No description available.

Hong Sin Yi, Grace. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1997. / Includes bibliographical references (leaves 50-51). / Acknowledgments / Abstract / Chapter Chapter 1 --- Introduction / Chapter 1.1 --- Metallic Glass and its application --- p.1 / Chapter 1.2 --- Glass Forming Ability (GFA) --- p.2 / Chapter 1.3 --- Equilibrium Phase --- p.3 / Chapter 1.4 --- Nucleation and Growth --- p.6 / Chapter 1.5 --- Spinodal Decomposition --- p.8 / Chapter 1.6 --- Morphology Comparison between Nucleation and Growth and Spinodal --- p.13 / Figures --- p.14 / References --- p.24 / Chapter Chapter 2 --- Experimental Method / Experimental Method --- p.25 / Figure --- p.29 / References --- p.30 / Chapter Chapter 3 --- Metastable liquid miscibility gap in Pd-Si and its glass forming ability / Introduction --- p.32 / Experimental --- p.33 / Results --- p.34 / Discussion --- p.36 / Figures --- p.40 / References --- p.49 / Bibliography --- p.50

Orientadores: Amauri Garcia, Noé Cheung / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica / Made available in DSpace on 2018-08-16T22:32:29Z (GMT). No. of bitstreams: 1 Bertelli_Felipe_M.pdf: 6382019 bytes, checksum: 7bc62ea83b7e721ef82e5669d236559f (MD5) Previous issue date: 2010 / Resumo: Neste trabalho, o software ANSYS, baseado no Método dos Elementos Finitos, é adaptado para a simulação tridimensional do fluxo de calor no processo de refusão superficial a laser. A análise numérica é validada com resultados simulados por outros modelos existentes na literatura para casos de refusão superficial a laser de alumínio puro e com resultados simulados e experimentais de uma liga Al-5%Ni. Ensaios experimentais próprios foram realizados em amostras de uma liga Al-1,5%Fe, utilizando um laser à fibra dopado com Itérbio, com potência máxima disponível de 2 kW. Para efeito comparativo, as trilhas foram feitas variando-se valores de velocidade de deslocamento do feixe laser para um mesmo valor de potência. Observou-se que a microestrutura tanto do substrato quanto da zona tratada apresentou morfologia tipicamente celular. As microestruturas resultantes dos tratamentos a laser foram analisadas através de microscopia eletrônica de varredura, sendo observados espaçamentos celulares extremamente refinados na área tratada a laser refletindo no aumento significativo da dureza confirmado por ensaios de microdureza Vickers. Uma técnica de dissolução parcial das amostras tratadas a laser foi aplicada para evidenciar os intermetálicos no substrato e na região tratada a laser, mostrando a modificação da redistribuição dos intermetálicos no interior da poça fundida e dando indicações de aumento da resistência à corrosão na região tratada / Abstract: In this work, the software ANSYS, based on the Finite Element Method, is adapted to simulate the three-dimensional heat flux during the laser remelting surface treatment. The numerical analysis is validated against theoretical results furnished by other models from the literature for laser surface remelting of aluminum and against theoretical and experimental results of Al-5wt%Ni alloy samples. Laser remelting experiments with Al-1,5%wtFe samples have been carried out by using a 2kW Yb fiber laser. For comparative effects, the laser tracks were performed with different laser beam velocities for a fixed value of power. It was observed that both the substrate and the treated region had a typical cellular morphology. The microstructures resulting from the laser treatment were analyzed by using electron scanning microscopy and very refined cell spacing has been observed, which can induce a significant hardness increase confirmed by Vickers microhardness tests. A partial dissolution technique has been performed to foreground the intermetallics at the substrate and at the laser treated zone, showing the intermetallics redistribution inside the molten pool and giving indications of increased corrosion resistance on the treated region / Mestrado / Materiais e Processos de Fabricação / Mestre em Engenharia Mecânica

Orientador: Amauri Garcia / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecanica / Made available in DSpace on 2018-08-14T11:35:39Z (GMT). No. of bitstreams: 1 Cante_ManuelVenceslau_D.pdf: 12795296 bytes, checksum: bd8bf91fcf479b66a28e7e3d6b5f7090 (MD5) Previous issue date: 2009 / Resumo: O desenvolvimento de microestruturas otimizadas durante o processo de solidificação são de fundamental importância nas propriedades e desempenho de produtos acabados baseados em ligas metálicas. Neste estudo é analisada a cinética envolvida no processo de solidificação, seus efeitos nos parâmetros macro e microestruturais e a sua conseqüente influência nas propriedades mecânicas. Com esse intuito, ligas hipoeutéticas do sistema binário são estudadas Al-Ni por meio de experimentos de solidificação vertical ascendente sob regime transitório de condução de calor. Os espaçamentos dendríticos primários (?1) e secundários(?2) foram medidos ao longo de todos os lingotes para cada uma das ligas analisadas e correlacionados com as variáveis térmicas de solidificação. Uma abordagem teórico-experimental é utilizada na determinação quantitativa de tais variáveis térmicas: coeficiente de transferência de calor na interface metal/molde, velocidade de deslocamento da isoterma liquidus, gradientes térmicos, taxa de resfriamento e tempo local de solidificação. Os dados experimentais referentes à solidificação das ligas são confrontados com os