Hydrogen is considered one of the best alternatives to fossil-fuels for the transportation sector because hydrogen can be burnt cleanly and efficiently in a fuel cell to drive an electric motor. However, due to the low density of H2 at ambient conditions the conventional H2 storage technologies (cryogenic liquid and compressed gas) cannot achieve energy densities comparable to to gasoline and diesel. A second energy storage challenge onboard electric fuel cell vehicles is fuel cell power management at peak current loads, which requires an auxiliary power source like a battery or supercapacitor. The development of efficient onboard energy storage systems for H2 and auxiliary power is critical to realisation of a hydrogen economy. Mesoporous carbons were investigated as H2 storage materials in composites with magnesium hydride (MgH2),and as electrode materials for electrochemical double-later capacitors. The mesoporous carbons were prepared by two methods: (1) from porous silica and alumina templates, and (2) by chemical activation of a waste carbon source (waste coffee grounds). The experimental approach targeted reducing the cost of mesoporous carbon preparation by using a cheaper template, where the cost of alumina template was one-fifth the cost of the silica template (at the laboratory scale), or by using a waste material as a carbon source. The alumina template was found to be suitable to produce a mesoporous carbon with an average pore size of 4.8 nm. Chemical activation of coffee grounds with ZnCl2 produced activated carbons with BET surface areas up to 1280 m2/g. Mesopore volume increased with ZnCl2 impregnation ratio, with mesopore size distributions in the range 2 - 20 nm. The theoretical H2 capacity of MgH2 is 7.6 % but MgH2 application in fuel cell vehicles is limited by slow hydrogenation kinetics and high temperatures (> 573 K) for H2 release. Magnesium was impregnated on activated carbon fibres (ACF) and mesoporous carbon (prepared from silica and alumina templates) to improve H2 storage kinetics and thermodynamics by reducing the magnesium hydride particle size. Thermal gravimetric analysis (TGA) and temperature programmed desorption (TPD) studies showed that thermal decomposition of MgCl2 supported on ACF at 1173 K in N2 and H2 can produce a Mg-ACF composite. At 573 K and 2 MPa H2 pressure a Mg-ACF composite, containing 11.2 %wt Mg, showed improved H2 adsorption kinetics compared to bulk Mg powder, but the total capacity of the Mg-ACF composite was only 0.4 % wt H2. To achieve a target of 6 %wt for onboard H2 storage higher Mg loadings are required. Attempts to impregnate Mg in mesoporous carbon via the MgCl2 thermal decomposition process highlighted the difficulties of avoiding MgO formation, and show that MgH2 loaded carbon is unlikely to be a practical high density onboard H2 storage technology. Activated carbons from waste coffee grounds (CGCs) were used as electrode materials in electrochemical double-layer capacitors. The specific capacitance of CGCs was as high as 368 F/g in 1 mol/L H2SO4, with good capacitance retention at fast charge rates and stable cycling performance. The good electrochemical performance of CGCs is attributed to a porous structure featuring both micropores 0.5 - 1.0 nm wide, which are effective for double-layer formation, and small mesopores, which facilitate electrolyte transport at fast charge rates. The capacitance of CGCs is enhanced by pseudo-Faradaic reactions involving nitrogen and oxygen functional groups. At fast charge-discharge rates the CGCs had higher energy density and better stability than a commercial benchmark activated carbon (Maxsorb). The ZnCl2 activation process can be optimised to develop mesopores for improved capacitance at fast charge rates and capacitance in organic electrolytes. In 1 mol/L tetra ethyl ammonium tetrafluoroborate (TEABF4) / acetonitrile the CGC with the most mesopores, which was prepared with a ZnCl2 to coffee ratio of 5:1, has the highest capacitance at high power density. CGCs with greater mesopore content retained higher specific capacitance at fast charge-discharge rates as the mesopores acts as channels or reservoirs for electrolyte transport. An improved model for evaluation of contributions to capacitance from micropore surfaces and mesopore surfaces is proposed. From this model the double-layer capacitance of mesopores surface area was found to be about 14 μF/cm2 and did not change considerably with increasing current load. The contribution of micropores to capacitance is dependent on the accessibility of ions to the micropores, and this accessibility is proportional to the mesopore surface area. An exponential function was found to describe the contribution of mesopores and micropore surfaces to capacitance. The effective double-layer capacitance of the micropore surface area drops at fast charge-discharge rates as a result of restricted ion transport, and this result highlights the importance of mesopores to retain energy density for high power supercapacitor applications.

