ........................................................................................ 21 3.1 Sampling ...................................................................................................... 21 3.2 Sample Digestion.......................................................................................... 22 3.3 Isotopic and Elemental Analysis by ICP-MS.................................................. 23 3.4 Statistical Analysis ........................................................................................ 25 3.5 Back Trajectory Calculation .......................................................................... 26 4 Study Locations .................................................................................................. 26 4.1 Study Location of Goiânia ............................................................................. 26 4.1.1 Climate of Goiânia ................................................................................. 27 4.1.2 Geology of Goiânia ................................................................................ 28 4.1.3 Pedology of Goiânia .............................................................................. 29 4.2 Study Location of Rio Claro .......................................................................... 29 4.2.1 Climate of Rio Claro ............................................................................... 30 4.2.2 Geology of Rio Claro ............................................................................. 31 4.2.3 Pedology of Rio Claro ............................................................................ 31 5 Results and discussions ...................................................................................... 32 5.1 Preliminary Studies ....................................................................................... 32 5.1.1 Acquisition Time Optimization ................................................................ 32 5.1.2 Dead Time Determination ...................................................................... 34 5.1.3 Mass Bias Correction ............................................................................. 35 5.2 Results and Discussion for Studies in Goiânia .............................................. 36 5.2.1 Elemental Determination of Lead in Samples from Goiânia ................... 36 5.2.2 Back Trajectory Analysis of Sampling Point in Goiânia .......................... 40 5.2.3 Isotopic Determination of Lead in Samples from Goiânia ....................... 46 5.3 Results and Discussion for Studies in Rio Claro ........................................... 53 5.3.1 Elemental Determination of Lead inSamples from Rio Claro .................. 53 5.3.2 Back Trajectory Analysis of Sampling Point in Rio Claro ........................ 58 5.3.3 Isotopic Determination of Lead in Samples from Rio Claro .................... 65 6 Conclusion .......................................................................................................... 72 7 References.......................................................................................................... 73 8 Rio Claro, 2016 1 Introduction and Objectives Air pollution caused by particulate matter produces a plurality of harmful effects on human health, as statistical associations in previous studies between health effects and particle concentration in the air, especially fine particles with an aerodynamic diameter of 10 μm (PM10), are showing (Schwartz et al., 1996; Pope, 2002). Particles larger than 10 μm can be filtered out in the human nasopharynx and trachea, but smaller particles penetrate deep into the lungs, may even getting into the bloodstream, which results in a number of health problems in extremely serious conditions. Furthermore, particulate matter is a catalyser for reaction of gas pollutants and thus having a significant influence on the climate (Zheng et al., 2004). On the other hand, lead is a poisonous metal, that produces adverse effects on human health in low concentrations, damaging nervous connections, especially on infants, and causing blood and brain diseases. Besides this, lead can also be used as an ambient pollution tracer, according to its four stable isotopes with the atomic masses of 204, 206, 207 and 208, which abundances depend on the emission source, known as isotopic signature. This principle has been used in a variety of ways to determine anthropogenic sources of Pb, for example in freshwater, sediments, aerosols and particulate matter (Mastral et al., 2009). For an accurate elemental and isotopic determination of lead in particulate matter, the use of appropriate analytical methods, such as Inductively Coupled Plasma Mass Spectrometry (ICP-MS), is required, due to its high detection capability and its ability to obtain isotopic information (EPA, 2011). The growth of the regions of Goiânia, located in the centre of Brazil and capital of the state of Goiás, and Rio Claro, located in the south of Brazil in the middle east of the state of São Paulo, is highly connected with environmental problems like air pollution, introducing a wide range of toxic gases, vapours and particulate matter in the atmosphere, causing serious effects on human health like mentioned above. The ceramic production in Rio Claro and emissions from motor vehicles representing two of the main contaminant sources (CETESB, 2014). 9 Rio Claro, 2016 1.1 Objectives The main objective of this study is to determine and compare contents and isotope ratios of lead in particulate matter of air in the cities of Goiânia (GO) and Rio Claro (SP) in the dry and rainy season, in order to identify possible Pb emission sources in both cities. For the city of Goiânia, it was expected that the Pb isotope ratios in particulate are related with the combustion of fossil fuels, while for the city of Rio Claro, it was assumed that the Pb isotope ratios are influenced by the burning process of the raw material in the industrial activity of the ceramic production. Furthermore, basic methods using ICP-MS technique had to be implement previously to ensure an accurate and precise determination of Pb isotope ratios. 2 Literature Review 2.1 Particulate Matter Air pollution is basically present as aerosols or particulate matter (PM). This particulate material is composed by solid or liquid particles in suspension, forming clusters and particles of different sizes and different degrees of toxicity and physicochemical properties. Particulate Matter can be characterized by the distribution of its particle sizes and its chemical compositions (Jacobson, 2002). It has varied chemical compositions, mainly due to the combination of fine particles with secondary pollutant gases (Godish, 1997). The particle size can vary from a few nanometres in diameter up to several micrometres. The finer particles are associated with health problems. (Wholgate et al. 1999; Pope et al. 2002; WHO 2006; Lewtas et al. 2007). The size of particles is directly linked to their potential for causing health problems. The fine particulate matter, which consists of particles suspended in air smaller than 2.5 μm in aerodynamic diameter (PM2.5), reaches the 10 Rio Claro, 2016 bronchioles through respiration, whereas coarse particles between 2.5 and 10 μm (PM10) are retained in the nostrils and nasopharyngeal tract and are subsequently eliminated from the body (Godish, 1997). Numerous scientific studies have linked particle pollution exposure to a variety of problems, including premature death in people with heart or lung diseases, nonfatal heart attacks, irregular heartbeat, aggravated asthma, decreased lung function and increased respiratory symptoms, such as irritation of the airways, coughing or difficulty of breathing (Pope et al., 1995; Schwartz et al., 1996, Zheng et al., 2004). Besides causing harmful effects on human health, airborne particles are pollutants that reduce visibility. Particulate matter can be carried over long distances by wind and then settle on ground or water. The effects of this settling can result in acidifying lakes and streams, changing the nutrient balance in coastal waters and large river basins, depleting the nutrients in soil, damaging sensitive forests and farm crops, and affecting the diversity of ecosystems. Moreover, it catalyses gas pollutant reactions and thus, influences the climate (Kyotani and Iwatsuki, 2002; EPA, 2015). Some particles, known as primary particles, are emitted directly from a source, such as construction sites, unpaved roads, fields, smokestacks or fires. Others are form in complicated reactions in the atmosphere by chemicals, such as sulphur dioxides and nitrogen oxides, that are emitted from power plants, industries and automobiles. The fine particulate matter (PM2.5) is highly connected to sources of combustion processes, like industrial chimneys, vehicle fuel and biomass burning. Chemical species commonly found in this kind of particles are nitrates, sulphates, heavy metals such as lead, organic compounds and carbon (Hopke, 1991). 2.2 Properties of Lead Lead (Pb) is one of the chemical elements which is frequently found in particulate matter, derived from contamination sources mentioned above. It is a poisonous metal, that can cause health damages by harming nervous connections, affecting kidney functions and causing blood and brain disorders. 