In der vorliegenden Arbeit wurden die mikromagnetische Struktur sowie das Ummagnetisierungsverhalten epitaktisch gewachsener MnAs Filme auf dem Substrat GaAs untersucht. Im Mittelpunkt steht die mikromagnetischen Struktur von anisotrop erspannten MnAs Filmen auf GaAs(001). Die Verspannung führt zur selbstorganisierten Anordnung ferromagnetischer Streifen. Ihre Domänenstruktur wurde mittels MFM (magnetischer Kraftmikroskopie) bestimmt und mit den Resultaten der XMCDPEEM (X-ray magnetic circular dichroism photoemission electron microscopy) verglichen. Um eine vollständige Charakterisierung der mikromagnetischen Eigenschaften der Streifenstruktur zu erreichen, wurden die MFM Experimente in einem äusseren Magnetfeld durchgeführt. Die Beantwortung der zentralen Frage nach der Domänenstruktur ist mit der Entwicklung eines mikromagnetischen Simulators für dreidimensionale magnetische Strukturen auf mesoskopischer Skala gelungen. Die Stabilität der dreidimensionalen mikromagnetischen Struktur hängt von den Eigenschaften der selbstorganisierten Streifenstruktur ab, d.h. sowohl von der Filmdicke als auch vom Verhältnis ihrer Breite zur Filmdicke - und damit der Temperatur. Durch die Erkenntnis, dass eine magnetische Struktur in der Tiefe des Streifens vorhanden ist, können die verbleibenden Unterschiede in den XMCDPEEM- und MFM-Resultaten erklärt werden. Durch die Simulationsergebnisse in Kombination mit den Experimenten wird eine widerspruchsfreie Deutung der mikromagnetischen Struktur sowie deren Ummagnetisierungsverhalten ermöglicht. Zudem wird die mikromagnetische Struktur von MnAs auf GaAs(111) simuliert und damit das Verständnis der mikromagnetischen Strukturen auf alle vorhandenen Substratorientierungen vervollständigt. / In the present thesis the micromagnetic structure, as well as the magnetization reversal, of epitaxial MnAs films on GaAs substrates are studied. The investigation is focused on the micromagnetic structure of anisotropically strained MnAs films on GaAs(001). The strain originates a selforganized array of ferromagnetic stripes. The magnetic domains were investigated using MFM (magnetic force microscopy) and the results were compared with XMCDPEEM (X-ray magnetic circular dichroism photoemission electron microscopy). To completely characterize the micromagnetic properties of the stripe structure, MFM experiments were performed in the presence of an external field. To unambiguously determine the domain structure a three-dimensional micromagnetic simulator was developed capable to calculate magnetic structures with mesoscopic dimensions. The stability of the three-dimensional micromagnetic structure depends on the properties of the selforganized stripe structure, i.e., on the film thickness as well as on the ratio of the stipe width to thickness - and thus the temperature. Taking into account the magnetization distribution in-depth, the remaining differences between the XMCDPEEM and the MFM results can be explained by the disturbing effect of the MFM tip. The results of the micromagnetic simulations, in combination with the experimental results, allow for a determination of the micromagnetic structure in an applied field throughout the phase coexistence regime. Moreover, the micromagnetic structure of MnAs films on GaAs(111) is simulated and thus the understanding of the micromagnetic properties have been extended on all substrate orientations.

