A research report submitted to the Faculty of Engineering and the Built Environment, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Master of Science in Engineering. August, 2017 / The foremost aim of potable water treatment is to produce water that does not pose a health risk when consumed and/or otherwise used. Nevertheless, research has established that the quality of treated water deteriorates during distribution. The nature and extent of this deterioration varies from system to system and from time to time. The aim of this research study was to monitor the parameters that are known to significantly affect biostability and biofilm formation potential in drinking water distribution systems. Biweekly water samples were collected from thirteen sites, across a section of Johannesburg Water’s network, between September 2015 and August 2016. All samples were assayed for a suite of fifteen water quality parameters using standard methods. Heightened temperature, dearth of chlorine residuals, availability of biodegradable dissolved organic carbon (BDOC), and advanced water age all engendered the loss of biostability (instability). Biostability controlling parameters varied seasonally and spatially. Samples collected during spring and summer, in general, were most likely to be characterized by instability than samples collected during winter and autumn. Samples collected from sites RW80, RW81, RW82, RW83, RW104 and RW253 were more prone to instability compared to samples from other sites. From the results, it is clear that chlorine residuals ought to be kept above 0.2 mg/l, and, BDOC below 0.3mg/l to prevent the loss of heterotrophic stability in distributed water. BDOC concentrations can be decreased by, flushing the pipes, cleaning reservoirs regularly and by further treating feed water before distributing. Booster disinfection can be relied upon to ensure that chlorine residuals are maintained throughout the network. Apart from potential health risks, biological instability and biofilm growth can result in non-compliance with regulations. / MT2018

Este trabalho apresenta a aplicação da Cromatografia Líquida de Ultra Performance acoplada à espectrometria de massas (UPLC-MS/MS) como uma técnica analítica em potencial para a determinação de contaminantes orgânicos em amostras de água para consumo humano. A técnica foi aplicada no desenvolvimento de um método analítico para a determinação de 28 praguicidas listados na Portaria 2914/11 como indicadores da qualidade de águas para o consumo humano. O método analítico apresentou limite de quantificação (LQ) (entre 0,2 e 50 g L-1, considerando todas as substâncias investigadas) capaz de monitorar as substâncias de interesse em concentrações inferiores aos limites máximos permitidos pela legislação, com taxas de recuperação superiores a 80% e coeficiente de variação (CV) menor do que 15% no nível de µg L-1. A filtração da amostra, empregando filtros de seringa de 0,22µm de diâmetro de poro, foi empregada como etapa de preparo de amostra, não sendo necessário o emprego de métodos de extração e concentração da amostra previamente à análise. O método analítico proposto foi aplicado em duas amostras de água potável coletadas na cidade de Paulínia-SP; entretanto, os praguicidas investigados não foram detectados em ambas as amostras analisadas. Adições de padrão foram realizadas sobre as amostras no nível de concentração do Valor Máximo Permitido (VMP) e três vezes o VMP pela Portaria 2914/11. Valores de recuperação entre 80 e 120% foram obtidos, demostrando a aplicabilidade e exatidão do método proposto. / This work describes the application of Ultra Performance Liquid Chromatography tandem mass spectrometry (UPLC-MS/MS) for determination of organic contaminants in drinking water. The tecnique was used for the development of an analytical method for the determination of 28 pesticides, listed in the Brazilian Administrative Rule 2914/11, these compounds are considered as quality indicators for drinking waters. The limit of quantification (LQ) of the analytical method was between 0,1 e 5 µg L-1 for all target substances, being able to quantify the target substances at concentrations lower than the maximum levels of contamination (MLC) established by the Rule, with accuracy above 80% and relative standard deviation below 15%, at g L-1 level. The filtration of the sample, using 0.22 µm syringe filters, was applied as sample preparation step. The analytical method based on UPLCMS/MS was used to analyze two drinking water samples collected in Paulínia-SP city. The target pesticides were not detected in any samples. The two samples were then enriched with analytical standards of the pesticides at the MLC and 3 times the MLC levels and analyzed by the proposed method. Recoveries ranged from 80 to 120%, showing the applicability and efficiency of the proposed analytical method.