principais modelos teóricos de crescimento dendrítico da literatura. Este estudo aborda, também, a influência do teor de soluto nos espaçamentos dendríticos para as ligas estudadas. Do ponto de vista macroestrutural, verifica-se que a transição colunar/equiaxial (TCE) ocorre para ligas hipoeutéticas Al-Ni para uma taxa crítica de resfriamento de 0,16 K/s. Por ensaios de tração as propriedades mecânicas das ligas do sistema Al-Ni são correlacionadas com parâmetros da micro-estrutura dendrítica resultante do processo de solidificação. Verifica-se que os limites de escoamento e de resistência à tração crescem com o aumento da concentração de soluto e decrescem com o aumento dos espaçamentos dendríticos, ?1 e ?2. O alongamento específico, por outro lado, mostra-se independente da composição e do arranjo dendrítico. Para a liga Al-5%Ni foi também realizado um estudo de solidificação rápida por refusão da superfície a laser para análise das variações microestruturais e de dureza entre as áreas não tratadas e tratadas superficialmente. / Abstract: The development of optimized microstructures during the solidification stage of processing is of fundamental importance to the mechanical properties and to the performance of finished products of metallic alloys. In this study the kinetics of solidification and its effects on macro and microstructural parameters, as well as the consequent influence on the final mechanical properties are analyzed. Hypoeutectic Al-Ni alloys are studied by upward unidirectional solidification experiments under transient heat flow conditions. Primary (?1) and secondary (?2) dendrite arm spacings are measured along the castings for all alloys analyzed and correlated with transient solidification thermal variables. A combined theoretical/ experimental approach is used to quantitatively determine such thermal variables, i.e., transient metal/mold heat transfer coefficients, tip growth rates, thermal gradients, tip cooling rates and local solidification time. The experimental data concerning the Al-Ni alloys solidification are compared to the main predictive dendritic models from the literature and the dependence of dendrite arm spacing on the alloy solute content is also analyzed. From the macrostructural point of view, it is found that the CET occurs for a critical value of cooling rate of about 0.16 K/s for hypoeutectic Al-Ni alloys.With a view to correlate mechanical properties to dendrite arm spacings, tensile testings were carried out. It is found that the ultimate tensile strength and the yield strength increase with increasing alloy solute content and with decreasing primary and secondary dendrite arm spacings. In contrast, the elongation is found to be independent of both alloy composition and dendritic arrangement. For the Al 5%Ni alloy a rapid solidification study is carried out by using laser surface remelting in order to permit microstructural and microhardness variations throughout the resulting treated and untreated zones, to be analysed. / Doutorado / Materiais e Processos de Fabricação / Doutor em Engenharia Mecânica

Melt convection and solid phase movement play an important role in solidification processes, which significantly influence the formation of grain structures and solute segregations. In general, the melt convection and grain movement are a result of buoyancy forces. The densities within melt are different due to the variation of temperature and concentration, leading to thermally and solutally driven melt convection. Similarly, the density differences between the grains and the bulk melt cause the grain movement, leading to solid sedimentation or grain floating, as the case may be. Free, unattached solid grains are produced by partial remelting and fragmentation of dendrites, by mechanical disturbances such as stirring or vibration and by heterogeneous nucleation of grains in solidification of grain-refined alloys. In this way, movement of solid crystals during solidification can be ascertained in the following two cases. In the first case, during columnar solidification of non-grain-refined alloys, solid movement is possible in the form of dendrite fragments detached from the columnar stalks by the process of remelting and fragmentation. Movement of grains during columnar solidification gives rise to altogether different microstructure from columnar to equiaxed. In the second case, during equiaxed solidification of grain-refined alloys, the movement of solid crystals is possible in the form of equiaxed dendrite crystals nucleated due to presence of grain refiners. The rate and manner by which the free solids settle (or float) will influence macrosegregation in metal castings. Control of the solidification process is possible through an understanding of the solid movement and its effect on macrosegregation and microstructure. With this viewpoint, the overall objective of the present thesis is to study, experimentally and numerically, the phenomenon of solid phase movement during solidification. Through this study, deeper insights of the role of solid phase movement in solidification are developed which can be used for possible control of quality in castings. Both columnar and equiaxed solidification are considered. Models for transport phenomena associated with columnar solidification with solid phase movement are rarely found in the literature, because of inherent difficulty associated with consideration of microscopic features such as remelting and fragmentation. To