Realizamos a caracterização elétrica de alexandrita (BeAl2O4:Cr3+), formas sintética e natural, através de medidas de Corrente de Despolarização Termicamente Estimulada (CDTE). Obtivemos evidências conclusivas do fenômeno de relaxação dipolar em ambos os tipos de amostra, e que as cwas experimentais devem ser ajustadas por urna distribuição contínua dos parârnetros de relaxação. Para a amostra sintética a banda de CDTE está centralizada em tomo de 179K e para as naturais em 187 a 195K. Utilizando o método de Havriliak-Negarni são necessárias duas distribuições continuas de Ea e &#9640 para ajustar as curvas experimentais, sendo que uma delas, em torno de 177 K, com Ea &#8773 0,56 e &#9640 &#8773 1,2x10-14s sente em ambos os tipos de amostras. As bandas de CDTE são atribuídas a dipolos do tipo impureza-vacância de oxigênio ou a deformação local da estrutura causada pela diferença de raio iônico entre os íons Cr3+ (0,615 &#197) e A13+(0,535 &#197). Também realizamos medidas de CDTE fotoinduzidas, onde as amostras são irradiadas com um laser sintonizado em comprimentos de onda entre 3373 e 676,4nm. Verificamos que as bandas de CDTE podem ser \"destruídas\" ou \"criadas\" com a incidência de luz com diferentes condições iniciais de polarização. Para ajudar a interpretação dos resultados de CDTE nós usamos outras técnicas de caracterização, tais como Absorção Óptica, Luminescência, Difração de Raios X e micro análises de EDX e WDX. Todas estas técnicas foram também aplicadas às amostras naturais após tratamentos térmicos consecutivos / We have done electrical characterization of natural and synthet ic alexandrite (BeAl2O4:Cr3+), usimg the thermally stimulated depolarization current (TSDC) technique. We have obtained conclusive evidences of dipole relaxation in both kinds of samples. Besides, the experimental data must be fitted by a continuous distribution of relaxation parameters. For the synthetic sample, TSDC band has a peak about 179K and for natural samples, TSDC bands have peaks about 187K at 195K. Using Havriliak-Negami method, we need two continuous distributions of activation energy (Ea) and relaxation time constant (&#9640) to fit experimental data. One of these two curves, centered at 177K, is present for both kinds of sarnples and has Ea &#8773 0,56 e &#9640 &#8773 1,2x10-14s. T SDC bands are attributed to impurity-oxygen vacancy dipoles or local structure deforrnation caused by the dserence between ionic radius of Cr3+ (0,615 &#197) and A13+(0,535 &#197) ions. We have also carried out photo-induced TSDC, where sarnples are irradiated with a tunable laser with wavelength fiom 337.5nm to 676.5nm. We have observed that TSDC bands rnay be destroyed or created with illumination fiom daerent polarization conditions. To help the interpretation of TSDC results we have used other techniques of characterization such as optical absorption, luminescence, X-ray difliaction, besides EDX and WDX rnicroanalyses. All of these techniques were also applied to natural samples afier consecutive annealing

Realizamos a caracterização elétrica de alexandrita (BeAl2O4:Cr3+), formas sintética e natural, através de medidas de Corrente de Despolarização Termicamente Estimulada (CDTE). Obtivemos evidências conclusivas do fenômeno de relaxação dipolar em ambos os tipos de amostra, e que as cwas experimentais devem ser ajustadas por urna distribuição contínua dos parârnetros de relaxação. Para a amostra sintética a banda de CDTE está centralizada em tomo de 179K e para as naturais em 187 a 195K. Utilizando o método de Havriliak-Negarni são necessárias duas distribuições continuas de Ea e &#9640 para ajustar as curvas experimentais, sendo que uma delas, em torno de 177 K, com Ea &#8773 0,56 e &#9640 &#8773 1,2x10-14s sente em ambos os tipos de amostras. As bandas de CDTE são atribuídas a dipolos do tipo impureza-vacância de oxigênio ou a deformação local da estrutura causada pela diferença de raio iônico entre os íons Cr3+ (0,615 &#197) e A13+(0,535 &#197). Também realizamos medidas de CDTE fotoinduzidas, onde as amostras são irradiadas com um laser sintonizado em comprimentos de onda entre 3373 e 676,4nm. Verificamos que as bandas de CDTE podem ser \"destruídas\" ou \"criadas\" com a incidência de luz com diferentes condições iniciais de polarização. Para ajudar a interpretação dos resultados de CDTE nós usamos outras técnicas de caracterização, tais como Absorção Óptica, Luminescência, Difração de Raios X e micro análises de EDX e WDX. Todas estas técnicas foram também aplicadas às amostras naturais após tratamentos térmicos consecutivos / We have done electrical characterization of natural and synthet ic