Fontes radioativas seladas são largamente empregadas no mundo. A Agência Internacional de Energia Atômica AIEA estima em dezenas de milhões de fontes radioativas no mundo. No Brasil, este número é próximo de 500 mil unidades, considerando-se as fontes de pára-raios e detectores de fumaça. Uma fonte selada pode tornar-se desnecessária, seja devido ao seu decaimento ou outro motivo, sendo classificada como fontes radioativas seladas fora de uso (FRS). No Brasil, a maioria das FRS é considerada rejeito radioativo e são encaminhadas a um dos institutos da CNEN. Sem uma estratégia definida para a deposição das FRS, estas são armazenadas aguardando uma solução. A Gerência de Rejeitos Radioativos - GRR do IPEN-CNEN/SP é o principal centro de recepção desse material e até meados de 2010 havia recebido cerca de 14.000 fontes. O processo proposto para a gestão dessas FRS consiste em retirá-las de suas blindagens originais e transferi-las a outra blindagem, projetada para esse fim. As operações de transferência são realizadas em uma cela quente, garantindo a segurança radiológica. Uma das exigências da CNEN é que todas as fontes seladas sejam caracterizadas. No processo estudado, a avaliação da atividade de cada fonte será feita utilizando-se um detector, do tipo câmara de ionização tipo poço. O presente trabalho tem por objetivo desenvolver uma metodologia para aferir ou determinar a atividade de FRS armazenadas na GRR de acordo com sua geometria e determinar suas incertezas. / Sealed radioactive sources are widely used in many applications of nuclear technology in industry, medicine, research and others. The International Atomic Energy Agency (IAEA) estimates tens of millions sources in the world. In Brazil, the number is about 500 thousand sources, if the Americium-241 sources present in radioactive lightning rods and smoke detectors are included in the inventory. At the end of the useful life, most sources become disused, constitute a radioactive waste, and are then termed spent sealed radioactive sources (SSRS). In Brazil, this waste is collected by the research institutes of the Nuclear Commission of Nuclear Energy and kept under centralized storage, awaiting definition of the final disposal route. The Waste Management Laboratory (WML) at the Nuclear and Energy Research Institute is the main storage center, having received until July 2010 about 14.000 disused sources, not including the tens of thousands of lightning rod and smoke detector sources. A program is underway in the WML to replacing the original shielding by a standard disposal package and to determining the radioisotope content and activity of each one. The identification of the radionuclides and the measurement of activities will be carried out with a well type ionization chamber. This work aims to develop a methodology for measuring or to determine the activity SSRS stored in the WML accordance with its geometry and determine their uncertainties.

Fontes radioativas seladas são largamente empregadas no mundo. A Agência Internacional de Energia Atômica AIEA estima em dezenas de milhões de fontes radioativas no mundo. No Brasil, este número é próximo de 500 mil unidades, considerando-se as fontes de pára-raios e detectores de fumaça. Uma fonte selada pode tornar-se desnecessária, seja devido ao seu decaimento ou outro motivo, sendo classificada como fontes radioativas seladas fora de uso (FRS). No Brasil, a maioria das FRS é considerada rejeito radioativo e são encaminhadas a um dos institutos da CNEN. Sem uma estratégia definida para a deposição das FRS, estas são armazenadas aguardando uma solução. A Gerência de Rejeitos Radioativos - GRR do IPEN-CNEN/SP é o principal centro de recepção desse material e até meados de 2010 havia recebido cerca de 14.000 fontes. O processo proposto para a gestão dessas FRS consiste em retirá-las de suas blindagens originais e transferi-las a outra blindagem, projetada para esse fim. As operações de transferência são realizadas em uma cela quente, garantindo a segurança radiológica. Uma das exigências da CNEN é que todas as fontes seladas sejam caracterizadas. No processo estudado, a avaliação da atividade de cada fonte será feita utilizando-se um detector, do tipo câmara de ionização tipo poço. O presente trabalho tem por objetivo desenvolver uma metodologia