As cianotoxinas, potencialmente prejudiciais à saúde dos homens e animais, são toxinas produzidas pelas cianobactérias e que podem estar presentes em reservatórios utilizados para captação de água para abastecimento, sobretudo em função do agravamento das condições sanitárias das bacias hidrográficas. Diante disso, torna-se fundamental que a água esteja adequada ao consumo humano em sua distribuição. Uma das formas de tratamento para remoção das cianotoxinas que vem sendo estudada é por meio do uso de carvão ativado granular (CAG), cuja eficiência depende de suas características intrínsecas e das condições operacionais. Outra possibilidade para remoção de tais toxinas se dá por meio da utilização de microrganismos capazes de promover sua biodegradação. A presente pesquisa analisou o potencial de remoção de microcistina por meio de adsorção por três diferentes CAGs comerciais (materiais de origem: casca de coco, mineral e osso) e por meio da degradação biológica por dois gêneros de bactérias. Para isso, foi utilizado um extrato de toxina produzido por meio do cultivo de Microcystis aeruginosa (cepa BB005) em meio WC, ao longo de 30 dias, com fotoperíodo de 12 h. Os carvões foram caracterizados de acordo com a massa específica aparente, umidade, pH, teor de cinzas, número de iodo, índice de azul de metileno e área superficial específica. Em seguida, foram realizados ensaios de adsorção com a microcistina (concentração inicial de 100 µg.L-1) para construção de isotermas com tempo de equilíbrio de 1 h. As análises de toxina foram realizadas pelo método ELISA e os dados foram ajustados aos modelos matemáticos de Langmuir e de Freundlich. Os melhores resultados de adsorção foram obtidos a partir do carvão de origem mineral (99% de remoção), que apresentou valores maiores para número de iodo (710 mgI2.g-1), índice de azul de metileno (169 mL.g-1), área superficial específica (911 m2.g-1) e porcentagem de microporos (70%), com melhor ajuste da isoterma obtido pelo modelo de Freundlich (R2 = 0,88). Os ensaios de biodegradação da microcistina com os gêneros Sphingomonas sp. e Brevundimonas sp foram realizados com a toxina esterilizada e não esterilizada (concentração inicial de 75 µg.L-1). Os resultados demonstraram que tais bactérias não foram capazes de degradar a toxina esterilizada. Entretanto, foi observada degradação nos ensaios em que a toxina não estava esterilizada (porcentagem de remoção de 98%), indicando a presença de algum microrganismo e/ou enzima atuante nesse processo, uma vez que o cultivo da cepa para obtenção do extrato não foi realizado sob condições de assepsia. Recomenda-se, para futuras pesquisas, a identificação do responsável pela biodegradação da microcistina, bem como a aplicação do carvão mineral em maior escala (filtro de leito fixo). Deste modo, será possível avaliar a formação de biofilme no leito de carvão, e comparar o desempenho da adsorção e da biodegradação na remoção do poluente-alvo em escala mais próxima à real. / The cyanotoxins, which are potentially harmful to the health of humans and animals, are toxins produced by cyanobacteria. They can be found in reservoirs used for water supply, especially due to the degradation of sanitary conditions within the watersheds. Therefore, it is essential to assure potable water for human consumption in its distribution. The use of granular activated carbon (GAC) has been studied for advanced water treatment and cyanotoxins\' removal. The efficiency of such technology depends on activated carbon intrinsic characteristics and operating conditions. Removing such toxins through biodegration by microrganisms is another possibility. The present study was perfomed in laboratory conditions and analyzed microcystin removal by adsorption by three commercial GACs (source materials: coconut shell, mineral and bone) and through the biological degradation by two genera of bacteria. An extract of toxin was produced by the cultivation of Microcystis aeruginosa (strain BB005), in medium WC throughout 30 days, with a 12 h photoperiod. The activated carbon samples were characterized regarding the apparent density, moisture, pH, ash content, iodine number, methylene blue index and specific surface area. Thereafter, adsorption experiments were conducted with microcystin (initial concentration of 100 µg.L-1) for estimating isotherms considering an equilibrium time of 1 h. The toxin analyses were performed by the ELISA (Enzyme-Linked Immunosorbent Assay) and the data were adjusted to the mathematical models of Langmuir and Freundlich. The best adsorption results were obtained with the mineral carbon (removal percentages of 98%), which also showed the highest values for the iodine number (710 mgI2.g-1), methylene blue index (169 mL.g-1), specific surface area (911 m2.g-1) and percentage of micropores (70%), with best adjustment of the isotherm through the Freundlich model (R2 = 0.88). The biodegradation tests of microcystin with Sphingomonas sp. and Brevundimonas sp. were carried out with the sterile and non-sterile toxin (initial concentration of 75 µg.L-1). The results showed that these bacteria were not able to promote degradation of the toxin when it was sterilized. However, the degradation was observed in those tests in which the toxin was not sterilized (reduction up to 98%), suggesting the presence of a microrganism and/or an enzyme responsible for this process, since the cultivation of the strain for obtaining the extract was not performed under sterile conditions. As for future research, it is recommendedthe identification of the factor responsible for the biodegradation of the microcystin, as well as the study of the use of the activated carbon from mineral source in a larger scale (fixed-bed filter). This would help to reach a scale closer to the real water treatment plants and evaluate the biofilm formation in the carbon bed, allowing the comparison of the performance of both adsorption and biodegradation processes in the removal of the target pollutant.