alexandrite (BeAl2O4:Cr3+), usimg the thermally stimulated depolarization current (TSDC) technique. We have obtained conclusive evidences of dipole relaxation in both kinds of samples. Besides, the experimental data must be fitted by a continuous distribution of relaxation parameters. For the synthetic sample, TSDC band has a peak about 179K and for natural samples, TSDC bands have peaks about 187K at 195K. Using Havriliak-Negami method, we need two continuous distributions of activation energy (Ea) and relaxation time constant (&#9640) to fit experimental data. One of these two curves, centered at 177K, is present for both kinds of sarnples and has Ea &#8773 0,56 e &#9640 &#8773 1,2x10-14s. T SDC bands are attributed to impurity-oxygen vacancy dipoles or local structure deforrnation caused by the dserence between ionic radius of Cr3+ (0,615 &#197) and A13+(0,535 &#197) ions. We have also carried out photo-induced TSDC, where sarnples are irradiated with a tunable laser with wavelength fiom 337.5nm to 676.5nm. We have observed that TSDC bands rnay be destroyed or created with illumination fiom daerent polarization conditions. To help the interpretation of TSDC results we have used other techniques of characterization such as optical absorption, luminescence, X-ray difliaction, besides EDX and WDX rnicroanalyses. All of these techniques were also applied to natural samples afier consecutive annealing

Dans ce mémoire nous présentons plusieurs études expérimentales des propriétés magnétiques d’un condensat de Bose-Einstein de Chrome chargé dans un réseau 3D, en nous focalisant sur les effets associés aux interactions dipolaires. Nous montrons que dans un réseau 3D, la relaxation dipolaire est un processus résonant du fait de la réduction de la densité d’états orbitaux accessibles. Les résonances sont observées à des champs magnétiques Bres tels que l’énergie Zeeman relâchée soit égale à l’énergie nécessaire à exciter les atomes dans une bande d’énergie supérieure du réseau. Nous pouvons inhiber ce processus en appliquant un champ différent de Bres. L’analyse des résonances a permis de sonder la structure de bande 3D du réseau, ainsi que la mise en évidence de l’effet des interactions entre atomes. Nous avons étudié la dynamique d’échange de spin dans un réseau 3D. Nous présentons en particulier la première observation d’échange de spin entre atomes localisés dans des sites séparés. Ces études permettent une exploration nouvelle du magnétisme en réseau. En variant la profondeur du réseau, nous étudions ces effets dans le régime superfluide, bien décrit par une théorie de champ moyen, et dans le régime fortement corrélé, dont la description théorique est difficile. Enfin, nous étudions l’évolution de deux spins géants interagissant par interaction dipolaire. Le condensat initialement divisé en deux, les atomes des deux nuages sont préparés dans des états de spin opposés formant ainsi deux spins géants ±3xN. Nous montrons que toute dynamique de spin est énergétiquement inhibée pour de grands spins ce qui est bien reproduit par une théorie classique. / This Thesis reports on several experimental studies of magnetic properties of a Chromium Bose-Einsteincondensate loaded into a 3D optical lattice, focusing on the effects induced by dipolar interactions.We show that in a 3D lattice dipolar relaxation is a resonant process due to the reduction of the density ofaccessible orbital states. These resonances are observed for magnetic fields Bres such that the Zeeman energyreleased matches an excitation towards higher-energy bands of the lattice. We can thus inhibit those processes byapplying a field different from Bres. Analyses of the resonances allowed us to probe the lattice 3D band structureas well as to demonstrate the effects of local interactions between atoms.We study spin exchange dynamics in a 3D lattice. We especially observed for the first time spin exchangebetween atoms localized in different lattice sites mediated by dipolar interactions. These studies are the firststep toward a new exploration of magnetism in lattice. Varying the depth of the lattice we study these effects inthe superfluid regime, well described by mean filed theories, as well as in the strongly correlated regime, whosetheoretical description is still challenging.Finally, we study the evolution dynamics of two giant spins interacting through dipolar interactions. Thecondensate being initially splitted in half, atoms from the two clouds are prepared in opposite spin states thusproducing two giant spins ±3×N. We show that any spin dynamics is energetically inhibited for large spinswhich is well accounted for by a classical theory