Table of contents

The 3rd International Conference on Advances in Solidification Processes was held in the Rolduc Abbey in the Netherlands a few kilometres away from Aachen. Around 200 scientists from 24 countries come in for the four day meeting. They found a stimulating but also relaxing environment and atmosphere, with beautiful weather and the medieval abbey inviting for walks, discussions, sitting outside and drinking a beer or wine.

The contributions given at the conference reflected recent advances in various topics of solidification processes, ranging from fundamental aspects to applied casting technologies. In 20 oral sessions and a large poster session innovative results of segregation phenomena, microstructure evolution, nucleation and growth, phase formation, polyphase solidification, rapid solidification and welding, casting technology, thermophysics of molten alloys, solidification with forced melt flow and growth of single crystals and superalloys together with innovative diagnostic techniques were presented. Thereby, findings from experiments as well as from numerical modeling on different lengths scales were jointly discussed and contribute to new insight in solidification behaviour.

The papers presented in this open access proceedings cover about half the oral and poster presentations given. They were carefully reviewed as in classical peer reviewed journals by two independent referees and most of them were revised and thus improved according to the reviewers comments. We think that this collection of papers presented at ICASP-3 gives an impression of the excellent contributions made. The papers embrace both the basic and applied aspects of solidification.

We especially wish to express our appreciation for the team around Georg Schmitz and Margret Nienhaus organising this event and giving us their valued advice and support at every stage in preparing the conference. We also thank Lokasenna Lektorat for taking the task of checking all language-associated issues and fixing the papers according to the templates given by IOP Conference Series. We also wish to express our gratitude to the IOP Conference Series publishers, who were always helpful and patient with us.

All papers published in this volume of IOP Conference Series: Materials Science and Engineering have been peer reviewed through processes administered by the proceedings Editors. Reviews were conducted by expert referees to the professional and scientific standards expected of a proceedings journal published by IOP Publishing.

The Growth Restriction Factor, Q, proved to be useful to analyse and control grain refinement during solidification of alloys [1-3]. It is known that in multicomponent alloys a simple summation of the Q_i values of the individual constituents taken from the binary phase diagrams can lead to grossly wrong results and that the ternary or higher-level phase diagram needs to be evaluated. This work demonstrates that the actual evaluation of Q using the liquidus gradient and partition coefficients of the multicomponent phase diagram requires some precautions and may be cumbersome. More importantly, this approach entirely fails if an intermetallic phase turns out to be the primary solidifying phase even in tiny amount. A very simple and general solution of this problem is illustrated for Al-Si-Mg-Cu alloys.

Ultrasonic melt treatment (UST) is known to induce grain refinement in aluminum alloys, especially when transition metals like Zr and Ti are present. The refinement of primary intermetallics (e.g. Al₃Zr, Al₃(Zr,Ti)) caused by UST may influence the subsequent solidification process when intermetallics act as nucleation sites. In this paper, an Al-Ti-Zr alloy is used to analyse the effect of different ultrasonic intensities on the formation of these primary intermetallics. The possible nucleation behaviour of Al₃Zr particles during UST is also discussed and an edge to edge matching model is used to make a preliminary analysis of lattice mismatch between aluminum oxide and Al₃Zr phase. The experimental results show that when UST is applied, the Al₃(Zr,Ti) particles might nucleate on aluminum oxides and remain fine during solidification.

Microstructure of Al-Si alloy castings depends most generally on melt preparation and on the cooling rate imposed by the thermal modulus of the component. In the case of Al-Si alloys, emphasis is put during melt preparation on refinement of pro-eutectic (Al) grains and on modification of the Al-Si eutectic. Thermal analysis has been used since long to check melt preparation before casting, i.e. by analysis of the cooling curve during solidification of a sample cast in an instrumented cup. The conclusions drawn from such analysis are however valid for the particular cooling conditions of the cups. It thus appeared of interest to investigate how these conclusions could extrapolate to predict microstructure in complicated cast parts showing local changes in the solidification conditions. For that purpose, thermal analysis cups and instrumented sand and die castings with different thermal moduli and thus cooling rates have been made, and the whole set of cooling curves thus recorded has been analysed. A statistical analysis of the characteristic features of the cooling curves related to grain refinement in sand and die castings allowed determining the most significant parameters and expressing the cube of grain size as a polynomial of these parameters. After introduction of a further parameter quantifying melt refining an excellent correlation, with a R² factor of 0.99 was obtained.

It has been widely considered that Al₃Ti is involved in the aluminium nucleation on TiB₂, although the mechanism has not been fully understood. In this paper molecular dynamics has been conducted to investigate this phenomenon at an atomistic scale. It was found that a two-dimensional Al₃Ti layer may remain on TiB₂ above the aluminium liquidus. In addition, the results showed that this 2D Al₃Ti undergoes interface reconstruction by forming a triangular pattern. This triangular pattern consists of different alternative stacking sequences. The transition region between the triangles forms an area of strain concentration. By means of this mechanism, this interfacial Al₃Ti layer stabilizes itself by localizing the large misfit strain between TiB₂ and Al₃Ti This reconstruction is similar to the hdp-fcc interface reconstruction in other systems which has been observed experimentally [1].

Simple casting experiments were set up to solve the question, if heterogeneous nucleation of the liquid-liquid decomposition in monotectic systems is possible. Al–Pb alloys with different inoculants were solidified, and the resulting microstructure was analysed by SEM and X-ray microtomography. Pronounced changes in the distribution of the lead precipitations indicate that it is possible to trigger the nucleation.

Heterogeneous nucleation is an important phenomenon in solidification processes. A smooth interface between the nucleus and its substrate is assumed for considerations on heterogeneous nucleation in classical nucleation theory, but this condition cannot always be satisfied for the factual heterogeneous nucleation process. The formation of a nucleus on a coarse substrate is investigated, and the effect of the roughness factor of the substrate on the nucleation barrier is discussed based on Wenzel's wetting model. It is shown that the heterogeneous nucleation barrier is generally less than that of the homogeneous nucleation barrier, and a lower intrinsic wetting angle between the nucleus and the substrate leads to easier nucleation. However, it is also found that, when the intrinsic wetting angle is less than 90°, higher roughness leads to easier heterogeneous nucleation; for a certain roughness factor, an intrinsic critical wetting angle exists which decreases the heterogeneous nucleation barrier to 0. On the other hand, when the intrinsic wetting angle is larger than 90°, higher roughness leads to a larger heterogeneous nucleation barrier; for a certain roughness factor, an intrinsic critical non-wetting angle exists which increases the heterogeneous nucleation barrier to the value of the homogeneous nucleation. It thus leads to the ineffectiveness of the foreign substrate for nucleation.

In this study, the effects of the misfit on in-plane structures of liquid Al and interfacial structure of solidified Al in contact with the heterophase substrates have been investigated, using molecular dynamics (MD) simulations. The MD simulations were conducted for Al/fcc (111) substrates with varied misfits. The order parameter and atomic arrangement indicated that the in-plane ordering of the liquid at the interface decreases significantly with an increase of the misfit, i.e., solid-like for small misfit and liquid-like for large misfit. Further, our MD simulation results revealed that a perfect orientation relationship forms at the interface between the substrate and the solidified Al for a misfit of less than -3% and the boundary is coherent. With an increase in the misfit, Shockley partial and extended dislocations form at the interface, and the boundary becomes a semi-coherent or low-angle twist boundary.