A intoxicação por chumbo tem conseqüências devastadoras para o ser humano, principalmente para crianças. Em Ribeirão Preto foram detectadas crianças com níveis de chumbo no dente tão elevados quanto de crianças com conhecida contaminação na cidade de Bauru (SP). Diante disso, este trabalho teve como objetivos investigar fatores físicos e químicos que possam elevar a concentração de chumbo na água de consumo, e avaliar a possível correlação entre a contaminação por chumbo nas crianças com a água de suas residências. O estudo da capacidade de corrosão da água utilizando um pedaço de cano a base de chumbo mostrou que ao diminuir o pH de 6,40 para 5,94 a lixiviação de chumbo aumentou 3 vezes, enquanto o aumento de 6,40 para 7,06 diminuiu a lixiviação em 20%, após 8 horas de experimento. Não foi observada influência na capacidade de corrosão em testes de 48 horas quando aumentou-se a dureza da água. Foram coletadas amostras (1L) de água das 40 casas estudadas, sendo as primeiras alíquotas coletadas após pelo menos 6 horas de repouso nos encanamentos. Amostras de água recém tratada de 11 postos de abastecimento de Ribeirão Preto serviram como controle. A faixa de condutividade da água das casas foi bastante ampla (50,4 - 116,9 S cm-1), e nenhum dado foi estatisticamente diferente do grupo controle (70,6 ± 40,2 S cm-1). Cerca de 60% dos valores de pH das amostras das casas estavam menores que do controle (6,20 ± 0,25), provavelmente devido à hidrólise de metais vindos da corrosão de encanamentos. Dos parâmetros estudados, as variações mais significativas que ocorreram entre a água da torneira do registro e da cozinha foi o pH, que teve a tendência em aumentar, principalmente em casas que tinham caixa d\'água. As concentrações de cálcio nas casas variaram de 2,36 - 7,23 mg L-1, e magnésio de 0,77 - 2,62 mg L-1, estando portanto dentro da faixa de concentração do grupo controle (1,94 - 10,74 mg L-1 Ca2+ e 1,01 - 3,62 mg L-1 Mg2+). Nas casas amostradas as concentrações de cobre (de 0,82 a 458 g L-1) e chumbo (de <0,28 a 3,19 g L-1) não ultrapassaram os máximos recomendados pela portaria 518 do Ministério da Saúde. No entanto, verificou-se um aumento na concentração desses metais em relação ao grupo controle ([Pb] = 0,16 ± 0,10 g L-1 e [Cu] = 2,17 ± 3,19 g L-1), reforçando a idéia de que ocorre lixiviação no percurso dos postos de abastecimento às residências. Após o escoamento de 5 minutos entre alíquotas, as concentrações de metais foram significativamente menores. Os coeficientes de correlação linear entre a concentração de chumbo na água e no sangue ou no esmalte (1ª e 2ª biópsia) das crianças não foram estatisticamente significativos (rmáximo = 0,2324; p = 0,17231). Porém, considerando as concentrações de chumbo na água maiores que 0,5 g L-1 o coeficiente de correlação linear com o sangue passou de 0,1517 para 0,8373 (p = 0,00129; n = 11). A relevância dessa correlação só poderá ser estabelecida uma vez que haja um maior número de amostras. / Lead poisoning has devastating consequences for humans, especially for children. It has been found that a significant percentage of children from Ribeirão Preto had as much lead in enamel teeth as the children from Bauru (SP), a city with a well described contamination by lead. Thus, this study aimed to investigate physical and chemical factors that can increase the concentration of lead in drinking water, and evaluate the possible correlation between lead poisoning in children with the water from their homes. The drinking water capability of corrosion was studied using a piece of lead-based pipe. The results showed that lowering the pH from 6.40 to 5.94 increased the leaching of lead three times, while the increase from 6.40 to 7.06, decreased the leaching by 20% after 8 hours of experiment. Increases on the hardness of the water had no effect on the capability of corrosion in tests performed during 48 hours. Drinking water samples (1L) from 40 houses were collected after at least 6 hours of stagnation in the pipes. Samples (1L) of water from 40 homes were studied, being the first fractions collected after at least six hours of stagnation in the pipes. Samples from 11 water treatment plants from Ribeirão Preto were collected after treatment and were used as the control group. The water conductivity from the studied houses had a large variation (50.4 to 116.9 S cm-1), however, these values were within those for the control group (70.6 ± 40.2 S cm-1). The pH values for 60% of the drinking water samples were bellow 6.20 ± 0.25 found for the control group. This is probably due to the hydrolysis of metals from the pipes and parts that may occur during the corrosion processes. When comparing the results found for the water from the back yard and that from the kitchen, the pH values had the most significant variations. For most of the houses which had a water reservoir the pH increased after standing in the reservoir. Calcium concentrations in the drinking water ranged from 2.36 to 7.23 mg L-1, and magnesium from 0.77 to 2.62 mg L-1, which is close to the range of the control group (1.94 - 10.74 mg L-1 Ca2+ and 1.01 to 3.62 mg L-1 Mg2+). Concentrations of copper (0.82 to 458 g L-1) and lead (from <0.28 to 3.19 g L-1) did not exceed the maximum recommended by the Brazilian Ministry of Health. However, there was an increase in the concentration of these metals compared to the control group ([Pb] = 0.16 ± 0.10 g L-1 and [Cu] = 2.17 ± 3.19 g L- 1), reinforcing the idea that an important leaching can occur from the source to the houses. Metal concentrations were significantly lower after 5 minutes of flushing. There was not a significant linear correlation (rmaximum = 0.2324; p = 0,17231) between the concentration of lead in drinking water and in blood or tooth enamel (1st and 2nd biopsy) from children. However, considering lead concentrations in drinking water higher than 0.5 g L-1 the linear correlation coefficient with lead in blood went from 0.1517 to 0.8373 (p = 0,00129; n = 11). The relevance of this correlation can only be established after the analysis of a larger number of samples.