This thesis deals with diffusion and flow of sub-critical hydrocarbons in activated carbon by using a differential permeation method. The hydrocarbons are selected according to the effect on environmental concerns and their unique characteristics such as polarity and affinity towards activated carbon. Although it has been known that transport processes in activated carbon consist of Knudsen diffusion, gaseous viscous flow, adsorbed phase diffusion (so called, surface diffusion) and condensate flow, there have been no rigorous models to describe the transport processes in activated carbon with a full range of pressures. In particular among the four processes, the mechanism of adsorbed phase diffusion in activated carbon is still far from complete understanding. Also due to the dispersion interactions between adsorbing molecules and the solid surface, one would expect that Knudsen diffusion is influenced by the dispersive forces. From intensive experimental observations with a great care over a full range of pressures, conventional methods (for example, direct estimation from inert gas experiments) to determine adsorbed phase diffusion are found to be inadequate for strongly adsorbing vapors in activated carbon. By incorporating the effect of adsorbate-adsorbent interactions into Knudsen diffusivity, the general behavior of adsorbed phase diffusion in terms of pressure (or surface loading) can be obtained, showing a significant role in transport at low pressures. For non-polar hydrocarbons such as benzene, carbon tetrachloride and n-hexane, a mathematical model, which accounts for the effects of adsorbate-adsorbent interactions and pore size distribution, is formulated and validated, resulting in a good agreement with experimental data. Moreover, the adsorption and dynamic behaviors of alcohol molecules (which are polar compounds) are investigated with an aim to compare their behaviors against those of non-polar compounds.

This thesis deals with diffusion and flow of sub-critical hydrocarbons in activated carbon by using a differential permeation method. The hydrocarbons are selected according to the effect on environmental concerns and their unique characteristics such as polarity and affinity towards activated carbon. Although it has been known that transport processes in activated carbon consist of Knudsen diffusion, gaseous viscous flow, adsorbed phase diffusion (so called, surface diffusion) and condensate flow, there have been no rigorous models to describe the transport processes in activated carbon with a full range of pressures. In particular among the four processes, the mechanism of adsorbed phase diffusion in activated carbon is still far from complete understanding. Also due to the dispersion interactions between adsorbing molecules and the solid surface, one would expect that Knudsen diffusion is influenced by the dispersive forces. From intensive experimental observations with a great care over a full range of pressures, conventional methods (for example, direct estimation from inert gas experiments) to determine adsorbed phase diffusion are found to be inadequate for strongly adsorbing vapors in activated carbon. By incorporating the effect of adsorbate-adsorbent interactions into Knudsen diffusivity, the general behavior of adsorbed phase diffusion in terms of pressure (or surface loading) can be obtained, showing a significant role in transport at low pressures. For non-polar hydrocarbons such as benzene, carbon tetrachloride and n-hexane, a mathematical model, which accounts for the effects of adsorbate-adsorbent interactions and pore size distribution, is formulated and validated, resulting in a good agreement with experimental data. Moreover, the adsorption and dynamic behaviors of alcohol molecules (which are polar compounds) are investigated with an aim to compare their behaviors against those of non-polar compounds.