Iron filled carbon nanotubes (FeCNT) can be described as carbon nanotubes which contain an iron nanowire of several micrometers length and a diameter of approximately 10-100 nm. The carbon shells protect the iron core from oxidation and mechanical damage thus enabling a wide range of applications that require a long-term stability. The magnetic properties of the enclosed nanowire are in part determined by its small size and elongated shape. Magnetic force microscopy (MFM) measurements show that the iron wire exhibits a single domain behavior. Due to the large shape anisotropy it is magnetized along the long wire axis in the remanent state. Two magnetic monopoles of opposing polarity are located at the wire extremities. Depending on the structure and geometry of the individual nanowire, switching fields in the range of 100-400 mT can be found when the external field is applied along the FeCNT’s easy axis. Cantilever magnetometry shows that the switching can be attributed to a thermally assisted magnetization reversal mechanism with the nucleation and propagation of a domain wall. The defined magnetic properties of individual FeCNT combined with their mechanical strength make them ideal candidates for an application as high resolution high stability MFM probes. The fabrication of such probes can be achieved with the help of a micromanipulation setup in a scanning electron microscope. FeCNT MFM probes achieve a sub 25 nm lateral magnetic resolution. MFM measurements with FeCNT MFM probes in external fields show that the magnetization of these probes is exceptionally stable compared to conventional coated MFM probes. This greatly simplifies the data evaluation of such applied field MFM measurements. The emphasis of this work was put on the calibration of FeCNT probes to enable straightforward quantitative MFM measurements. The defined shape of the magnetically active iron nanowire allows an application of a point monopole description. Microscale parallel current carrying lines that produce a defined magnetic field are used as calibration structures to determine the effective magnetic moment of different MFM probes. The line geometry is varied in order to produce multiple magnetic field decay lengths and investigate the influence on the effective probe moment. The results show that while the effective magnetic monopole moment of a conventional MFM probe increases with an increasing sample stray field decay length, the effective moment of a FeCNT MFM probe remains constant. This enables a MFM probe calibration that stays valid for a large variety of magnetic samples. Furthermore, the fitted monopole moment of a FeCNT probe (in the order of 10E-9 Am) is consistent with the moment calculated from the nanowire geometry and the saturation magnetization of iron.

La radiographie X, synchrotron et avec une source de laboratoire, a été appliquée pour étudier des phénomènes en relation avec la solidification directionnelle d’alliages Al-Cu sous champs externes et dépendants du temps. La formation des microstructures de solidification a été analysée durant des campagnes des vols paraboliques dans le cadre du projet XRMON. Nous avons montré que les variations du niveau de gravité modifiait la couche solutale, et donc la surfusion du liquide devant l’interface solide/liquide. En particulier, l’accélération du niveau de gravité induit une transition Colonnaire-Equiaxe pour un alliage affiné. Pour un alliage non-affiné, la vitesse de croissance des dendrites varie suivant les variations du niveau de gravité durant chaque parabole. Le second volet de ma thèse concerne l'effet d'un champ magnétique permanent sur la solidification équiaxe dans un gradient de température. Nous avons établi que le couplage entre le champ magnétique et le gradient de température génère une force Thermo-Electro-Magnétique qui modifie la trajectoire des grains équiaxes au cours de leur sédimentation. L’action de cette force est bien décrite par le modèle analytique de Wang et al., moyennant des corrections associées au confinement et à la morphologie des grains. Certains effets du champ magnétique sur le liquide ont été également révélés. / X-ray radiography with synchrotron and laboratory source was applied to study phenomena related to the directional solidification of Al-Cu alloys under external fields and time-dependent. The solidification microstructures formation has been analyzed during parabolic flight campaigns in the framework of XRMON project. We have shown that the gravity level variations modified the solutal layer and then the liquid undercooling ahead of the solid/liquid interface. In particular, the acceleration of gravity level induces the Columnar-to-Equiaxed transition in a refined alloy. For a non-refined alloy, the dendrite growth velocity changes with respect to the gravity level variations during each parabola. The second part of this thesis relates to the effect of a permanent magnetic field on the equiaxed solidification in a temperature gradient. The coupling between the magnetic field and the temperature gradient generates a Thermo-Electro-Magnetic force that modifies the equiaxed grains trajectory during their sedimentation. The action of this force is well described by an analytical model proposed by Wang et al. with additional corrections associated to the wall confinement and the grain morphology. Effects of the magnetic field on the liquid have been also revealed.

