This thesis reports on the setup of a new ion trap apparatus designed for experiments with single ⁴⁰Ca⁺ ions to perform molecular spectroscopy. The calcium ion is laser cooled, allowing for sympathetic cooling of the nonfluorescing molecular ion. The aim of these experiments is to explore loading and identifying molecular ions in RF-Paul traps, as well as developing new spectroscopic tools to measure transitions of molecular ions via the fluorescence of co-trapped ⁴⁰Ca⁺ ions. Ground state cooling of a mixed ion pair is implemented as a first step towards increasing the sensitivity of our technique to the level necessary to measure transitions with low scattering rates (like those present in molecular ions). Doppler cooling on the S(1/2)->P(1/2) transition of the calcium ion results in the formation of a Coulomb crystal, the behavior of which may be used to infer properties of the molecular ion. Following cooling, sideband spectroscopy on the narrow S(1/2)->D(5/2) quadrupole transition of calcium may be used to identify the mass of single molecular ions. This method is verified via a non-destructive measurement on ⁴⁰CaH⁺ and ⁴⁰Ca¹⁶O⁺. The normal modes of the Coulomb crystal can also be used to extract information from the target ion to the control ion. By driving the blue side of a transition, laser induced heating can be put into the two ion system, which leads to changes in fluorescence of the ⁴⁰Ca⁺ ion, first demonstrated with two Ca⁺ isotopes. Increasing the sensitivity of this technique requires ground state cooling of both the ⁴⁰Ca⁺ ion and the ion of interest, enabling the transfer of the ion's motional state into the ground state with high probability. This thesis demonstrates ground state cooling of the atomic ion and sympathetic cooling of a second ion (⁴⁴Ca⁺). Once in the ground state, heating of the Coulomb crystal by scattering photons off of the spectroscopy ion can be measured by monitoring the resolved motional sidebands of the S(1/2)->D(5/2) transition of ⁴⁰Ca⁺, allowing for spectral lines to be inferred. Future experiments will investigate this technique with molecular ions.

With only two ingredients, atomic nuclei exhibit a rich structure depending on the ordering of the different proton- and neutron-occupied states. This ordering can give rise to excited states with exceptionally long half-lives, also known as isomers, especially near shell closures. On-line mass spectrometry can often be compromised by the existence of such states that may even be produced in higher proportion than the ground state. This thesis presents the first results obtained from a nuclear spectroscopy setup coupled with the high-resolution Penning-trap mass spectrometer ISOLTRAP, at CERN’s radioactive ion beam facility ISOLDE. The isomerism in the neutron-deficient thallium isotopes was investigated. The data on 184,190,193−195 Tl allow an improvement of existing mass values as well as a mass-spin-state assignment in 190,193,194 Tl. Due to the presence of the ground and isomeric state for 194 Tl the excitation energy of the latter was determined for the first time experimentally. Systematic trends in the vicinity of the Z = 82 shell closure have been discussed.

Les noyaux les plus complexes sont situés entre les nombres magiques et les médianes des espaces de valence, dans des régions connues pour les changements abrupts des observables nucléaires. Dans ces régions appelées de transition de forme, le paradigme nucléaire change entre la goutte liquide vibrationnelle et le rotor statique. Sauf quelques exceptions, les noyaux de ces régions sont radioactifs, avec des demi-vies qui chutent dans les millisecondes. Complémentaires aux propriétés des états excités à basse énergie, les énergies de liaison et les rayons de charge nucléaires sont parmi les observables les plus sensibles à ces changements de structure nucléaire. Dans ce travail, une étude du phénomène de transition de forme est effectuée, par des mesures de nucléides radioactifs produits dans le laboratoire ISOLDE au CERN. Les masses des isotopes de rubidium riches en neutrons 98-100Rb et des isotopes d’or riches en protons 180,185,188,190,191Au sont mesurées avec le spectromètre de masse de type Penning ISOLTRAP. La masse de 100Rb est déterminée pour la première fois. Des déviations significatives par rapport à la littérature sont trouvées pour les isotopes 188,190Au. Une nouvelle méthode expérimentale est présentée, utilisant un spectromètre de masse à multi-réflexion comme analyseur de faisceau pour la spectroscopie laser d’ionisation résonante. La nouvelle méthode donne la possibilité d’effectuer des études de structure hyperfine atomique avec ISOLTRAP, dont on peut extraire les rayons de charge et les moments électromagnétiques nucléaires. / The most complex nuclei are situated between the magic and the mid-shell ones, in regions known for sudden changes of the trends of nuclear observables. These are the so-called shape-transition regions, where the nuclear paradigm changes from the vibrational liquid drop to the static rotor. With few exceptions, nuclei in these regions are radioactive, with half-lives dropping into the millisecond range.Complementing the information obtained from the low-lying excitation spectrum, nuclear binding energies and mean-square charge radii are among the observables most sensitive to these changes of nuclear structure. In the present work, a study of the shape-transition phenomenon is performed by measurements of radioactive nuclides produced by the ISOLDE facility at CERN. The masses of the neutron-rich rubidium isotopes 98−100Rb and of the neutron-deficient gold isotopes 180,185,188,190,191Au are determined using the Penning-trap mass spectrometer ISOLTRAP. The mass of 100Rb is determined for the first time. Significant deviations from the literature values are found for the isotopes 188,190Au. A new experimental method is presented, using a recently developed multi-reflection time-of-flight mass spectrometer as a beam-analysis tool for resonance-ionization laser spectroscopy. The new method opens the path to measurements of atomic hyperfine spectra with ISOLTRAP, from which charge radii and electromagnetic moments of radioactive nuclides can be extracted. The properties of the studied nuclides map the borders of two prominent regions of quadrupole deformation, which constrain the fine balance between pairing and quadrupole correlations in the nuclear ground states. This balance is studied by the Hartree-Fock- Bogoliubov (HFB) approach. The sensitivity of the shape-transition phenomenon to the strength of pairing correlations is demonstrated. In particular, the strong odd-even staggering of charge radii in the mercury isotopic chain is shown to result in the HFB approach from the fine interplay between pairing, quadrupole correlations and quasi-particle blocking.

