Será explorada a versão hidrodinâmica da equação de Schrödinger não-linear e não-local, descrevendo condensados de Bose-Einstein com auto-interações de longo alcance. Tais sistemas têm despertado interesse tendo em vista a busca da realização da condensação de Bose-Einstein sem necessidade de um potencial externo confinante e nos quais as interações atômicas locais não são suficientes. Para obter a descrição hidrodinâmica, a transformação de Madelung para a função de onda será utilizada, reduzindo o problema a uma equação da continuidade e a uma equação de transporte de momentum. Esta última é similar à equação de Euler em fluidos ideais, porém contendo um potencial quântico efetivo e um termo não local, o qual advém da interação atômica. Tais equações de fluido traduzem, respectivamente, a conservação da probabilidade e do momentum total. O método hidrodinâmico permitirá o estudo de excitações elementares, entre os quais os modos de Bogoliubov, segundo uma abordagem macroscópica. / The hydrodynamic version of the Schrödinger equation nonlinear and nonlocal will be explored, describing Bose-Einstein condensates with long-range self-interactions. Such systems have aroused interest with a view to pursuing the realization of Bose-Einstein condensation without an external confining potential and in which local atomic interactions are not enough. For the hydrodynamic description, the eikonal decomposition of the wave function is used, reducing the problem to one equation of continuity and to a transport of momentum equation. The latter is similar to the Euler equation in ideal fluid but containing an effective quantum potential and a nonlocal term, which comes from the atomic interaction. Such fluid equations translate, respectively, conservation of probability and total momentum. The hydrodynamic method will allow the study of elementary excitations, including Bogoliubov modes according to a macroscopic approach.

Made available in DSpace on 2015-05-15T11:46:18Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 899692 bytes, checksum: d08b5615c89c2ad179f1f6dda0a8c410 (MD5) Previous issue date: 2014-06-14 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / In this work, we study issues related the existence, nonexistence and regularity of solutions to semilinear Schrödinger equations of type u + a(x)u = jujp2u; u 2 H1(RN); where N 2, p > 2 if N = 2 and 2 < p < 2N=(N 2) if N 3 and the potential a(x) is a positive function that belongs to L1(RN). To obtain the results, we use a Linking Theorem and the Principle of Symmetric Criticality. / Neste trabalho, estudamos questões relacionadas a existência, não-existência e regularidade de soluções para equações de Schrödinger semilineares do tipo u + a(x)u = jujp2u; u 2 H1(RN); onde N 2, p > 2 se N = 2 e 2 < p < 2N=(N 2) se N 3 e o potencial a(x) é uma função positiva que pertence a L1(RN). Para obtenção dos resultados, usamos um Teorema de Linking e o Princípio da Criticalidade Simétrica.

Made available in DSpace on 2015-05-15T11:46:23Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 1404779 bytes, checksum: fa23dbf1324f104548ef91fcbbf20fba (MD5) Previous issue date: 2014-09-05 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / In this work we study questions related to the existence of positive solutions for some classes of quasilinear Schrödinger equations, with hypotheses on the potential that permit this potential to vanish at infinity. In order to use variational methods to obtain our results, we make some changes of variables to obtain some semilinear equations, whose associated functionals are well defined in a classical Sobolev spaces. We also work with these equations on an Orlicz type space whose energy functional satisfy the geometric properties of the Mountain Pass Theorem. We still use the penalty technique due to Del Pino and Felmer and the Moser iteration method to obtain estimates in L1 norm, which are fundamental to our study. / Neste trabalho, estudamos questões relacionadas à existência de soluções positivas para algumas classes de equações de Schrödinger quasilineares, com hipóteses sobre o potencial que o possibilita se anular no infinito. Afim de usarmos métodos variacionais na obtenção de nossos resultados, aplicamos mudança de variáveis para reduzirmos as equações quasilineares a equações semilineares. Os funcionais associados a essas novas equações estão bem definidos em espaços de Sobolev clássicos e em espaços tipo Orlicz e satisfazem as propriedades geométricas do Teorema do Passo da Montanha. Ainda utilizamos a técnica de penalização de Del Pino e Felmer e o método de iteração de Moser para obtenção de estimativas, fundamentais para o nosso estudo, na norma L1.