Welding of duplex stainless steel pipeline material for the oil and gas industry is now common practice. To date, research has been conducted primarily on the parent material and heat affected zones in terms of its susceptibility to various forms of corrosion. However, there has been little research conducted on the degree of sensitisation of the various successive weld layers, namely the root, fill and cap layers. The focus of this research study was to: (i) provide an in-depth microstructural analysis of the various weld passes, (ii) study the mechanical properties of the weld regions; (iii) determine degree of sensitisation of the various weld passes; and (iv) investigate the residual stress levels within the various regions/ phases of the welds. Four test conditions were prepared using manual Gas Tungsten Arc Welding with 'V' and 'U' bevel configuration. Structural analysis consisted of (i) optical microscopy, scanning electron microscopy and magnetic force microscopy; (ii) ferrite determination using Magna-Gauge, Fischer Ferrite-scope and Point Count method. Mechanical testing consisted of Vickers hardness measurements, Charpy impact studies and transverse tensile testing. The degree of sensitisation was determined by three test methods: a modified ASTM A262, ASTM A923 and a modified Double Loop Electrochemical Potentiodynamic Reactivation (DL-EPR) test. Residual stress levels were determined using two neutron diffraction techniques: a reactor source and a time of flight spallation source. Microstructure observed by optical microscopy and magnetic force microscopy shows the formation of both fine and coarse structures within the weld metal. There was no evidence of secondary austenite, being present in any of the weld metal conditions examined. In addition, no detrimental intermetallic phases or carbides were present. The DL-EPR test results revealed that the fill layer regions for all four conditions and the base material showed the highest values for Ir/Ia and Qr/Qa. All four test conditions passed the ASTM A262 and A923 qualitative type tests, even under restricted and modified conditions. Residual stress measurements by neutron diffraction conducted at Lucas Heights Hi-Flux Reactor revealed that the ferrite phase stress was tensile in the heat affected zones and weld, and appeared to be balanced by a local compressive austenite phase stresses in the normal and transverse directions. Residual stress measurements by neutron diffraction conducted at Los Alamos Nuclear Science Centre revealed that in the hoop direction, ferrite (211) and austenite (311) exhibit tensile strains in the weld. In the axial and radial direction, the strains for both phases were more compressive. Correlations between the degree of sensitization and microstructural changes / ferrite content were observed. Higher degrees of sensitization (Ir/Ia and Qr/Qa) were associated with reduced ferrite (increased austenite) content. Correlations between the stresses generated, the evolved microstructures and degree of sensitization were evident. Stresses within the cap region were generally shown to be of a tensile nature in the transverse and longitudinal direction. In summary, the study has shown that correlations exist between the weld microstructure, susceptibility to sensitisation and levels / distribution of internal stresses within the weld regions.

科学研究費補助金 研究種目:試験研究(B) 課題番号:06555177 研究代表者:佐々木 勝寛 研究期間:1994-1995年度