tackle this problem, solidification modules for remelting and fragmentation are developed first, followed by integration of these molecules in a macroscopic solidification model. A Rayleigh number based fragmentation criterion is developed for detachment of dendrite fragments from the developing mushy zone, which determines the conditions favorable for fragmentation of dendrites. The criterion developed is a function of net concentration difference, liquid fraction, permeability, growth rate of mushy layer, and thermophysical properties of the material. The effect of various solidification parameters on fragmentation is highlighted. The integrated continuum model developed is applied to stimulate the solidification of aqua-ammonia system in a side-cooled rectangular cavity. The numerical results are in good qualitative agreement with those of experiments reported in literature. A gentle ramp of the mushy zone due to settling of solid crystals, as also noticed in experimental literature, is observed towards the bottom of the cavity. The influence of various modeling parameters on solid phase movement and resulting macrosegregation is investigated through a parametric study. Movement of grains during columnar solidification gives rise to altogether different microstructure and sometimes may initiate a morphological transition of the microstructure from columnar to equiaxed if the number and size of equiaxed grains ahead of the columnar front become sufficient to arrest the columnar growth. The generalised model developed, considering solid phase movement during columnar solidification is used to predict columnar-to-equiaxed transition (CET) based on a prescribed cooling rate criterion. It is found that presence of convection significantly affects the solidification behaviour. Moreover, the movement of dendrite fragments and their accumulation at the columnar front further trigger the occurrence of CET. Cooling configuration, too significantly affects the nature of CET. In unidirectional solidification cases, the locations of CET are found to be in a plane parallel to the chill face. However, for the case of the non-unidirectional solidification (as in side-cooled cavity), the locations of CET need not be in a plane parallel to the chill face. In contrast to fixed columnar solidification, equiaxed solidification is poorly understood; in particular, the phenomena associated with solid crystal movement. Movement of unattached solid crystals, formed due to heterogeneous nucleation on grain-refiners, is induced by the convective currents as well as by buoyancy effects, causing the solid to sediment or to float, depending on density of solid compared to that of the bulk melt. While moving in the bulk melt these crystals can also remelt or grow. A series of casting experiments with AI-based alloys are performed to investigate the role and influence of movement of solid crystals on macrosegregation and microstructure evolution during equiaxed solidification. Controlled experiments are designed for studying, separately, settling and floatation of equiaxed crystals for different cooling conditions and configurations. Further, these experiments are carried out in convective and non-convective cases to understand the effect of convection on solid phase movement. Temperature measurements are performed at various locations in the mould during the experiments. After the cavity is solidified, microstructural and chemical analyses of the experimental samples are carried out, several notable features are observed in temperature histories, macrosegregation pattern, and microstructures due to settling/flotation phenomenon of solid crystals. It is found that the flow behavior of solid grains has a profound influence on the progress of solidification (in terms of grain size distribution and fraction eutectic) and macrosegregation distribution. In some cases, the induced flow due to solid phase movement can cause a flow reversal. The observations and quantitative data obtained from experiments, with the help of detailed solidification conditions provided, can be used for future validations of models for equiaxed solidification. Subsequently, numerical studies are carried out, using a modified version of the macroscopic model developed for columnar solidification with motion of solid crystals, to predict the transport phenomena during equiaxed solidification. The model is applied to simulate the solidification processes corresponding to each of the experimental cases performed in this study. For a better understanding of the phenomenon of movement of solid crystals, the following two special cases of solidification are also presented: 1) without movement of solid crystals and 2) movement of solid crystals without any relative velocity between solid and liquid phases. The numerical predictions showing nature of flow field and progress of solidification are substantiated by the experimental data for the thermal analysis, qualitative microstructural Images and quantitative microstructural analysis. It is concluded, with the help of various experiments and simulations, that movement of solid crystals influences the casting quality appreciably, in terms of macrosegregation and microstructures. It is expected that the improved understanding of the role and influence of solid phase movement during solidification processes (both columnar and equiaxed) obtained through this thesis will be useful for possible control of quality of as-cast products.