Será explorada a versão hidrodinâmica da equação de Schrödinger não-linear e não-local, descrevendo condensados de Bose-Einstein com auto-interações de longo alcance. Tais sistemas têm despertado interesse tendo em vista a busca da realização da condensação de Bose-Einstein sem necessidade de um potencial externo confinante e nos quais as interações atômicas locais não são suficientes. Para obter a descrição hidrodinâmica, a transformação de Madelung para a função de onda será utilizada, reduzindo o problema a uma equação da continuidade e a uma equação de transporte de momentum. Esta última é similar à equação de Euler em fluidos ideais, porém contendo um potencial quântico efetivo e um termo não local, o qual advém da interação atômica. Tais equações de fluido traduzem, respectivamente, a conservação da probabilidade e do momentum total. O método hidrodinâmico permitirá o estudo de excitações elementares, entre os quais os modos de Bogoliubov, segundo uma abordagem macroscópica. / The hydrodynamic version of the Schrödinger equation nonlinear and nonlocal will be explored, describing Bose-Einstein condensates with long-range self-interactions. Such systems have aroused interest with a view to pursuing the realization of Bose-Einstein condensation without an external confining potential and in which local atomic interactions are not enough. For the hydrodynamic description, the eikonal decomposition of the wave function is used, reducing the problem to one equation of continuity and to a transport of momentum equation. The latter is similar to the Euler equation in ideal fluid but containing an effective quantum potential and a nonlocal term, which comes from the atomic interaction. Such fluid equations translate, respectively, conservation of probability and total momentum. The hydrodynamic method will allow the study of elementary excitations, including Bogoliubov modes according to a macroscopic approach.

L’objet de cette thèse est l’étude spectrale et dynamique de systèmes de la mécanique quantique en utilisant des techniques de commutateurs. Deux parmi les trois articles présentés traitent l’opérateur de Schrödinger discret sur un réseau. Dans le premier article, un principe d’absorption limite est établi pour le Laplacien discret multidimensionnel perturbé par la somme d’un potentiel de type Wigner-von Neumann et d’un potentiel de type longue portée. Ce résultat implique notamment l’absolue continuité du spectre de cet Hamiltonien à certaines énergies. Dans le second article, nous considérons à nouveau l’opérateur de Schrödinger discret multidimensionnel dont le potentiel est de type longue portée. Il est démontré que les fonctions propres correspondant à des valeurs propres de l’Hamiltonien décroissent sous-exponentiellement lorsque ces dernières ne sont pas un seuil. En dimension un, il est démontré de surcroît que ces fonctions propres décroissent exponentiellement. Une conséquence de ceci est l’absence de valeurs propres dans la partie centrale du spectre délimité aux extrémités par des seuils. Le troisième article étudie des propriétés dynamiques d’Hamiltoniens vérifiant des hypothèses minimales dans la théorie des commutateurs. En se basant sur une estimation des vitesses minimales d’une part et une version améliorée du théorème du RAGE d’autre part, nous dérivons deux estimations de propagation pour cette famille d’Hamiltoniens. Ces estimations indiquent que les états du système se comportent dynamiquement de façon très similaire aux états de diffusion. Toutefois, ceci n’écarte pas la possibilité de spectre singulier continu. / This thesis deals with the analysis of spectral and dynamical properties of quantum mechanical systems using techniques of operator commutators. Two of the three research papers that are presented deal exclusively with the discrete Schrödinger operators on the lattice. The first article proves a limiting absorption principle for the multi-dimensional discrete Laplacian perturbed by the sum of a Wigner-von Neumann potential and long-range potential. This result notably implies the absolute continuity of the spectrum of this Hamiltonian at certain energies. The second article proves that eigenfunctions corresponding to non-threshold eigenvalues of multidimensional discrete Schrödinger operators decay sub-exponentially. In one dimension, it is further proven that these eigenfunctions decay exponentially. A consequence of this is the absence of eigenvalues when the middle portion of the spectrum does not contain any thresholds. The third article investigates dynamical properties of Hamiltonians under very minimal assumptions in the theory of commutators. Based on minimal escape velocities and an improved version of the RAGE Theorem, we derive propagation estimates for these types of Hamiltonians. These estimates indicate that the states of the system behave dynamically very much like scattering states. Nonetheless, the existence of singularly continuous states cannot be disproved.