La future installation SPIRAL2 du GANIL, à Caen, permettra de produire une gamme étendue de noyaux exotiques avec des intensités très importantes. Cependant, ces faisceaux ne pourront pas être directement utilisés pour réaliser certaines études de haute précision et devront d'abord être purifiés. C'est pour réaliser ce travail que des équipes du CENBG, du MPIK (Heidelberg), du CSNSM, du LPC Caen, du GANIL et de l'IPNO développent un double-piège de Penning dans le cadre du projet PIPERADE. Ce double-piège nécessitant un travail de mise en forme préalable du faisceau (diminution de l'émittance transverse et mise en paquet), le CENBG est en charge de la réalisation d'un refroidisseur-regroupeur quadripolaire radiofréquence : le GPIB. C'est ce développement qui a constitué la majeure partie de mon travail de thèse, notamment en ce qui concerne les simulations de ce refroidisseur-regroupeur dont les résultats ont permis de trouver une méthode innovante pour la mise en paquet et de valider sa conception mécanique. Nous disposons également d'une source d'ions afin de pouvoir tester le GPIB et le double-piège et il m'a fallu la remonter, la comprendre et la caractériser pour que ces tests soient ensuite possibles. Dans un second temps j'ai aussi participer à l'analyse de l'expérience E622S menée au GANIL et qui avait pour but de déterminer précisément la demi-vie de 17F. Ce travail n'a pas permis d'améliorer la précision sur la demi-vie de 17F mais nous avons cependant quelques doutes sur les 2 précédentes mesures et sur leur détermination du taux de contamination de leurs échantillons. / The future SPIRAL2 installation of GANIL, at Caen, will produce large range of exotic nuclei with very high intensities. Nevertheless, these beams could not be used directly for some high precision studies and will have to be purified first. To achieve this work, teams of CENBG, MPIK (Heidelberg), CSNSM, LPC Caen, GANIL and IPNO develop a double Penning trap in the framework of the PIPERADE project. Because this double Penning-trap needs some shaping work (reduction of transverse emittance and bunching), the CENBG team is in charge of the realisation of a radio-frequency quadrupole cooler and buncher : the GPIB. This developpement work was the main part of my PhD work, especially for the simulations of this cooler buncher whose results lead us to find a new bunching method and allowed us to approve its mechanical design. We also have an ion source to be able to test both GPIB and Penning trap and I had to reassemble it, to understand it and to characterise it before these tests could be achieved. In a second time I also took part to the analysis of the E622S experiment which aimed to determined precisely the 17F half-life. This work did not lead to a more precise determination of this half-life but we now have some doubts concerning the 2 last measurements and their way to determine the contamination rate of their radioactive samples.