It is well known that in aluminium alloys containing Zr, grain refiner additions do not function as desired, producing an effect often referred to as nuclei poisoning. This paper investigates the structure of direct chill-cast ingots of commercial AA3003 aluminium alloys, with and without Zr, at various addition levels of Al5Ti1B master alloy. In Bridgman experiments simulating ingot solidification, Zr-containing alloys were studied after the addition of various amounts of Ti. It could be demonstrated, in both ingot casting and simulation experiments, that Zr poisoning can be compensated for by adding more Ti and/or Al5Ti1B. The results confirm better refinement behaviour with the addition of Ti + B than of only Ti. The various combinations of Zr and Ti also influenced the formation of AlFeMn phases, and the precipitation of large Al₆(Mn,Fe) particles was revealed. AlZrTiSi intermetallic compounds were also detected.

Initial sidebranch spacing is one of the most important characteristics during directional dendrite growth, which is of interest for both scientific and technologic aspects. Effects of primary spacing, thermal gradient, pulling velocity, and surface tension anisotropy on the initial sidebranch spacing are investigated by using the quantitative phase field model. Simulation results indicate that the initial sidebranch spacing has a scaling law relationship with pulling velocity but is almost independent of primary spacing, thermal gradient and surface tension anisotropy.

This paper presents a numerical simulation of a directional solidification experiment, which was conducted in microgravity conditions in the Material Science Lab (MSL) onboard the International Space Station. Solidification of an Al-7wt.%Si alloy in the Low Gradient Furnace (LGF) was investigated. The LGF is a Bridgman-type furnace insert for the MSL, consisting of two heated cavities separated by an insulated adiabatic zone. The simulation results include the prediction of Columnar to Equiaxed Transition (CET) and average as-cast equiaxed grain diameters. A front tracking algorithm was employed to track the growing columnar dendrite front while a volume averaging method was used to model equiaxed nucleation, growth and impingement. The thermal boundary conditions for the simulation domain were defined and computed via temperature readings that were recorded during the experiment. The experimental data were obtained from a number of thermocouples that were attached to the crucible of the sample cartridge assembly. To conclude, the microgravity experimental results and the model simulation results, including the CET, are compared.

The mechanical properties of a cast product, and therefore its application, depend strongly on its inner microstructure. During the solidification step a change from columnar to equiaxed grain structure can occur. It is thus critical to understand the physical mechanisms of this transition in order to accurately predict and control its occurrence and the final grain structure morphology. This article reports on observations of the CET (Columnar to Equiaxed Transition) induced by a sudden increase of the pulling velocity during the directional solidification on a refined Al-3.5wt%Ni alloy by using synchrotron X-ray radiography. The influence of the pulling velocity on the blocking of the columnar structure is described. Next, the distribution of surface area and longitudinal asymmetry of the grains after CET are quantitatively characterized and discussed.

The primary dendrite arm spacing (PDAS) is one of the main characteristic parameters of directionally solidified dendritic structures. It is widely used to characterise as-cast structures and to correlate them with the local solidification conditions. Although, common use seems to refer to a unique relationship between solidification conditions and PDAS, it has been widely accepted that there exists a distribution of spacings in a dendritic array. In the present work a novel method for the determination of the PDAS is proposed, which yields the average PDAS, its distribution and the number of nearest neighbours. These quantities - which can be used to characterise dendritic arrays - are obtained by determining the centre of gravity of all dendrites in the array under consideration and then applying a specially designed algorithm to identify the nearest neighbours of each dendrite. The characteristics of the method are demonstrated by applying it to test structures and dendritic structures obtained from directional solidification experiments.

Using the Bridgman technique, the solidification of aluminum alloys of types 6005, 6063, and 6082, was studied. The solidification process differs both between the alloys and between billet surface and bulk, due to higher concentrations of Fe, Si, Mn and Mg near the surface. Previously determined surface concentrations were used to calculate the Fe, Si, Mn and Mg additions needed for Bridgman experiments that simulate surface region solidification. Microstructures were studied and grain size, secondary dendrite arm spacing, intermetallic particles, segregation, and coarsening were evaluated. The increased alloy element concentrations at the surface were found to influence both the structure coarseness and the kind of intermetallic precipitation. In addition, clear differences could be determined between the alloy types, depending on their pull rate in the furnace.

The effect of a medium-density DC-current and its direction on the directional solidification microstructure and cooling behavior of Pb-80wt%Sn alloy were studied. Experimental results indicated that the microstructure was refined with DC compared to the one without. With the increase of DC density, the primary (λ₁) and secondary (λ₂) dendrite arm spacing of the primary phase decrease clearly. It is also found that not only the intensity of the DC current but also its direction affects the microstructure. The size of λ₁ and λ₂, in samples treated with positive DC, decreases more clearly than in those with negative DC. The interface morphology in samples processed without and with DC is different when the current direction is changed. Finally, the mechanism of microstructure evolution is discussed.

A three-phase Eulerian approach is used to model the columnar-to-equiaxed transition (CET) during solidification in DC casting of technical bronze. The three phases are the melt, the solidifying columnar dendrites and the equiaxed grains. They are considered as spatially interpenetrating and interacting continua by solving the conservation equations of mass, momentum, species and enthalpy for all three phases. The so defined solidification model is applied to a binary CuSn6 DC casting process as a benchmark to demonstrate the model potentials. Two cases are studied: one considering only feeding flow and one including both feeding flow and equiaxed sedimentation. The simulated results of mixed columnar and equiaxed solidification are presented and discussed including the occurrence of CET, phase distribution, feeding flow, equiaxed sedimentation and their influence on macrosegregation.

A central parameter to describe the formation of porosity and macrosegregation during casting processes is the permeability of the dendritic mushy zone. To determine this specific feature for a binary Al-18wt.%Cu alloy, flow simulations based on the Lattice Boltz-mann (LB) method were performed. The LB method allows an efficient solving of fluid flow problems dealing with complex shapes within an acceptable period of time. The 3D structure required as input for the simulations was captured with X-ray microtomography, which enables the generation of representative geometries for permeability investigations. Removing the eutectic phase from the measured dataset generated a remaining network of solid primary dendrites. In the simulations, a pressure gradient was applied to force the liquid through the free interdendritic channels. The permeability of the structure was then calculated from the resulting flow velocity pattern using Darcy's law. To examine the influence of different boundary conditions on the results obtained, several simulations were conducted.

Bridgman experiments for various withdrawal rates were conducted in an Al - 7 wt % Si alloy. The dendrite arm spacing and the grain size of the primary dendritic aluminium-rich phase were measured, together with the spatial distribution of the interdendritic eutectic micro structure. A fine grain structure is observed for large pulling rates at constant temperature gradient, corresponding to a uniform distribution of the eutectic structure very close to the Gulliver-Scheil approximation. When decreasing the cooling rate, the average fraction of the eutectic structures decreases as it is expected based on back-diffusion. However, the distribution of the eutectic structure progressively looses uniformity and the grain size increases. Detailed analyses of these observations are conducted using a cellular automaton – finite element model. It allows the simulation of the local solidification path for an open system coupled with a solution of the heat and solute mass balances in the presence of fluid flow for the entire Bridgman sample. Evolutions of the size and density of structures are retrieved when increasing the cooling rates. This demonstrates the key role played by the thermosolutal buoyancy forces and their interaction with the solidifying dendritic grain structure to explain the distributions of structures and solute segregation.

This work evaluates a mixed columnar-equiaxed solidification model [Wu et al. 2010 Comp. Mater. Sci.50 32] by comparison with the laboratory castings. In the numerical model nucleation of equiaxed crystals, tracking of the position of columnar primary dendrite tips, transition from non-dendritic to dendritic growth, columnar-to-equiaxed transition (CET), melt flow and grain sedimentation are taken into account. As modeling result the as-cast structure, macrosegregation, volume-averaged inter- and extra-dendritic eutectic phase can be calculated. In the laboratory experiment a series of Al-Cu ingots with different pouring parameters and compositions were cast, and the corresponding structural information and macrosegregation were analyzed. The primary goals are (1) to explore the uncertainties and limitation of the numerical model; (2) to identify the sensitive parameters influencing the casting and modeling results; finally (3) to further justify the model assumptions.