11 Rio Claro, 2016 This is particularly a risk for young children, due to the increased sensitivity of young tissues and organs (Wang and Zhang, 2006; Mastral et al. 2009). Having an atomic mass of 207.2, lead has four naturally occurring isotopes: 204Pb (1.4%), 206Pb (24.1%), 207Pb (22.1%) and 208Pb (52.4%). However, the ratio between the different lead isotopes vary in different geological environments, e.g. 206Pb and 207Pb are formed by the radioactive decay of 238U (T1/2 = 4.47 109 years) and 235U (T1/2 = 7.04 108 years), while 208Pb is the product of the decay of 232Th (T1/2 = 1.4 1010 years). 204Pb is the only lead isotope which is not a product of radioactive decay. Thus, the isotope ratio of lead depends on the concentration of Pb, Th and U in the starting mixture of the particular environment, the half-life of the decay process and the elapsed time of this process. Hence, every lead ore deposit shows its own characteristic lead composition, due to its age, which represents the time that the lead separated from its source rocks, and the condition of its genesis. Thus, the isotopic composition of Pb in environmental materials can be coupled with the ore bodies from which it was derived. Therefore, if each source of lead has a distinctive isotopic composition, the mixing processes of local Pb and anthropogenic inputs in the environment can be quantified by the Pb isotope ratios (Monna et al., 1995). The isotope ratio of lead is highly representative for its origin and so can be used to identify sources of contamination in soils, sediments, water, marshes, tree trunks, marine organisms and others (Margui et al. 2007). Furthermore, the isotopic composition of lead in the atmosphere is characteristic for a particular region, mainly due to the anthropogenic emission of the element, having a relatively long atmospheric residence time of 5-10 days. Therefore, the Pb isotopic signature is considered as an ideal tool for tracing sources of atmospheric pollution (Settle and Patterson, 1991). Lead isotopes have the advantage of not being fractionated in the part of any natural chemical or biological process on the earth's surface. The environmental sources of lead vary widely and in order to use the isotopic signature as a discriminatory tool, it is necessary that the isotopic ratios are different for each source (Bellis et al., 2005). Lead is naturally abundant in the terrestrial crust, but anthropogenic Pb 12 Rio Claro, 2016 has a relatively higher abundance in the urban atmosphere than in the natural fraction, which comes from sources such as soil dust or sea salt. For a long time, one of the main lead anthropogenic sources was related to traffic emissions due to alkyl lead derivatives used as antiknock agents in gasoline, accounting for 90 percent of airborne lead pollution. In order to reduce its concentration, restrictive normatives in several countries like the USA or the European Union (Directive 98/70/EC, 2000) were established. New non-leaded gasoline with a high octane index was introduced into the market and more resistant materials in the engine industry were used, which not needed lead lubrication for its maintenance. However, lead is still introduced into the atmosphere by traffic, as the current used gasoline still possesses variable lead amounts due to its presence in the crude oil (10–300 μg l-1). In contrast, the concentration of Pb in diesel emissions is usually smaller than 1 μg l-1 (Mastral et al., 2009). Other anthropogenic lead sources are smelting, steel mills, incineration of residues, wood and coal combustion, resuspension of contaminated grounds and general production activities like production of paintings, chemical agents, welds, etc. (Flament et al., 2002; Widory et al., 2004). 2.3 Legislation and Standard Values for Particulate Matter and Lead The Brazilian National Council of Environment (CONAMA) established in the Resolution No. 003 of 1990 parameters for air quality in case of PM10 and TSP (total suspended particles), but no regimentation exists for PM2.5. This national resolution sets a standard for the annual geometric mean concentration of total suspend particles of 80 µg m-3 along with a mean concentration of 240 µg m-3 that should not be exceeded in more than one time per year. Whereas, for inhalable particles (PM10), it is set a standard value of 50 µg m-³ for the arithmetic mean concentration and the mean concentration of 150 µg m-³ in 24 hours should not be exceeded more than one time per year (CONAMA, 1990). A standard for lead is not included in the resolution, but was established along with PM2.5 and PM10 standards by the Environmental State Company of 13 Rio Claro, 2016 São Paulo (CETESB) in the Decree No. 59113/2013 (Table 1). There, the air quality standards were launched in two criteria. Intermediate goals (MI) were set up as temporary values to be achieved in stages, aiming the gradual improvement of air quality in São Paulo state, based on efforts to reduce emissions from stationary and mobile sources, in line with the principles of a sustainable development. Final standards (PF) are standards determined by the best scientific knowledge for preserving people's health as much as possible in relation to effects on human health caused by air pollution (CETESB, 2013). Table 1: Standard values for PM10; PM2.5; TSP and Pb established by CETESB (Source: CETESB, 2013) Pollutent Sampling time MI1 MI2 MI3 PF PM10 24 hours MAA1 120 µg m-³ 40 µg m-³ 100 µg m-³ 35 µg m-³ 75 µg m-³ 30 µg m-³ 50 µg m-³ 20 µg m-³ PM2.5 24 hours MAA1 60 µg m-³ 20 µg m-³ 50 µg m-³ 17 µg m-³ 37 µg m-³ 15 µg m-³ 25 µg m-³ 10 µg m-³ Total suspended particles (TSP) 24 hours MGA² - - - - - - 240 µg m-³ 80 µg m-³ Lead MAA1 - - - 0.5 µg m-³ Meta Intermediate Stage 1 - (MI1) - concentration values of air pollutants that must be respected since 24.4.2013; Intermediate target Stage 2 - (MI2) - concentration values of atmospheric pollutants to be subsequently adhered to MI1, which will come into force after assessments made in Stage 1, developed by technical studies submitted by the State Environmental Agency, validated by CONSEMA; Meta Intermediate Stage 3 - (MI3) - concentration values of air pollutants that must be respected in subsequent years to MI2, and the period of validity will be set by CONSEMA, from the beginning of its term, based on the assessments carried out in Stage 2. The final standards (PF) are applied without intermediate stages when milestones are not established, as in the case of total particulate matter and lead. For other pollutants, the final standards go into effect from the end of the MI3 term (adopted from CETESB, 2013) Furthermore, the Decree No. 59113/2013 also established criteria for acute episodes of air pollution. The decree states the terms "Attention", "Alert" and "Emergency" in addition to exceeded concentration levels (Table 2). 14 Rio Claro, 2016 Table 2: Criteria of acute episodes of air pollution for PM10 and PM2.5 established by CETESB (Source: CETESB, 2013) Parameters Attention Alert Emergency PM2.5 (µg m-³) – 24 h 125 210 250 PM10 (µg m-³) – 24 h 250 420 500 For each measured pollutant, an index is calculated, which is a dimensionless value. Depending on the obtained ratio, the air receives a qualification, which is a score for the air quality as shown in Table 3. Table 3: Air quality indexes for PM10 and PM2.5 established by CETESB (Source: CETESB, 2013) Quality Index PM10 (µg m-3) 24h PM2.5 (µg m-3) 24h N1 – Good 0 – 40 0 – 50 0 – 25 N2 – Moderate 41 – 80 >50 – 100 >25 – 50 N3 – Bad 81 – 120 >100 – 150 >50 – 75 N4 – Very Bad 121 – 200 >150 – 250 >75 – 125 N5 – Miserable >200 >250 >125 However, the World Health Organization (WHO) recommends quality standards for PM2.5 (WHO, 2006). Currently countries like United States, the European Union countries, Canada, Mexico and others, included the monitoring of PM2.5, considering that this particle size causes short and long-term effects on human health (WHO, 2006). The European Union has developed a detailed legislation body which established air quality standards and health based objectives for a number of pollutants in air. These standard values are listed in Table 4. 15 Rio Claro, 2016 Table 4: Standard values for PM10; PM2.5 and Pb established by the European Union (Source: Directive 2008/50/EC, 2008) Pollutent Concentration Averaging period Permitted exceedences each year PM2.5 25 µg m-³ 1 year n/a PM10 50 µg m-³ 24 hours 35 40 µg m-³ 1 year n/a Lead 0.5 µg m-³ 1 year n/a With the Directive 2008/50/EC, additional objectives for PM2.5 were introduced, which target the exposure of the population to fine particles. These objectives are set at the national level and are based on the average exposure indicator (AEI). For the best assess of PM2.5 exposures to the general population, the AEI indicator is determined as 3-year running annual mean PM2.5 concentrations and averaged over several selected monitoring sites in urban agglomeration areas, set in urban background locations (Directive 2008/50/EC, 2008). In the USA, the Clean Air Act requires the United States Environmental Protection Agency (EPA) to set National Ambient Air Quality Standards (NAAQS) for pollutants that are considered as harmful to public health and the environment (Table 5). The Clean Air Act differentiates two types of national ambient air quality standards. Primary standards have the aim to provide public health protection, which includes the protection of the health of "sensitive" populations such as asthmatics, elderly people and infants. Secondary standards are set with the goal to provide public welfare protection, which involves the protection against decreased visibility and damage to animals, crops, vegetation and buildings (EPA, 2015). 