L’Inconel 718 est un superalliage base nickel, élaboré dans les années 60, utilisé dans la fabrication de pièces pour les parties chaudes des moteurs d’avion. Il acquiert ses propriétés mécaniques et sa microstructure finale au cours du procédé de mise en forme appelé forgeage à chaud. La maîtrise de ce procédé nécessite de comprendre l’interaction entre les phénomènes d’écrouissage et de recristallisation dynamique tout en intégrant l’influence de diverses conditions thermomécaniques. Cette étude s’est focalisée, expérimentalement, sur les phénomènes liés au forgeage à chaud en mise en forme unipasses et multipasses super-δ-solvus (1050 °C). Afin de les caractériser, des essais de compression ont été réalisés à l’échelle de pions. Des trempes à l’hélium, après déformation, ont permis de figer les microstructures dans le but de comprendre leur évolution en fonction des paramètres thermomécaniques (ε, et T). Des observations ont ensuite été réalisées expérimentalement : microscopie optique et à balayage, EBSD et diffraction des neutrons. Une attention particulière a été portée sur l’évolution de la phase δ, influençant indirectement les propriétés mécaniques de l’alliage, au cours de traitements thermiques puis thermomécaniques. La quantification ainsi que la détermination des cinétiques d’évolution statiques et dynamiques de cette phase a permis de mieux comprendre son influence au cours du forgeage sub-δ-solvus (980 °C). Un chaînage séquentiel a été développé entre deux modèles, l’un de plasticité cristalline implémenté dans un code éléments finis (CPFEM) et l’autre de recristallisation, implémenté dans un code automates cellulaires. Ce chaînage séquentiel permet de décrire les évolutions de champs mécaniques et de microstructures au cours du forgeage à chaud et a été validé par une comparaison avec les résultats expérimentaux. MOTS CLÉS : forgeage à chaud, Inconel 718, recristallisation dynamique, mise en forme multipasses, phase δ, plasticité cristalline (CPFEM), automates cellulaires, chaînage séquentiel, agrégats polycristallins 3D. / Developed in the 60’s, the nickel-base superalloy, Inconel 718, is widely used for hot parts of aircraft engines. The hot forging process confers to the alloy its final microstructure and its mechanical properties. The control of the process requires a deep knowledge of the interactions between the hardening phenomena and the dynamic recrystallization for the various thermomechanical conditions which are used. The present study mainly focuses on the experimental characterization of the phenomena linked to hot forging in the super-δ-solvus domain (1050 °C), as well for a single pass process as for a multipass one. Hot compression tests are used to simulate forging. After deformation, samples are helium quenched in order to freeze the microstructure that allows understanding its evolution as a function of the thermomechanical parameters (ε, and T). Microstructure analyses have then been performed using optical and scanning microscopy, EBSD, and neutron diffraction. A specific attention is paid on the study of the -phase evolution as it has a direct influence on the mechanical properties of the alloy. Its evolution is followed along thermal and thermomechanical treatments. The measure of the static and dynamic precipitation kinetics has led to a better understanding of the -phase role during hot forging at temperatures below solvus (980 °C). A sequential coupling is developed, based on two models; the first one is a crystal plasticity model implemented in a finite element code (CPFEM), the second one being a modeling of recrystallization using a cellular automata approach. The coupling allows the evolutions of the mechanical fields and the microstructure to be simulated during hot forging. The numerical results fit correctly most of the experimental data, mechanical and structural.