Rods of a homogenized coarse-grained Al-4wt.%Cu alloy were subjected for a short time (in the order of seconds) to a steep temperature gradient and (partial) melting. θ-phase or eutectic particles are visible in the resolidified micros true ture. These microstructural features are interpreted as traces of former liquid regions, particularly intragranular liquid droplets. The former droplets were characterized with respect to their morphology and size. The bimodal size distribution indicates two classes of nucleation sites in the solid.

On 26th March 2010 the MAXUS-8 sounding rocket was launched from the Esrange Space Center in Sweden. As part of the Intermetallic Materials Processing in Relation to Earth and Space Solidification (IMPRESS) project, a solidification experiment was conducted on a Ti-45.5at.%Al-8at.%Nb intermetallic alloy in a module on this rocket. The experiment was designed to investigate columnar and equiaxed microstructures in the alloy. A furnace model of the MAXUS 8 experiment with a Front Tracking Model of solidification has been developed to determine the macrostructure and thermal history of the samples in the experiment. This paper gives details of results of the front tracking model applied to the MAXUS 8 microgravity experiment. A model for columnar growth is presented and compared to experimental results for furnace A of the experiment module.

The solidification and age hardening behaviour of an Mg-6Zn-2Gd (wt. %) alloy has been investigated. It was found that the microstructure of the as-cast samples was composed of equiaxed α-Mg grains surrounded by some eutectic compound both at triple points and along grain boundary. The eutectic compound was composed of MgZn₂, Mg₅Gd phases and α-Mg matrix. After a solution treatment at 500 °C for 18 h, the eutectic compound almost dissolved, but some discontinuous second phase particles still survived at grain boundaries. The Mg₃Gd₂Zn₃ phase also formed during subsequent ageing. The addition of Gd not only improved the thermal stability of the second phase formed during solidification, but also postponed the overaging during ageing at 200 °C up to 100 h. The precipitates with three different morphologies: [0001] _α rods/lath, (0001) _α plates and blocky particles, were observed in the α-Mg matrix. The solute elements of Zn and Gd were found to significantly partition into the precipitates, especially for blocky particles. This significant partition can be correlated directly to improved mechanical properties at elevated temperature.

Solidified microstructures and phases have been investigated at three compositions in the Cu-Hf-Ti alloy system The examined compositions were found to be the best glass formers in the ternary Cu-Hf-Ti system. However, neither the stable phases that are suppressed by the amorphous phase during rapid quenching, nor the solidifying structure have been in the focus of interest. Phases and the solidifying structures formed at different cooling rates (5 K/min, ∼150 K/s) were examined for different compositions. Solidification sequence, morphology of the phases, and phase analyses were performed by SEM-EDX, XRD and thermal analyses (DSC).

Magnesium matrix composite is one of the advanced lightweight materials with high potential to be used in automotive and aircraft industries due to its low density and high specific mechanical properties. The magnesium composites can be fabricated by adding the reinforcements of fibers or/and particles. In the previous literature, extensive studies have been performed on the development of matrix grain structure of aluminum-based metal matrix composites. However, there is limited information available on the development of grain structure during the solidification of particulate-reinforced magnesium. In this work, a 5 vol.% SiC_p particulate-reinforced magnesium (AM50A) matrix composite (AM50A/SiC_p) was prepared by stir casting. The solidification behavior of the cast AM50A/SiC_p composite was investigated by computer-based thermal analysis. Optical and scanning electron microscopies (SEM) were employed to examine the occurrence of nucleation and grain refinement involved. The results indicate that the addition of SiC_p particulates leads to a finer grain structure in the composite compared with the matrix alloy. The refinement of grain structure should be attributed to both the heterogeneous nucleation and the restricted primary crystal growth.

In secondary AlSi alloys, the presence of small amounts of Fe causes the formation of intermetallic phases, which have a negative effect on mechanical and physical properties of castings. To understand the effect of fluid flow on the microstructure and intermetallic phases, Al-5/7/9 wt pet Si 0.2/0.5/1.0 wt pet Fe alloys have been directionally solidified under defined thermal (gradient 3 K/mm, solidification velocity 0.04 mm/s) and fluid flow (rotating magnetic field 6 mT) conditions. The primary α-Al phase and intermetallic phases were studied using light microscopy and SEM with EDX. The influence of fluid flow and intermetallic phases (β-Al₅FeSi) on microstructure was characterized by changes of primary and secondary dendrite arm spacing and specific surface area of the dendrites. We observe a pronounced effect of flow on the length of the intermetallic precipitates, a macro-segregation Fe and Si and even small amounts of iron and thus intermetallics reduce possible effects of flow on microstructural parameters.

The present investigation is aimed at developing ultrafine eutectic/dendrite Ti-Fe-Sn in-situ composite with balanced combination of strength and plasticity. It also studies the microstructure evolution in the series of hypereutectic Ti-Fe-Sn ternary alloys. Sn concentration of these alloys has been varied from 0 – 10 atom% in the binary alloy (Ti₇₁Fe₂₉) keeping the Ti concentration fixed. These alloys have been prepared by arc melting under an Ar atmosphere on a water-cooled Cu hearth, which are subsequently suction cast in a split Cu-mold under an Ar atmosphere. Detailed X-ray diffraction (XRD) study shows the presence of TiFe, β-Ti, and Ti₃Sn phases. The SEM micrographs reveal that the microstructures consist of fine scale eutectic matrix (β-Ti and TiFe) with primary dendrite phases (TiFe and/or Ti₃Sn) depending on concentration of Sn. α -Ti forms as a eutectoid reaction product of β-Ti. The room temperature uniaxial compressive test reveals simultaneous improvement in the strength (1942 MPa) and plasticity (13.1 %) for Ti₇₁Fe₂₆Sn₃ ternary alloy. The fracture surface indicates a ductile mode of fracture for the alloy.

Five spheroidal graphite cast irons were investigated, a usual ferritic grade and four pearlitic alloys containing Cu and doped with Sb, Sn and Ti. These alloys were remelted in a graphite crucible, leading to volatilization of the magnesium added for spheroidization and to carbon saturation of the liquid. The alloys were then cooled down and maintained at a temperature above the eutectic temperature. During this step, primary graphite could develop showing various features depending on the doping elements added. The largest effects were that of Ti which greatly reduces graphite nucleation and growth, and that of Sb which leads to rounded agglomerates instead of lamellar graphite. The samples have been investigated with secondary ion mass spectrometry to enlighten distribution of elements in primary graphite. SIMS analysis showed almost even distribution of elements, including Mg and Al (from the inoculant) in the ferritic grade, while uneven distribution was evident in all doped alloys. Investigations are going on to clarify if the uneven distribution is associated with structural defects in the graphite precipitates.

We model the Fe-Sn system by using a higher order polynomial to describe the free energy of the liquid, and study three different aspects in morphological evolution in the monotectic alloy. Firstly, phase separation, in which case the liquid decomposes into two, is investigated inside of the spinodal decomposition region. Secondly, we study the core-shell morphology in the Fe-Sn alloy, which arises by spinodal decomposition in 2D. Finally, stable lamellar and unstable droplet morphologies in directional solidication are investigated.

Systematic directional solidification experiments were done with TRIS-NPG organic peritectic alloys to investigate peritectic solidification close to the limit of constitution undercooling. The experiments were carried out as in-situ observation in a horizontal micro Bridgman furnace. Under specific conditions isothermal peritectic coupled growth (PCG) were obtained for alloys within the hypo-peritectic region. Isothermal PCG was obtained either by (i) reducing the growth velocity from above the critical value for morphological stability of both solid phases to a value below, or by (ii) long-time growing with a constant velocity below the critical value for morphological stability of both solid phases. The later happens via island banding, where nucleation and lateral growth of the peritectic phase compete with growth of the primary phase. Interphase spacings and the widths of α and β lamellae were measured as function of growth velocity for a fixed composition and as function of composition for a fix growth velocity.