16 Rio Claro, 2016 Table 5: Standard values for PM10; PM2.5 and Pb established by the United States Environmental Protection (Source: EPA, 2015) Pollutant Primary/ Secondary Averaging time Level Form PM2.5 primary 1 year 12.0 μg m-³ Annual mean, averaged over 3 years secondary 1 year 15.0 μg m-³ Annual mean, averaged over 3 years primary and secondary 24 hours 35.0 μg m-³ 98th percentile, averaged over 3 years PM10 primary and secondary 24 hours 150.0 μg m-³ Not to be exceeded more than once per year on average over 3 years Lead Primary and secondary Rolling 3 month period 0.15 μg m-³ Not to be exceeded 2.4 Related Studies Several studies on atmospheric isotope ratio of lead and its application to trace air pollution sources have been realized in Europe (Monna et al., 1995; Veron 1999; Mastral et al., 2009), in North and Central America (Sturges and Barrie, 1989; Bollhöfer and Rosman, 2001; Salcedo et al., 2016) and Asia, especially in China as urbanization and rapid economic growth has led to a profound deterioration of urban air quality (Mukai et al., 1993; Zheng et al., 2004; Xu et al., 2011; Lee et al., 2015). ). In contrast, just a few studies on lead isotope ratios in particulate matter were realized in Brazil. The group of Mukai et al. (1993) was one of the first research groups, which characterized the sources of lead in the air using stable isotope ratios of lead. The group collected particulate matter of various urban sites in China, Japan, Korea, Thailand, Sri Lanka and Indonesia and analysed them for Pb isotope ratios along with materials, that were related to the source, such as coal and leaded gasoline, using ICP-MS. A high variation of the lead isotope ratios in the air was determined, as well as that these differences highly corresponded with regional differences in the lead sources. Furthermore, the group discovered, that leaded gasoline was still a primary source in several Asian 17 Rio Claro, 2016 cities, having a major effect on the lead isotope ratios in the air. However, also coal combustion and industrial activity were observed as a considerable effect in Chinese and Korean cities. In 1995, the group of Monna et al. sampled particulate matter, incinerator ashes and gasoline from urban areas in 12 French and two British cities to determine the lead isotope ratios in order to identify Pb pollution sources, using TIMS and ICP-MS. The ratios of 206Pb/207Pb in gasoline varied from 1.061 to 1.094 (mean values were 1.084 for France and 1.067 for the United Kingdom), while for industrially derived lead, ratios of 206Pb/207Pb ranged from 1.143 to 1.155. The isotopic signature of Pb of atmospheric particulate matter reflected the relative importance of each of these sources. The study depicted that lead additives (tetraethyl lead) in gasoline were typically still the main source of Pb in urban particulate matter. The importance of Pb derived gas decreased comparing to results from other sources (natural and industrial), as a result of recent environmental legislation in the European Union, which significantly limited Pb concentrations in gasoline and increased market penetration of unleaded gasoline. Furthermore, the urban Pb isotope ratios may varied considerably due to seasonal variations in sources (e.g. fluctuations in traffic density) and with the wind direction. The study of Mastral et al. (2009) was focused to the concentrations of lead and its isotope ratios of PM10 of Zaragoza, Spain. The objective of that study was to assess the main sources of the pollution of Pb and to verify if the influence of the prohibition of leaded gasoline inside the European Union was appropriate. PM10 were sampled in two campaigns, the first when leaded gasoline still was permitted and the second when leaded gasoline already were prohibited. The sampling was performed with a high volume air sampler to trap PM10 with glass fibre filters coated with Teflon. The filters were analysed with ICP-OES (concentration of Pb) and subsequently with ICP-MS (isotope ratios of Pb), for those samples that had higher concentrations than the quantification limit of Pb. The ICP-MS analysis resulted with differences in respect of lead isotope ratios for both sampling periods, but also showing that anthropogenic sources related to industrial processes were reflected on both campaigns. The contribution of gasoline in relation to the first sampling campaign, as for the 18 Rio Claro, 2016 second campaign just an insignificant contribution was determined, confirming the success of the lead policies of the European Union on the quality of the environment. Zheng et al. (2004) evaluated Pb concentrations and Pb isotopic ratios in samples of PM10, which were collected in seven monitoring locations in Shanghai, China, to investigate the characteristics of Pb isotope ratios of the local air. Several samples related to sources of cement, coal and dust from oil combustion, metallurgic powder, exhaust particles from vehicles derived from leaded gasoline, non-leaded gasoline and polluted soil were analysed for their lead content and isotope ratios and were compared to samples of local PM10. The concentrations of Pb varied between 167 and 854 ng m-3 in seven monitored sites with an average of 515 ng m-3, indicating that a high concentration of Pb remained in the air after the gradual phasing out of leaded gasoline. Isotope ratios in PM10 (207Pb/206Pb = 0.86087 ± 0.0018; 208Pb/206Pb = 2.1057 ± 0.005) were clearly distinct from the exhaust particles of vehicles (207Pb/206Pb = 0.88547 ± 0.0075; 208Pb/206Pb = 2.1457 ± 0.006). A binary mixture equation and a source apportionment based on 207Pb/206Pb ratios were used, indicating that the contribution from automotive emission to the airborne Pb amounted around 20 %. Lee et al. (2006) published a study describing the analysis of TSP (Total Suspended Particles) in the urban and suburban areas of the two large urban centres in South China, Hong Kong and Guangzhou. TSP samples were analysed for major elements (Al, Fe, Mg and Mn), trace elements (Cd, Cr, Cu, Pb, V and Zn) and as well for their Pb isotopic composition, focusing on the isotope ratios of 204Pb/207Pb, 206Pb/207Pb and 208Pb/207Pb. Besides this, seven- day air mass backward trajectories, calculated by the HYSPLIT model (HYbrid Single-Particle Lagrangian Integrated Trajectory, Version 4.7), were used in that research to assess sources of heavy metals in the analysed samples, ending at height levels of 500 m, 1000 m and 1500 m above ground level (AGL). 206Pb/207Pb and 208Pb/207Pb ratios showed lower values in winter, indicating the influence of Pb from the northern inland areas of China and the Pearl River Delta region, and 206Pb/207Pb and 208Pb/207Pb ratios were lower in summer, representing the influence of Pb from the South Asian region and from marine 19 Rio Claro, 2016 sources. Analysation of air mass back trajectories assumed that enrichment of heavy metals in Hong Kong and Guangzhou was highly connected with the air masses originated from northern China, reflecting the long-range transport of heavy metal contaminants from the northern inland areas of China to the South Chinese coast. Xu et al. (2011) determined lead concentrations as well as 207Pb/206Pb and 208Pb/206Pb isotope ratios of PM2.5 between 2007 and 2009 in the Chinese city of Xi`an in order to investigate the effects of the phasing out of leaded gasoline in 2000. PM2.5 samples were collected every 24 hours by using pre- fired 47 mm Whatman quartz-fibre filters. The elemental analysis was conducted by Energy Dispersive X-Ray Fluorescence and the isotopic determination by Thermo Elemental X-Series ICP-MS. Pb concentrations ranged from under the detection limit to 2.631 ng m-3 with an annual average of 0.306 ng m-3, representing a decline since the phasing out of leaded gasoline. Pb isotope ratios were significantly lower in winter (average 207Pb/206Pb ratio of 0.843 ± 0.032; 208Pb/206Pb of 1.908 ± 0.058) than in summer (207Pb/206Pb of 0.860 ± 0.032; 208Pb/206Pb of 2.039 ± 0.057), suggesting that coal combustion was the main Pb source in winter and vehicular emission the main Pb source in summer. The group of Lee et al. (2015) sampled particulate matter PM10 and PM2.5 in the city of Daejeon, Korea, at Asian dust periods and non-Asian dust periods. Asian dust particles, transported over long distances by westerly winds in the late winter and spring seasons, influence the ambient aerosol concentrations in Korea. Samples, daily taken by high-volume air samplers (HV-1000F, Sibata Inc., Japan) using a 20 x 25 cm2 fibre quartz filter, were evaluated on their Pb concentration along with other elements and their Pb isotope ratios. Analysis were performed by Inductively Coupled Plasma – Atomic Emission Spectrometry (ICP-AES) and a Multi-Collector