As superligas à base de cobalto são bem conhecidas por sua excelente resistência ao desgaste. Muitas pesquisas reportadas na literatura abordam o comportamento do desgaste destas ligas, seja no desgaste por deslizamento, erosivo ou abrasivo. Não obstante, o desgaste microabrasivo destas ligas não tem sido muito investigado, apesar dos danos causados por este tipo de solicitação. O comportamento do desgaste microabrasivo de três superligas à base de cobalto: a) 48% Co, 29 %Cr, 19 % Fe; b) uma liga com composição química próxima à da liga comercial Tribaloy T400 (Co 56 %, Cr 8.5%, Mo 29% Si 3.3 %) e c) uma liga com composição próxima à da liga comercial Stellite 6 (Co 64%, Cr 24 %, W 4.2 %, C 2,3%) foram investigadas. Os ensaios de microabrasão foram conduzidos com três abrasivos SiO2, Al2O3, e SiC em suspensão em água destilada, com concentração de 0,1 g/cm3. A carga aplicada foi de 0,3 N, a velocidade angular 20 rpm e a distância total de deslizamento, 48 metros. A análise das superfícies desgastadas por microscopia óptica, eletrônica de varredura e por perfilometria de contato mostraram que o tamanho, forma e dureza dos abrasivos podem influenciar significativamente os coeficientes de desgaste. Os ensaios conduzidos com partículas abrasivas de SiC e Al2O3 apresentaram maiores coeficientes de desgaste que os conduzidos com partículas de SiO2. A Liga Co-Cr-Fe mostrou os maiores coeficientes de desgaste quando comparada com as demais ligas, devido à baixa fração volumétrica de partículas de segunda fase, duras, precipitadas em sua microestrutura. Durante os ensaios, as três ligas, ensaiadas com os três diferentes abrasivos, apresentaram coeficientes de desgaste crescentes com o aumento da dureza do abrasivo; observou-se uma variação linear dos coeficientes de desgaste com a razão entre a dureza do abrasivo (Ha) e a dureza composta da liga (Hs), com R2 = 0.74. O micromecanismo dominante em todos os ensaios foi o desgaste abrasivo a dois corpos (grooving wear). A liga com composição próxima à da liga comercial Tribaloy T400, contendo fases de Laves dispersas em sua microestrutura, apresentou uma transição de micromecanismo de desgaste dúctil para frágil, quando submetida a ensaios com partículas abrasivas de Al2O3. Assim sendo, o volume de material removido nesta liga foi ligeiramente maior que o observado no ensaio com partículas de SiC. Na liga contendo baixa fração volumétrica de partículas de segunda fase, com matriz constituída por Co (CFC), observou-se uma camada subsuperficial nanocristalina de aproximadamente 1 µm de espessura, severamente deformada, imediatamente abaixo da superfície desgastada. Concluiu-se que o desgaste microabrasivo induziu a recristalização a frio do material encruado, com formação de grãos equiaxiais de dimensões nanométricas. / Cobalt alloys are well known for their excellent wear resistance. Many investigations are reported in literature related to the behavior of erosive, abrasive or sliding wear of these alloys. Nevertheless, the micro-abrasive wear of these alloys has not been thoroughly investigated, despite the damage caused by this type of wear. The microabrasive wear behavior of three cobalt alloys: a) 48 wt.% Co, 29 wt.% Cr, 19 wt.% Fe; b) an alloy with chemical composition close to Tribaloy T400 (56 wt.% Co, 8.5 wt.% Cr, 29% wt. Mo, 3.3 wt. %Si) and c) an alloy with chemical composition close to Stellite 6 (64wt.% Co 24 wt.% Cr, 4.2 wt.% W, 2,3 wt.% C were investigated. The tests were carried out using three 0,1 g/cm3 slurries composed by SiO2, Al2O3, and SiC particles, in suspension in distilled water. The applied load was 0.3 N, the rotational speed 20 rpm and the total sliding distance 48 m. Analysis of the worn surfaces of the tested alloys by Optical Microscopy, Scanning Electron Microscopy and Contact Stylus Profilometry showed that abrasive size, shape and hardness could significantly influence the wear coefficients. The tests carried out with SiC and Al2O3 slurries resulted in greater wear rates than those carried out in SiO2 slurry. Stellite 250, showed the greatest wear coefficient, compared to the two other experimental alloys, due to a very low volume fraction of hard second phase particles in the microstructure. Wear coefficients decreased with increasing abrasive particles hardness. An approximate linear correlation with the ratio between the hardness of the abrasives (Ha) and the compound hardness of the alloys (Ha) with a correlation factor R2= 0.74. The dominant wear micromechanism observed in all tests was two-body abrasion (grooving wear). The modified T400 alloy, containing Laves phase showed a transition from ductile to brittle wear mechanisms when testing with alumina slurries. The worn volume was slightly greater than the one observed with SiC. A severely deformed nanocrystalline layer was identified, immediately below the worn surface. It was concluded that cold recrystallization of the work-hardened material occurred, with the formation of nano sized equiaxed grains.