The aim of the present thesis is to understand the phase transformation behavior of embedded alloy nanoparticles embedded in immiscible matrices. Embedded alloy inclusions have been dispersed in immiscible matrix via rapid solidification method. The present work deals with synthesis of embedded particles, evolution of microstructure, morphology and crystallographic orientation relation relationships among different phases, phase transformation and phase stability behavior of embedded alloy inclusions in different matrices. In the present investigation the systems chosen are Bi-Sn and Bi-Pb in Zn matrix and Cd-Sn in Al matrix. Chapter 1 gives the brief introduction of present work Chapter 2 gives a brief review of nanoscale materials, various synthesis techniques, microstructure evolution, solidification and melting theories. Chapter 3 discusses the processing and experimental techniques used for characterization of the different samples in the present work. Melt-spinning technique used to synthesize the rapidly solidified ribbons. The structural characterization is carried out using X-ray diffraction and transmission electron microscopy. Chapter 4 illustrates the size dependent solubility and phase transformation behavior of Sn-Cd alloy nanoparticles embedded in aluminum matrix. X-ray diffraction study shows the presence of fcc Al, bct Sn, hcp Cd solid solution and hcp Cd phases. Based on Zen’s law, the amount of Sn present Cd solid solution is estimated. Using overlapped sterograms, the orientational relationships among various phases are found. Microscopy studies reveal that majority of the alloy nano inclusions exhibit a cuboctahedral shape with 111 and 100 facets and they are bicrystalline. STEM-EDS analysis shows that both phases exhibit size dependent solubility behavior and for particles size smaller than 18 nm, single phase solid solution could only be observed. Calorimetric studies reveal a depression in eutectic melting point of bimetallic particles. In situ heating studies show that melting initiates at triple line junction corner and melt first grows into the interior of the Sn rich phase of the particle and then later the melt grows into the interior of the Cd phase of the particle. During cooling first Cd phase solidifies later Sn phase solidifies and on further cooling at low temperatures entire particle transforming into complete solid solution phase particle. Size dependent melting studies show that during heating smaller particles melted first, later bigger particles melted. During cooling first bigger particle solidified later smaller particles solidified. High resolution imaging indicates presence of steps across particle-matrix interface that may get annihilated during heating. During cooling, molten particles in the size range of 16-30 nm solidify as solid solution which for molten particles greater than 30 nm solidify as biphasic particle. Insitu heating studies indicates that for solid particles less than 15 nm get dissolved in the Al matrix at temperatures at around 135°C. Differential scanning calorimetry (DSC) studies show in the first heating cycle most of the particles melt with an onset of melting of at 166.8°C which is close to the bulk eutectic temperature of Sn-Cd alooy. The heating cycle reveals that the melting event is not sharp which can be understood from in-situ microscopy heating studies. In the second and the third cycles, the onset of melting observed at still lower temperatures 164.3°C and 158.5°C .The decrease in onset melting point in subsequent heating cycles is attributed to solid solution formation of all small particles whose size range below 30 nm during cooling. cooling cycles exhibit an undercooling of 90°C with respect to Cd liquidus temperature. Thermal cycling experiments using DSC were carried out by arresting the run at certain pre-determined temperatures during cooling and reheating the sample to observe the change in the melting peak position and area under the peak. The areas of these endothermic peaks give us an estimate of the fraction of the particles solidified upto the temperature when the cycling is reversed. Based on experimental observations, a thermodynamic model is developed, to understand the solubility behavior and to describe the eutectic melting