Thermal Ionization Mass Spectrometer (TIMS). Obtained patterns demonstrated that monthly mean concentrations of trace elements in winter were 2 to 13 times higher than in other seasons. The Pb isotope signature in the particulate matter from the non- Asian dust period showed lower 206Pb/207Pb, 207Pb/204Pb, and 206Pb/204Pb ratios in the autumn, winter, and spring seasons, demonstrating that Pb isotope ratios 20 Rio Claro, 2016 were highly influenced by the industrial areas of China. The higher Pb isotope ratios in the summer indicate the dilution and/or mixing effect of marine air masses from the south. Studies of atmospheric lead in Brazil were conducted among others by Babinski et al. (2003). The group determined Pb isotope ratios from PM10, collected with Teflon filters, and rain water samples at the University of São Paulo. Isotopic lead signature of particulate matter and rain water as well as possible lead sources like gasoline, soot from vehicle exhaust pipes and PM from industrial emissions were analysed by thermal ionization mass spectrometry (TIMS). It was suggested that industrial emissions were the main contribution factor for Pb in the PM10 samples, since the contribution of Pb in vehicular fuels was insignificant. In the work of Mirlean et al. (2005), lead concentrations and isotope ratios in atmospheric deposits and potential lead sources in the city of Rio Grande in South Brazil were investigated. Isotope ratios of 206Pb/207Pb and 208Pb/207Pb were measured by VG-Sector Mass Spectrometer and used trace sources of atmospheric lead from different sampling locations distributed over the city. Gioia et al. (2010) provided the first results of a study in São Paulo City, which investigated the processes that control concentrations and sources of Pb in the atmosphere. Pb concentrations and Pb isotope ratios from PM were determined by high temporal resolution (3 hours) collection and Thermal Ionization Mass Spectrometry (TIMS) analysis. Results highlighted the possible impact of a singular short-lived pollution event, present as fireworks, on the urban air quality. Four methods for the calculation of pulse counting detector dead time for quadrupole ICP-MS were described and analysed by the group of Nelms et al. (2000), focusing on the criteria of ease of application and accuracy. While three methods were based on the measurement of 204Pb/208Pb isotope ratio, one method were based on comparison of pulse counting signals versus the corresponding analogue signals across a range of signal intensities. The most suitable method was found out as the method based on 204Pb/208Pb measurement for different Pb concentration solutions, prepared with the NIST 21 Rio Claro, 2016 Standard Reference Material NBS-982, using different dead time settings. The uncertainty in the dead time result of this method yielded minor uncertainties than the other evaluated methods. 3 Materials and Methods 3.1 Sampling Particulate matter samples were collected in the cities of Goiânia and Rio Claro during the rainy and dry season between 2014 and 2016. 32 particulate matter samples from Goiânia (16 from April/May 2014 that represent rainy season, 16 from August 2014 that represent dry season) and 24 samples from Rio Claro (eight from July 2015 that represent dry season, 16 from December, February, March 2015/2016 that represent rainy season) were gathered. To sample particulate matter, a low volume sampler of the type STACKER, which contains a filter holder device was used. This device was connected to a diaphragm pump at the entrance and thus capturing particulate matter with a clean Teflon filter (Figure 1). These filters were FluoroporeTM membrane filters of the type PTFE with a diameter of 47 mm and a pore diameter of 0.22 µm. The entrance and the AFG device were connected to a diaphragm pump, having a volume meter and maintaining an airflow of about 40 m³ per day in Goiânia and 30 m³ per day in Rio Claro. (Castanho and Artaxo, 2001) The sampling time was 24 (Goiania samples, July 2015 samples from Rio Claro) and 48 hours (rainy season samples from Rio Claro). Sampling time of 48 hours in rainy time was chosen in order to accumulate higher amounts of particulate matter and therefore obtain higher precisions at the elemental and isotopic determination of lead, as in rainy season markedly lower amounts of particulate matter were recorded. All filter membranes were weighed (Figure 2) before and after collection in order to obtain the mass of the particulate matter. 22 Rio Claro, 2016 Figure 1: Photograph of the low volume sampler STACKER (left) and the integrated filter holder device with Teflon filter (right) for particulate matter sampling 3.2 Sample Digestion The sample digestion method was adopted from Method IO-3.1: "Selection, Preparation and Extraction of Filter Material" (Compendium of Methods for the Determination of Inorganic Compounds in Ambient Air, U.S. EPA, 1999). For each filter, digestion was realized in 10 mL of aqua regia at a concentration of 22.3 % (v v-1) in a microwave tube (Figure 2) using the microwave digester Speedwave®four by BERGHOF with the parameter settings as shown in Table 6. Table 6: Parameter settings of Speedwave®four by BERGHOF for the digestion of the particulate matter samples Steps Temperature Time Ramp Power Pressure 1 140 °C 10 min 5 min 45 % 35 bar 2 160 °C 10 min 5 min 45 % 40 bar 3 200 °C 5 min 5 min 45 % 35 bar The standard solutions were prepared on the basis of the certified solid monoelemental standard SRM 981 by the National Institute of Standards & Technlogy (NIST) containing 1 g of Pb with a purity of 99,9+ %. The concentrations of aqua regia in the standards and solutions after the digestion in the microwave digester were 5.3 % (v v-1). 23 Rio Claro, 2016 Figure 2: Gravimetric measurement of Teflon filter before sampling (left), filter with sampled particulate matter before digestion in front of microwave tube (right) The quantification of Pb requires the use of appropriate rooms, properly decontaminated material and ultrapure reagents. To mitigate any lead interference, all decontamination procedures and sample preparations were realized under an exhaust hood and a laminar flow hood. All solutions were prepared with analytical grade chemicals, bidestillated acids below the boiling temperature and ultrapure water (18 MΩ cm). The decontamination procedure of the microwave tubes was based on the method by BERGHOF (Cleaning Vessel Conditioning), using 10 ml of agua regia and parameters as shown in Table 7 in the microwave digester. Table 7: Parameter settings of Speedwave®four by BERGHOF for the cleaning of the microwave tubes Steps Temperature Time Ramp Power Pressure 1 170 °C 5 min 5 min 90 % 50 bar 2 200 °C 15 min 1 min 90 % 50 bar 3 50 °C 15 min 1 min 0 0 4 50 °C 1 min 1 min 0 0 5 50 °C 1 min 1 min 0 0 3.3 Isotopic and Elemental Analysis by ICP-MS Inductively Coupled Plasma Mass Spectrometry (ICP-MS) is an 24 Rio Claro, 2016 instrumental analysis technique, which provides the determination of elements in a relatively short time. Because of its high detection capability, this technique is nowadays the most widely used technique for analysing ultra-trace elements. In addition to these characteristics, the technique can also be used for isotopic measurements with accuracies to 0.05 % in case of devices based on quadrupole type mass analysers (Skoog and Leary, 2009; Taylor, 2001; Thomas, 2004). The components of ICP-MS are highlighted by the sample introduction system, usually using a nebulizer, the ionization source in form of the inductively coupled plasma, the mass analyser (quadrupole) and the ion detection system (Becker and Dietze, 1998). The technique is based on the fact that each element has at least one isotope, whose ratio between the mass and the charge thereof is exclusive. The chemical compounds contained in the solution are decomposed into their atomic constitution by inductively coupled plasma and ionized with a high degree of ionization (90 % for majority of the elements), forming a fraction of ions of very small multiple loads (< 1 %). The mass analyser separates the ions by their mass/charge ratios, which are then detected by an electron multiplier. However, a factor that may significantly affect the analysis by ICP-MS is the occurrence of interference from various sources, which can affect the accuracy and the detection limits. These interferences are commonly classified to their origin, in spectral and non-spectral interferences (Becker and Dietze, 1998). The elemental and isotopic analysis of the digested samples were performed by ICP-MS of the type X Series II (Thermo Fisher Scientific Inc.) using the instrumental and data acquisition parameters as shown in Table 8 and 9. The produced lead standard solutions had a concentration of 0 ng mL-1, 0.5 ng mL-1, 2.5 ng mL-1 and 5 ng mL-1. Lead concentration was calculated by adding up the measured counts per second of the Pb isotopes (206; 207; 208) per sample and inserting them in the generated linear function of the concentration of the standard solutions. 