The subject of this thesis is high-precision mass-measurements performed with Penning trap mass spectrometers (PTMS). In particular it describes the SMILETRAP I PTMS and the final results obtained with it, the masses of 40Ca and that of the proton. The mass of 40Ca is an indispensible input in the evaluation of measurements of the bound electron g-factor, used to test quantum electrodynamical calculations in strong fields. The value obtained agrees with available literature values but has a ten times higher precision. The measurement of the proton mass, considered a fundamental physical constant, was performed with the aim of validating other Penning trap results and to test the limit of SMILETRAP I. It was also anticipated that a measurement at a relative precision close to 10-10 would give insight in how to treat certain systematic uncertainties. The result is a value of the proton mass in agreement with earlier measurements and with an unprecedented precision of 1.8×10-10. Vital for the achieved precision of the proton mass measurement was the use of the Ramsey excitation technique. This technique, how it was implemented at SMILETRAP I and the benefits from it is discussed in the thesis and in one of the included papers. The second part of the thesis describes the improved SMILETRAP II setup at the S-EBIT laboratory, AlbaNova. All major changes and upgrades compared to SMILETRAP I are discussed. This includes, apart from the Ramsey excitation technique, higher ionic charge states, improved temperature stabilization, longer run times, different reference ions, stronger and more stable magnetic field and a more efficient ion detection. Altogether these changes should reduce the uncertainty in future mass determinations by an order of magnitude, possibly down to 10-11. / At the time of the doctoral defense, the following paper was unpublished and had a status as follows: Paper 9: Accepted.

La thèse porte sur l’étude des supernovae à effondrement de coeur (CCSN). D’abord, nous avons cherché à modéliser la composition du coeur de l’étoile massive lors de son effondrement. Nous avons élaboré un nouveau traitement de l’équilibre statistique nucléaire à partir d’une équation d’état basée sur l’approximation de noyau unique (Lattimer et Swesty, LS). Cela permet d’obtenir une description plus réaliste de la distribution des noyaux constitutifs du coeur et de quantifier plus spécifiquement le rôle des masses nucléaires. Les distributions obtenues avec la fonctionnelle de masse originelle (LS) et celles obtenues à partir des modèles de masse tels que HFB-24 et DZ10 ont été comparées pour plusieurs conditions thermodynamiques d’une trajectoire typique de CCSN. Les différences de composition pouvant conduire à des déviations sur le taux de capture électronique allant jusqu’à ∼25%, il est apparu important de définir un modèle de masse réaliste à utiliser pour la simulation de CCSN. Pour cela, nous avons réalisé des mesures de masse de précision dans la région d’intérêt, avec un double piège de Penning auprès de l’installation IGISOL (située à Jyväskylä, Finlande). Cinq nouveaux excès de masse ont été déterminés pour les noyaux suivants : 69m,70Co, 74,75Ni et 76mCu. La précision a été améliorée pour cinq autres : 67Fe, 69Co, 76,78Cu, 79mZn. Enfin, nous avons confirmé les valeurs obtenues par des études récentes pour le 77Cu et le 79Zn. Dès lors, les valeurs expérimentales des gaps nucléaires Z=28 et N=50 ont été comparées aux résultats prédits par DZ10 et HFB-24. Ce dernier modèle reproduit mieux l’évolution des gaps en question. De fait, HFB-24 a été utilisé dans notre nouveau traitement statistique, que nous avons implémenté dans une simulation hydrodynamique existante de CCSN. Au final, nous avons observé un impact modéré du modèle de masse sur la composition du coeur en effondrement. Par ailleurs, nous avons montré que ces différences de composition ont peu d’effet sur la dynamique d’effondrement, celle-ci semble plus affectée par le taux de capture électronique. Les études futures devraient se focaliser plutôt sur ce paramètre. / The thesis sets itself in the framework of the study of core-collapse supernovae (CCSN). First, the modeling of the composition of the core of a massive star during its collapse has been investigated. To this aim, we have built a new treatment of the nuclear statistical equilibrium starting from a single-nucleus approximation equation of state (Lattimer and Swesty, LS). This allows a more realistic description of the nuclear distribution inside the core and, more specifically, to quantify the role of the nuclear masses. The distributions obtained with the original mass functional (LS) and those obtained with HFB-24 and DZ10 mass models have been compared for several thermodynamic conditions of a typical CCSN trajectory. The differences in the composition could lead up to ∼25% deviations in the electron-capture rate, thus showing the need to identify a proper mass model to use in CCSN simulations. Therefore, we performed high precision mass measurements in the nuclear mass region of interest, via a double Penning trap at the IGISOL facility (Jyväskylä, Finland). Five new mass excess were determined for the following nuclei : 69m,70Co, 74,75Ni and 76mCu. The precision has been improved for five others : 67Fe, 69Co, 76,78Cu and 79mZn. Finally, we have confirmed the values obtained by recent studies for 77Cu and 79Zn. The experimental values of the nuclear gaps for Z=28 and N=50 have been compared with the results predicted by DZ10 and HFB-24. The latter model better reproduces the evolution of these gaps. Therefore, HFB-24 was used in our new statistical treatment, that we implemented in an existing CCSN hydrodynamical simulation. We have observed a moderated impact of the mass model on the composition of the collapsing core. Moreover, we found that the differences in composition have small effect on the collapse dynamics, which appears to be more sensitive to the electron-capture model. Further studies should thus focus on this parameter.