Submitted by Luciana Ferreira (lucgeral@gmail.com) on 2018-07-17T13:07:11Z No. of bitstreams: 2 Tese - Luciano Calaça Alves - 2018.pdf: 4699371 bytes, checksum: 706846314ebb74648be890b03e18d4cc (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2018-07-17T13:55:46Z (GMT) No. of bitstreams: 2 Tese - Luciano Calaça Alves - 2018.pdf: 4699371 bytes, checksum: 706846314ebb74648be890b03e18d4cc (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-07-17T13:55:46Z (GMT). No. of bitstreams: 2 Tese - Luciano Calaça Alves - 2018.pdf: 4699371 bytes, checksum: 706846314ebb74648be890b03e18d4cc (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2018-06-07 / Fundação de Amparo à Pesquisa do Estado de Goiás - FAPEG / This work presents the study of solitary wave solutions, known as solitons, in non-linear and non- homogeneous media using non-linear Schrödinger equations. Three cases are studied: first considering a logarithmic nonlinear term; second with saturation effect and finally including effects of high orders (Raman scattering). Solutions are modulated by three different types of potential. First, linear in the spatial and oscillatory coordinate in the temporal coordinate. The second, quadratic in the spatial and oscillatory in the temporal coordinates. Finally, it is also modulated using a mixed potential, which is the junction of the two potentials presented above. After including inomogeneities in linear and nonlinear coefficients, the similarity transformation technique is used to convert the non-linear, non-autonomous equation into an autonomous one that will be solved analytically. This field of study has potential applications in crystals, optical fibers and in Bose- Einstein condensates, also serving to understand the fundamentals related to this state of matter. The stability of the solutions are checked by numerical simulations. / Este trabalho apresenta o estudo de soluções de ondas solitárias, conhecidas como sólitons, em meios não lineares e não homogêneos por meio de equações não lineares de Schrödinger. São estudados três casos: primeiro considerando um termo não linear do tipo logarítmico; segundo com efeito de saturação e por último incluindo efeitos de altas ordens (espalhamento Raman). As soluções são moduladas por três tipos diferentes de potencial. O primeiro, linear na coordenada espacial e oscilatório na coordenada temporal. O segundo, quadrático na coordenada espacial e oscilatório na temporal. Por fim, modula-se também utilizando um potencial misto, que é a junção dos dois potenciais apresentados anteriormente. Depois de incluir heterogeneidades nos coeficientes lineares e não lineares, é utilizada a técnica da transformação de similaridade para converter a equação não linear, não autônoma em uma autônoma que será resolvida analiticamente. Esse campo de estudo tem potenciais aplicações em cristais, fibras ópticas e em condensados de Bose-Einstein, servindo também para o entendimento dos fundamentos relacionados a esse estado da matéria. A estabilidade das soluções são checadas por meio de simulações numéricas.

Cette thèse est consacrée à l’étude de problèmes inverses associés à des équations aux dérivées partielles hyperboliques et de type Schrödinger.La première partie de la thèse est consacrée à l’étude de problèmes inverses pour l’équation des ondes. Il s’agit d’examiner les propriétés de stabilité et d’unicité dans l’identification de certains coefficients apparaissant dans l’équation des ondes, à partir de différents types d’observation.La deuxième partie de cette thèse, traite du problème de l’identification du champ magnétique et du potentiel électrique apparaissant dans l’équation du Schrödinger. Nous prouvons que ces coefficients peuvent être déterminés de façon stable dans tout le domaine, à partir de données de type Neumann. La dérivation de ces résultats est basée sur la construction d’un ensemble de solutions de type optique géométrique, adaptées au système étudié. Il existe une méthode alternative pour l’analyse de ce type de problèmes inverses, celle de Bukhgeim-Klibanov, qui utilise une estimation de Carleman spécifique à l’opérateur con-sidéré. Elle nous a permis de montrer qu’il est possible de récupérer de façon stable et simultanée, la partie spatiale des potentiels électrique et magnétique de l’équation de Schrödinger magnétique, à partir d’un nombre fini de mesures partielles de la solution. / This thesis is devoted to the study of inverse problems associated to hyperbolic and Schrödinger equations. The first part of the thesis is devoted to the study of inverse problemsfor the wave equation. The aim is to examine the stability andthe uniqueness issues in the identification of certain coefficients appearing in the wave equation from different types of observation. The second part of this thesis deals with the problem of the identification of a magnetic field and an electric potential appearing in the Schrödinger equation. We prove that these coefficients can be stably determined throughout the domain, using Neumann data. The derivation of these results is based on the construction of a set of geometric optics solutions adapted to the system studied. There is an alternative method for the analysis of this type of inverse problem, which is due to Bukhgeim-Klibanov, and which uses a Carleman estimate. We show that it is possible to stably and simultaneously recover the spatial part of the electrical and magnetic potentialsappearing in the magnetic Schrödinger equation, from a finite number of measurements.