25 Rio Claro, 2016 Table 8: ICP-MS instrumental parameters Nebulizer Mira Mist RF Power [W] 1450 Flow rate nebulizer gas [l min-1] 0.8 Flow rate coolant gas [l min-1] 13 Flow rate auxiliary gas [l min-1] 0.7 Peristaltic pump rotation [RPM] 15 Torch Sample depth [mm] 150 Vertical [mm] 93 Horizontal [mm] 53 Table 9: Data acquisition parameters Measuring mode Peak jump Points per mass 3 Dwell time 4 ns for 206 and 207 Pb; 2 ns for 208 Sweeps 20450 Acquisition time per main run 300 s Replicates 3 In one analysation campaign, always 8 sample filters were digested and their respective sample solutions analysed along with two filter blanks and two acid blanks. For the determination of the isotopic ratios of Pb, equipment optimization tests in case of acquisition time, dead time and mass discrimination had to be performed before the analysis of lead isotope ratios in the samples could have been realized. 3.4 Statistical Analysis Results were summarized into a multielemental database using Microsoft Excel 2016 and statistically characterized and analysed with the BioEstat 5.3 and ORIGIN 8 statistical softwares. As the data was not normally distributed, the nonparametric Kruskal–Wallis test was employed to investigate the 26 Rio Claro, 2016 differences in lead concentrations and isotope ratios. The Kruskal-Wallis test statistic, H, was calculated as equation 1: � = �� �(���) ∑ ��² �� − 3(� + 1)���� (1) where ni is the number of observations in group i, N the total number of observations in all k groups), and Ri the sum of the ranks of the ni observations in group i. To estimate which groups are differing significantly from each other, a Post-hoc test has to be performed. If the test results in a P-value of P<0.05, the resumption that the two groups are statistically equal to each other, has to be rejected (Zar, 1999). 3.5 Back Trajectory Calculation In order to investigate the sources of lead in the particulate matter samples, air mass backward trajectories were calculated using the HYSPLIT model (Hybrid Single-Particle Lagrangian Integrated Trajectory, Version 4.7), a comprehensive modelling system provided by the National Oceanic and Atmospheric Administration (NOAA) Air Resource Laboratory (Draxler and Rolph, 2003). For detailed episode studies, trajectories ended at different height levels, i.e. 0 m; 100 m; 250 m AGL (above ground level), at 0900 UTC (1000 UTC for samples taken in daylight saving time period) representing 12:00 local time, arriving at the sampling point at the end of each sampling period. Run time of each back trajectory was 24 h or rather 48 h depending on the sampling time of each sampling period. 4 Study Locations 4.1 Study Location of Goiânia The city of Goiânia, the capital of the Brazilian state of Goiás, is located in the Mid-West of Brazil and possessed 1.43 mi. inhabitants in 2015, owing a fleet of 1.103 mi. vehicles, which is a proportion comparable to highly 27 Rio Claro, 2016 industrialized countries (IBGE, 2015). The city of Goiânia is located in the central region of Goiás, the sampling point is placed in the central area of the city along the old railway station under the following geographical coordinates: 16°66’31.86’’S; 49°26‘15.75’’W (Figure 3). Sampling equipment was donated by SEMARH (Secretaria do Meio Ambiente e dos Recursos Hídricos). The location of the sampling point has a strategic importance. The surrounding area is densely occupied by public buildings (town hall of Goiânia, Court of Auditors of the Municipality, main bus station and hospitals), private buildings intended for trade in general (Goiânia Station, Araguaia Shopping, supermarket and hotels) and periodically the largest street market in the city of Goiânia, which attracts thousands of people from the city and the region. The near Praça da Estação is lined by avenues and streets with a high traffic flow (Avenida Goiás Norte, Avenida Independência and Rua 44). Figure 3: Location of the sampling point in Goiania (Source: COSTA e SILVA, 2015) 4.1.1 Climate of Goiânia The climate in the municipality of Goiânia according to the Köppen climate classification is type Aw, tropical humid with two distinct seasons, an almost cold and dry winter from May to October and hot and rainy summer from November to April. In the hottest months, the maximum temperatures vary 28 Rio Claro, 2016 between 29 °C and 35 °C, while in the coldest months of the year, the minimum temperatures can reach 13 °C (Alvares et al., 2005). The summers are characterized by rapid changes in daily weather conditions, with rains of short duration and high intensity, accompanied by thunderstorms and high wind gusts. But also short dry seasons with periods of drought lasting 7 to 15 days can occur in this time of the year. The amount of precipitation can oscillate between 1,200 and 2,500 mm between the months of September to April. In the fall, temperatures become milder due to the entry of cold air masses with minor temperatures between 12 ºC and 18 ºC and average maximums of 28 ºC. The relative humidity is high with values reaching up to 98 %. The winter brings the typically dry climate of the Cerrado biome, with low moisture content, reaching extreme low humidity values. In September, with the beginning of spring, the rains become more intense and frequent again (Governo Goiás, 2016). 4.1.2 Geology of Goiânia Goiânia is located on the Brasília Fold Belt, which is situated between the Craton of São Francisco and the Massive of Goiás. This fold-and-thrust-belt is formed by a greenstone belt and high metamorphic rocks (Fonseca et al., 1995; Ulhein et al., 2012; Pimentel and Fuck, 1992). Granite-greenstone terrains occupy a substantial part of Goiás, including large tracts of granitic-gneissic rocks of a granodioritic to tonalitic composition, metamorphosed in the green schist facies to amphibolite, generally mylonitized and ultramylonitized (Danni et al., 1982). The Annapolis-Itauçu granulitic complex, located in the north-west and the center of the Folha Goiânia, is formed by basic to ultrabasic and metasedimentary components, represented by impure quartzites, calcisilicatic rocks, impure marbles and aluminous to hyper-aluminous gneisses. This granolithic complex is genetically related with the Orthogneissic-Migmatic Formation (Associação Ortognáissica-Migmatítica), which forms a belt from the north of the Folha Goiânia to its south-east, where calcium-sodic, tonalitic and granodioritic orthogneisses are dominating. The metasediments of the Faixa Uruaçu range, distributed in the south-west and 29 Rio Claro, 2016 north-east edges of the Folha Goiânia, are defined as Middle Proterozoic, present as a recumbent fold structure with metamorphism of the "Barrowianan" type, showing an increasing intensity from east to west (Fuck and Marini, 1981). 4.1.3 Pedology of Goiânia In Goiania, as in the biggest part of the state of Goiás, latosols with a low till medium fertility are the predominant soils, especially in form of dystrophic dark yellow latosol and alic red-yellow latosol. In this soils, clays of the kaolinite type are typical, whose particles are coated with iron oxides, which are responsible for the typical reddish colours. The transition between the soil horizons is gradual or diffuse, except the darkening in the A horizon. The B horizon structure is composed of rounded and very small (0.5 to 3.0 mm) aggregates, leaving a lot macro pores between them, which provides a high water permeability even with high levels of clay (Silva and de Castro, 2002). Another soil type in this region is the alic yellow-red podsol. These soils have a sub superficial horizon of accumulated clay, as clay particles migrate from the A horizon and deposit in the B horizon. Thus, the most superficial layers (horizons A and E) are sandier. Compared to latosols, yellow-red podsol shows a less shallower, slightly larger proportions of silt, minerals with little resistant to weathering and a remarkable differentiation of their horizons (Silva and de Castro, 2002). 4.2 Study Location of Rio Claro The city of Rio Claro, owing 199.961 inhabitants (IBGE, 2015), is located in the middle east of the Brazilian state of São Paulo, belonging to the large ceramic center of Santa Gertrudes, which also includes the cities of Limeira, Ipeuna, Cordeirópolis, Araras and Piracicaba, integrated in the association of ASPACER (Associação Paulista das Cerâmicas de Revestimento). The production of ceramic in this region reached 760 million m² in 2010, representing 83% of the total production in São Paulo and 56 % in Brazil, considered as the biggest ceramic production centre in Latin America and the 30 Rio Claro, 2016 second biggest in the world. Besides ceramic production, the region faces environmental problems derived from the burning of sugar cane plantations, mainly in the dry season, when the air quality decreases drastically (ASPACER, 2011). The particulate matter sampling point was installed at the Campus Bela Vista, near CEAPLA, at the terrain of the Instituto de Geociências e Ciências Exatas (IGCE) of the State University of São Paulo (UNESP) in Rio Claro (Figure 4), located at the geographic coordinates with a latitude of 22º39’ 22.30’’S, a longitude of 47º54’ 59.11’’W and an altitude of 626.5 m. Figure 4: Location of the sampling point in Rio Claro (Source: Lorenzeto de Abreu, 2014) 4.2.1 Climate of Rio Claro The climate classification system of Köppen ranks the region of Rio Claro as Cwa (temperate with cold winters). The region is subjected to two distinct seasons, dry season from April to September and rainy season from October to March. Dry season shows an average temperature of 17 °C, while in rainy season average temperature is above 22 °C (Troppmair, 1992). Annual rainfall is above 1500 mm, of which 77 % are concentrated in the rainiest months (October to March) (Alvares et al., 2005). Considering the year 2005, the maximum precipitation occurred in January (482.2 mm) and the minimum in 31 Rio Claro, 2016 July (3.6 mm), the total annual precipitation amounted 1398.3 mm (Estação Meteorológica de Rio Claro (CEAPLA), 2005). The prevailing winds of the region of Rio Claro come from the south and south-east, justifying the location of the industrial district in the northern part of the city in order to avoid air pollution (Troppmair, 1992). 