Resumo: O Bisfenol A (BFA) é um monômero formado por dois anéis fenólicos e empregado na produção de plástico policarbonato e resinas epóxi. O policarbonato é amplamente empregado na fabricação de vários produtos dos meios digitais (CD e DVD), recipientes reusáveis de alimentos e bebidas e muitos outros produtos. As resinas epóxi são usadas em circuitos elétricos, pinturas, adesivos, revestimentos protetores de latas de alimentos e bebidas e complexos dentários para obturações. Estudos revelaram que o BFA age nos organismos vivos com efeitos de desregulação endócrina, como um xenobiótico com ação estrogênica, provocando aumento da incidência de câncer de mama, queda da quantidade de esperma, diminuição da fertilidade, defeitos congênitos secundários à exposição fetal e outras alterações. O BFA chega ao ambiente aquático por inúmeras vias distintas como, por exemplo, vindo do lodo proveniente de esgoto, da descarga de esgotos domésticos e industriais, dos lixões e aterros sanitários não controlados. Neste trabalho, desenvolveu-se e validou-se um método para a determinação de BFA em águas superficiais e de abastecimento à população de Campo Grande - MS, utilizando-se extração em fase sólida EFS (cartuchos de fase polimérica Strata-X), derivação com N O-Bis-(trimetilsilil) trifluroacetamida (BSTFA) e detecção por cromatografia a gás acoplada à espectrometria de massas. A precisão e a exatidão do método foram avaliadas em de ensaios de recuperação. Recuperações variando de 91 a 113% com coeficientes de variação percentuais inferiores a 8,9% foram obtidos com limites de detecção e quantificação do método de 2,4 e 10 ng L-1, respectivamente. Também foi avaliada a conservação do BFA nos próprios cartuchos de EFS onde se observou que este analito pode ser armazenado por até 90 dias sem diferenças significativas entre o valor... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Bisphenol A (BPA) is a monomer consisting of two phenolic rings and employed in the production of polycarbonate plastic and epoxy resins. The polycarbonate is widely used in many manufactured products from digital media (CD and DVD), reusable containers food and beverages and many other products. Epoxy resins are used in electrical circuits, paints, adhesives, protective coatings for cans of food, drinks and complex dental fillings. Studies have shown that BPA acts in living organisms with the purpose of endocrine disrupting as a xenobiotic with estrogenic action, causing an increased incidence of breast cancer, decrease the amount sperm, fertility decreased, birth defects secondary to fetal exposure and other changes. BPA reaches the aquatic environment by several distinct pathways, for example, coming from the sludge sewage, discharge of domestic sewage and industrial waste dumps and the landfills. In this work, we developed and validated a method for the determination of BPA in surface water and supply population of Campo Grande - MS was developmente ande validated using solid phase extraction SPE (polymeric phase cartridges Strata-X), derivatized with N-O-Bis (trimethylsilyl) trifluroacetamide (BSTFA) and detection by gas chromatography coupled to mass spectrometry. The precision and accuracy of method were evaluated through recovery experiments. Recoveries ranging from 91-113% with coefficients of variation lower to 8.9% were obtained with limits of detection and quantification of 2.4 and 10 ng L-1, respectively. We also analyzed the estability of BPA in individual cartridges SPE where it was observed that the analyte can be stored for up to 90 days without significant differences between the determined and the value taken as true. The leaching of BPA from PVC pipes was investigated coming to conclusions that this is a source contamination of water by this important endocrine... (Complete abstract click electronic access below) / Orientador: Maria Lucia Ribeiro / Coorientador: Marcia Helena de Rizzo da Matta / Banca: Luciana Polese / Banca: Carolina Lourencetti / Banca: Raquel Fernandes Pupo Nogueira / Banca: Wilson de Figueiredo Jardim / Doutor

L'objectif de ce travail a consisté à (i) déterminer l'occurrence et (ii) à étudier le comportement de différents micro-organismes pathogènes pour l'Homme reconnus comme responsables d'épidémies d'origine hydrique dans des ressources (eaux de surface et eaux de nappes) et des eaux traitées de plusieurs collectivités situées le long de la Moselle. Trois agents microbiologiques aux propriétés morphologiques et intrinsèques très différentes ont été étudiés: des bactéries (C. jejuni et C. coli), des virus entériques (adénovirus humains dont les adénovirus entériques), des parasites (C. parvum, C. hominis, C. meleagridis). Après avoir mis en place différentes méthodologies permettant leur détection, leur quantification et la révélation de leur viabilité et/ou leur infectiosité, nous avons montré que l'eau de la Moselle était largement contaminée par ces pathogènes à des concentrations