transition of a binary Sn-Cd alloy particle embedded in Al matrix. Chapter 5 discusses the phase stability and phase transformation behavior of nanoscaled Bi-Sn alloys in Zn matrix. Bi-Sn alloys with eutectic composition embedded in Zn matrix using melt spinning technique. X-ray diffraction study shows the presence of rhombohedral Bi, pure BCT Sn and hcp Zn phases. In X-ray diffractogram, there are also other new peaks observed, whose peak positions (interplanar spacings) do not coincide either with rhombohedral Bi or bct Sn or hcp Zn. Assuming these new phase peaks belong to bct Sn rich solid solution(based on earlier work on Bi-Sn rapidly solidified metastable alloys) whole pattern fitting done on x-ray diffractogram using Lebail method. The new phase peaks indicated as bct M1(metastable phase1), bct M2(metastable phase2) phases. The amount of Bi present in M1, M2 solid solution is estimated using Zens law. Two sets of inclusions were found, one contains equilibrium bismuth and tin phases and the other set contains equilibrium bismuth and a metastable phase. In-situ TEM experiments suggest that as temperature increases bismuth diffuses into tin and becomes complete solid solution. Melting intiates along the matrix–particle interface leading to a core shell microstructure. During cooling the entire inclusion solidify as solid solution and decomposes at lower temperatures. High temperature XRD studies show that as temperature increases M1, M2 phases peaks merge with Sn phase peaks and Bi phase peak intensities slowly disappear and on further increasing temperature Sn solid solution phase peaks also disappear. During cooling diffraction studies show that first Sn solid solution phase peaks appear and later Bi phase peaks appear. But, the peaks belong to metstable phases not appeared while cooling. Chapter 6 presents morphology and phase transformation of nanoscaled bismuth-lead alloys with eutectic (Pb44.5-Bi55.5) and peritectic (Pb70-Bi30) compositions embedded in zinc matrix. using melt spinning technique. In alloy1[ Zn-2at%(Pb44.5-Bi55.5)] inclusions were found to be phase separated into two parts one is rhombohedral Bi and the other is hcp Pb7Bi3 phase. X-ray diffraction study shows the presence of rhombohedral Bi, hcp Pb7Bi3 and hcp Zn phases in Zn-2at%(Pb44.5-Bi55.5) melt spun sample. The morphology and orientation relationships among various phases have been found. In-situ microscpy heating studies show that melt initially spreads along the matrix–particle interface leading to a core-shell microstructure. And in the core of the core-sell particles, first Bi phase melts later Pb7Bi3 phase will melt and during cooling the whole particle solidify as biphase particle with large undercooling. In-situ heating studies carried out to study the size dependent melting and solidification behavior of biphase particles. During heating smaller particles melt melt first later bigger particle will melt. In contrast, while cooling smaller particles solidifies first, later bigger particles will solidify. Detailed high temperature x-ray diffraction studies indicate there increases first Bi phase peaks disappear later Pb7Bi3 phase peaks disappear and during cooling first Pb7Bi3 phase peaks appear and later Bi phase peaks appear. In alloy2[ Zn-2at%(Pb70-Bi30)] inclusions were found to be single phase particles. X-ray diffraction study shows the presence of hcp Pb7Bi3 and hcp Zn phases in Zn-2at%(Pb70-Bi30) melt spun sample. The crystallographic orientation relationship between hcp Pb7Bi3 and hcp Zn phases. In-situ microscpy heating studies show that melting initiates across the matrix–particle interface grows gradually into the interior of the particle. Three phase equilibrium at peritectic reaction temperature is not observed during insitu heating TEM studies. Size dependent melting point depression of single phase particles is not observed from in-situ heating studies. Detailed high temperature x-ray diffraction studies show that while heating the Pb7Bi3 phase peak intensities start decreasing after 170°C and become zero at 234°C. And during cooling Pb7Bi3 phase peaks starts appearing at 200°C and on further cooling the Pb7Bi3 phase peak intensities increase upto 150°C, below this temperature peak intensities remain constant.

