Neste trabalho estudamos algumas interações entre átomos de Rydberg em uma armadilha ótica de dipolo do tipo QUEST. Com esta armadilha obtemos uma amostra de 1,2 × 106 átomos de 85Rb no estado fundamental, à uma densidade de ∼ 1012 átomos/cm3 e temperatura média de 60 µK. Os átomos de Rydberg foram preparados utilizando uma transição de dois fótons a partir do estado fundamental 5S1/2, passando pelo estado intermediário 5P3/2 e em seguida para o estado de Rydberg desejado. Estudamos a interação entre pares de átomos, em diferentes níveis energéticos, através de duas técnicas diferentes. Na primeira, monitoramos os efeitos de transferência de população em estados nD oriundas de uma ressonância Förster. Estudamos a ressonância nD5/2 + nD5/2 → (n+2)P3/2 + (n2)F7/2 onde 37 ≤ n ≤ 47, em função da densidade da amostra atômica; e para o estado 37D5/2 manipulamos a mesma ressonância com a aplicação de um campo elétrico externo. Os resultados mostraram que este é um processo binário, evidenciado pela dependência quadrática da população no estado produto com relação ao estado excitado. Num outro experimento, estudamos a interação entre estados nS através do monitoramento, e controle, do efeito de bloqueio de excitação. Aqui mostramos que é possível controlar as interações interatômicas em uma amostra quasi-unidimensional de átomos de Rydberg variando a orientação de um campo elétrico externo. Mostramos que ao polarizar uma amostra de átomos no estado 50S1/2 com um campo, esta passa a interagir de modo semelhante ao de dipolos elétricos clássicos, onde a interação pode ser controlada com a variação da orientação dos dipolos atômicos. Tal interação pode, inclusive, ser cancelada quando os dipolos elétricos são alinhados em um ângulo de 54,7° com relação ao eixo internuclear. / In this work we have studied some interactions between Rydberg-atoms in a QUEST type optical dipole trap. With this trap we obtained a sample of 1,2 × 106 85Rb atoms in the ground state, in a density of ∼ 1012 atoms/cm3 and average temperature of 60 µK. The Rydberg-atoms were prepared using a two-photon transition from the ground state 5S1/2, through an intermediate state 5P3/2 and then to the desired Rydberg state. We have studied interactions between pairs of atoms at several energy levels, using two different techniques. In the first one, we have monitored the effects of the population transfer in nD states derived from a Förster resonance. We have studied the resonance nD5/2 + nD5/2 → (n + 2)P3/2 + (n 2)F7/2 for states of 37 ≤ n ≤ 47 as a function of of the samples atomic density. For the 37D5/2 state we have also manipulated the same resonance with the application of an external electric field. Our results have shown that this is a binary process, indicated by the quadratic dependence of the transferred population in relation with the excited state. In another experiment, we have studied the interaction between nS1/2 states by monitoring, and controlling, the excitation blockade effect. Here we have shown that it is possible to control the inter-atomic interactions in a quasi-one-dimensional sample of Rydberg-atoms by varying the orientation of an external electric field. We have demonstrated that when polarizing a sample of atoms, in the 50S1/2 state with a field, it starts to interact in a similar way as classic electric-dipoles, where the interaction can be controlled by varying the orientation of the atomic dipoles. Such interaction may even be canceled when the electric dipoles are aligned at an angle of 54,7° related to the internuclear axis.