variables selon la période de l'année. Ainsi par exemple, il a été retrouvé des bactéries C. jejuni plus fréquemment en été que pendant la période hivernale alors que l'inverse a été observé pour les adénovirus humains. Les résultats d'occurrence au niveau des nappes alluviales étudiées ont montré que malgré une protection liée à la filtration naturelle de l'eau, elles pouvaient être contaminées par des campylobacters viables et des adénovirus entériques. Concernant les eaux traitées, plusieurs signatures génomiques liées à ces deux derniers agents (bactérien et viral) ont été révélées par des outils de biologie moléculaire sans qu'un risque en lien avec la viabilité et l'infectiosité puisse être établi. Dans ces eaux, les expérimentations d'inactivation ont clairement montré la grande sensibilité de ces deux agents au traitement chloré mais pas celle des parasites Cryptosporidium qui s'avèrent être les pathogènes les plus résistants quel que soit le milieu ou la température d'incubation / The aim of this study was (i) to determine the occurrence and (ii) to study the behavior of different microorganisms known to cause human waterborne outbreaks in raw waters (surface water and groundwater) and treated water of several reservoirs along the Moselle river. Three pathogens were studied: bacteria (C. jejuni and C. coli), enteric viruses (human adenoviruses which enteric adenoviruses), parasites (C. parvum, C. hominis, C. meleagridis). Because of these micro-organisms have different properties such as size, various methodologies were used for their detection, quantification and viability and/or infectivity. In the study of occurrence we showed that the Moselle river was heavily contaminated by these pathogens in varying concentrations depending on time of the year. For example it was found bacteria C. jejuni more frequently in summer than during winter while the reverse was observed for human adenoviruses. The results obtained in groundwater showed that despite a protection linked to the natural filtration of water, they could be contaminated with viable campylobacters and enteric adenoviruses. In the treated water, several genomic signatures associated with these two agents (bacterial and viral) were revealed by molecular biology tools without any risk related to the viability and infectivity can be established. In these waters, the inactivation experiments clearly showed the high sensitivity of these two agents to chlorination but not the parasite Cryptosporidium, which known to be the most resistant pathogens, whatever the type of water or incubation temperature

The persistent and toxic perfluoroalkyl acids (PFAAs) are ubiquitously present in the environment and reach humans predominantly via food and drinking water. The aim of the present study was to investigate the effect of low-grade (&lt;10 ng L-1 of single PFAAs) contaminated PFAAs drinking water on serum PFAA concentrations in a representative Swedish adolescent population, and to examine the influence of potential determinants on the variation of the PFAAs serum concentrations. This was done by using multivariate regression analysis on the possible determinants of blood serum PFAA concentrations in 479 Swedish adolescents, 10 to 21 years of age, who had left complete dietary and life style information in 2016-17 in the nation-wide food consumption survey Riksmaten Ungdom. Raw and drinking water samples (DW) from water treatment plants (WTPs) that delivered DW to participants schools were sampled in 2018, analysed for PFAAs, and used for assessing the participants DW PFAA exposure. Maternal education level and maternal birth country, consumption of fish, as well as age and sex were significantly associated with the participants PFAAs serum concentrations. DW concentrations as low as &lt;1 ng L for PFOA and PFHxS, &lt;0.45 ng L-1 for PFNA and &lt;4 ng L-1 for PFOS were significantly associated with increased adolescent serum concentrations of the PFAAs in question, which suggests that low-grade contaminated drinking water is an important exposure route for Swedish adolescents. For risk assessment purposes, it was investigated whether parts of the adolescent population exceeded the serum PFOS and PFOA concentrations corresponding to the current health-based reference intakes as assessed by the European Food Safety Authority (EFSA) and the U.S. Agency for Toxic Substances and Disease Registry (ATSDR). Around 1.7% and 2.7% of participants had PFOS serum concentrations exceeding serum levels used to derive the tolerable daily intake (TDI) (EFSA) and the minimum risk level (MRL) (ATSDR), respectively and a cause for concern was consequently identified. However, the high serum concentrations of participants exceeding the TDI and MRL serum concentrations belong to participants suspected to have been previously exposed to highly contaminated drinking water and not from consuming foods and beverages containing background concentrations of PFAAs.