This thesis deals with diffusion and flow of sub-critical hydrocarbons in activated carbon by using a differential permeation method. The hydrocarbons are selected according to the effect on environmental concerns and their unique characteristics such as polarity and affinity towards activated carbon. Although it has been known that transport processes in activated carbon consist of Knudsen diffusion, gaseous viscous flow, adsorbed phase diffusion (so called, surface diffusion) and condensate flow, there have been no rigorous models to describe the transport processes in activated carbon with a full range of pressures. In particular among the four processes, the mechanism of adsorbed phase diffusion in activated carbon is still far from complete understanding. Also due to the dispersion interactions between adsorbing molecules and the solid surface, one would expect that Knudsen diffusion is influenced by the dispersive forces. From intensive experimental observations with a great care over a full range of pressures, conventional methods (for example, direct estimation from inert gas experiments) to determine adsorbed phase diffusion are found to be inadequate for strongly adsorbing vapors in activated carbon. By incorporating the effect of adsorbate-adsorbent interactions into Knudsen diffusivity, the general behavior of adsorbed phase diffusion in terms of pressure (or surface loading) can be obtained, showing a significant role in transport at low pressures. For non-polar hydrocarbons such as benzene, carbon tetrachloride and n-hexane, a mathematical model, which accounts for the effects of adsorbate-adsorbent interactions and pore size distribution, is formulated and validated, resulting in a good agreement with experimental data. Moreover, the adsorption and dynamic behaviors of alcohol molecules (which are polar compounds) are investigated with an aim to compare their behaviors against those of non-polar compounds.

This thesis deals with diffusion and flow of sub-critical hydrocarbons in activated carbon by using a differential permeation method. The hydrocarbons are selected according to the effect on environmental concerns and their unique characteristics such as polarity and affinity towards activated carbon. Although it has been known that transport processes in activated carbon consist of Knudsen diffusion, gaseous viscous flow, adsorbed phase diffusion (so called, surface diffusion) and condensate flow, there have been no rigorous models to describe the transport processes in activated carbon with a full range of pressures. In particular among the four processes, the mechanism of adsorbed phase diffusion in activated carbon is still far from complete understanding. Also due to the dispersion interactions between adsorbing molecules and the solid surface, one would expect that Knudsen diffusion is influenced by the dispersive forces. From intensive experimental observations with a great care over a full range of pressures, conventional methods (for example, direct estimation from inert gas experiments) to determine adsorbed phase diffusion are found to be inadequate for strongly adsorbing vapors in activated carbon. By incorporating the effect of adsorbate-adsorbent interactions into Knudsen diffusivity, the general behavior of adsorbed phase diffusion in terms of pressure (or surface loading) can be obtained, showing a significant role in transport at low pressures. For non-polar hydrocarbons such as benzene, carbon tetrachloride and n-hexane, a mathematical model, which accounts for the effects of adsorbate-adsorbent interactions and pore size distribution, is formulated and validated, resulting in a good agreement with experimental data. Moreover, the adsorption and dynamic behaviors of alcohol molecules (which are polar compounds) are investigated with an aim to compare their behaviors against those of non-polar compounds.

This thesis deals with diffusion and flow of sub-critical hydrocarbons in activated carbon by using a differential permeation method. The hydrocarbons are selected according to the effect on environmental concerns and their unique characteristics such as polarity and affinity towards activated carbon. Although it has been known that transport processes in activated carbon consist of Knudsen diffusion, gaseous viscous flow, adsorbed phase diffusion (so called, surface diffusion) and condensate flow, there have been no rigorous models to describe the transport processes in activated carbon with a full range of pressures. In particular among the four processes, the mechanism of adsorbed phase diffusion in activated carbon is still far from complete understanding. Also due to the dispersion interactions between adsorbing molecules and the solid surface, one would expect that Knudsen diffusion is influenced by the dispersive forces. From intensive experimental observations with a great care over a full range of pressures, conventional methods (for example, direct estimation from inert gas experiments) to determine adsorbed phase diffusion are found to be inadequate for strongly adsorbing vapors in activated carbon. By incorporating the effect of adsorbate-adsorbent interactions into Knudsen diffusivity, the general behavior of adsorbed phase diffusion in terms of pressure (or surface loading) can be obtained, showing a significant role in transport at low pressures. For non-polar hydrocarbons such as benzene, carbon tetrachloride and n-hexane, a mathematical model, which accounts for the effects of adsorbate-adsorbent interactions and pore size distribution, is formulated and validated, resulting in a good agreement with experimental data. Moreover, the adsorption and dynamic behaviors of alcohol molecules (which are polar compounds) are investigated with an aim to compare their behaviors against those of non-polar compounds.