4.2.2 Geology of Rio Claro Rio Claro is located in the center of the Corumbataí river basin, situated in the north-east part of the Paraná Sedimentary Basin, which belongs to the Peripheral Paulista Depression in the middle-east of the state of São Paulo. This depression belongs to the outcrop area of Palaeozoic (Itararé group, Tatuí formation and Passa Dois group – Irati and Corumbataí formation), Mesozoic (Pirambóia formation, São Bento group – Botucatu and Serra Geral formation, intrusive magmatic rocks – dykes and sills, and Itaqueri formation) and Cenozoic (Rio Claro formation) rocks (Almeida, 1964; Melo, 1995). The Corumbataí river flows from its source in the region of Analândia, to its mouth, the Piracicaba river, along on various rock types as mentioned with the stratigraphic units above. Due to the structure caused by the Pitanga Dome, the stratigraphic units of the upper river course (Botucatu, Pirambóia and Corumbataí formation) and the lower river course (Corumbataí formation) are younger than the average course (Itararé Group and Tatuí formation). The Pitanga Dome is a Mesozoical structural high, today topographically downgraded by the drainage of the Corumbataí river basin, which promotes the appearance of Paleozoic rocks from the base of the stratigraphic column of the Paraná Sedimentary Basin in this region (Almeida, 1964; Penteado, 1976). 4.2.3 Pedology of Rio Claro In the city of Rio Claro, sandstone rocks and consequently sandy soils are predominant. This soils have a fast rainwater infiltration and a high percolation to the horizon. Thus, leaching is dominant in the soil horizons, turning the soils acid and poor in nutrients. These soils are predominantly 32 Rio Claro, 2016 occupied by grasslands and the forestation of eucalyptus (Troppmair, 1992). In Rio Claro occur basically three types of sandy soils. The yellow-red podsol with a medium-clayey texture is present at low topographies like in the valleys of Corumbataí River and Ribeirão Claro. This soil is formed from fine sediments of the Passa Dois and Tubarão group (Cottas, 1983). Red-yellow latosols are also present in this region as well as purple latosols, which have a clayey to very clayey texture. Another type of soil of this region, but present in little distribution, are the lithic soils, whose main feature is its small profoundness (less than 30 cm). These soils are associated with siltstones and argillites from the Corumbataí formation and with diabase intrusions from the Serra Geral. Furthermore, hydromorphic soils occur related to alluvial sediments in the area, restricted to the valleys of Corumbataí River and Ribeirão Claro (Cottas, 1983; Oliveira et al., 1999). 5 Results and discussions 5.1 Preliminary Studies 5.1.1 Acquisition Time Optimization To obtain the optimal acquisition time for the subsequent analysis of the particulate matter samples, an experiment was conducted, where lead standards with different concentrations (0.25 ng mL-1, 0.75 ng mL-1, 1 ng mL-1) were analysed by ICP-MS for their 206Pb/207Pb and 208Pb/207Pb ratios at two different acquisition times (5 min and 10 min) in five main runs. Subsequently, the precision in form of the relative standard deviations (RSD) of the five main runs for each obtained ratio was compared to the appropriate theoretical values, which were calculated by the following equation: RSD =�1 c1 + 1 c2 x 100 (2) where c1 is the measured incoming counts per second for the first isotope and c2 is the measured incoming counts per second for the second isotope obtained 33 Rio Claro, 2016 by ICP-MS. As expected, for the calculated Pb isotope ratios, the relative standard deviations obtained by an acquisition time of 10 min were always lower than the standard deviations of the same samples at 5 min of acquisition time. In contrast to this, the experimental obtained values differ highly from the theoretical standard deviations, as in four cases the relative standard deviations of 5 mins of acquisition time showed a lower value then the values of 10 min of acquisition times for the same sample (Figure 5). This results emphasize that an application of a high acquisition time of 10 min for the further isotopic determination of the particulate matter samples is unessential with regard to the precision and also to the saving of time and costs. Thus, an acquisition time of 5 mins was applicated for the further isotopic analysis of the particulate matter samples. Figure 5: Theoretically calculated and experimentally obtained relative standard deviations of 206Pb/207Pb and 208Pb/207Pb isotope ratios for different concentrations at 5 min and 10 min of acquisition time 34 Rio Claro, 2016 5.1.2 Dead Time Determination An accurate determination of the detector dead time is indispensable, if accurate isotope ratio measurements with ICP-MS are to be made. Detector dead time is considered as the time for which the ion counting system is unable to record incoming ions after the impact and detection of a previous ion (Knoll, 2000). This results in counting losses increase in magnitude as the counting rate increases. Because of this fact, inconsistencies in isotope abundance ratio measurements occur, which are independent of mass discrimination effects and on account of this, has to be accurately corrected prior to correcting mass discrimination. The dead time is a feature of the ion counting electronics and the detector itself. (Jarvis, 1992). In order to find the optimal dead time adjustment for the analysis of the particulate matter samples by ICP-MS, four lead standard solutions with different concentrations (2 ng mL-1, 4 ng mL-1, 8 ng mL-1, 10 ng mL-1), produced from NIST Standard Reference Material 981, were prepared and analysed by ICP-MS on their 208Pb/207Pb ratios at different dead time adjustments (20 ns, 30 ns, 40 ns, 50 ns and 60 ns). The measured 208Pb/207Pb isotope ratios for each standard concentration were normalised (=(208Pb/207Pb)measured/true value) and recorded in a diagram with the 208Pb/207Pb ratios plotted against the detector dead time (Figure 6). The polynomial trendlines of the 208Pb/207Pb ratios for each concentration were added to the diagram for obtaining the optimal dead time of the detector, which is shown by the point of intersection of the lines. In this case, the dead time of the detection system was observed as 34 ns. This value was adopted to the parameter settings of the ICP-MS and applicated for the further elemental and isotopic determination of the particulate matter samples. 35 Rio Claro, 2016 Figure 6: Normalized 208Pb/207Pb isotope ratio plotted as a function of the value applied for dead time correction of the 'raw' results obtained for solutions with a Pb concentration ranging from 2 to 10 ppb 5.1.3 Mass Bias Correction The differential transmission of ions of different masses from the point at where the ions enter the sampling device until they finally are detected by the electron multiplier is called mass bias effect. This effects occur principally in the interface region of the magnetic sector of the ICP. It is considered, that several processes conduce to this effect, e.g. the space charge effect in the plasma or vacuum interface regions (Jarvis, 1992). The mass bias was corrected by measuring an isotopic standard of Pb (2.5 ng mL-1) and calculating a correction factor by using the following mathematical expression: Rtrue Robs =(1+εlinear.