Este trabalho teve por objetivos estudar a distribuição das concentrações de atividade dos isótopos de rádio de meias-vidas longas, 226Ra e 228Ra, em amostras de águas minerais coletadas no Balneário de Caxambu (MG) e no Balneário de Águas de Lindóia (SP). Foram observadas concentrações de atividade de 228Ra relativamente aumentadas em relação às de 226Ra nas águas minerais estudadas no Balneário de Caxambu. As concentrações elevadas de carbonatos e bicarbonatos destas águas podem resultar em um conseqüente aumento da solubilidade dos isótopos de Ra e pode ter um papel importante para a distribuição do 228Ra e para as reações de equilíbrio entre as fases sólida e líquida das soluções. No Balneário de Caxambu, as concentrações de atividade médias aritméticas variaram de 83 mBq L-1 a 3599 mBq L-1 e de 60 mBq L-1 a 4481 mBq L-1 para 226Ra e 228Ra, respectivamente. O maior valor de atividade de 226Ra foi observado nas águas da Fonte Venâncio, enquanto que a concentração máxima de 228Ra foi determinada na Fonte Ernestina Guedes. As razões de atividade 228Ra/ 226Ra variaram de 0,079 (Fonte Conde DEau e Princesa Isabel) a 4,2 (Fonte Mairink II). Em Águas de Lindóia as concentrações de atividade médias aritméticas variaram de 4,6 mBq L-1 a 41 mBq L-1 e de 30 mBq L-1 a 54 mBq L-1 para 226Ra e 228Ra, respectivamente. O maior valor de atividade de 226Ra foi observado nas águas da Fonte Engarrafada São Jorge, enquanto que a concentração máxima de 228Ra foi determinada na Fonte Santa Filomena (torneira pública 2). As razões de atividade 228Ra/ 226Ra variaram de 1,2 (Fonte Engarrafada São Jorge) a 9,1 (Fonte Engarrafada Jatobá 1). A exposição de indivíduos do público à radiação natural em virtude do consumo das águas minerais também foi avaliada neste trabalho. As doses efetivas foram estimadas utilizando-se um modelo dosimétrico conservativo, e integrando-se os resultados ao longo de 70 anos de vida de um indivíduo adulto, após a incorporação de 226Ra e 228Ra. Os resultados desta avaliação radiossanitária demonstraram que o nível de referência de dose efetiva comprometida recomendado pela Organização Mundial de Saúde de 0,1 mSv a-1 foi ultrapassado em quase todas as fontes de Caxambu, exceto para a Fonte Viotti, Fonte Dom Pedro e Fonte Dona Leopoldina. Em Águas de Lindóia e Lindóia as maiores doses efetivas comprometidas obtidas foram 8,4 10-3 mSv a-1 para o 226Ra (Fonte engarrafada São Jorge) e 2,6 10-2 mSv a-1 para o 228Ra (Fonte Santa Filomena torneira pública 2). Estes valores são menores que o limite recomendado pela Organização Mundial de Saúde. Doses efetivas comprometidas de até 7,4 x 10-1 mSv a-1 para o 226Ra (Fonte Venâncio, Caxambu) e 2,2 mSv a-1 para o 228Ra (Fonte Ernestina Guedes, Caxambu) foram estimadas no presente trabalho. No caso do 226Ra, este valor representa menos que 31% da dose efetiva média anual atribuída a radioatividade natural. Contudo, no caso do 228Ra, o limite anual de dose efetiva comprometida recomendado pela ICRP (considerando-se todas as fontes de exposição da população mundial à radiação) foi excedido não só na Fonte Ernestina Guedes, mas também na Fonte Beleza. / The aim of this work was to study the long-lived Ra isotopes, 226Ra and 228Ra, natural distribution in mineral and spring waters from Caxambu (MG) and Águas de Lindóia (SP) waterparks. In Caxambu mineral waters it was observed 228Ra acitivity concentrations slightly higher than those of 226Ra. The elevated content of carbonates and bicarbonates of these waters can result in an increased solubility of the both Ra isotopes and may play an important role for the fate of 228Ra and its equilibrium distribution between solid and liquid phases. In Caxambu Thermal Spa, arithmetic mean activities ranged from 83 mBq L-1 to 3599 mBq L-1 and from 60 mBq L-1 to 4481 mBq L-1 for 226Ra and 228Ra, respectively. The highest 226Ra activity was found in Venancio Spring, while the maximum 228Ra activity value was determined in Ernestina Guedes. 228Ra/ 226Ra activity ratios varied from 0.079 (Conde DEau and Princesa Isabel Spring) to 4.2 (Mairink II Spring). In Águas de Lindóia, arithmetic mean activities ranged from 4.6 mBq L-1 to 41 mBq L-1 and from 30 mBq L-1 to 54 mBq L-1 for 226Ra and 228Ra, respectively. The maximum 226Ra activity concentration was found in the bottled mineral water São Jorge, while the higher 228Ra activity concentration was determined in Santa Filomena Spring (public station 2). 