Coupled eutectic growth of Al and Al₂Cu was investigated in univariant Al-Cu-Ag alloys during solidification with planar and cellular morphology. Experiments reveal the dynamic selection of small spacings, below the minimum undercooling spacing and show that distinct morphological features pertain to nearly isotropic or anisotropic Al-Al₂Cu interfaces.

The dynamics of rod-like eutectics are examined using a directional solidification setup, which allows real-time observation of the whole solidification front in specimens of transparent eutectic alloys -here, succinonitrile-(D)camphor. In steady-state, rod eutectic growth patterns consist of triangular arrays, more or less disturbed by topological defects. In the absence of strong convection and of crystallographic anisotropy, the long-time evolution of the pattern is dominated by "imperfections" of the system, such as misalignment of the temperature gradient, and finite-size. In this study, we present experimental results on the finite-size effects on rod eutectics and show that a rod to lamella transition takes place as a result of finite-size effect only, at a given alloy concentration.

The solidification behaviour of multi-component and multi-phase systems has been largely investigated in binary and ternary alloys. In the present study, a quaternary model system is proposed based on the well known Al-Cu-Ag and Al-Cu-Mg ternary eutectic alloys. The quaternary eutectic composition and temperature were determined by EDS (Energy Dispersive Spectrometry) and DSC (Differential Scanning Calorimetry) analysis, respectively. The microstructure was then characterised by SEM (Scanning Electron Microscope). In the DSC experiments, two types of quaternary eutectics were determined according to their phase composition. For each type of eutectic, various microstructures were observed, which result in different eutectic compositions. Only one of the determined eutectic compositions was further studied by the controlled growth technique in a vertical Bridgeman type furnace. In the initial part of the directionally solidified sample, competing growth between two-phase dendrites and three-phase eutectics was obtained, which was later transformed to competing growth between three-phase and four-phase eutectics. Moreover, silver enrichment was measured at the solidification front, which is possibly caused by Ag sedimentation due to gravity and Ag rejection from dendritic and three-phase eutectic growth, and its accumulation at the solidification front.

Ternary eutectics provide a unique opportunity for studying the effects of complex microstructure formation, as three distinct phases must be formed simultaneously from the melt. In order to produce fully coupled three-phase growth, Al-Ag-Cu at the ternary eutectic composition was directionally solidified in a constant temperature gradient of 3 K/mm at velocities between 0.2 and 5.0 μm/sec. Under these conditions, the two intermetallic phases appear to grow as closely coupled rods in an α(Al) matrix, with the solidification velocity affecting the specific morphologies chosen by the rods and the general degree of alignment of the structure. Crystal orientations were examined by EBSD to determine if variations in morphology within a single sample are due to specific orientation relationships. Although no conclusive connection to morphology has yet been found, two different sets of orientation relationships between the three phases have thus far been identified.

An addition of small amounts of Na and Sr is commonly used in the industry to modify the eutectic in Al-Si alloys. Both Na and Sr suppress nucleation of the eutectic forcing nucleation and growth to take place at higher undercooling than in the unmodified material. Thus the scale of the eutectic and the shape of the Si crystals are modified to a fine fibrous form so that the ductility of the material is increased. In the present work a one-dimensional numerical model is proposed that describes nucleation and growth of both primary dendrites and eutectic grains as a function of cooling conditions and modification. The model assumes that dendrites nucleate easily when the liquidus temperature is reached and that they grow as heat is extracted by the mould. Nucleation of the eutectic grains depends on local undercooling and growth is governed by a balance between growth of the eutectic grains and the rate at which heat is extracted by the mould. Experimental data is used to determine constants in the nucleation function. It is shown how cooling conditions and mode of modification influence nucleation and growth conditions.

In the present work the influence of a cross-section transition on freckle formation in the directionally solidified superalloy CMSX-4 was investigated. It was found that in stepwise expanding specimens the new freckle chains were not formed immediately at the bottom edge of the steps, but after an incubation height of about 10 mm. The flow behaviour through a mushy zone was simulated in a similar system, in which the cross section was changed stepwise. After an expansion of the cross section, the flow velocity slowed down significantly. This explained the suppression effect of the expanding step on the interdendritic convection and furthermore on the freckle formation. The step effect was successfully applied to avoid the freckle formation in expanding specimens by applying a step length smaller than the critical incubation length.

Stray grains which may appear during production of single-crystal blades, can severely deprave the performance of turbine blades. This investigation primarily explores the conditions of stray grains formed in DZ417G superalloy by using samples with varying cross sections. The effect of withdrawal velocities on the microstructure of DZ417G superalloy in directional solidification is investigated. It is observed that the growth direction of dendrites did not change, and no stray grain appears in the re-entrant corner at low withdrawal velocity. Nevertheless, the stray grains grow up when the withdrawal velocity is higher than 150μm/s. The temperature profiles in the region of the cross section change at different velocities are examined. The increase of undercooling in the corner of the cross section change leads to the formation of stray grains.

In this paper a novel directional solidification process - Thin Shell Casting (TSC) is presented. According to this new technology a thin shell mould mounted to a chill is dipped through the floating baffle into the melt bath. By pulling up the shell mould, a directional solidification can be performed in a downward direction. Due to the use of extremely thin shell moulds and the application of gas cooling, an efficient heat extraction can be achieved in comparison to the conventional and modified Bridgman techniques. In the experiments with Cu- and Al-alloys, not only directional solidification (DS), but also single crystal (SC) growth can be successfully performed. This confirms the feasibility of the new technology, which will be further applied to the production of SC components of superalloys.

In the present study, the non-uniformity of the thermal condition and the corresponding grain defect formation in the customary Bridgman process were investigated. The casting clusters in radial alignment were directionally solidified in a Bridgman furnace. It was found that in the casting cluster, the shadow side facing the central rod was ineffectively heated in the hot zone and ineffectively cooled in the cooling zone during withdrawal, compared with the heater side facing the furnace heater. The metallographic examination of the simplified turbine blades exhibited that the platforms on the shadow side are very prone to stray grain formation, while the heater side reveals a markedly lower tendency for that. The asymmetric thermal condition causes the asymmetrical formation of these grain defects. This non-uniformity of the thermal condition should be minimized as far as possible, in order to effectively optimize the quality of the SC superalloy components.

There is a continuing demand to raise the operating temperature of jet engine turbine blades to meet the need for higher turbine entry temperatures (TET) in order to increase thermal efficiency and thrust. Modern, high-pressure turbine blades are made from Ni-based superalloys in single-crystal form via the investment casting process. One important post-cast surface defect, known as 'surface scale', has been investigated on the alloy CMSX-10N. This is an area of distinct discolouration of the aerofoil seen after casting. Auger electron and X-ray photoelectron spectroscopy analysis were carried out on both scaled and un-scaled areas. In the scaled region, a thin layer (∼800nm) of Ni oxide is evident. In the un-scaled regions there is a thicker Al₂O₃ layer. It is shown that, as the blade cools during casting, differential thermal contraction of mould and alloy causes the solid blade to 'detach' from the mould in these scaled areas. The formation of Ni Oxides is facilitated by this separation.