Δm) (3) where Rtrue is the corrected isotope ratio; Robs is the experimental obtained isotope ratio; Δm is the difference of the atomic weights between the two isotopes and εlinear is the correction factor. R² = 0.8 R² = 0.9793 R² = 0.9893 R² = 0.9903 0.996 0.998 1 1.002 1.004 1.006 1.008 1.01 1.012 20 25 30 35 40 45 50 55 60 2 0 8 P b /2 0 7 P b Detector dead time applied for correction (ns) 'Normalised' (208Pb/207Pb) isotope ratio 2 ppb 4 ppb 8 ppb 10 ppb Linear (2 ppb) Linear (4 ppb) Linear (8 ppb) Linear (10 ppb) 36 Rio Claro, 2016 5.2 Results and Discussion for Studies in Goiânia 5.2.1 Elemental Determination of Lead in Samples from Goiânia Figure 7 demonstrates one of the analytical curves of the standard solutions used for the quantification of lead in Goiânia, showing a coefficient of correlation of R² = 0.9999, which represents a very high linear relationship. All analytical curves of the used standard solutions had coefficients of determination values of at least R² = 0.9995. Figure 7: Analytical curve for the quantification of Pb via ICP-MS Figures 8 (April/May) and 9 (August) represent the obtained lead concentrations from Goiânia in ng mL-1 in the sample dilutions along with the lead concentrations in the respective filter blank dilutions and acid blanks. The standard deviation of the lead concentrations of the three main runs is amounted for each sample or blank on the respective bar in the diagram. Instrumental detection limit, which is equal to three times the standard deviation of the blank signal at the selected analytical masses, was identified as 0.0035 ng mL-1. The method detection limit had a value of 0.014 ng mL-1. y = 82933x + 2686.3 R² = 0.9999 0 50000 100000 150000 200000 250000 300000 350000 400000 450000 0 2 4 6 S ig n a l [I C P S ] Concentration [ng ml-1] 37 Rio Claro, 2016 Figure 8: Pb concentrations (Conc.) and respective standard deviations (s) in ng ml-1 in sample and blank dilutions of Goiânia samples from April and May 2014 (BL Ac = Acid blank, BL F = Filter blank) Figure 9: Pb concentrations (Conc.) and respective standard deviations (s) in ng ml-1 in sample and blank dilutions of Goiânia samples from August 2014 (BL Ac = Acid blank, BL F= Filter Blank) In order to receive the lead concentration per volume of air, the determined volume of the Pb concentration in the sample solutions in ng mL-1 were calculated to the mass of Pb per m³ of air by multiplicating the Pb concentration in the sample dilution with the final volume of the solution and dividing it through the respective sampled volume of air. The results are presented in the Figures 10 and 11. -0.1 0.9 1.9 2.9 3.9 4.9 1 1 /0 4 /1 4 1 3 /0 4 /1 4 1 4 /0 4 /1 4 1 6 /0 4 /1 4 1 7 /0 4 /1 4 1 8 /0 4 /1 4 1 9 /0 4 /1 4 2 1 /0 4 /1 4 2 3 /0 4 /1 4 2 4 /0 4 /1 4 2 5 /0 4 /1 4 2 7 /0 4 /1 4 2 9 /0 4 /1 4 3 0 /0 4 /1 4 0 7 /0 5 /1 4 1 4 /0 5 /1 4 B L A c 1 B L A c 2 B L A c 3 B L A c 4 B L F 1 B L F 2 B L F 3 B L F 4P b c o n ce n tr a ti o n [ n g m L-1 ] April/ May 2014, Goiânia Conc. (ppb) s -0.1 0.9 1.9 2.9 3.9 4.9 0 4 /0 8 /1 4 0 6 /0 8 /1 4 0 7 /0 8 /1 4 0 8 /0 8 /1 4 0 9 /0 8 /1 4 1 0 /0 8 /1 4 1 2 /0 8 /1 4 1 5 /0 8 /1 4 1 6 /0 8 /1 4 1 8 /0 8 /1 4 2 0 /0 8 /1 4 2 1 /0 8 /1 4 2 2 /0 8 /1 4 2 4 /0 8 /1 4 2 6 /0 8 /1 4 2 7 /0 8 /1 4 B L A c 1 B L A c 2 B L A c 3 B L A c 4 B L F 1 B L F 2 B L F 3 B L F 4 P b c o n ce n tr a ti o n [ n g m L-1 ] August 2014, Goiânia Conc (ppb) s 38 Rio Claro, 2016 Figure 11: Pb concentration in ng m-3 for Goiânia samples from August 2014 (X = mean value; s = standard deviation; RSD% = relative standard deviation; n = quantity) 3.70 1.79 1.36 2.92 1.23 1.74 2.41 1.16 0.90 0.72 0.98 2.47 1.50 2.66 2.75 1.91 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 P b c o n ce n tr a ti o n [ n g m -3 ] August 2014, Goiânia X = 1.89 ng m-3 s = 0.85 ng m-3 RSD% = 45.06 % Figure 10: Pb concentration in ng m-3 for Goiânia samples from April and May 2014 (X = mean value; s = standard deviation; RSD% = relative standard deviation; n = quantity) The Pb concentrations of the Goiânia samples show a high disparity between rainy season samples (April/May) and dry season samples (August). The average Pb concentration from April is lower with a mean value of 1.28 ng m-3 than the mean Pb concentration from August with 1.89 ng m-3. Furthermore, samples of rainy season show a considerably higher coefficient of variation with 97.31 % compared to 45.06 % of the samples from dry season, demonstrating a high divergence within the rainy season samples. Samples from April/May differ from 0.36 ng m-3 to 4.53 ng m-3. Two samples have a higher concentration than 3 ng m-3, no other sample of this campaign shows a higher value than 2 ng m-3. This outlier values may could be attributed to a possible sample contamination. 0.61 0.58 1.98 3.89 1.33 1.65 0.63 0.53 0.52 0.75 0.36 0.37 0.38 1.51 0.92 4.53 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 P b c o n ce n tr a ti o n [ n g m -3 ] April/ May 2014, Goiânia X = 1.28 ng m-3 s = 1.25 ng m-3 RSD% = 97.31 % n = 16 39 Rio Claro, 2016 In contrast, August samples differ from 0.72 ng m-3 till 3.70 ng m-3. Six samples from August have a higher Pb concentration than 2 ng m-3, one sample show a higher concentration than 3 ng m-3. Median and arithmetic average of the rainy season samples differ considerably while median and arithmetic average of dry season samples are quite close to each other (Table 10). The high variation of April samples is also expressed in the asymmetry, where this samples show a value of 1.871, whereas August shows a value closer to 0 with 0.532. Table 10: Descriptive statistics of rainy season (April 2014) and dry season (August 2014) samples from Goiânia Rainy Season 2014 Dry Season 2014 Sample size 16 16 Minimum 0.360 0.720 Maximum 4.530 3.700 Total Amplitude 4.170 2.980 Median 0.690 1.765 First Quartile (25%) 0.528 1.213 Third Quartile (75%) 1.545 2.518 Interquartile Deviation 1.018 1.305 Arithmetic Average 1.284 1.888 Variance 1.564 0.726 Standard Deviation 1.250 0.852 Standard Error 0.313 0.213 Coefficient of Variation 97.40% 45.14% Asymmetry 1.871 0.532 Also, both sample periods show differences in their quartiles as shown in the generated boxplot diagram (Fig. 12). While 25 % of April samples are lower than 0.528 ng m-3, just 25 % of August samples are lower than 1.765 ng m-3. Even with the two high outlier values, just 25 % of the April samples show higher Pb concentrations than 1.545 ng m-3, while 25 % of the August samples reach higher Pb concentrations than 2.518 ng m-3. Due to the two outlier values, arithmetic average of rainy season samples is stretched to a higher value. This data demonstrates that in general dry season samples show markedly higher Pb concentrations and differ in a much lower scope than samples from rainy season. 40 Rio Claro, 2016 Figure 12: Boxplot diagram for Pb concentrations of samples from rainy and dry season 2014 from Goiânia To further examine differences in the Pb concentrations between the two sample periods, a Kruskal-Wallis test was performed. The result of the test demonstrated a H-value of 5.8182 and P-value lower than 0.05 (P= 0.0143), rejecting the H0 hypothesis that both sample groups are equal and showing that Pb concentrations from April and August are statistically different to each other. No lead concentration value exceeded the guide values established by CETESB, the European Union or United States Environmental Protection Agency. In comparison with a previous study by Bollhöfer and Rosman (2000), where Pb concentration in particulate matter from São Paulo City ranged between 3 ± 1 and 52 ± 14 ng m-3 with an average of 29 ng m-3, Pb concentrations from Goiania are relatively low. Also, the study from AILY (2001) shows clearly higher Pb concentrations ranging from 10.23 ng m-3 to even 254.52 ng m-3 in particulate matter sampled in São Paulo City. In that study, dry season samples with an average value of 60.56 ± 43.38 ng m-3 showed higher Pb concentrations than in rainy season with an average of 36.14 ± 37.80 ng m- 3. 5.2.2 Back Trajectory Analysis of Sampling Point in Goiânia Back trajectories with a run time of 24 hours were calculated for sampling dates in rainy and dry season with markedly elevated (higher than or near 75 % quartile) and low concentration levels (lower than or near 25 % quartile) of lead 41 Rio Claro, 2016 in order to assess directions of air masses and hence possible sources of Pb in Goiânia samples. Trajectory plots ended respectively at the end of each sampling date at three different heights (0 m, 100 m, 250 m AGL). In rainy season at sampling dates with elevated Pb concentrations (14th, 16th, 18th April, 14th May) all air masses came from eastern direction, originated in the hilly landscape in the north of the state of Minas Gerais, passing the Serra dos Christais and the eastern suburbans of Goiânia, mainly through low heights up to 500 m AGL (Fig. 13). At sampling days in rainy season with low Pb concentrations (11th, 21th, 23th and 29th April), direction of air masses varied between east, north-east, and south-east direction (Fig. 14). Air masses from eastern direction showed similar trajectories as at sampling days with high Pb concentration, but passing greater heights up to 800 m AGL. Trajectories from north-eastern direction originated in the mountain ridges of the Serra Geral do Parana and moved across the city of Brasilia to Goiânia, while air masses from south-eastern direction passed from the hills of Serra do Conforte to Goiânia in heights up to 500 m AGL. In dry season for sampling days with high Pb concentrations (4th, 8th, 21th and 26th August), no pattern of direction is distinguishable (Fig. 15). Air masses coming from southern and south-eastern direction from the elevated plains, passing low heights between 50 and 250 m AGL, and from eastern direction, where trajectories travelled in great heights up to 1400 m AGL from the Serra dos Christais westward to Goiânia. On sampling days with low Pb concentrations in dry season (7th, 9th, 18th and 20th August), a pattern that all trajectory plots came from east direction is recognizable. While trajectories originated in low heights in the north of Minas Gerais, they increased their height up to 500 m before arriving in Goiânia (Fig. 16). Hence, back trajectory analysis shows that on days with elevated lead concentrations in rainy season, air masses came from eastern direction