228Ra/ 226Ra activity ratios varied from 1.2 (bottled mineral water São Jorge) to 9.1 (bottled mineral water Jatobá 1). This work also performed the dose assessment due to the ingestion of 226Ra and 228Ra in Caxambu and Águas de Lindóia mineral and spring waters. The committed effective doses were estimated by using a conservative dosimetric model and taking into account the results over a lifetime (70 years) following intake of both long-lived Ra isotopes. The results from this radiological evaluation showed that the guindance committed effective dose level of 0.1 mSv y-1 recommended by World Health Organization was exceeded in almost all samples studied in Caxambu, except for Viotti Spring, Dom Pedro Spring and Dona Leopoldina Spring. In Águas de Lindóia and Lindóia, the highest committed effective doses estimated were 8.4 10-3 mSv y-1 for the intake of 226Ra (bottled mineral water São Jorge) and 2.6 10-2 mSv y-1 for 226Ra (Santa Filomena spring station 2). Both values are below the limit recommended by World Health Organization. The maximum commited effective doses per year obtained in this work were 7.4 10-1 mSv y-1 for 226Ra (Venancio Spring) and 2.2 mSv y-1 for 228Ra (Ernestina Guedes Spring). Considering the case of 226Ra intake via mineral water, the dose obtained represents less than 31 % of the average effective dose attributable annually from natural background radiation. However, in the case of 228Ra the annual committed effective dose limit recommended by ICRP (considering all sources of radiation exposure for the world population) was exceeded not only in Ernestina Guedes Spring, but also in Beleza Spring.

Une étude a été menée sur la faisabilité technique du traitement de potabilisation des eaux de forage contaminées à l’arsenic par procédé membranaire. Pour ce faire, l’efficacité de deux membranes commerciales (NF270 et TW30) vis-à-vis de la rétention des espèces d’arsenic, arsénite [As (III)] et arséniate [As (V)], a été d’abord comparée sous différentes conditions expérimentales. Ensuite, la contribution des différents mécanismes intervenant dans l’élimination des espèces d’arsenic a été identifiée. Enfin, la membrane la plus performante a été utilisée pour le traitement d’une eau souterraine naturellement riche en arsenic. Il ressort de cette étude que la pression opératoire et le pH sont les paramètres importants sur lesquels il faut agir pour améliorer l’efficacité du système membranaire pour le traitement de l’arsenic. L’As(V) est facilement éliminé par rapport à l’As (III) aussi bien avec la membrane de nanofiltration que d’osmose inverse. Toutefois, la membrane d’osmose inverse (TW30) permet une meilleure élimination des différentes espèces d’arsenic que la membrane NF270. Cela est dû au fait que la membrane TW30 est une membrane dense. L’application de cette membrane pour la production d’eau potable à partir d’une eau de forage contenant 413,74µg/L d’arsenic total dont 405,63µg/L d’As (V) et 8,11 µg/L d’As (III) a donné une concentration d’arsenic total dans l’eau traitée de 9,99 µg/L soit un taux de rétention de 97,6% et un volume d’eau traité correspondant à 70% du volume d’eau brute. Il est donc possible de mettre en place un procédé de traitement de ces eaux de forage à partir de la membrane TW30 afin de produire de l’eau potable pour les populations du nord du Burkina Faso. La performance de ce procédé peut être encore améliorée en oxydant l’As (III) en As (V) avant filtration. / In order to produce potable water for human consumption, a study was carried out on the technical feasibility of arsenic-rich borehole water treatment by membrane process. Firstly the efficiency of two commercial membranes (NF2Une étude a été menée sur la faisabilité technique du traitement de potabilisation des eaux de forage contaminées à l’arsenic par procédé membranaire. Pour ce faire, l’efficacité de deux membranes commerciales (NF270 et TW30) vis-à-vis de la rétention des espèces d’arsenic, arsénite [As (III)] et arséniate [As (V)], a été d’abord comparée sous différentes conditions expérimentales. Ensuite, la contribution des différents mécanismes intervenant dans l’élimination des espèces d’arsenic a été identifiée. Enfin, la membrane la plus performante a été utilisée pour le traitement d’une eau souterraine naturellement riche en arsenic. Il ressort de cette étude que la pression opératoire et le pH sont les paramètres importants sur lesquels il faut agir pour améliorer l’efficacité du système membranaire pour le traitement de l’arsenic. L’As(V) est facilement éliminé par rapport à l’As (III) aussi bien avec la membrane de nanofiltration que d’osmose inverse. Toutefois, la membrane d’osmose inverse (TW30) permet une meilleure élimination des différentes espèces d’arsenic que la membrane NF270. Cela est dû au fait que la membrane TW30 est une membrane dense. L’application de cette membrane pour la production d’eau potable à partir d’une eau de forage contenant 413,74µg/L d’arsenic total dont 405,63µg/L d’As (V) et 8,11 µg/L d’As (III) a donné une concentration d’arsenic total dans l’eau traitée de 9,99 µg/L soit un taux de rétention de 97,6% et un volume d’eau traité correspondant à 70% du volume d’eau brute. Il est donc possible de mettre en place un procédé de traitement de ces eaux de forage à partir de la membrane TW30 afin de produire de l’eau potable pour les populations du nord du Burkina Faso. La performance de ce procédé peut être encore améliorée en oxydant l’As (III) en As (V) avant filtration.70 and TW30) for different arsenic species