Fundação de Amparo a Pesquisa do Estado de Minas Gerais / In this work we have employed the Melting-Nucleation method in order to synthesize semiconductor and semimagnetic quantum dots (QDs), CdSe, CdSxSe1-x, CdS e Cd1-xMnxS, in a glass matrix. We have investigated these QDs by using several experimental techniques and theoretical models, reaching a comprehensive understanding of their fundamental properties. The lattice contraction in CdSe QDs was confirmed by Raman spectroscopy, evidencing that the glass matrix (host material) plays an important role on the vibrational modes of nanocrystals (NCs). Advancement in the synthesis of pseudo-bynary CdSxSe1-x QDs was achieved by obtaining a good control on the alloy composition in two variants of precursor dopings, while the QD size is tuned by annealing. Resonant Raman spectra of these CdSxSe1-x QDs were very well described by the continuum lattice dynamics model, suggesting the propagation of optical phonons within the NC. By studying CdS QDs by temperaturedependent Raman spectroscopy, we have demonstrated that there is a large difference between the thermal expansion coefficients of QD e of bulk material. The same conclusions were obtained for the modal Grüneisen parameters and for the anharmonicity coupling constants, so that they have to be estimated independently of bulk parameters in the study of thermal properties of QDs. We have employed the Electron Paramagnetic Resonance in order to confirm that the migration of Mn2+ ions in Cd1-xMnxS NCs, from the core to the surface, can be controlled by a thermal annealing. We have proved yet that the luminescence emitted by Cd1-xMnxS NCs can be controlled by changing the x concentration and by the thermal annealing, in which the emission of Mn2+ ions (4T1 6A1) is only observed when this magnetic impurity is substitutionally into the core of Cd1-xMnxS NC. From temperature-dependent photoluminescence measurements, we have come up with a model based on rate equations that describes well the energy transfers involving two coupled groups of Cd1-xMnxS NPs. Further emissions from deep defect levels were attributed to two energetically different divacancies, VCd VS, in the wurtzite structure of Cd1-xMnxS NPs, which can be controlled by the magnetic doping. By studying the magneto-optical properties of Cd1-xMnxS NPs, we have demonstrated that the self-purification is the dominant mechanism controlling the doping process in semiconductor QDs grown by the melting-nucleation method. We have demonstrated that the multiphonon Raman scattering in the coupled Cd1-xMnxS NPs, as well as the coupling strength between electrons and optical phonons, can satisfactorily be tuned by the magnetic doping with Mn2+ ions and by an appropriate thermal annealing. Furthermore, we have verified that the magnetic doping have induced variations on the sp-d exchange interaction and in the crystalline quality of NPs. / Neste trabalho nós utilizamos o método de Fusão-Nucleação para sintetizar pontos quânticos (PQs) semicondutores e semimagnéticos, CdSe, CdSxSe1-x, CdS e Cd1-xMnxS, em uma matriz vítrea. Nós investigamos estes PQs utilizando várias técnicas experimentais e modelos teóricos, obtendo um entendimento compreensivo das suas propriedades fundamentais. A contração da rede cristalina nos PQs de CdSe foi confirmada pela espectroscopia Raman, evidenciando que a matriz vítrea (material hospedeiro) desempenha um papel importante sobre os modos vibracionais dos nanocristais (NCs). Um avanço na síntese de PQs pseudo-binários de CdSxSe1-x foi alcançado, obtendo um bom controle sobre a composição da liga em duas variações de dopantes precursores, enquanto o tamanho do PQ é controlado pelo recozimento. Os espectros Raman ressonantes destes PQs de CdSxSe1-x foram muito bem descritos pelo modelo da dinâmica de rede contínua, sugerindo a propagação de fônons ópticos dentro do NC. Ao estudar PQs de CdS por espectroscopia Raman dependente da temperatura, nós demonstramos que existe uma grande diferença entre os coeficientes de expansão térmica do PQ e do material bulk. As mesmas conclusões foram obtidas para os parâmetros de Grüneisen modais e para as constantes de acoplamento da anarmonicidade, de maneira que eles devem ser estimados independentemente dos parâmetros do bulk no estudo das propriedades térmicas dos PQs. Nós utilizamos a Ressonância Paramagnética Eletrônica para confirmar que a migração de íons Mn2+ em NCs de Cd1-xMnxS, do núcleo para a superfície, pode ser controlada por um recozimento térmico. Comprovamos ainda que a luminescência emitida pelos NCs de Cd1-xMnxS pode ser controlada pela modificação da concentração x e pelo