Macro- and meso-segregations correspond to compositional heterogeneities at the scale of a casting. They develop during the solidification process. One of the parameters that have an essential effect on these segregations is the mush permeability, which is highly nonlinear, and varies over a wide range of magnitudes. We present simulation results for solidification of a Sn-Pb alloy in a two-dimensional cavity, highlighting the role of (i) the numerical interpolation schemes used for the finite-volume discretization of the highly-nonlinear permeability term and (ii) of the mesh size on the prediction of mesosegregations and macrosegregation. We observe that solute-rich liquid flowing through the mushy zone due to thermo-solutal convection results in patches of thin channel structures, which develop into mesosegregations. We notice little sensitivity of the predicted macrosegregation to different discretization schemes for the permeability term. However, we found their influence on the prediction of channel segregates to be significant when using coarse computational grids, customary in the simulation of industrial castings. Mesh refinement is crucial for capturing the complex phenomena in the formation of channel segregates. With a very fine mesh channels have been captured with more than one grid point along their width, allowing the determination of their width.

We study the effect of solidification kinetics, driven by local limited diffusion in the liquid, on macrosegregation. If the diffusion in the liquid surrounding a growing grain is slow, the local average liquid concentration is lower than the thermodynamic equilibrium concentration at the interface. The redistribution of solute by the flow of intergranular liquid on the macroscopic scale is affected by the modified microsegregation in the liquid. We study this phenomenon using a two-phase model based on the volume-averaging method, describing macroscopic transport coupled to a microscopic grain growth model. The growth kinetics is resolved by accounting for finite diffusion in the liquid and solid phases, assuming an equiaxed globular morphology. To accurately model the diffusion field around the grain, we propose an improved approximation for the solutal boundary layer thickness accounting for the growth conditions and liquid convection. The effect of growth kinetics on macrosegregation is then investigated in the case of solidification of a binary alloy in a small cavity where the solid phase is fixed and fluid flow is driven by natural convection. We show that it is important to accurately model the diffusion field around the grain to capture correctly the effect of growth kinetics on the weakening of macrosegregation.

Cylindrical Sn-Pb alloy samples of different compositions (10, 20 and 30 wt.-% Pb) were prepared from high purity (4N) components. After metals have been melted, a rotating magnetic field (RMF) with an induction of 150 mT and a frequency of 50 Hz was switched on in order to homogenize the liquid. The electromagnetic field was generated by a 3-phase, 2-pole inductor. Just before the start of the solidification process, the magnetic field was switched off to achieve a microstructure free of melt flow influence. The sample translation velocity was constant (0.05 mm/s), and the temperature gradient changed from 7 to 3 K/mm during the solidification process. The first half part of each sample solidified without influence of rotating magnetic field while solidification of the second half part proceeded under the action of the RMF. The columnar microstructure formed in the absence of RMF induced fluid flow was replaced after switching on the RMF by a characteristic "Christmas tree"- like macro-segregated structure with equiaxed dendrites. The secondary dendrite arm spacing and the volume fraction of primary tin phase (dendrite) were measured by an automatic image analyzer on the longitudinal polished sections along the whole length of the samples. The effect of the forced melt flow and alloy composition on its micro- and macrostructure development was investigated.

Hypereutectic Al-Si based alloys are gaining popularity for applications where a combination of light weight and high wear resistance is required. The high wear resistance arising from the hard primary Si particles comes at the price of extremely poor machine tool life. To minimize machining problems while exploiting outstanding wear resistance, the primary Si particles must be controlled to a uniform small size and uniform spatial distribution. The current industrial means of refining primary Si chemically by the addition of phosphorous suffers from a number of problems. In the present paper an alternative, physical means of refining primary Si by intensive shearing of the melt prior to casting is investigated. Al-15wt%Si alloy has been solidified under varying casting conditions (cooling rate) and the resulting microstructures have been studied using microscopy and quantitative image analysis. Primary Si particles were finer, more compact in shape and more numerous with increasing cooling rate. Intensive melt shearing led to greater refinement and more enhanced nucleation of primary Si than was achieved by adding phosphorous. The mechanism of enhanced nucleation is discussed.

A new direct chill (DC) casting process, melt conditioned DC (MC-DC) process, has been developed for the production of high quality billets/slabs of light alloys by application of intensive melt shearing through a rotor-stator high shear device during the DC casting process. The rotor-stator high shear device provides intensive melt shearing to disperse the naturally occurring oxide films, and other inclusions, while creating a microscopic flow pattern to homogenize the temperature and composition fields in the sump. In this paper, we report the grain refining effect of intensive melt shearing in the MC-DC casting processing. Experimental results on DC casting of Mg-alloys with and without intensive melt shearing have demonstrated that the MC-DC casting process can produce magnesium alloy billets with significantly refined microstructure. Such grain refinement in the MC-DC casting process can be attributed to enhanced heterogeneous nucleation by dispersed naturally occurring oxide particles, increased nuclei survival rate in uniform temperature and compositional fields in the sump, and potential contribution from dendrite arm fragmentation.

The precipitation of Mn-rich intermetallics in AlSiMn alloys during solidification ahead of the mushy zone affects the solidification microstructure, especially if fluid flow is present. Recently Steinbach and Ratke reported a barrier effect of α-AlMnSi, meaning these intermetallics prevent fluid flow to enter the mush. To investigate this effect further we studied the solidification of AlSi7Mn1 with a fluid flow field induced by a traveling magnetic field (TMF). Samples were molten and directionally solidified within a silica aerogel crucible at various constant solidification velocities between 0.03 and 0.24 mm/s. The application of two separate heaters allowed the fixation of constant temperature gradients in the solid and liquid parts of the samples, the use of a transparent aerogel as crucible material permitted direct optical verification of the desired solidification velocity using an infrared line camera. Three collinear coils induced a TMF of approximately 5 mT strength, traveling either up or downward in the direction of the sample axis. The microstructures of the processed samples were studied using light microscopy and SEM-EDX and characterised by the primary and secondary dendrite arm spacing, the distribution of intermetallic phases as well as the radial segregation of primary phase and eutectic. Results are presented which show differences between samples with and without TMF-induced fluid flow. We noticed a radial macro-segregation dependent on the orientation of the TMF and the effects of the induced fluid flow on the primary and secondary dendrite arm spacing are examined.

The free dendrite growth in a pure substance, supposed to a parallel forced convection which is modulated in time, is studied using two-dimensional phase-field simulations. We show that the response of the tip velocity to the velocity pulses exhibits a first-order time delay. Most notably, when the pulse duration is chosen appropriately a resonant side-branching can be achieved. This provides a way to manipulate the dendrite morphology and consequently its mechanical properties.

Solidification structure of a High Strength Low Alloy (HSLA) steel, in terms of dendrite arm spacing distribution across the shell thickness, is studied in a breakout shell from a thin-slab caster at Tata Steel in IJmuiden. Columnar dendrites were found to be the predominant morphology throughout the shell with size variations across the shell thickness. Primary Dendrite Arm Spacing (PDAS) increases by increasing the distance from meniscus or slab surface. Subsequently, a model is proposed to describe the variation of the PDAS with the shell thickness (the distance from slab surface) under solidifiction conditions experienced in the primary cooling zone of thin-slab casting. The proposed relationship related the PDAS to the shell thickness and, hence, can be used as a tool for predicting solidifcation structure and optimizing the thin-slab casting of low alloy steels.

Oxides, in liquid aluminium alloys, can cause severe difficulties during casting, contribute to the formation of cast defects and degrade the mechanical properties of cast components. In this paper, microstructural characteristics of naturally occurring oxides in the melts of commercial purity aluminium and Al-Mg binary alloys have been investigated. They are characterised by densely populated oxide particles within liquid oxide films. With intensive shearing, the particle agglomerates are dispersed into uniformly distributed individual particles. It was found that with intensive melt shearing, grain refinement of α-Al can be achieved by the dispersed oxide particles. The smaller lattice misfit between the oxide particles and the α-Al phase is characterised by a well defined crystallographic orientation relationship. And the mechanisms of grain refinement are discussed.