Les atomes de Rydberg sont des états très excités, dans lesquels un électron est placé sur une orbite éloignée du noyau. Leur grand dipôle électrique les rend très sensibles à leur environnement électromagnétique. En utilisant des champs microondes et radiofréquences, nous préparons des états quantiques non-classiques spécialement conçus pour exploiter au mieux cette sensibilité et mesurer des champs électriques et magnétiques avec une grande précision. Dans la première partie, nous préparons des états chats de Schrödinger, superpositions d'orbitales de polarisabilités très différentes, qui nous permettent de mesurer de petites variations du champ électrique statique avec une sensibilité bien supérieure à la limite quantique standard et proche de la limite Heisenberg fondamentale. Nous atteignons une sensibilité par atome de 30mV/m pour un temps d'interrogation de 200ns, faisant de notre système l'un des électromètres les plus sensibles à ce jour. Nous implémentons ensuite des manipulations plus complexes de l'atome. Grâce à une technique d'écho de spin qui exploite la richesse de la multiplicité Rydberg, nous mesurons les corrélations temporelles du champ électrique avec une bande passante de l'ordre du MHz. Dans la partie finale, nous préparons une superposition quantique de deux états circulaires de nombres quantiques magnétiques opposés. Cet état très non-classique correspond à un électron tournant à la fois dans des directions opposées sur la même orbite. La grande différence de moment magnétique entre les deux composantes de la superposition, de l'ordre de 100muB, ouvre la voie à la mesure de petites variations du champ magnétique avec une grande bande passante. / Rydberg atoms are highly excited states, in which the electron is orbiting far from the nucleus. Their large electric dipole makes them very sensitive to their electromagnetic environment. Using a combination of microwave and radio-frequency fields, we engineer non-classical quantum states specifically designed to exploit at best this sensitivity for electric and magnetic field metrology. In the first part, we prepare non-classical states, similar to Schrödinger cat states, superpositions of two orbitals with very different polarizabilities, that allow us to measure small variations of the static electric field with a sensitivity well beyond the standard quantum limit and close to the fundamental Heisenberg limit. We reach a single atom sensitivity of 30mV/m for a 200ns interrogation time. It makes our system one of the most sensitive electrometers to date. We then implement more complex manipulations of the atom. Using a spin-echo technique taking advantage of the full extent of the Rydberg manifold, we perform a correlation function measurement of the electric field with a MHz bandwidth.In the final part, we prepare a quantum superposition of two circular states with opposite magnetic quantum numbers. It corresponds to an electron rotating at the same time in opposite directions on the same orbit, a rather non-classical situation. The huge difference of magnetic moment between the two components of the superposition, in the order of 100muB, opens the way to the measurement of small variations of the magnetic field with a high bandwidth.

Cette thèse concerne l'étude spectrale de l'opérateur de Schrödinger avec champ magnétique et paramètre semi-classique, sur un domaine borné et régulier en dimension 2, avec condition de Neumann au bord. On s'intéresse plus particulièrement au cas où le champ magnétique s'annule sur une union de courbes régulières. L'objectif est de comprendre l'influence d'une annulation du champ et d'expliciter le comportement des basses valeurs propres et des fonctions propres associées lorsque le paramètre semi-classique tend vers 0. Dans cette limite - dite semi-classique - la description précise des éléments propres passe par la compréhension de différents opérateurs modèles sous-jacents. La première partie est consacrée au cas d'un champ magnétique qui s'annule de manière non dégénérée le long d'une courbe régulière simple intersectant le bord du domaine. La deuxième partie concerne le cas d'une annulation quadratique à l'intérieur du domaine. Dans de ces deux cas d'étude, on donne dans un premier temps un équivalent asymptotique de la première valeur propre. La majoration s'obtient par une construction de fonctions tests appropriées tandis que la minoration s'obtient par une méthode de localisation quantique. Ce dernier aspect est délicat car il s'agit de gérer la transition entre des modèles ayant des homogénéités différentes. Dans un second temps, on examine les propriétés de localisation des premières fonctions propres, via des estimées d'Agmon semi-classiques. Ceci permet d'obtenir un développement asymptotique complet des premières valeurs propres, à n'importe quel ordre. Dans le cas d'une annulation quadratique, la thèse est complétée par une étude de l'opérateur modèle pour lequel le lieu d'annulation est une union de deux droites sécantes faisant un angle non nul. Dans la limite petit angle, la structure du spectre est gouvernée un symbole opérateur à deux paramètres. On établit différentes propriétés de ce symbole opérateur et de la fonction de bande associée. Des simulations numériques basées sur la librairie éléments finis Mélina++ ont guidé l'analyse et illustrent les résultats obtenus. Les difficultés numériques - dues aux fortes oscillations de la phase dans l'expression des fonctions propres - sont gérées grâce à une interpolation polynomiale de haut degré. / This thesis is devoted to the spectral analysis of the Schrödinger operator with magnetic field and semiclassical parameter, on a bounded regular domain in dimension two, with Neumann boundary condition. We investigate the case when the magnetic field vanishes along a union of smooth curves. The aim is to understand the influence of the cancellation and to study the behaviour of the lowest eigenvalues and the associated eigenfunctions when the semiclassical parameter tends to 0. In this regime - called the semiclassical limit - the precise description of the eigenpairs requires the understanding of underlying models. In the first part, we consider a magnetic field which vanishes linearly along a smooth simple curve intersecting the boundary. The second part is devoted to the case when the magnetic field vanishes quadratically. In both cases, we firstly give a one term asymptotics of the lowest eigenvalue. The upper bound is obtained by using appropriate test functions whereas the lower bound results from a localisation process. This last aspect constitutes the most difficult part because of the different scales involved. Then we investigate the localisation properties of the first eigenfunctions thanks to semiclassical Agmon estimates. This leads to a full asymptotic expansion of the first eigenvalues. In the case when the magnetic field vanishes quadratically, we study in addition the model operator for which the cancellation set is a union of two straight lines, whose intersection form a non-zero angle. In the small angle regime, the structure of the spectrum is governed by an operator symbol with two parameters. We establish different properties of this symbol and the associated band function. Numerical simulations based on the finite elements library Mélina++ have guided the analysis and illustrate the obtained results. The difficulties of the numerical computations - induced by the high phase oscillations of the eigenfunctions - are circumvented by polynomial interpolation of high degree.