Amostras de nióbio de alta pureza na forma de monocristais, bicristais e policristais foram retiradas de seções longitudinais de lingotes fundidos por feixe eletrônico. As amostras foram deformadas via extrusão em canal angular (ECAE - Equal Channel Angular Extrusion) em temperatura ambiente até 8 passes, utilizando a rota Bc numa matriz com ângulo de intersecção entre os canais de  = 90º. As amostras foram caracterizadas em termos da evolução microestrutural e da textura de deformação. A caracterização microestrutural foi realizada com o auxílio de microscopias ótica, eletrônica de varredura e eletrônica de transmissão, além de medidas de difração de elétrons retroespalhados (EBSD) para determinação da microtextura e da mesotextura. Medidas de microdureza Vickers foram realizadas para acompanhar o encruamento e o amolecimento das amostras. Um outro monocristal de nióbio foi deformado em 1 passe via ensaio interrompido, utilizando uma matriz com ângulo  = 120º, a fim de estudar a evolução da textura durante a passagem pelo canal de ECAE. A textura foi determinada por meio de difração de raios X e comparada com os dados da literatura para materiais deformados via ECAE com estrutura CCC e também com as texturas simuladas pelo modelo VPSC (visco-plastic self-consistent). No estudo comparativo numa escala maior (monocristal e policristal), verificou-se que houve um refinamento microestrutural significativo após 8 passes. O espaçamento médio entre os contornos de alto ângulo medido perpendicular à direção de extrusão foi próximo nos dois casos (500 nm), maior que o observado para o monocristal deformado numa escala menor (440 nm). Os resultados mostram ainda que os grãos do policristal deformado são mais equiaxiais que os do monocristal. Amostras foram recozidas isotermicamente para avaliar o comportamento frente ao engrossamento microestrutural. Os resultados mostram que o engrossamento torna-se apreciável, em geral, a partir de 500oC com a ocorrência de recristalização descontínua. Acima de 700oC, o crescimento normal de grão passa a ser o principal mecanismo de engrossamento microestrutural. Efeitos de orientação importantes foram observados no bicristal nos estados encruado e recozido. / High-purity niobium single crystals, bicrystals and polycrystals were cut out from longitudinal sections of ingots processed by electron beam melting. Samples were deformed by Equal Angular Channel Extrusion (ECAE) at room temperature up to 8 passes, using the route Bc with a die angle  = 90o. Samples were characterized in terms of their microstructural evolution and deformation textures. Microstructural characterization was performed using optical, scanning electron, and transmission electron microscopies, as well as electron-backscatter diffraction measurements (EBSD) to determine both microtexture and mesotexture. Vickers microhardness testing was performed to follow hardening and softening behaviors in the samples. Another single crystal was deformed by 1 pass in an interrupted ECAE experiment using a die angle  = 120o to follow the changes in texture through the extrusion channel. Texture was determined by X ray diffraction and compared with those reported in the literature for deformed bcc materials and also with those predicted using the viscoplastic self-consistent model (VPSC). A comparative study in a larger scale (single and polycrystals) was also performed. It was observed that there is a significant refinement of the microstructure after 8 passes. The average spacing between high angle boundaries perpendicular to extrusion direction was close in the two cases (500 nm), larger than observed in the single crystal deformed in a smaller scale (440 nm). Results also show that ultrafine grains of the deformed polycrystal are more equiaxial compared to those found in the deformed single crystal. Samples were annealed to evaluate their behavior regarding microstructural coarsening. Results show that coarsening becomes noticeable at temperatures higher than 500oC by means of discontinuous recrystallization. Above 700oC, normal grain growth becomes the main microstructure coarsening mechanism. Important orientation effects were observed in the bicrystal in both deformed and annealed states.

The objective of the research presented in this thesis has been to develop a physics-based dislocation density-based numerical framework to simulate microstructure evolution in severe plastic deformation (SPD) manufacturing processes for different materials. Different mechanisms of microstructure evolution in SPD manufacturing processes were investigated and summarized for different materials under dynamic or high strain rates over a wide temperature range. Thorough literature reviews were performed to clarify discrepancies of the mechanism responsible for the formation of nanocrystalline structure in the machined surface layer under both low-temperature and high-temperature conditions. Under this framework, metallo-thermo-mechanically (MTM) coupled finite element (FE) models were developed to predict the microstructure evolution during different SPD manufacturing processes. Different material flow stress responses were modeled subject to responsible plastic deformation mechanisms. These MTM coupled FE models successfully captured the microstructure evolution process for various materials subjected to multiple mechanisms. Cellular automaton models were developed for SPD manufacturing processes under intermediate to high strain rates for the first time to simulate the microstructure evolution subjected to discontinuous dynamic recrystallization and thermally driven grain growth. The cellular automaton simulations revealed that the recrystallization process usually cannot be completed by the end of the plastic deformation under intermediate to high strain rates. The completion of the recrystallization process during the cooling stage after the plastic deformation process was modeled for the first time for SPD manufacturing processes at elevated temperatures.