recozimento térmico, em que a emissão dos íons Mn2+ (4T1 6A1) é somente observada quando esta impureza magnética está substitucionalmente incorporada no núcleo do NC de Cd1-xMnxS. A partir de medidas de fotoluminescência dependente da temperatura, nós desenvolvemos um modelo baseado em equações de taxa que descreve bem as transferências de energia envolvendo dois grupos acoplados de nanopartículas (NPs) de Cd1-xMnxS. Emissões adicionais a partir de níveis de defeitos profundos que foram atribuídos duas divacâncias energeticamente diferentes, VCd VS, na estrutura wurtzita das NPs de Cd1-xMnxS, que podem ser controladas pela dopagem magnética. Ao estudar as propriedades magneto-ópticas das NPs de Cd1-xMnxS, nós demonstramos que a autopurificação é o mecanismo dominante controlando o processo de dopagem em PQs semicondutores crescidos pelo método de fusão-nucleação. Nós demonstramos que o espalhamento Raman multi-fônon nas NPs de Cd1-xMnxS acopladas, bem como a força do acoplamento entre elétrons e fônons ópticos, podem ser adequadamente sintonizados pela dopagem magnética com íons Mn2+ e por um recozimento térmico apropriado. Além disso, verificamos que a dopagem magnética induziu variações na interação de troca sp-d e na qualidade cristalina das NPs. / Doutor em Física

Ces dernières années, la Méthode Intégrale de Volume (MIV) a reçu une attention particulière pour lamodélisation des problèmes électromagnétiques en basse fréquence. Son intérêt principal est l’absencedu maillage de la région air, ce qui rend la méthode légère et rapide. Associée aux méthodes decompression matricielle la MIV devient aujourd'hui une alternative compétitive à la méthode deséléments finis pour la modélisation de dispositifs électromagnétiques ayant un volume d'airprépondérant.Ce rapport porte sur le développement de deux formulations intégrales de volume pour la résolution deproblèmes magnétostatiques avec prise en compte des matériaux non linéaires, des aimants, desbobines, des circuits magnétiques avec ou sans entrefer et des régions minces magnétiques. Lapremière est une formulation en flux de mailles indépendantes basée sur l'interpolation par éléments defacette. La deuxième est une formulation en potentiel vecteur magnétique basée sur l'interpolation paréléments d'arête. L'application de ces formulations permet d’une part d'obtenir des résultats précismême en présence d’un faible maillage et d’autre part de résoudre aisément des problèmes nonlinéaires. Des méthodes de calcul de la force magnétique globale ainsi que du flux magnétique dansles bobines ont été également mises en oeuvre. Les développements informatiques ont été réalisés dansla plateforme MIPSE et ont été validés sur des problèmes académiques ainsi que sur quelquesdispositifs industriels. / In recent years, the Volume Integral Method (VIM) has been received particular attention formodeling of low frequency electromagnetic problems. The main advantage of this method is thatinactive regions do not to be discretized, which makes it light and rapid. Associated with matrixcompression methods, the VIM is a competitive alternative to the finite element method for modelingelectromagnetic devices containing a predominant air volume.This PhD thesis focuses on the development of two volume integral formulations for solvingmagnetostatic problems, in the presence of nonlinear materials, magnets, coils, multiply connectedmagnetic regions, and the presence of magnetic shielding. The first one is a mesh magnetic fluxformulation based on the interpolation of facet elements and the second one is a magnetic vectorpotential formulation based on the interpolation of edge elements. The application of theseformulations provides accurate results even with coarse meshes and allows solving straightforwardnonlinear magnetostatic problems. Methods for computing global magnetic force and magnetic fluxthrough a coil were also implemented as part of this work. Developments performed in the MIPSEplatform were validated on academic case-tests as well as some industrial devices.