abstract: The world is grappling with two serious issues related to energy and climate change. The use of solar energy is receiving much attention due to its potential as one of the solutions. Air conditioning is particularly attractive as a solar energy application because of the near coincidence of peak cooling loads with the available solar power. Recently, researchers have started serious discussions of using adsorptive processes for refrigeration and heat pumps. There is some success for the >100 ton adsorption systems but none exists in the <10 ton size range required for residential air conditioning. There are myriad reasons for the lack of small-scale systems such as low Coefficient of Performance (COP), high capital cost, scalability, and limited performance data. A numerical model to simulate an adsorption system was developed and its performance was compared with similar thermal-powered systems. Results showed that both the adsorption and absorption systems provide equal cooling capacity for a driving temperature range of 70-120 ºC, but the adsorption system is the only system to deliver cooling at temperatures below 65 ºC. Additionally, the absorption and desiccant systems provide better COP at low temperatures, but the COP's of the three systems converge at higher regeneration temperatures. To further investigate the viability of solar-powered heat pump systems, an hourly building load simulation was developed for a single-family house in the Phoenix metropolitan area. Thermal as well as economic performance comparison was conducted for adsorption, absorption, and solar photovoltaic (PV) powered vapor compression systems for a range of solar collector area and storage capacity. The results showed that for a small collector area, solar PV is more cost-effective whereas adsorption is better than absorption for larger collector area. The optimum solar collector area and the storage size were determined for each type of solar system. As part of this dissertation work, a small-scale proof-of-concept prototype of the adsorption system was assembled using some novel heat transfer enhancement strategies. Activated carbon and butane was chosen as the adsorbent-refrigerant pair. It was found that a COP of 0.12 and a cooling capacity of 89.6 W can be achieved. / Dissertation/Thesis / Ph.D. Mechanical Engineering 2011

O zinco é um dos metais mais utilizados nos processos de tratamento de superfícies. Os efluentes contaminados com zinco, gerados nestes processos, trazem preocupação devido ao grande volume, elevada toxicidade, acumulação e persistência. O presente estudo avaliou a remoção de íons zinco de soluções aquosas, por adsorção em carvão ativado comercial, através de experimentos em batelada e em escala piloto. Ensaios em batelada objetivaram estudar a influência dos parâmetros de processo pH, tempo de contato e concentração de sólido adsorvente na remoção de íons Zn2+ a partir de uma solução com concentração inicial de 10 mg.L-1. A partir das melhores condições encontradas, a isoterma de equilíbrio do sistema foi construída e o ajuste dos dados experimentais aos modelos não lineares de Langmuir, Freundlich, Redlich-Peterson e Sips foi verificado. Ensaios em planta piloto foram realizados a fim de identificar a curva de ruptura do sistema. No processo de adsorção contínuo em leito fixo, foram avaliadas diferentes condições operacionais, variando-se a vazão de alimentação (15 e 20 mL.min-1) e a massa de adsorvente utilizada como recheio do leito (10, 13, 20 e 40 g). Os resultados apontaram como parâmetros ótimos de adsorção: pH igual a 6, tempo de contato de 30 min e concentração de sólido adsorvente de 20 g.L-1. O modelo de Sips apresentou o melhor ajuste dentre os modelos matemáticos estudados. O experimento realizado em planta piloto utilizando-se uma vazão de 15 mL.min-1 e massa de sólido sorvente de 40 g, obteve os melhores resultados dentre as condições analisadas, apresentando tempo de ruptura e exaustão do leito aos 5 e 90 min, respectivamente. / Zinc is one of the most used metals in surface treatment processes. The effluents contaminated with zinc, generated in these processes, bring concern due to the large volumes, high toxicity, persistence and accumulation. This study evaluated the removal of zinc ions from aqueous solutions, through adsorption by commercial activated carbon, using continuous (pilot plant) and bench scale. The batch experiments studied the influence of pH, contact time and adsorbent concentration in the removal of Zn 2+ ions from a solution with initial concentration of 10 mg.L-1. After achieving the best process parameters, the system’s equilibrium isotherm was constructed, and the fit of experimental data to the nonlinear Langmuir, Freundlich, Redlich-Peterson and Sips models was verified. Pilot plant tests were carried out in order to identify the system’s breakthrough curve. In the continuous fixed bed adsorption process, different operating conditions were evaluated by varying the feed flow (15 and 20 mL.min-1) and the mass of adsorbent material used as filling (10, 13, 20 and 40 g). The results showed as the optimum parameters for adsorption: pH 6, contact time of 30 min and adsorbent concentration of 20 g.L-1. The Sips model presented the best fit among the mathematical models studied. The experiment conducted in a pilot plant, using a flow rate of 15 mL.min-1 and solid adsorbent mass of 40 g, had the best results among the examined conditions, with breakthrough and exhaust times of 5 and 90 min, respectively.