para aferir ou determinar a atividade de FRS armazenadas na GRR de acordo com sua geometria e determinar suas incertezas. / Sealed radioactive sources are widely used in many applications of nuclear technology in industry, medicine, research and others. The International Atomic Energy Agency (IAEA) estimates tens of millions sources in the world. In Brazil, the number is about 500 thousand sources, if the Americium-241 sources present in radioactive lightning rods and smoke detectors are included in the inventory. At the end of the useful life, most sources become disused, constitute a radioactive waste, and are then termed spent sealed radioactive sources (SSRS). In Brazil, this waste is collected by the research institutes of the Nuclear Commission of Nuclear Energy and kept under centralized storage, awaiting definition of the final disposal route. The Waste Management Laboratory (WML) at the Nuclear and Energy Research Institute is the main storage center, having received until July 2010 about 14.000 disused sources, not including the tens of thousands of lightning rod and smoke detector sources. A program is underway in the WML to replacing the original shielding by a standard disposal package and to determining the radioisotope content and activity of each one. The identification of the radionuclides and the measurement of activities will be carried out with a well type ionization chamber. This work aims to develop a methodology for measuring or to determine the activity SSRS stored in the WML accordance with its geometry and determine their uncertainties.

Les activités nucléaires génèrent des déchets radioactifs classés selon leur niveau d’activité et la durée de vie des radioéléments présents. La garantie d’un classement et d’une gestion optimale nécessite une caractérisation précise. Les déchets de moyenne et haute activité, contenant des radioéléments à vie très longue, seront stockés en profondeur pendant plusieurs centaines de milliers d’années, à l’issue desquelles il est nécessaire de pouvoir garantir l’absence de risques pour l’homme et l’environnement, non seulement sur le plan radiologique, mais aussi en ce qui concerne des éléments stables, toxiques du point de vue chimique. Cette thèse concerne la caractérisation par activation neutronique de ces éléments toxiques, ainsi que celle des matières nucléaires présentes dans les colis. Elle a été réalisée dans le cadre d’une collaboration entre le Laboratoire de Mesures Nucléaires du CEA Cadarache, en France, et l’institut de Gestion des Déchets Radioactifs et de Sûreté des Réacteurs du centre de recherche FZJ (Forschungszentrum Jülich), en Allemagne. La première étude a consisté à valider le modèle numérique de la cellule d’activation neutronique MEDINA (FZJ) avec le code de transport Monte Carlo MCNP. Les rayonnements gamma prompts de capture radiative d’échantillons contenant des éléments d’intérêt (béryllium, aluminium, chlore, cuivre, sélénium, strontium et tantale) ont été mesurés et comparés aux simulations avec diverses bases de données nucléaires, permettant d’aboutir à un accord satisfaisant et validant le schéma de calcul en vue des études suivantes. Ensuite, la mesure des rayonnements gamma retardés de fissions induites sur les isotopes 235U et 239Pu a été étudiée pour des fûts de 225 L contenant des enrobés bitumineux ou une matrice béton, représentatifs de déchets produits en France et en Allemagne. Les rendements d’émission des rayonnements gamma retardés de fission d’intérêt, cohérents avec ceux publiés dans la littérature, ont été déterminés à partir des mesures d’échantillons métalliques d’uranium et de plutonium dans la cellule d’activation neutronique REGAIN du LMN. Le signal utile a ensuite été extrapolé par simulation MCNP pour une répartition homogène d’isotopes 239Pu ou 235U dans les matrices considérées, en utilisant le modèle numérique de MEDINA. Des signaux faibles, de l’ordre de 100 coups par gramme d’isotope 239Pu ou 235U, ont été obtenus. Pour le colis d’enrobés bitumineux, le niveau d’irradiation gamma très élevé, dû à une activité en 137Cs de l’ordre de 1 TBq par fût, nécessiterait l’utilisation d’une collimation et/ou d’écrans pour éviter la saturation de l’électronique de mesure, rendant indétectables les rayonnements gamma retardés de fission. Les