The reaction between aqueous Fe(II) and Fe(III) oxides is extremely complex, and can catalyze Fe(II)-Fe(III) electron transfer, exchange of Fe atoms between the aqueous and solid phases, mineral transformation, and contaminant reduction. Together, these processes represent a phenomenon referred to as Fe(II)-catalyzed Fe oxide recrystallization, which has been observed under controlled conditions in the laboratory for numerous Fe oxides. In the environment, Fe oxides are likely surrounded by organic carbon in various forms, but their potential to interfere with Fe(II)-catalyzed Fe oxide recrystallization, and its subsequent environmental relevance has not been well studied. The Fe(II)-catalyzed recrystallization of stable Fe oxides goethite and magnetite was studied in the presence of several environmentally relevant classes of organic carbon. For both goethite and magnetite, Fe(II)-catalyzed recrystallization continued relatively undeterred in the presence of electron shuttling compounds, natural organic matter isolates, and extracellular polysaccharides. Slight inhibition was observed when spent media from dissimilatory iron-reducing cultures was present, but only by sorbing a long-chain phospholipid to the oxides was significant inhibition observed. The lack of interference by organic carbon indicates that Fe(II)-catalyzed Fe oxide recrystallization is likely to be relevant throughout a wide range of environments, and represents a significant process with regards to the geochemical cycling of Fe atoms, a claim supported by evidence of Fe(II)-driven isotope mixing in real soils. The movement of atoms during Fe(II)-catalyzed Fe oxide recrystallization is not limited to just Fe however. Multiple trace elements have been shown to exchange between the aqueous and solid phases along with Fe during the Fe(II)-catalyzed recrystallization of Fe oxides. The effect of organic carbon, both sorbed to the oxide surface and coprecipitated with the oxide, on Fe(II)-catalyzed atom exchange and transformation of ferrihydrite was studied. Again, the presence of organic carbon did not appear to influence Fe atom exchange kinetics. It also did not appear to influence the rapid transformation of ferrihydrite to lepidocrocite. The presence of organic carbon does appear to ultimately have implications for mineral transformation, as over longer time periods it stabilized lepidocrocite, preventing its subsequent transformation to magnetite or goethite.

Human exposure to arsenic in groundwater is a global concern, and arsenic mobility in groundwater is often controlled by Fe mineral dissolution and precipitation. Additionally, Fe(II)-catalyzed recrystallization of Fe oxides has been shown to enable trace element release from and incorporation into Fe oxides. However, the effect of As(V) on the Fe(II)-catalyzed recrystallization of Fe oxides such as goethite, magnetite, and ferrihydrite remains unclear. Here, we measured the extent of Fe atom exchange between aqueous Fe(II) and magnetite, goethite, or ferrihydrite in the presence of As(V) by reacting isotopically "normal" Fe oxides with 57Fe-enriched aqueous Fe(II). At lower levels of adsorption (≤13.3 μM), As(V) had little influence on the rate or extent of Fe(II)-catalyzed Fe atom exchange in goethite or magnetite. However, Fe atom exchange was increasingly inhibited as As(V) concentration increased above 100 μM. Additionally, adsorbed As(V) may be incorporated into magnetite over time in the presence and absence of added aqueous Fe(II) as indicated by X-ray absorption spectroscopy (XAS) and chemical extraction data, with more rapid incorporation in the absence of added Fe(II). XAS and chemical extraction data are also consistent with the incorporation of As(V) during goethite and magnetite precipitation. Additionally, atom exchange data indicated that low levels of As(V) coprecipitation (As:Fe = 0.0005-0.0155) had little influence on the rate or extent of Fe(II)-catalyzed Fe atom exchange in goethite or magnetite. Atom exchange data indicated that ferrihydrite likely transforms via a dissolution-reprecipitation mechanism both to lepidocrocite at 0.2 mM Fe(II) and to magnetite at 5 mM Fe(II). The presence of 206 μM As(V) slowed the transformation of ferrihydrite to more crystalline iron minerals and slowed the rate of atom exchange between aqueous Fe(II) and ferrihydrite. However, the degree of atom exchange did not directly correlate with the amount of ferrihydrite transformed. In summary, Fe oxide recrystallization processes may affect As(V) uptake and release in the environment, and As(V) may inhibit Fe(II)-catalyzed Fe oxide recrystallization.