O presente trabalho estuda a remoção dos fármacos amoxicilina (AMX), diclofenaco sódico (DCF) e paracetamol (PAR) em solução aquosa pelos processos de adsorção em batelada e coluna de leito fixo utilizando o carvão ativado granulado (CAG) como adsorvente. Os experimentos foram realizados para cada fármaco de forma independente. Na adsorção em batelada foram avaliadas as influências do pH (2 – 10), concentração de adsorvente (5 – 20 g L-1) e tempo de contato (5 – 350 min). Foi realizada a investigação da cinética de adsorção e também do equilíbrio de adsorção através de isotermas nas temperaturas de 25, 35 e 45 ºC. A adsorção em leito fixo foi estudada através de planejamentos experimentais, onde foram avaliados os efeitos da concentração inicial do poluente (C0: 20 – 100 mg L-1), massa do leito (W: 0,5 – 1,5 g) e vazão (Q: 3 – 5 mL min-1) sobre o tempo de saturação (tsat) e a quantidade adsorvida (qsat). Os modelos de Thomas, Bohart-Adams e Yan, além de um modelo desenvolvido no software EMSO foram utilizados para análise das curvas de ruptura. O CAG utilizado apresentou área BET de 463 m² g-1 e maior volume de microporos, de 0,20 cm³ g-1. Os experimentos em batelada mostraram que o pH não teve influência significativa sobre a remoção dos três fármacos. O equilíbrio de adsorção da AMX e do DCF foi atingido após 150 min e do PAR após 180 min. O modelo de PSO foi o que melhor representou a cinética de adsorção dos três fármacos. A isoterma de Langmuir descreveu o equilíbrio da AMX a 25 e 35 ºC, e o modelo de Sips a 45 ºC. Já a adsorção do DCF foi representada pela isoterma de Freundlich e o PAR pela de Redlich-Peterson. O estudo termodinâmico indicou que a adsorção dos três fármacos foi espontânea e favorável, além de aumentar com o aumento da temperatura. Na adsorção em leito fixo, foi observado menores valores de tsat com o aumento de C0 e de Q e diminuição de W. Foi verificado que qsat aumentou com o aumento de C0 e diminuição da Q para o planejamento do PAR, onde essa variável foi significativa. Já o aumento de W aumentou qsat no planejamento do PAR e diminuiu nos casos da AMX e DCF. A AMX foi o poluente que apresentou os menores tempos de saturação, seguido do DCF e do PAR, na adsorção em leito fixo. Foi constatado que o modelo de Yan foi o que melhor reproduziu o comportamento das curvas de ruptura para os três fármacos, na comparação com os outros modelos analíticos e com o modelo numérico proposto no software EMSO. De forma geral, foi verificado que os processos de adsorção tanto em batelada quanto em leito fixo apresentam potencial de aplicação como alternativa de tratamento avançado de água e efluentes que contenham fármacos. / The present work studies the removal of amoxicillin (AMX), sodium diclofenac (DCF), and paracetamol (PAR) from water by adsorption onto granular activated carbon in batch process and fixed bed column. Batch adsorption experiments were performed to evaluate the influence of pH (2 – 10), adsorbent concentration (5 – 20 g L-1) and contact time (5 – 350 min). Pseudo-first order, pseudo-second order and intraparticle diffusion models were evaluated in the kinetics investigation. Equilibrium adsorption was investigated using Langmuir, Freundlich, Sips and Redlich-Peterson equations. Fixed bed adsorption was studied through experimental design to evaluate initial contaminant concentration (C0, 20 – 100 mg L-1), amount of adsorbent (W, 0.5 – 1.5 g) and feed flow rate (Q, 3 – 5 mL min-1) effects. The analytical models of Thomas, Bohart-Adams and Yan were selected to investigate the breakthrough curves behavior. In addition, a numerical model was developed and solved using EMSO software. The granular activated carbon (GAC) used had BET surface area of 463 m² g-1 and volume of 0.20 cm³ g-1 of micropores. The pH had no significant effect on the adsorption removal of the three drugs. Adsorption equilibrium of AMX and DCF was reached after 150 min and 180 min for PAR. Pseudo second order model best represented