and heights up to 500 m, while on days with low Pb concentration no trend was recognizable. In contrast to that, in dry season on days with high Pb concentration, no pattern of air mass direction was estimated, whereas on days with lower Pb concentrations air mass directions showed the trend to come from 42 Rio Claro, 2016 Figure 13: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with elevated Pb concentrations in rainy season arriving at sampling point in Goiânia eastern direction, travelling in heights up to 500 m. 43 Rio Claro, 2016 Figure 14: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with low Pb concentrations in rainy season arriving at sampling point in Goiânia 44 Rio Claro, 2016 Figure 15: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with elevated Pb concentrations in dry season arriving at sampling point in Goiânia 45 Rio Claro, 2016 Figure 16: Back trajectory plots ending at three different height levels (0 m; 100 m; 250 m AGL) on episodic dates with low Pb concentrations in dry season arriving at sampling point in Goiânia 46 Rio Claro, 2016 5.2.3 Isotopic Determination of Lead in Samples from Goiânia The tables 11 and 12 show the mass bias corrected 206Pb/207Pb and 208Pb/207Pb isotope ratios with the afore mentioned equation (2) along with their relative standard deviations (RSD) for three main analysing runs. For an accurate lead isotope evaluation, it is intended to have a RSD of <0.2 % for the isotope ratios. Table 11: Corrected lead isotope ratios of Goiânia samples from rainy season (April/May 2014) with their relative standard deviations (X = mean value; s = standard deviation; RSD% = relative standard deviation) Date 206Pb/207Pb RSD% 208Pb/207Pb RSD% 11/04/14 1.1711 0.077 2.4495 0.162 13/04/14 1.1757 0.112 2.4462 0.111 14/04/14 1.1731 0.048 2.4465 0.158 16/04/14 1.1751 0.038 2.4505 0.030 17/04/14 1.1737 0.040 2.4522 0.089 18/04/14 1.1721 0.210 2.4475 0.081 19/04/14 1.1737 0.093 2.4512 0.075 21/04/14 1.1620 0.190 2.4406 0.304 23/04/14 1.1717 0.247 2.4492 0.117 24/04/14 1.1660 0.240 2.4436 0.264 25/04/14 1.1727 0.379 2.4512 0.190 27/04/14 1.1761 0.212 2.4495 0.106 29/04/14 1.1757 0.246 2.4482 0.117 30/04/14 1.1711 0.026 2.4505 0.135 07/05/14 1.1667 0.173 2.4452 0.107 14/05/14 1.1717 0.191 2.4442 0.147 X 1.1718 0.158 2.4479 0.137 s 0.0039 0.0033 RSD% 0.3323 0.1331 47 Rio Claro, 2016 Table 12: Corrected lead isotope ratios of Goiânia samples from dry season (August 2014) with their relative standard deviations (X = mean value; s = standard deviation; RSD% = relative standard deviation) Date 206Pb/207Pb RSD% 208Pb/207Pb RSD% 04/08/14 1.1608 0.243 2.4370 0.182 06/08/14 1.1839 0.054 2.4460 0.008 07/08/14 1.1689 0.078 2.4380 0.040 08/08/14 1.1688 0.075 2.4401 0.034 09/08/14 1.1749 0.080 2.4459 0.107 10/08/14 1.1779 0.044 2.4520 0.075 12/08/14 1.1729 0.117 2.4499 0.092 15/08/14 1.1760 0.099 2.4459 0.073 16/08/14 1.1779 0.200 2.4480 0.169 18/08/14 1.1749 0.060 2.4499 0.063 20/08/14 1.1839 0.154 2.4490 0.127 21/08/14 1.1729 0.112 2.4489 0.071 22/08/14 1.1789 0.163 2.4510 0.153 24/08/14 1.1739 0.205 2.4460 0.079 26/08/14 1.1749 0.079 2.4459 0.108 27/08/14 1.1789 0.064 2.4460 0.108 X 1.1750 0.114 2.4462 0.093 s 0.0057 0.0044 RSD% 0.4880 0.1808 Measured 208Pb/207Pb isotope ratios from April/May samples show a good precision with just one RSD value higher than 0.2 %, while there are six RSD values higher than 0.2 % at the 206Pb/207Pb isotope ratios. In contrast, August samples show a higher precision with just three RSD values higher than 0.2 % at the 206Pb/207Pb isotope ratios and no RSD value higher than 0.2 % at 208Pb/207Pb isotope ratios. The higher precision in dry season samples is explicable with the higher Pb concentrations in this sampling period. Pb–Pb diagrams were represented in previous studies with the aim of analysing, interpreting and comparing data. In this work, it has been used to plot the difference of 208Pb/207Pb and the 206Pb/207Pb ratios between both sampling periods (Fig. 17). 48 Rio Claro, 2016 Figure 17: Pb–Pb diagram according to the Pb isotope ratios for samples taken in Goiânia The 206Pb/207Pb isotope ratio varied in the rainy season samples from 1.1620 to 1.1761 and the variation of the 208Pb/207Pb isotope ratios accounted from 2.4460 to 2.4522. The 206Pb/207Pb isotope ratios from dry season samples differed from 1.1608 to 1.1839 and 208Pb/207Pb isotope ratios from 2.4370 to 2.4520. Arithmetic averages of 206Pb/207Pb isotope ratios of August samples are slightly higher (1.1750) than in April/May (average of 1.1718), while the 208Pb/207Pb ratios show just a marginally divergence (2.4462 in dry season and 2.4479 in rainy season). The relative standard deviations of all ratios demonstrate a higher variation of the 206Pb/207Pb isotope ratios in dry season than in rainy season, while the 208Pb/207Pb isotope ratios show very close relative standard deviations in both months. Kruskal-Wallis tests show, that 206Pb/207Pb isotope ratios from rainy season vary significantly from dry season (H= 4.3063, P= 0.0382), whereas 208Pb/207Pb isotope ratios between both sample periods show no significant statistical disparities (H= 1.3681, P= 0.2421). Thus, while 208Pb/207Pb isotope ratios in rainy and dry season demonstrate quite similar values, 206Pb/207Pb isotope ratios of dry season are in average higher than 206Pb/207Pb isotope ratios in rainy season. This observation supposes an enrichment of 206Pb isotopes in dry season. In order to assess possible emission sources, measured 208Pb/207Pb and 1.10 1.12 1.14 1.16 1.18 1.20 1.22 1.24 2.40 2.42 2.44 2.46 2.48 2.50 2 0 6 P b /2 0 7 P b 208Pb/207Pb April/ May 2014 August 2014 49 Rio Claro, 2016 206Pb/207Pb isotope ratios were plotted on a Pb-Pb diagram (Fig. 18) together with data from previous studies from Brazil, which contain Pb isotope ratios of aerosol samples from São Paulo (Bollhöfer and Rosman, 2000), gasoline, industrial emission and exhaust fume samples from São Paulo (Aily, 2001) and atmospheric deposit samples from Rio Grande (Mirlean et al., 2005). Figure 18: Pb–Pb diagram according to the Pb isotope ratios for samples taken in Goiânia compared with Pb isotope ratios from recent studies from Brazil Pb isotope ratios in aerosol samples from São Paulo and atmospheric deposit samples from Rio Grande are highly different from Pb isotope ratios in Goiânia samples. Both, Pb isotope ratios from rainy season and dry season samples of Goiania, are slightly located on the line of gasoline Pb isotope ratios from São Paulo. Thus, gasoline can be considered as a possible influence on Pb isotope ratios of Goiânia, even though Pb isotope ratios from gasoline samples are distributed in a broader range with a low quantity of samples. Pb isotope ratio data of exhaust fumes from São Paulo is also distributed near isotope ratios from Goiânia samples, supporting the assumption that Pb isotope ratios of Goiânia may could be characterised by traffic emissions. Industrial emissions show a high variability grade in its isotope ratios, but also could be a factor of influence on Pb isotope ratios from Goiânia, especially for August 1.06 1.08 1.10 1.12 1.14 1.16 1.18 1.20 1.22 1.24 1.26 2.38 2.40 2.42 2.44 2.46 2.48 2.50 2 0 6 P b /2 0 7 P b 208Pb/207Pb April/ May 2014 August 2014 Aerosols (São Paulo) Atm. Deposit (Rio Grande) Gasoline (São Paulo) Exhaust fumes (São Paulo) Industrial Emission (São Paulo) Linear (Gasoline (São Paulo)) Linear (Exhaust fumes (São Paulo)) 50 Rio Claro, 2016 1.85 1.90 1.95 2.00 2.05 2.10 2.15 2.20 2.25 2.30 0.82 0.83 0.84 0.85 0.86 0.87 0.88 0.89 2 0 8 P b /2 0 6 P b 207Pb/206Pb April/ May 2014 August 2014 Figure 19: 208Pb/206Pb-207Pb/206Pb diagram according to the Pb isotope ratios for samples taken in Goiânia samples, which show higher 206Pb/207Pb isotope ratios and thus, are closer located to the Pb isotope data of industrial emissions. In order to investigate contribution from the decay of Thorium to the Pb in PM in Goiânia, ratios of 208Pb/206Pb and 207Pb/206Pb were calculated and plotted against each other in a diagram (Fig. 19). 208Pb/206Pb and 207Pb/206Pb isotope ratios show significant differences between rainy and dry season. 207Pb/206Pb isotope ratios from rainy season are higher than in dry season. The significant variance of 207Pb/206Pb ratios between both sampling periods is proved by the consulted Kruskal-Wallis-Test (H= 4.3063, P= 0.0382). 208Pb/206Pb ratios demonstrate a highly significant variance between rainy and dry season (H= 8.8716, P= 0.0029) with distinct higher 208Pb/206Pb ratios in rainy season, indicating an enrichment of 208Pb isotopes in this period derived from Th-rich environments. In Fig. 20, 208Pb/206Pb and 207Pb/206Pb isotope ratios from PM in Goiânia were adopted in a Pb-Pb-diagram together with data 208Pb/206Pb and 207Pb/206Pb isotope ratios from Pb sources taken from the study of Aily (2001). Results are demonstrating the location of Pb isotope ratios from PM in Goiânia in the conformity of isotopic signature of gasoline, exhaust fumes and street dust, indicating a high contribution of traffic 51 Rio Claro, 2016 1.90 1.95 2.00 2.05 2.10 2.15 2.20 2.25 0.77 0.79 0.81 0.83 0.85 0.87 0.89 0.91 0.93 0.95 2 0 8 P b /2 0 6 P b 207Pb/206Pb April/ May 2014 August 2014 Exhaust fumes (São Paulo) Feldspat (São Paulo region) Gasoline (São Paul