The influence of transverse magnetic field on the liquid-solid interface stability and morphology has been investigated in directionally solidified Al-0.85wt%Cu alloy. Experimental results show that the transverse magnetic field causes the interface to be instable and the interface shape to be depressed on one side along the radius. The interface instability increases with increasing magnetic field. Increasing the solidification velocity reduced extent of interface destabilization by the magnetic field. The depression of the interface with the magnetic field is more dramatic at low solidification velocities. These phenomena are attributed to the thermoelectromagnetic convection (TEMC) on the interface and cellular scale.

In the course of the unidirectional solidification performed in a rotating magnetic field (RMF), it can be observed that the flows, developing in the metallic alloy, melting under the influence of the magnetic field, have a significant effect on the developing structure. In the references [1, 2], the values of induction and frequency of magnetic fields creating the stirring, as well as the geometrical sizes of solidified samples, are used as parameters in case of identical alloys. An exact comparison can only be made in this case if the melt flows are similar during the procedures. However, the intensity of flow developing in the melt is influenced, not only by the geometrical parameters of magnetic field and that of the solidified samples, but by other various parameters as well.

An Al-20Mg-4Si high Mg containing alloy has been produced and its characteristics investigated. The as-cast alloy revealed primary Mg₂Si particles evenly distributed throughout an α-Al matrix with a β-Al₃Mg₂ fully divorced eutectic phase observed in interdendritic regions. The Mg₂Si particles displayed octahedral, truncated octahedral, and hopper morphologies. Additions of Sb, Ti and Zr had a refining influence reducing the size of the Mg₂Si from 52 ± 4 μm to 25 ± 0.1 μm, 35 ± 1 μm and 34 ± 1 μm respectively. HPDC tensile test samples could be produced with a 0.6 wt.% Mn addition which prevented die soldering. Solution heating for 1 hr was found to dissolve the majority of the Al₃Mg₂ eutectic phase with no evidence of any effect on the primary Mg₂Si. Preliminary results indicate that the heat treatment has a beneficial effect on the elongation and the UTS.

Directional solidification experiments were performed in PbSn and AlSi alloys under the influence of a rotating magnetic field. The solidification process was started without forced melt convection before applying a sudden onset of the electromagnetic force. Just before and after the initiation of the magnetic field the grain structure was analysed. Electrolytic etching was used to identify grains having different orientations and related EBSD measurements provided quantitative values for the crystallographic orientation of the grains. Whereas a distinct asymmetric grain growth or a deflection of the columnar grains becomes obvious in case of forced melt flow, a perceivable modification of the crystal orientation cannot be detected.

The effect of a high magnetic field on dendritic morphology in Al-10at%Cu alloy was investigated with help of differential thermal analysis (DTA). The DTA curves showed that nucleation temperature of primary α-Al phase was reduced and its growth rates increased in magnetic fields. The reduction of nucleation temperature was due to the increase of interfacial free energy between liquid and Al nucleus in a magnetic field, while imposition of a magnetic field modified the effective distribution coefficient and further altered the solutal and curvature undercooling, which made the driving force for dendritic growth becoming larger. It was shown that the dendritic morphology was changed from disordered dendrites to regular ones and the angle between preferred growth direction and magnetic field is about 55° in 12T. The formation of regular primary dendrites could be attributed to magnetic orientation and the suppression of free convection.

We present an experimental study devoted to the directional solidification of Sn-Pb alloys affected by electromagnetically-driven melt convection. Within this paper the influence of rotating, travelling and the superposition of both field types is compared. The combination of rotating (RMF) and travelling magnetic fields (TMF) was realised in form of subsequent pulses. The analysis was focused on flow-induced modifications of the temperature field and the solidified structure.

This work is devoted to the two- and three-dimensional modeling of pure Gallium melting in a cuboid cavity, as has been studied experimentally by Gau and Viskanta (C. Gau & R. Viskanta. J. Heat Transfer 108, pp. 171-174, 1986). The geometry and boundary conditions used in this work are taken from the benchmark experiment. The fixed-grid technique with artificial porosity was implemented into a finite-volume code, where the pressure-velocity coupling was treated using the SIMPLE algorithm. The results of simulations showed that the so-called multicellular flow pattern occurs only in two-dimensional (2D) simulations at the beginning of melting. However, an analysis of three-dimensional simulations (3D) showed the absence of a multivortex structure at any time step. This fact is explained by the presence of walls in the third direction which suppress the flow and by the occurrence of weak turbulence in the bulk of the melt preventing the formation of the long-living multicellular flow pattern known from previous 2D simulations. Additionally, we studied the influence of the cuboid width on the flow structure and solid front shapes. The results are discussed.

A two-phase columnar solidification model was used to study the formation mechanism of the channel segregation in a Sn-10 wt.% Pb benchmark ingot. The two phases considered in the current model are the melt and columnar phase. The morphology of the columnar dendrite trunks is approximated as step-wise cylinders with constant primary dendrite arm spacing. The growth kinetics of the columnar trunks is governed by the diffusion of the rejected solute element surrounding the columnar trunks. From the 3D modelling result one can 'visualize' the dynamic formation sequence of the channel segregation. The final channel segregation pattern shows a discontinuous and lamellar- or plume-structure. The formation mechanism of the channels is due to the disturbed flow, which can accelerate and retard the local solidification rate in the mushy zone and near the solidification front. This study has also verified that the channel segregation can form in the condition without remelting, namely, remelting is not a necessary condition for channel segregation.

Macrosegregation occurs in continuously cast products due to shrinkage, thermo-solutal natural and forced convection, floating of the solid grains and the deformation of the mushy zone, induced by the supporting rolls. A simple Lagrange-an traveling slice model has been successfully used in the past for prediction of the relations between the process parameters and the temperature field [1], grain structure [2], optimization of process parameters, and calculation of caster regulation coefficients. It is the purpose of this paper to demonstrate the applicability of this simple model for assessment of the macrosegregation. The main advantage of the slice model is its very fast calculation time in comparison with the complete 3D heat, solute and fluid flow model. The macroscopic model is based on continuum mixture theory by solving the enthalpy and concentration equations, while the microsegregation model is based on the Scheil rule. The fluid flow effects in the melt have been approximated by the increased solute diffusivity in the liquid phase of the slice. Ordinary, binary carbon steel is used for demonstration in this study. A sensitivity study of the approach on the recently introduced standard continuous casting configuration is performed with respect to inlet superheat, casting speed and cooling rates. The model surprisingly properly predicts the qualitative features of the macrosegregation pattern. Possible refinements of the model with respect to the additional physical mechanisms are discussed.

The aim of this paper is the simulation of dendritic growth in steel in two dimensions by a coupled deterministic continuum mechanics heat and species transfer model and a stochastic localized phase change kinetics model taking into account the undercooling, curvature, kinetic, and thermodynamic anisotropy. The stochastic model receives temperature and concentration information from the deterministic model and the deterministic heat, and species diffusion equations receive the solid fraction information from the stochastic model. The heat and species transfer models are solved on a regular grid by the standard explicit Finite Difference Method (FDM). The phase-change kinetics model is solved by a novel Point Automata (PA) approach. The PA method was developed [1] in order to circumvent the mesh anisotropy problem, associated with the classical Cellular Automata (CA) method. The PA approach is established on randomly distributed points and neighbourhood configuration, similar as appears in meshless methods. A comparison of the PA and CA methods is shown. It is demonstrated that the results with the new PA method are not sensitive to the crystallographic orientations of the dendrite.