L’expérience I.C.E a pour objectif de tester le principe d’équivalence faible (WEP) à la base de la théorie de la relativité générale d’Einstein et postulant l’équivalence entre masse inertielle et masse grave. Si ce principe a toujours été vérifié jusqu’à aujourd’hui, il est d’un intérêt fondamental pour la physique moderne de poursuivre les mesures avec une précision accrue. En effet, de nouvelles théories d’unification de la mécanique quantique et de la relativité générale prévoient une violation de ce principe. Pour réaliser un test du WEP, il suffit de comparer les accélérations de deux objets en chute libre dans un même champ de gravitation, et c’est ce que réalise l’expérience I.C.E à l’échelle quantique (à la différence de la mission spatiale Microscope qui à ce jour a pu vérifier le WEP avec des objets macroscopiques avec une sensibilité sur le paramètre de 2.10−14). Ainsi, l’expérience consiste à réaliser, par une méthode interférométrique, la mesure de l’accélération de deux espèces atomiques (87Rb et 39K) de masses et de compositions différentes, en chute libre dans une enceinte à vide. La sensibilité de la mesure des effets inertiels auxquels les atomes sont sensibles (accélérations et rotations) est d’autant plus grande que la durée de chute libre des atomes est élevée et que la température des nuages est faible. Or, sur Terre au laboratoire, les atomes finissent par tomber au fond de l’enceinte les contenant sous l’effet de la gravité, ce qui limite grandement la sensibilité de la mesure. C’est pourquoi il est intéressant de placer l’expérience dans un environnement de micropesanteur dans lequel les atomes restent au centre de la chambre à vide afin d’atteindre des temps d’interrogation beaucoup plus longs. A ce titre, l’expérience est embarquée jusqu’à plusieurs fois par an, à bord de l’avion Zéro-g de la société Novespace. Les durées de micropesanteur proposées permettent d’atteindre des temps d’interrogation théoriques de l’ordre de la seconde ce qui doit porter le niveau de sensibilité à 10−11. Cependant, nous sommes aujourd’hui très fortement limités par le niveau élevé de vibrations et de rotations de l’avion : la perte de contraste des franges d’interférence engendrée ainsi que le bruit de phase introduit, ne nous permettent pas de dépasser des temps d’interrogation de 5 ms en 0 g. En parallèle, le laboratoire s’est récemment doté d’un simulateur de microgravité sur lequel est montée l’expérience, donnant accès à des temps d’interrogation de plus de 200 ms avec des trajectoires paraboliques d’une très bonne répétabilité (de l’ordre de 3 mg). La cohérence d’une source atomique étant directement reliée à sa température, l’utilisation de nuages ultra-froids est d’un grand intérêt pour améliorer le contraste des franges d’interférence, d’autant plus pour les longs temps d’interrogation visés. Le présent manuscrit synthétise les travaux ayant permis de produire le tout premier condensat de Bose-Einstein (la source atomique ultime) de 87Rb en microgravité par une méthode tout optique, et ce, de manière répétable toutes les 13,5 secondes. Nous démontrons l’efficacité de note méthode de chargement du piège dipolaire basée sur l’association d’un refroidissement par mélasse grise et d’une modulation spatiale des faisceaux dipolaires. Ces résultats ouvrent la voie vers de futures mesures interférométriques très sensibles à grand facteur d’échelle. / The I.C.E experiment aims at testing the weak equivalence principle (WEP) underlying Einstein’s theory of general relativity and which postulates the equivalence between inertial mass and gravitationnal mass. If this principle has always been verified until today, it is of fundamental interest for physics to continue the measurements with greater precision. Indeed, new unifying theories of quantum mechanics and general relativity predict a violation of this principle. To carry out a test of the WEP, it suffices to compare the accelerations of two objects in free fall in the same gravitationnal field. This is what the I.C.E experiment, on the quantum scale, achieves (unlike the spatial Microscope mission, which to date has been able to verify the principle of equivalence with macroscopic objects with a sensitivity on of 2.10−14). Thus, the experiment consists in performing, by an interferometric method, the measurement of the acceleration of two atomic species (87Rb and 39K) of different mass and composition in free fall in a vacuum chamber. The measurement sensitivity of the inertial effects to which the atoms are sensitive (accelerations and rotations) is all the greater as the free fall time of the atoms is high and their temperature is low. But on Earth, in the laboratory, the atoms eventually fall to the bottom of the vacuum chamber containing them under the effect of gravity, which greatly limits the measurement sensitivity achievable. This is why it is interesting to place the experiment in a microgravity environment in which the atoms stay in the center of the vacuum chamber in order to reach much longer interrogation times. As such, several times a year, the experiment is put aboard the aircraft Zero-g of the Novespace company. The available microgravity durations make it possible to reach theoretical interrogation times of the order of one second, which should raise the sensitivity level to 10−11. However, we are today very strongly limited by the high level of vibrations of the aircraft as well as its rotations : the loss of contrast of the interference fringes and the phase noise caused, do not allow us to exceed 5 ms of interrogation times in 0 g. Since the coherence of an atomic source is directly related to its temperature, the use of ultra-cold clouds is of great interest to improve the contrast of the interference fringes, especially for the long interrogation times targeted. In parallel, the laboratory is now equipped with a microgravity simulator on which is mounted the experiment, giving access to interrogation times of more than 250 ms with parabolic trajectories of a very good repeatability (of the order of 3 mg). This manuscript synthesizes the work that produced the very first 87Rb Bose-Einstein condensate in microgravity by all-optical methods, with a repetition rate of 13,5 seconds. We demonstrate the efficiency of our dipole trap loading method based on the association of a grey molasses cooling and a spatial modulation of the dipole beams. These results pave the way for future highly sensitive interferometric measurements with a large scale factor.