kinetic adsorption of the three compounds. At best conditions in batch process, adsorbent concentration was 12.5 g L-1 for AMX and DCF and 10 g L-1 for PAR. Langmuir isotherm best described AMX adsorption equilibrium at 25 and 35 ºC, and Sips model at 45 ºC. DCF and PAR adsorption followed the Freundlich isotherm and Redlich-Peterson model, respectively. Thermodynamic study indicated that the three drugs adsorption were spontaneous and favorable processes. In addition, adsorption increased at higher temperatures. In fixed bed adsorption experiments, saturation time (tsat) decreased with the increase of initial concentration and flow rate for both drugs. W had positive effect on tsat. The amount adsorbed (qsat) was enhanced at higher C0 and lower Q. qsat was higher at higher Q for AMX and DCF and lower Q for PAR adsorption. Yan model best reproduced breakthrough curves behavior for all drugs among the analytical models and the numerical model developed on EMSO software. Thus, adsorption processes in batch mode and fixed bed column showed to be effective for the removal of drugs of different therapeutic classes from water.

A eutrofização acelerada dos mananciais superficiais, devido aos despejos de águas residuárias não tratadas, vem comprometendo a qualidade das águas utilizadas no abastecimento público. Como conseqüência desse processo, inúmeros registros de florações de cianobactérias são relatados em todo o mundo. Esses microrganismos são potenciais produtores de toxinas, as quais, presentes na água bruta que abastece uma Estação de Tratamento de Água (ETA), precisam ser removidas. As toxinas podem ter efeitos adversos à saúde, podendo causar danos hepatotóxicos, neurotóxicos e dermatotóxicos. A maioria das ETAs brasileiras tratam a água através do processo convencional de tratamento, compreendendo as etapas de coagulação, floculação, sedimentação, filtração e cloração. Esse tratamento é considerado eficiente para remover células de microrganismos, incluindo as cianobactérias. No entanto, suas toxinas não são afetadas, permanecendo na água tratada. Em função disso, alternativas de tratamento devem ser incorporadas ao tratamento convencional, visando remover a hepatotoxina microcistina a concentrações menores ou iguais a 1 μg.L-1, pois esse é o valor máximo permitido (VMP) pela Portaria nº 518/2004, do Ministério da Saúde, a qual define as diretrizes relativas ao controle e vigilância da qualidade da água para consumo humano. Nesse contexto, o objetivo do trabalho foi avaliar a eficiência do carvão ativado pulverizado (CAP), na remoção da cianotoxina microcistina presente na água utilizada para abastecimento público. Cinco amostras de CAP, produzidas a partir de madeira, osso, antracito e coco, foram caracterizadas e, para cada uma, determinadas as Isotermas de Freundlich. As Isotermas mostraram que os CAPs produzidos a partir da madeira apresentaram maior capacidade de remover microcistina. Os residuais de microcistina obtidos nestes ensaios foram ajustados ao modelo de decaimento logarítmico. Assim, para cada CAP, foi estabelecida uma equação geral do processo adsortivo, com a qual foi possível estabelecer dosagens de CAP, variando a concentração de microcistina inicial, para atingir uma concentração residual de 1 μg.L-1. A aplicabilidade da equação foi testada para uma amostra de CAP, em água natural, utilizando o coagulante policloreto de alumínio e concentrações iniciais de microcistina de 1, 10 e 100 μg.L-1. Os resultados mostraram que as dosagens de CAP calculadas foram suficientes para atingir o residual desejado. Além disso, através de ensaios de adsorção, em equipamento de jarros, foram simuladas as etapas de coagulação (utilizando coagulantes sulfato de alumínio e cloreto férrico), floculação, sedimentação e filtração do tratamento convencional para água natural acrescida de 100 μg.L-1 de microcistina. A aplicação do CAP foi realizada em dois pontos do tratamento (1) entrada da água bruta e (2) antes da