In this work, I present nuclear probe spectroscopy studies, in detail, µ+SR and 57Fe-Mössbauer spectroscopy on solid-state systems with localized magnetic moments of 3d transition-metal ions supported by density functional theory calculations. Local probes are able to extract local quantities, e.g. the spin dynamics of the 57Fe site or the local, mostly interstitial µ+ site to distinguish between di_erent magnetic phases. The density functional theory calculations help to identify the muon site position from which the local quantity depends. My µ+SR studies on frustrated 3d magnets with quenched disorder concern the physics of phase transitions, avoided order-by-disorder, quantum uctuations or the appearance of spin-liquid-by-disorder. µ+SR is able to identify quantum spinliquid-like ground states without symmetry breaking or static magnetic order by the magnetic field at the muon site. BaTi0.5Mn0.5O3 is a magnetically highly-frustrated double perovskite with quenched disorder.It shows no freezing temperature or no frequency dependence of x1as expected for a spin glass. Microscopically, it is proposed that local interactions between magnetic orphan spins, dimers, and magnetic trimers of Mn4+ play an important role. The µ+SR experiment on BaTi0.5Mn0.5O3 shows an increase of the dynamical muon spin relaxation rate below 3 K which saturates down to 0.019 K coexisting with residual short-range magnetic order (<20% of the signal). A clear difference is observed in comparison with the classical cluster-spin glass SrTi0.5Mn0.5O3 which shows a peak of the zero-field muon spin relaxation rate: a persistent low-energy spin dynamics is present in BaTi0.5Mn0.5O3 down to 20 K. My DFT calculations propose a positive muon site insight the Ba plane close to O atoms. Here, a slight preference of the muon site close to Mn4+ is possible which could put the muon close the orphan spins, dimers, and magnetic trimers, respectively, avoiding the nonmagnetic Ti4+ face-sharing octahedra. Theoretically, a specific ground state of BaTi0.5Mn0.5O3 is not proposed. A clear discrimination between a quantum spin liquid ground state and a mimicry state with the appearance of spin-liquid-by-disorder is not possible from the existing data. I present a µ+SR study on the bond-disordered magnetically highly frustrated pyrochlore fluoride NaCaCo2F7. Neutron spectroscopy studies on NaCaCo2F7 revealed static short-range order consistent with a continuous manifold of cluster-like states being a superposition of noncoplanar ψ2(m3z2-r2) and coplanar ψ3(mx2-y2) states with a correlation length of around 16Å. No evidence for static magnetic long-range order is found in NaCaCo2F7 probed by µ+SR confirming the absence of an order-by-disorder mechanism. The experimental results are not consistent with a classical local-planar XY cluster-spin glassiness. In these µSR experiments, two muon sites are observed. The relative occupancy of both muon sites is nearly temperature independent. Muon site I is a collinear diamagnetic F-µ+-F bound state pulling two F- close towards the muon revealed by the muon spin time evolution. To investigate the pure F-µ+-F bound state in a broad temperature range I have performed an additional µ+SR study on CaF2. This study solved open questions of muon diffusion around 290 K which was observed in NaCaCo2F7 as well. The F-µ+-F spin relaxation indicates the slowing down of the magnetic Co2+ spin fluctuations upon cooling towards the NMR spin freezing temperature Tf≈ 2.4 K. The relaxation rate saturates below 800 mK and remains constant down to 20 mK. The dominant part of the magnetic short-range relaxation signal is a dynamical relaxation as probed by longitudinal magnetic-field experiments. Muon site II exhibits a strong dynamical relaxation rate at 290 K and below and shows persistent µ+ spin dynamics down to 20 mK. Qualitatively, muon site II shows persistent µ+ spin dynamics with one order of magnitude higher dynamical relaxation rates compared to muon site I. DFT calculations of a comparison of the unperturbed unit cells of NaCaCo2F7 and NaCaNi2F7, which has shown just one muon site experimentally, are consistent with a decrease of the energy differences of energy minima and support the experimentally observed muon site ambivalence. In summary, the µ+SR studies propose NaCaCo2F7 as a quantum cluster-spin glass candidate. I present a systematic 57Fe-Mössbauer study on highly diluted Fe centers in Li2(Li1-xFex)N as a function of temperature and magnetic field applied transverse and longitudinal with respect to the single-ion anisotropy axis. Here, Fe is embedded in an α-Li3N matrix. The oxidation state of Fe and possible ferromagnetic nature are in controversial discussions in the literature. Below 30 K the Fe centers exhibit a giant magnetic hyperfine field of BA=70.25(2) T parallel to the axis of strongest electric field gradient Vzz=-154.0(1) V / Å 2. This observation is consistent with a Fe1+d7 charge state with unquenched orbital moment and J=7/2. Fluctuations of the magnetic hyperfine field are observed between 50 K and 300 K and described by the Blume two-level relaxation model consistent with single-atomic magnetism as proven by the invariance of Blume relaxation parameters for the concentration tuning x< 0.025 excluding a ferromagnetic nature. From the temperature dependence of the fluctuation rate an Orbach spin-lattice relaxation process is deduced. An Arrhenius analysis yields a single thermal-activation barrier of EA=570(6) K and an attempt frequency v0=309(10) GHz. Mössbauer spectroscopy studies with applied transverse magnetic fields up to 5 T reveal a large increase of the fluctuation rate by two orders of magnitude. In longitudinal magnetic fields a splitting of the fluctuation rate into two branches is observed. The experimental observations are qualitatively reproduced by a single-ion spin Hamiltonian analysis. It demonstrates that for dominant magnetic quantum tunneling relaxation processes a weak axial single-ion anisotropy D of the order of a few Kelvin can cause a two orders of magnitude larger energy barrier for longitudinal spin fluctuations.