Les aciers Dual Phase (DP) ferrito-martensitiques sont largement utilisés sous la forme de tôles minces dans la construction automobile en raison de leur excellent compromis résistance/ductilité et donc pour leur potentiel d’allègement. Ils sont élaborés par coulée continue, laminage à chaud et à froid suivis d’un recuit continu. Durant l’étape de chauffage et de maintien de ce recuit, la microstructure ferrito-perlitique déformée issue des étapes de laminage se transforme en microstructure ferrito-austénitique recristallisée. L’expérience montre que les cinétiques de recristallisation et de transformation ainsi que la distribution spatiale et morphologique des microstructures résultantes sont très sensibles aux vitesses de chauffage. Ce travail de thèse s’intéresse aux différents mécanismes expliquant cette sensibilité comme la maturation des carbures, la restauration, la recristallisation de la ferrite et la transformation austénitique et toutes leurs interactions. Ces mécanismes métallurgiques ont été caractérisés à différentes échelles et par des approches in situ sur un acier industriel puis modéliser par des approches à base physique pour guider une possible production. Après un premier chapitre dédié aux techniques expérimentales et de modélisations utilisées, le second chapitre de ce travail s’intéresse principalement à la caractérisation de la morphogénèse des microstructures ferrito-austénitique en microscopie électronique à balayage (MEB). Le troisième chapitre est une étude détaillée en Microscopie Electronique à Transmission (MET) et par modélisation thermocinétique (ThermoCalc, DICTRA) de la composition des carbures tout au long du processus, du laminage à chaud au recuit. Restauration et recristallisation sont étudiées au chapitre 4 principalement par des expériences in situ en Diffraction des Rayons X à Haute Energie (DRXHE) sur ligne de lumière synchrotron et modélisées par une approche originale à champs moyen. Enfin, le chapitre 5 propose une étude sous DICTRA pour comprendre les cinétiques de transformation austénitique en fonction des vitesses de chauffe. Cette approche est novatrice car elle prend en compte les carbures intergranulaires de la ferrite, a été conduite en conditions anisothermes et propose une analyse fine des modes de croissance de l’austénite associées au manganèse, élément clef de la composition de ces alliages. / Ferrite/Martensite Dual-Phase steels are largely used in the form of thin sheets in the automotive industry for their excellent balance between resistance and strength and thus for their lightening potential. They are elaborated by continuous casting, hot- and cold- rolling, followed by a continuous annealing. During the heating and the soaking stages of this latter process, the deformed ferrite/pearlite microstructure obtained after rolling evolves is transformed into a recrystallized ferrite-austenite microstructure. The experiments show that recrystallization and austenite transformation kinetics as well as the resulting spatial and morphological distribution of the phases are highly sensitive to the heating rate. This PhD thesis aims at understanding the different metallurgical mechanisms explaining this particular sensitivity as carbides ripening, recovery, recrystallization and austenite transformation and all their possible interactions. The mechanisms were characterized at different scales and by in situ technics on an industrial steel and model by physical based approaches in order to drive future production lines. After a first chapter dedicated to the experimental and modeling methods, the second chapter deals with the characterization of the morphogenesis of ferrite-austenite microstructures by Scanning Electron Microscopy (SEM). Chapter 3 is a study by Transmission Electron Microscopy (TEM) and by thermokinetic modeling (ThermoCalc, DICTRA) of the chemical composition of carbides along with manufacturing, from hot-rolling to annealing. Recovery and recrystallization are studied in chapter 4 by the means of in situ High Energy X-Ray Diffraction (HEXRD) experiments conducted on a synchrotron beamline and modeled by an original mean-field approach. Finally, chapter 5 proposes an analysis with DICTRA to understand austenite transformation kinetics as function of heating rates. The proposed approach is innovative as it accounts for intergranular carbides in the ferrite matrix, is conducted in non-isothermal conditions and propose a fine analysis of growth modes of austenite associated to manganese, a key alloying element of the studied steels.