L'auteur n'a pas fourni de résumé en français / Nowadays in medicine and biotechnology a wide range of applications involves magnetic micro/nano-object manipulation including remote control of magnetic beads, trapping of drug vectors, magnetic separation of labelled cells and so on. Handling and positioning magnetic particles and elements functionalized with these particles has greatly benefited from advances in microfabrication. Indeed reduction in size of the magnet while maintaining its field strength increases the field gradient. In this context, arrays made of permanent micromagnets are good candidates for magnetic handling devices. They are autonomous, suitable for integration into complex systems and their magnetic action is restricted to the region of interest.In this thesis we have elaborated an original approach based on AFM and MFM for quantitative study of the magnetic force and associated force gradients induced by TMP micromagnet array on an individual magnetic micro/nano-object. For this purpose, we have fabricated smart MFM probes where a single magnetic (sub)micronic sphere was fixed at the tip apex of a non-magnetic probe thanks to a dual beam FIB/SEM machine equipped with a micromanipulator.Scanning Force Microscopy conducted with such probes, the so-called Magnetic Particle Scanning Force Microscopy (MPSFM) was employed for 3D mapping of TMP micromagnets. This procedure involves two main aspects: (i) the quantification of magnetic interaction between micromagnet array and attached microsphere according to the distance between them and (ii) the complementary information about micromagnet array structure. The main advantage of MPSFM is the use of a probe with known magnetization and magnetic volume that in combination with modelling allows interpreting the results ably.We conducted MPSFM on TMP sample with two types of microparticle probes: with superparamagnetic and NdFeB microspheres. The measurements carried out with superparamagnetic microsphere probes reveal attractive forces (up to few tens of nN) while MFM maps obtained with NdFeB microsphere probes reveal attractive and repulsive forces (up to one hundred of nN) for which the nature of interaction is defined by superposition of microsphere and micromagnet array magnetizations. The derived force and its gradient from MFM measurements are in agreement with experiments on microparticle trapping confirming that the strongest magnetic interaction is observed above the TMP sample interfaces, between the areas with opposite magnetization. Thanks to 3D MFM maps, we demonstrated that intensity of magnetic signal decays fast with the distance and depends on micromagnet array and microsphere properties.Besides the magnetic interaction quantification, we obtained new information relevant to TMP sample structure: we observed and quantified the local magnetic roughness and associated fluctuations, in particular in zones of reversed magnetization. The variation of detected signal can reach the same order of magnitude as the signal above the micromagnet interfaces. These results complete the experiments on particle trapping explaining why magnetic microparticles are captured not only above the interfaces, but also inside the zones of reversed magnetization.Quantitative measurements of the force acting on a single (sub)microsphere associated to the modelling approach improve the understanding of processes involved in handling of magnetic objects in microfluidic devices. This could be employed to optimize the parameters of sorting devices and to define the quantity of magnetic nanoparticles required for labelling of biological cells according to their size. More generally these experimental and modelling approaches of magnetic interaction can meet a high interest in all sorts of applications where a well-known and controlled non-contact interaction is required at micro and nano-scale.

Uměle vytvořené systémy spinového ledu jsou vhodným nástrojem pro zkoumání neobvyklých jevů, které se v přírodě dají jen těžko pozorovat. Speciálním případem umělého spinového ledu je kagome mřížka, která umožňuje zkoumat kolektivní chování spinů v látce. Tento systém má řadu předpovězených exotických magnetických fází, které zatím nebyly změřeny a prozkoumány v reálném prostoru. V rámci této práce se zabýváme úpravou kagome mřížky tak, aby mohla být využita ke zkoumání exotických stavů v reálném prostoru. Experimenty provedené na naší upravené mřížce ukazují, že jsme schopni detekovat nízko i vysoko energiové stavy, a tedy, že námi navržená úprava kagome mřížky je vhodná pro zkoumání exotických stavů v reálném prostoru.

Silné prostorové omezení materiálů způsobuje jejich nové vlastnosti, které mohou najit uplatnění v mnoha vědeckých i technických odvětvích. Snaha zmenšit velikosti součástek, zvětšit hustotu zápisu a zefektivnit procesy je současným trendem elektronického průmyslu. V této práci je studován vliv prostorového omezení na vlastnosti metamagnetického železo-rhodia (FeRh) během fázové přeměny. FeRh je materiál vykazující fázovou přeměnu prvního druhu mezi antiferomagnetickou a feromagnetickou fází. Metodou mikroskopie magnetických sil v magnetickém poli kolmém na rovinu vzorku je zobrazeni a analyzována struktura fázových domén behem fázové přeměny. Kvantitativní analýza naměřených dat je provedena užitím výškové korelační funkce a její výsledky jsou porovnány pro různé velikosti struktur a tloušťky tenkých vrstev.