aplicação do coagulante. A aplicação na entrada da água bruta possibilitou remoção da toxina abaixo do VMP, correspondendo à redução de, aproximadamente, 99% da concentração inicial de toxina. Já no ponto de aplicação antes do coagulante, não foi atingido o VMP. A partir dos estudos aqui realizados, concluiu-se que a melhor maneira de escolher o CAP para remoção de microcistina é a realização de ensaios específicos, como a Isoterma de Freundlich. Para uma remoção eficiente da toxina, o tratamento convencional mostra-se eficaz, desde que a etapa de adsorção seja incorporada ao tratamento. / The accelerated surface waters eutrophication, due to non treated residual waters discharges, has been harming the water quality utilized in the public supply. As a consequence of this process, countless records of cyanobacterias’ bloom are reported all over the world. These microorganisms potentially producers of toxins, which, when presented in the raw water that supplies the Water Treatment Plant (WTP), needs to be removed. The toxins can have harmfull effects to the health causing hepatotoxic, neurotoxic and dermatotoxic damage. Most of the Brazilian WTPs treat the water through the conventional water treatment process, covering the stages of coagulation, flocculation, sedimentation, filtering and chlorination. This treatment is considered efficient to remove microorganisms’ cells, including cyanobacterias. However its toxins are not affected, remaining in the treated water. Due to this fact, alternative treatment must be incorporated to the conventional one, trying to remove the microcystin hepatotoxin to concentrations lower or equal to 1 μg.L-1, the maximum allowed value (MAV) according to the Brazilian Ministry of Health Administrative Ruling 518/2004, which defines the rules related to the control and vigilance of the water quality for human consumption. In this context, the objective of the present work was to evaluate the efficiency of the powdered activated carbon’s (PAC), in the removal of the microcystin cyanotoxin present in the water for public supply. Five samples of PAC, made of wood, bone, antracite and coconut, were characterized and, for each one, were determined Freundlich adsorption isotherm . The isotherm showed that the PACs produced from wood presented the highest capacity of microcystin removal. The residuals of microcystin obtained in these tests were adjusted to the logarithmic decay model. Thus, for each PAC, it was established a general equation of the adsorptive process, in order to establish PAC dosages, varying the initial microcystin concentration, to achieve the residual concentration of 1 μg.L-1. The usage of this equation was tested for one sample of PAC, in natural water, using poly-aluminum chloride as a coagulant and the initial concentrations of microcystins of 1, 10 and 100 μg.L-1. The results showed that the dosage of PAC calculated was enough to achieve the desired residual. Besides this, through adsorptive tests, using jar’s equipments, were simulated the stages of coagulation (using alum sulphate and ferric chloride as coagulants), flocculation, sedimentation and filtering of the conventional treatment for natural water and with 100 μg.L-1 of microcystin added. The application of the PAC was performed in two points of the treatment: (1) entrance of the raw water and; (2) before the application of the coagulant. The application in the entrance of the raw water made possible the removal of the toxin under MAV, corresponding to a reduction of, approximately, 99% of the initial concentration of the toxin. Nevertheless, in the point of application before the coagulant, the MAV was not achieved. According to these studies, it was concluded that the best manner of choosing the PAC for microcystin removal is the realization of specific tests like a Freundlich isotherm. For an efficient removal of the toxin, the conventional treatment is efficient, once the adsorption stage is incorporated to the treatment.