The effect of initial microstructures of an 0.05 wt.% C low carbon steel, acicular ferrite (AF), Bainite (B), polygonal ferrite (PF), fine polygonal ferrite (FPF), and a microstructure produced by direct strip casting (DSC) (termed SC), on the deformation and recrystallization behaviour of cold rolled low carbon (LC) steel, was investigated. The initially prepared samples with the initial microstructures were cold rolled to 50, 70 and 90% reductions, then annealed isothermally in the temperature range 580 ??? 680 oC. The microstructures and textures produced by deformation and annealing were studied by optical microscopy, XRD, TEM, SEM and EBSD. The initial microstructures were characterized mainly by optical microscopy and EBSD. Using EBSD, the ferrite grain size of the AF, B and SC samples was considerably larger than that found by optical microscopy with a large fraction of low angle grain boundaries (LAGBs) observed within prior austenite grains. All samples exhibited a very weak texture close to random. After cold rolling, the microstructures of AF and SC contained shear bands with PF and FPF generating deformation bands. For AF and SC, the pearlite phase was more extensively elongated in rolling direction compared with PF and FPF. After 90% cold rolling reduction, PF, FPF and SC consist mainly of the texture component and AF and B . It was found that FPF recrystallized most rapidly followed by B, PF and AF with SC recrystallizing orders of magnitude more slowly due to the solution drag caused by its uniformly distributed higher Mn content. Very strong (???-fibre) texture was generated in cold rolled PF followed by FPF, with AF, SC and B generating very weak textures. The texture evolution during annealing 90% reduction PF was examined in further detail. The behaviour of nucleation and grain growth provides strong evidence of orientated nucleation as the dominant factor for CRA texture development in this material.

With increasing miniaturization in microelectronics the wirebonds used in IC packages are witnessing a thrust towards fine pitch wirebonding. To have a precise control over loop height of the wirebond for fine pitch wirebonding, it is imperative to do mechanical characterization of the wirebond. The present work studies the mechanical properties of gold wire and wirebond using nanoindentation. The wirebond specimen surface was planarized using mechanical polishing. The loop height of the gold wirebond is directly proportional to the length of the heat affected zone (HAZ) above the ball of gold wirebond. Metallographic preparation of gold wirebond cross section reveals the presence of undesirable coarse grain structure in HAZ due to recrystallization and grain growth in the gold wire adjacent to the ball. The recrystallization temperature of our gold wire was found using D.S.C. to be 340.66°C. The doping elements present in the gold wire used, were identified using TOF-SIMS. Nanoindentation of the gold wire was done at different maximum loads to observe the hardness variation with load. The nanoindentation of gold wirebond has confirmed a v-shaped hardness profile in the HAZ. The hardness minima for the particular gold wire used with a ball size ratio of 2.4 was observed at distance of 160-170 µm from the neck of the ball. The elastic modulus was found to vary randomly and to be independent of the microstructure in the wirebond. A yield stress profile based on empirical hardness-yield strength correlation has been predicted for the gold wirebond. / Singapore-MIT Alliance (SMA)

In this thesis, the relationship among friction stir processing (FSP) parameters, microstructure evolution, texture development, and mechanical hehavior of AZ31B Mg alloy was investigated. First of all, in order to reveal the correlation among the deformation conditions, dynamic recrystallization (DRX) mechanisms, and microstructure evolution in the Mg alloy, hot compression tests at a wide range of Zener-Hollomon parameter (Z) values were conducted. Through optical microscopic examination, it was found out that above a critical Z value, twinning influences the DRX process resulting in a more effective grain refinement, which is manifested in a significant change in the slope of the Z-drec relationship, where drec is the recrystallized grain size. Moreover, EBSD examination revealed that the twinning also contributed to a distinct change in the recrystallization texture. Compression tests were performed along both through-thickness and in-rolling-plane directions of the plate to study the orientation dependency of twinning activities and its influence on the DRX process. X-ray line profile analysis (XLPA) provides further insights by highlighting the differences in the dislocation density/types, subgrain sizes, and twin densities during the DRX processes operating with or without the twinning. Secondly, the constitutive behaviour study was applied to the investigation of microstructure evolution during FSP. By varying the key FSP parameters systematically, i.e. rotation and travel rates of the tool, a series of FSP specimens were prepared with a wide range of thermo-mechanical inputs in terms of Z. The resulting tensile behavior in the stir zone (SZ) showed a dramatic change as a function of Z, caused by a systematic change in the texture within SZ measured by neutron diffraction. A three-dimensional transient model was developed to investigate the detailed deformation history including the temperature and strain rate profiles and material flow pattern during FSP of the Mg alloy. Such deformation history can be combined with the constitutive study from the compression tests in order to analyze the developments of micro-texture and DRX grains during FSP, which will, in turn, dominate the mechanical properties. Based on the studies above, new fundamental understandings were gained on the governing mechanisms for the deformation and recrystallization processes during FSP and the influence of thermo-mechanical input during FSP on ductility enhancement in the Mg alloy.