The primary focus of the present work is the development of macro-models for numerical simulation of binary alloy solidification processes, consistent with microscopic phase-change considerations, with a particular emphasis on capturing the effects of non-equilibrium species redistribution on overall macrosegregation behaviour. As a first step, a generalised macroscopic framework is developed for mathematical modelling of the process. The complete set of equivalent single-phase governing equations (mass, momentum, energy and species conservation) are solved following a pressure-based Finite Volume Method according to the SIMPLER algorithm. An algorithm is also developed for the prescription of the coupling between temperature and the melt-fraction. Based on the above unified approach of solidification modelling, a macroscopic numerical model is devised that is capable of capturing the interaction between the double-diffusive convective field and a localised fluid flow on account of solutal undercooling during non-equilibrium solidification of binary alloys. Numerical simulations are performed for the case of two-dimensional transient solidification of Pb-Sn alloys, and the simulation results are also compared with the corresponding experimental results quoted in the literature. It is observed that non-equilibrium effects on account of solutal undercooling result in an enhanced macrosegregation. Next, the model is extended to capture the effects of dendritic arm coarsening on the macroscopic transport phenomena occurring during a binary alloy solidification process. The numerical results are first tested against experimental results quoted in the literature, corresponding to the solidification of an Al-Cu alloy in a bottom-cooled cavity. It is concluded that dendritic arm coarsening leads to an increased effective permeability of the mushy region as well as an enhanced eutectic fraction of the solidified ingot. Consequently, an enhanced macrosegregation can be predicted as compared to that dictated by shrinkage-induced fluid flow alone. For an order-of-magnitude assessment of predictions from the numerical models, a systematic approach is subsequently developed for scaling analysis of momentum, heat and species conservation equations pertaining to the case of solidification of a binary mixture. A characteristic velocity scale inside the mushy region is derived, in terms of the morphological parameters of the two-phase region. A subsequent analysis of the energy equation results in an estimation of the solid layer thickness. It is also shown from scaling principles that non-equilibrium effects result in an enhanced macro-segregation compared to the case of an equilibrium model For the sake of assessment of the scaling analysis, the predictions are validated against computational results corresponding to the simulation of a full set of governing equations, thus confirming the trends suggested by the scale analysis. In order to analytically investigate certain limiting cases of unidirectional alloy solidification, a fully analytical solution technique is established for the solution of unidirectional, conduction-dominated, alloy solidification problems. The results are tested for the problem of solidification of an ammonium chloride-water solution, and are compared with those from existing analytical models as well as with the corresponding results from a fully numerical simulation. The effects of different microscopic models on solidification behaviour are illustrated, and transients in temperature and heat flux distribution are also analysed. An excellent agreement between the present solutions and results from the computational simulation can be observed. The generalised numerical model is subsequently utilised to investigate the effects of laminar double-diffusive Rayleigh-Benard convection on directional solidification of binary fluids, when cooled and solidified from the top. A series of experiments is also performed with ammonium chloride-water solutions of hypoeutectic and hypereutectic composition, so as to facilitate comparisons with numerical predictions. While excellent agreements can be obtained for the first case, the second case results in a peculiar situation, where crystals nucleated on the inner roof of the cavity start descending through the bulk fluid, and finally settle down at the bottom of the cavity in the form of a sedimented solid layer. An eutectic solidification front subsequently progresses from the top surface vertically downwards, and eventually meets the heap of solid crystals collected on the floor of the cavity. However, comparison of experimental observations with corresponding numerical results from the present model is not possible under this situation, since the associated transport process involves a complex combination of a number of closely interconnected physical mechanisms, many of which are yet to be resolved. Subsequent to the development of the mathematical model and experimental arrangements for macroscopic transport processes during an alloy solidification process, some of the important modes of double-diffusive instability are analytically investigated, as a binary alloy of any specified initial composition is directionally solidified from the top. By employing a close-formed solution technique, the critical liquid layer heights corresponding to the onset of direct mode of instability are identified, corresponding two a binary alloy with three different initial compositions. In order to simulate turbulent transport during non-equilibrium solidification processes of binary alloys, a modified k-8 model is subsequently developed. Particular emphasis is given for appropriate modelling of turbulence parameters, so that the model merges with single-phase turbulence closure equations in the pure liquid region in a smooth manner. Laboratory experiments are performed using an ammonium chloride-water solution that is solidified by cooling from the top of a rectangular cavity. A good agreement between numerical and experimental results is observed. Finally, in order to study the effects of three-dimensionality in fluid flow on overall macrosegregation behaviour, the interaction between double-diffusive convection and non-equilibrium solidification of a binary mixture in a cubic enclosure (cooled from a side) is numerically investigated using a three-dimensional transient mathematical model. Investigations are carried out for two separate model systems, one corresponding to a typical metal-ally analogue system and other corresponding to an actual metal-alloy system. As a result of three-dimensional convective flow-patterns, a significant solute macrosegregation is observed in the transverse sections of the cavity, which cannot be captured by two-dimensional simulations.