The main aim of this paper is to demonstrate the applicability and the advantages of a novel meshless method for simulation of macrosegregation in steel billets. The physical model is established on a set of macroscopic equations for mass, energy, momentum, species, turbulent kinetic energy, and dissipation rate in two dimensions. The mixture continuum model is used to treat the solidification system. The mushy zone is modelled as a Darcy porous media with Kozeny-Karman permeability relation, where the morphology of the porous media is modelled by a constant value. The incompressible turbulent flow of the molten steel is described by the Low-Reynolds-Number (LRN) k-epsilon turbulence model, closed by the Launder and Sharma closure coefficients and damping functions. The microsegregation equations rely on lever rule. The numerical method is established on explicit timestepping, and collocation with multiquadrics radial basis functions on non-uniform five-noded influence domains, and adaptive upwinding technique. The velocity-pressure coupling of the incompressible flow is resolved by the explicit Chorin's fractional step method, with the intermediate velocity field, calculated without the pressure term. A recently proposed standard continuous casting configuration with Fe-C system has been used for verification of the model. The advantages of the method are its simplicity and efficiency, since no polygonisation is involved, easy adaptation of the nodal points in areas with high gradients, almost the same formulation in two and three dimensions, high accuracy and low numerical diffusion.

Low Carbon (LC) steel is not expected to be sensitive to hot tearing and/or cracking while microalloyed steels are known for their high cracking sensitivity during continuous casting. Experience of the Direct Sheet Plant caster at Tata Steel in Ijmuiden (the Netherlands), seems to contradict this statement. It is observed that a LC steel grade has a high risk of cracking alias hot tearing, while a High Strength Low Alloyed (HSLA) steel has a very low cracking occurrence. Another HSLA steel grade, with a similar composition but less N and V is however very sensitive to hot tearing. An extreme crack results in a breakout. A previous statistical analysis of the breakout occurrence reveals a one and a half times higher possibility of a breakout for the HSLA grade compared to the LC grade. HSLA with extra N, V shows a four times smaller possibility of breakout than LC. This study assigns the unexpected effect of the chemical composition on the hot tearing sensitivity to the role of some alloying elements such as V and N as structure refiners.

The solidification of hypoeutectic gray cast iron starts with the nucleation of primary austenite crystals. Before graphite is nucleated, and the eutectic structure is formed, these crystals start to grow as columnar or equiaxed dendrites. However, very little is known about these dendrites, and especially how they influence the subsequent eutectic structure. Besides, it has previously been shown that the primary solidification structure influences the formation of defects. Shrinkage porosity was found between the dendrites, in the grain boundaries, and the formation of the primary solidification structure was found to influence problems related to metal expansion penetration. Therefore a better understanding about the formation of this structure is of importance. In this work, different inoculants and their influence on the formation of the micro- and macrostructures has been investigated. The inoculants considered are commercially used inoculants, i.e. inoculants used in the foundries, as well as different iron powders. The addition of iron powder is used to promote the primary solidification structure. It is shown that the nucleation of the dendrites is influenced by the amount of iron powder. Secondary dendrite arm spacing is a quantitative measurement in the microstructure related to these dendrites, which in turn depends on the solidification time. Eutectic cell size, on the other hand, is found to depend on secondary dendrite arm spacing. It is shown how the addition of inoculants influences both primary and eutectic solidification structures, and how they are related to each other.

Many advanced steels fall within the peritectic composition range, which are notoriously difficult to cast due to cracking and breakout problems in the continuous casting process especially at high casting speeds. In this study an attempt was made to obtain practical understanding of the solidification and the δ→γ phase transformation of various commercial steels using high-temperature laser-scanning confocal microscopy. Under rapid cooling conditions the transformation morphology showed a massive-type of transition rather than a classical diffusion-controlled transformation.

Heterogeneous nucleation of nodular graphite at inclusions in ductile iron during eutectic solidification has been investigated. The experimental part of this work deals with casting of ductile iron samples with two different inoculants in four different thicknesses. Chemical analysis, metallographic investigation and thermal analysis of the specimens have been carried out. A numerical model has been implemented and the results (i.e. cooling curve, cooling rate, nodule count and solid fraction) have shown a good agreement with experimental studies; following this, inoculation parameters in the model have been studied and discussed.

To enhance the quality of tool steels it is necessary to analyse all stages of the production process. During the ingot- or continuous casting processes and the following solidification, material and geometry depending reactions cause defects such as macro segregations or porosities. In former times the trial and error approach, together with the experience and creativity of the steelworks engineers was used to improve the as-cast quality, with a high amount of test procedures and a high demand of research time and costs. Further development in software and algorithms has allowed modern simulation techniques to find their way into industrial steel production and casting-simulations are widely used to achieve an accurate prediction of the ingot quality. To improve the as-cast quality, several ingot casting processes of tool steels were studied at the R&D department of Deutsche Edelstahlwerke GmbH by using the numerical casting simulation software MAGMASOFT^®. In this paper some results extracted from the simulation software are shown and compared to experimental investigations.

Pipeline systems for hydraulic networks are obtained via centrifugal casting of spheroidal graphite cast iron. The very high cooling rate that is achieved in the skin of the product can sometimes lead to carbide instead of graphite in cast iron. An experimental device has been built in the laboratory that allows reproducing the extreme thermal conditions encountered during formation of skin of centrifugally cast pipes. Liquid metal droplets fall on a cold substrate. Rapid directional solidification occurs. The temperature evolution of the lower surface of the droplet is recorded during the very first moment of the solidification (t < 200 ms) thanks to a photodiode, which is located below the substrate. The microstructures that are obtained in laboratory are characterised in both the as-cast state and the heat-treated state. They are compared to the centrifugally cast ones. A model of directional solidification of cast iron under a very large temperature gradient has been built. It allows explaining the transition from stable to metastable micro structure that was observed in the products and reproduced in the laboratory samples.

The existing twin roll casting technique for magnesium alloys suffers heterogeneity in both microstructure and chemistry and downstream processing is required to improve the strip quality, resulting in cost rise. In the present work, twin roll casting was carried out using an AZ31 magnesium alloy, with the application of intensive shearing melt conditioning prior to casting. The effect of process parameters such as pouring temperature and casting speed on microstructure control during casting and subsequent downstream processing was studied. Experimental results showed that the melt conditioning treatment allowed the production of AZ31 strips with uniform and refined microstructure free of centreline segregations. It was also shown that an optimized combination of pouring temperature and casting speed, in conjunction with a strip thickness control operation, resulted in uniformly distributed stored energies due to enhanced plastic deformation, which promoted recrystallization during casting and subsequent heat treatment. Strips prepared by twin roll casting and homogenization developed similar microstructural features to those prepared by twin roll casting followed by lengthy downstream processing by homogenization, hot rolling and annealing and displayed a weaker basal texture, exhibiting a potentially better formability.

Micro-shrinkage porosity in aluminum casting is analysed by computer simulation using three criteria functions and a fully-coupled shrinkage porosity model. Three process simulations of different precision were executed as basis for the porosity prediction for investigation of the impact of simulation precision on porosity prediction. To validate the simulation predictions, three identical blocks were cast in a special experimental setup. Chill plates enforce shrinkage porosity, which was analyzed using computer tomography. The results demonstrate that both precise numerical simulations and precise porosity models are needed for reliable porosity prediction.