Cette thèse traite de la manipulation d'atomes de rubidium 87 piégés dans un piège optique dipolaire microscopique. Les expériences sont réalisées dans divers régimes de chargement du piège allant d'un atome unique à quelques milliers d'atomes en moyenne.Le régime à un seul atome permet de calibrer le dispositif expérimental. Nous utilisons l'atome unique comme bit quantique dont nous pouvons préparer et lire l'état avec une efficacité de 99.97% et 98.6%, respectivement. Lorsque plusieurs atomes sont chargés dans le piège microscopique, nous observons une distribution sub-Poissonienne du nombre d'atomes, liée aux collisions assistées par la présence de lumière quasi résonante. Une étude de ces collisions dans notre cas particulier (piège microscopique) révèle des taux de pertes extrêmement élevés, proches de la limite théorique de Langevin. Enfin, nous montrons que le chargement du piège microscopique avec plusieurs atomes est plus efficace lorsque nous superposons sur ce piège un deuxième piège, macroscopique, qui joue le rôle de réservoir d'atomes. Ce réservoir permet de charger le micro-piège à partir du macro-piège en l'absence de lumière quasi résonante et donc d'éviter les collisions assistées par la lumière.Le chargement du micro-piège à partir du macro-piège conduit à des conditions initiales optimales pour l'évaporation forcée dans la perspective d'atteindre la condensation de Bose-Einstein avec seulement une dizaine d'atomes. Après évaporation du gaz nous atteignons des densités dans l'espace des phases proches du régime de dégénérescence. / This thesis focuses on the manipulation of rubidium 87 atoms in a microscopic optical dipole trap. The experiments are performed in various regimes where the number of atoms in the microscopic trap ranges from exactly one atom to several thousands on average.The single atom regime allows us to calibrate the experimental setup. We use it a quantum bit, which state we can prepare and read out with efficiencies of 99.97% and 98.6%, respectively. When several atoms are loaded in the microscopic trap we observe a sub-Poissonian distribution of the number of atoms due to light-assisted collisions in the presence of near-resonant light. A study of these collisions in our particular case (microscopic trap) reveals extremely high loss rates approaching the theoretical Langevin limit. Finally, we demonstrate that the loading of the microscopic trap is more efficient when we superimpose on this trap a second macroscopic trap, which we use as an atom reservoir. This reservoir allows us to load the micro trap from the macro trap in the absence of any near-resonant light, thus avoiding light-assisted collisions.The loading of the micro trap from the macro trap leads to optimal initial conditions for forced evaporation towards Bose-Einstein condensation with about ten atoms only. After evaporation we reach phase-space densities approaching the degenerate regime.

The quantum properties of matter waves, in particular quantum correlations and entanglement are an important frontier in atom optics with applications in quantum metrology and quantum information. In this thesis, we report the first observation of sub-Poissonian fluctuations in the magnetization of a spinor 87Rb condensate. The fluctuations in the magnetization are reduced up to 10 dB below the classical shot noise limit. This relative number squeezing is indicative of the predicted pair-correlations in a spinor condensate and lay the foundation for future experiments involving spin-squeezing and entanglement measurements. We have investigated the limits of the imaging techniques used in our lab, absorption and fluorescence imaging, and have developed the capability to measure atoms numbers with an uncertainly < 10 atoms. Condensates as small as ≈ 10 atoms were imaged and the measured fluctuations agree well with the theoretical predictions. Furthermore, we implement a reliable calibration method of our imaging system based on quantum projection noise measurements. We have resolved the individual lattice sites of a standing-wave potential created by a CO2 laser, which has a lattice spacing of 5.3 µm. Using microwaves, we site-selectively address and manipulate the condensate and therefore demonstrate the ability to perturb the lattice condensate of a local level. Interference between condensates in adjacent lattice sites and lattice sites separated by a lattice site are observed.