Cette thèse se situe à la croisée de trois domaines : la spintronique qui s'attache à utiliser le degré de liberté du spin de l'électron afin de fabriquer de nouveaux dispositifs électroniques; l'électronique moléculaire qui cherche à profiter des progrès de la chimie moderne afin de fournir des alternatives au tout semi-conducteur de la micro-électronique; le magnétisme moléculaire qui cherche à synthétiser des aimants moléculaires aux propriétés toujours plus riches. Notre travail a consisté à produire un dispositif électronique à base d'aimant moléculaire et d'utiliser le spin de l'électron afin d'étudier les propriétés magnétiques à l'échelle d'une molécule. Des dispositifs semblables pourraient, dans l'avenir, constituer l'une des briques élémentaires de l'information quantique. Nous avons pour cela opté pour un transistor moléculaire à effet de champ, ayant pour canal un aimant moléculaire aux propriétés magnétiques bien connues : le Terbium double-decker ou TbPc2. Grâce à ce dispositif, nous avons, dans un premier temps, mis en évidence le retournement de l'aimantation d'une molécule unique par effet tunnel ou QTM (quantum tunneling of the magnetization). En effet, nous avons démontré que ce retournement entraînait une modification soudaine de la conductance de notre système. En effectuant une étude statistique sur les valeurs du champ de retournement, nous avons mis en évidence la présence de résonances que nous avons pu attribuer au phénomène de QTM. Nous avons également mesuré l'état d'un spin nucléaire unique : chaque résonance étant associée à un état de spin nucléaire. Nous avons étudié la température du spin nucléaire et montré que celle-ci pouvait être influencée par l'environnement électrostatique du système. En outre, le temps de vie d'un état de spin nucléaire a été extrait et estimé à quelques secondes, vérifiant que le système était faiblement perturbé par notre technique de mesure. Ces travaux jettent les bases de la construction du premier Qbit à base d'aimants moléculaires. Par des techniques de radiofréquence, le spin nucléaire pourrait être manipulé, la lecture se faisant ensuite par une mesure en conductance. / This PhD thesis is at a cross-road between three different fields : the spintronics which uses the spin degree of freedom of the electron to build new devices ; the molecular electronics which tries to take advantage of the new development of the chemistry, to give a workaround to the all semiconductor paradigm of the microelectronics industry; and the molecular magnetism which synthesizes molecular magnet with properties of an increasing richness. Our work has been dedicated to the fabrication of a molecular magnet based electronic device with which we could use the spin of the electron to study the magnetic properties at a single molecule level. Such device could, in the future, be used in the field of quantum information. We have decided to fabricate a field effect molecular transistor in which a well known molecular magnet, the Terbium double-decker or TbPc2, acts as a channel. Thanks to this device, we evidenced the quantum tunnelling of the magnetization (QTM) at single molecule level. We demonstrated that the magnetic moment reversal induces an abrupt change in the differential conductance of the system. By performing a statistical study, we highlighted four resonances that were attributed to QTM. We also measured a single nuclear spin state : each resonance being directly associated with one particular nuclear spin state. We studied the nuclear spin temperature and showed that it could be influenced by the electrostatic environment. Furthermore, the spin state lifetime was assessed and estimated to few seconds, highlighting the low invasive character of our measurement technique. This work give the foundation of the first molecular magnet based Qbit. With radio frequency techniques, the nuclear spin could be manipulated, the readout being performed through conductance measurement.