Dans une série de travaux Zied Ammari et Francis Nier ont développé des méthodes pour étudier la dynamique de champ moyen bosonique pour des états quantiques généraux pouvant présenter des corrélations. Ils ont obtenu des formules pour décrire la dynamique des corrélations, ou plus généralement des matrices densité réduites d'ordre arbitraire. Cette thématique a été largement développée ces dernières années. Norbert Mauser en a été un des contributeurs, ainsi que sur la notion de mesure de Wigner qui est la clé de l'analyse développée par Z. Ammari et F. Nier. En général, il est admis que l'asymptotique de champ moyen est une bonne approximation du problème à N particules quand N dépasse la dizaine. Cela concerne l'asymptotique de la matrice densité réduite à une particule qui ne décrit pas la dynamique des corrélations. Un objectif est de tester la validité de la dynamique de champ moyen pour les matrices densité réduites à 2-particules. Pour des tests numériques, les modèles discrets qui n'ont pas été vraiment traités en détail dans les travaux précédents de Z. Ammari et F. Nier semblent bien adaptés. La thèse comprendra donc plusieurs étapes: adapter les résultats précédents de Z. Ammari et F. Nier à des modèles discrets , développer des méthodes numériques pour des systèmes simples mais pertinents, permettant de valider l'approximation de champ moyen et les formules pour la dynamique des corrélations. Au niveau numérique, on utilise des schémas numériques symplectiques, développés spécifiquement ces dernières années pour la discrétisation des équations hamiltoniennes. Une dernière étape concerne la combinaison des deux asymptotiques, champ moyen et approximation des modèles continus par les modèles discrets. / In a serie of works Z. Ammari and F. Nier developed methods to study the dynamics of bosonic mean field for general quantum states which can present correlations. They obtained formulas to describe the dynamics of the correlations, or more generally reduced density matrices with an arbitrary order. This topic was widely developed these last years. N.J. Mauser was one of contributors, as well as on the notion of Wigner measure which is the key of the analysis developed by Z. Ammari and F. Nier. Generally, the mean field asymptotic is admitted is a good approximation of the N-body problem when N exceed about ten. It concerns the asymptotics of the reduced density matrices for one particle which does not describe the dynamics of the correlations. An objective is to test the validity of the mean field dynamics for reduced density matrices for 2 particles. For numerical tests, the discrete models which were not really handled in detail in the previous works of Z. Ammari and F. Nier seem adapted well. The thesis will thus include several steps: adapt the previous results from Z. Ammari and F. Nier to discrete models , develop numerical methods, for simple but relevant systems, allowing to validate the approximation of mean field and the formulas for the dynamics of the correlations. About numerics, symplectic numerical scheme are used, developed specifically these last years for the discretization of the hamiltonian equations. A last possible step concerns the combination of both asymptotics, that is mean field and approximation of the continuous models by the discrete models.