In dieser Arbeit untersuchen wir kollektive Strahlungseffekte in Nanofaser-gekoppelten atomaren Ensembles, die sich über Tausende von optischen Wellenlängen erstrecken. Wir koppeln bis zu 1000 Atome optisch an die geführten Moden einer optischen Nanofaser, die langreichweitige Dipol-Dipol Wechselwirkungen zwischen den Atomen vermittelt. Wir realisieren eine unidirektionale Kopplung und damit ein kaskadiertes Quantensystem, in dem die Dynamik jedes Atoms ausschließlich durch die Dynamik der vorgelagerten Atome bestimmt wird. Wir regen die Atome mit nanofasergeführten optischen Pulsen kohärent an, was uns ermöglicht, den gesamten Parameterbereich von schwacher Anregung bis hin zur voll-ständigen Inversion zu erforschen. Wir stellen fest, dass die kohärente Vorwärtsstreuung, die für die Superradianz im Regime der schwachen Anregung verantwortlich ist, auch nahe voller Inversion eine wichtige Rolle für die Dynamik spielt. Wir beobachten superradiante Puls-Dynamik, die in unserem System trotz des makroskopischen Abstands zwischen den Atomen und einer asymmetrischen Kopplung auftritt. Wir stellen fest, dass die emittierte Spitzenleistung noch schneller mit der Anzahl der Atome skaliert als im Fall der idealen Dicke Superradianz, was auf eine kollektiv erhöhte Sammeleffizienz der nanofasergeführten Mode zurückzuführen ist. Die Analyse der Kohärenz-Eigenschaften des superradianten Pulses erlaubt es uns, zwei Regime der Puls-Dynamik zu identifizieren. Wir entwickeln ein kaskadiertes Wechselwirkungsmodell und zeigen, dass es die kollektive Dynamik unseres Systems über den gesamten in dieser Arbeit untersuchten Parameterbereich akkurat beschreibt. Schließlich untersuchen wir die getriebene Dynamik eines Nanofaser-gekoppelten Ensembles von Drei-Niveau-Atomen. Wir treiben Zwei-Photonen-Rabi-Oszillationen zwischen den beiden Grundzuständen eines $\Lambda$-Systems und beobachten die damit verbundene oszillatorische Raman-Verstärkung und -Absorption. / In this thesis, we study collective radiative effects in nanofiber-coupled atomic ensembles that extend over thousands of optical wavelengths. We optically couple up to 1000 atoms to the guided modes of an optical nanofiber, which mediates long-range dipole-dipole interactions between the atoms. We engineer the coupling to be unidirectional, realizing a cascaded quantum system in which the dynamics of each atom is solely determined by the dynamics of upstream atoms. We coherently excite the atoms using nanofiber-guided optical pulses, allowing us to explore the entire parameter regime from weak excitation to full inversion. We find that coherent forward scattering, which is responsible for superradiance in the weak excitation regime, plays an important role for the dynamics even close to full inversion. We observe superradiant burst dynamics, which occurs in our system despite the macroscopic separation between the atoms and an asymmetric coupling. We find that the peak-emitted power scales even faster with the number of atoms than in the case of ideal Dicke superradiance due to a collectively enhanced channeling efficiency into the nanofiber-guided mode. By analyzing the coherence properties of the superradiant burst, we directly identify two regimes of burst dynamics. In the second regime, there is no initial coherence, and the superradiant burst is seeded by vacuum fluctuations. We introduce a cascaded interaction model and find that it accurately describes the collective dynamics of our system over the entire parameter regime explored in this thesis. Finally, we study the driven dynamics of a nanofiber-coupled ensemble of three-level atoms. We drive two-photon Rabi oscillations between the two ground states of a $\Lambda$ system and observe the associated oscillatory Raman gain and absorption.