colis de déchets bétonnés produits en Allemagne présentant un niveau d’activité plus faible, il a été possible d’estimer des limites de détection allant de 10 à 290 g d’isotope fissile 235U ou 239Pu, selon la raie gamma considérée, suite à la mesure du bruit de fond actif dans MEDINA avec une matrice béton maquette. Afin d’améliorer ces performances, le blindage du détecteur germanium de MEDINA a été optimisé à l’aide de simulations MCNP, montrant la possibilité de réduire les bruits de fond gamma et neutron d’un facteur 4 et 5, respectivement. La validation expérimentale de l’efficacité du blindage a été effectuée à partir de configuration simples à implémenter dans MEDINA, confirmant les facteurs de réduction attendus. Un blindage du détecteur optimal permettrait d’améliorer les limites de détection et aussi d’utiliser une source de neutrons d’intensité supérieure, comme un générateur de neutron à haut flux ou un accélérateur linéaire d’électrons avec une cible de conversion appropriée. / Nuclear activities produce radioactive wastes classified following their radioactive level and decay time. An accurate characterization is necessary for efficient classification and management. Medium and high level wastes containing long lived radioactive isotopes will be stored in deep geological storage for hundreds of thousands years. At the end of this period, it is essential to ensure that the wastes do not represent any risk for humans and environment, not only from radioactive point of view, but also from stable toxic chemicals. This PhD thesis concerns the characterization of toxic chemicals and nuclear material in radioactive waste, by using neutron activation analysis, in the frame of collaboration between the Nuclear Measurement Laboratory of CEA Cadarache, France, and the Institute of Nuclear Waste Management and Reactor Safety of the research center, FZJ (Forschungszentrum Jülich GmbH), Germany. The first study is about the validation of the numerical model of the neutron activation cell MEDINA (FZJ), using MCNP Monte Carlo transport code. Simulations and measurements of prompt capture gamma rays from small samples measured in MEDINA have been compared for a number of elements of interest (beryllium, aluminum, chlorine, copper, selenium, strontium, and tantalum). The comparison was performed using different nuclear databases, resulting in satisfactory agreement and validating simulation in view of following studies. Then, the feasibility of fission delayed gamma-ray measurements of 239Pu and 235U in 225 L waste drums has been studied, considering bituminized or concrete matrixes representative of wastes produced in France and Germany. The delayed gamma emission yields were first determined from uranium and plutonium metallic samples measurements in REGAIN, the neutron activation cell of LMN, showing satisfactory consistency with published data. The useful delayed gamma signals of 239Pu and 235U, homogeneously distributed in the 225 L matrixes, were then determined by MCNP simulations using MEDINA numerical model. Weak signals of about one hundred counts per gram of 239Pu or 235U after 7200 s irradiation were obtained. Because of the high gamma emission in the bituminized waste produced in France (about 1 TBq of 137Cs per drum), the use of collimator and/or shielding is mandatory to avoid electronic saturation, making fission delayed gamma rays undetectable. However, German concrete drums being of lower activity, their corresponding active background was measured in MEDINA with a concrete mock-up, leading to detection limits between 10 and 290 g of 235U or239Pu, depending on the delayed gamma line. In order to improve these performances, the shielding of MEDINA germanium detector was optimized using MCNP calculations, resulting in gamma and neutron background reduction factors of 4 and 5, respectively. The experimental validation of the shielding efficiency was performed by implementing easy-to-build configurations in MEDINA, which confirmed the expected background reduction factors predicted by MCNP. Thanks to an optimized detector shielding, it will also be possible to use a higher neutron emission source, like a high flux neutron generator or an electron LINAC with appropriate conversion targets, in view to further reduce detection limits.