removal, arsenite [As (III)] and arsenate [As (V)], was compared under various experimental conditions. Afterward, the contribution of mechanisms in As species removal was determined. Finally, the most efficient membrane was used for As-rich groundwater treatment. Results show that, applied pressure and feed pH are the important parameters which must be considered if As removal wants to be improved. As (V) is much easily removed than As (III) both NF270 and TW30 membranes. Nevertheless, TW30 membrane has a better removal of As species than NF270 membrane. It is due to the fact that TW30 membrane is a compact membrane. The application of this membrane for potable water production from naturally contaminated groundwater containing 413.74 µg/L of total arsenic (405.63 µg/L of As (V) and 8.11 µg/L of As (III)) gave a treated water of 9.99 µg/L (97.6% of rejection) of total As concentration and water produced volume equivalent to 70% of raw water volume. Therefore, it is possible to implement a treatment process with TW30 membrane to remove arsenic from naturally contaminated groundwater coming from the northern part of Burkina Faso. The efficiency of this process can be improved again by oxidizing As (III) in As (V) before filtration.

The worldwide natural occurrence of high levels of arsenic (As) in groundwater and its deleterious effects on human health have inspired a great amount of related research in public health and geosciences internationally. With >100 million people in South and Southeast Asia exposed to >10 µg/L As in shallow groundwater that they use for drinking, the installation of deeper, low-As wells has emerged as a major strategy for lowering the exposure. As the magnitude of deep pumping continues to increase, this work focuses on the geochemical and hydrologic questions surrounding the vulnerability and sustainability of low-As aquifers in Bangladesh, the country most affected by As crisis. In an effort to better understand the residence time of groundwater in low-As aquifers at depth, radiocarbon (14C) and 13C in dissolved inorganic carbon, tritium (3H), stable isotopes of hydrogen (2H) and oxygen (18O), and noble gas concentrations were measured across a ~25 km2 area of Araihazar, ~30 km east of Dhaka. Groundwater from >120 m depth is shown to be ~10,000 years old and its isotopic signatures indicate that recharge occurred at the time of changing climate from the late Pleistocene to early Holocene, with little recharge occurring since. In contrast, the intermediate depth low-As aquifers (<120 m) have a heterogeneous distribution of groundwater chemistry and ages, and contain groundwater recharged <60 years ago in certain locations. In one such area surrounding a small village, the effects that subsurface clay layer distribution has on recharge patterns and redox status of the intermediate aquifer was investigated. The relevant hydrogeologic and geochemical processes that led to documented failures of a community well at the site were assessed using a combination of solid and water phase geochemistry with tritium-helium (3H/3He) dating, hydraulic head monitoring, and pumping tests. Organic matter seeping from a compressible clay layer, which is subject to a pumping-induced, downward hydraulic gradient, reduces iron oxides and helps release As in the grey, upper part of the intermediate aquifer. No recent recharge was detected by 3H measurements in the upper, grey sand layer, however a layer of orange sand beneath it contains groundwater that was recharged 10-60 years ago. This groundwater laterally bypasses the confining clay layer to recharge the middle of the aquifer and contains dissolved As levels of <10 µg/L. In this particular case, the pore water that leaches from clay layers contributes to As contamination, whereas the lateral recharge with shallow groundwater coincides with the low-As depth. Thus, clay layers may not always protect the low-As aquifers from As contamination, even if they can block direct vertical recharge with shallow groundwater enriched in As and organics. Finally, the adsorption of As to aquifer sediments, as a natural mechanism of the low-As aquifer defense against contamination, was assessed in the field via a column study. The column experiments were conducted by pumping shallow, high-As groundwater through freshly collected sediment cores to quantify the retardation of As transport through the aquifer. This study demonstrated an elegant method of assessing contaminant transport under nearly in situ conditions that resulted in sorption estimates similar to those made by field studies using more challenging methods or located at hard-to-find sites with convenient flow patterns. My work, therefore, contributed to a better understanding of low-As aquifers in Bangladesh from the perspectives of both the groundwater flow and water-sediment interactions on various scales, and it integrated methods that can be employed elsewhere to characterize aquifers and study contaminant transport.