The flow and filling characteristics during injection of liquid aluminum during high-pressure die-casting is studied threefoldly: a) analytically, b) experimentally and c) numerically. A planar jet of liquid aluminum is formed at the ingate due to its small width (≈O(10⁻³) m), its high aspect ratio (≈ 100) and high inlet velocity (up to 60 m/s). On the one hand, wavy disintegration of such a jet can inevitably lead to cold runs in the final casting. On the other hand, a high degree of atomization may strongly increase the porosity of the casting part. Both processes can highly reduce the mechanical stability of the product. Analytical investigations of Ohnesorge (or equivalently Weber) and Reynolds numbers show that the process of drop formation at the liquid planar free jet is dominated by atomization assuming an orifice nozzle geometry at the ingate. From a simple experimental investigation of an equivalent free jet of water, however, it is deduced that the process of drop formation can be changed to wavy disintegration by the nozzle geometry. Numerically, high-pressure die-casting is attacked by a Volume of Fluid approach. Although the drop formation at the phase interphase can not be captured by the numerical model since the drops are an order of magnitude smaller than feasible grid spacings, the global spreading of the free jet in the casting mold is well pictured by this first numerical simulation. In addition, a new approach is presented to detect cold runs at the final casting. Finally, the studies presented lead to an increased understanding of high pressure die casting and can help to improve the quality of casting products.

Rapid slurry formation is a semi-solid metal forming technique, which is based on a so-called solid enthalpy exchange material (EEM). It is a fascinating technology offering the opportunity to manufacture net-shaped metal components of complex geometry in a single forming operation. At the same time, high mechanical properties can be achieved due to the unique microstructure and flow behaviour. The major process parameters used in the RSF process are rotation speed of the EEM, melt superheat, amount of EEM added (determining fs), and holding time. The process parameters can be well controlled with clear effects on the microstructure. There is a lack of theoretical modelling of the morphological evolution in these two-phase slurries.

Macrosegregation in direct chill casting processes is controlled by fluid flow due to the thermosolutal natural and forced convection, shrinkage, and transport of unattached solid grains. Because grain refinement is usually used in aluminum direct chill casting, some effort must be made to model free-floating solid grains, and their attachment to a rigid mushy zone. Criteria for attachment vary, but many are based on using a critical solid packing fraction, which is treated as uniform and constant throughout the domain. In the case of horizontal casting (HDC), gravity acts perpendicularly to the casting direction, and the assumption of a uniform packing fraction cannot be applied because the solid particles attach to some surfaces by settling and others by being swept into the rigid solid from below. In this simulation of HDC casting of an Al-Cu-Mg alloy, the rigid and unattached solid is tracked separately, and a rule set is developed to determine the attachment of free-floating solid. Comparison between cases with and without unattached solid movement shows qualitatively different results, particularly in bottom part of slab. Non-uniform packing fractions cause very different segregation patterns in the lower half of the ingot compared to the cases with no solid movement, less segregation near centerline compared to uniform packing fraction cases, and positive segregation near the place where inlet jet impinges on the mushy zone.

The grain structure formation in direct chill (DC) casting is directly linked to nucleation, which is generally promoted by inoculation. Inoculation prevents defects, but also modifies the physical properties by changing the microstructure. We studied the coupling of the nucleation on inoculant particles and the grain growth in the presence of melt flow induced by thermosolutal convection and of the transport of free-floating equiaxed grains. We used a volume-averaged two-phase multiscale model with a fully coupled description of phenomena on the grain scale (nucleation on grain refiner particles and grain growth) and on the product scale (macroscopic transport). The transport of inoculant particles is also modeled, which accounts for the inhomogeneous distribution of inoculant particles in the melt. The model was applied to an industrial sized (350mm thick) DC cast aluminium alloy ingot. A discretised nuclei size distribution was defined and the impact of different macroscopic phenomena on the grain structure formation was studied: the zone and intensity of nucleation and the resulting grain size distribution. It is shown that nucleation in the presence of macroscopic transport cannot be explained only in terms of cooling rate, but variations of composition, nuclei density and grain density, all affected by transport, must be accounted for.

The objective of this work is to determine the microstructural characteristics of investment cast cobalt alloy as the cross-sectional area is varied, thus changing the local effective cooling rates and solidification times. The extent of published work on the as-cast properties of cobalt alloys is minimal. The primary aim of this work is therefore to extend knowledge of the behaviour of such alloys as they solidify, which will influence the design of new products as well as the industrial optimisation of the casting process. Wedge-shaped parts were cast from a biomedical grade cobalt alloy employing the method of lost wax investment casting. Analytical techniques such as optical microscopy, image analysis and microhardness testing were used to characterise the as-cast parts. Parameters studied include variations in grain structure, nature of the columnar and equiaxed zones and the spread of porosity (both shrinkage and gas). Changes in microstructure were compared to microhardness values obtained. The solidification profile of the alloy through the prototype cast component was investigated based on measurement of the dendrite arm spacings. A discussion on the physical phenomena controlling the microstructural variations is presented.

Investment casting of magnesium is a well suited process for the production of aeronautic and automotive components. But still, this process has not been properly developed. One reason for that are the reactions between the Mg melt and the ceramics of the mould that produce a non-desired oxide layer on the part surface. These reactions can be inhibited by the use of silica-free slurries with a higher stability than conventional ones. Another way is using inhibitors, chemical compounds based in fluorides that react with the melt, creating a protective surface layer in the casting. With the aim of developing a reaction-free process, alumina moulds with a stepped geometry have been constructed. These provide different interface conditions. Conventional SF₆, non-conventional KBF₄ and NaBF₄ and environmentally friendly FK inhibitors have been tested on. As a result, KBF₄ has been identified as the most suitable inhibitor for magnesium investment casting. Furthermore, the analysis of the cooling curve of different interfaces has provided essential information about the reaction mechanism of the inhibitors.

Five-step castings of aluminum alloy A443 with different section thicknesses (2, 6, 8, 10, 20 mm) were squeezed under a hydraulic pressure of 60 MPa. Temperatures inside the P20 steel die mold and at the casting surface were recorded by fine type-K thermocouples. A numerical solution, i.e. inverse method was employed to determine the casting-die interfacial heat transfer coefficients (IHTCs). The results show the IHTCs initially reached a maximum peak value followed by a gradually decline to a lower level. With the applied pressure of 60MPa, the peak IHTC values from steps 1 to 5 varied from 5629 W/m²K to 9419 W/m²K. The section thickness affected IHTC peak values significantly. Compared to the thin steps at upper cavity, the lower thick steps obtained higher peak IHTCs and heat flux values due to high local pressures and high melt temperature.

Aluminum alloy 7050 is of interest for aerospace industries due to its superior mechanical properties. However, its inherent solidification behaviour may augment the accumulation of residual stresses due to uneven cooling conditions upon direct-chill (DC) casting. This can increase the propensity of cold cracking (CC), which is a potentially catastrophic phenomenon in casting ingots. To predict the outcome of the aluminum casting process, ALSIM software is utilised. This software has the capability to predict CC susceptibility during the casting process. However, at the moment, ALSIM lacks the information regarding material constitutive behaviour in the sub-solidus temperature range, which is considered important for studying CC phenomenon. At the moment, ALSIM only has a partial constitutive database for AA7050 and misses data, especially in the vicinity of non-equilibrium solidus (NES) point. The present work presents measurements of tensile constitutive parameters in the temperature range between 400 °C and NES, which is for this alloy defined as 465 °C. The mechanical behaviour is tested in a Gleeble 3800 thermo-mechanical simulator. Constitutive parameters such as stress-strain curves, strain-rate sensitivity and ductility of the alloy have been measured at different test temperatures. With these constitutive data, we expect to improve the accuracy of ALSIM simulations in terms of CC prediction, and gain more insight into the evolution of mechanical properties of AA7050 in the temperature nearby the NES.

Fe is one of the inevitable and detrimental impurities in aluminium alloys that degrade the mechanical performance of castings. In the present work, intensive melt shearing has been demonstrated to modify the morphology of Fe-containing intermetallic compounds by promoting the formation of compact α-Al(Fe,Mn)Si at the expense of needle-shaped β-AlFeSi, leading to an improved mechanical properties of LM24 alloy processed by MC-HPDC process. The promotion of the formation of α -Al(Fe, Mn)Si phase is resulted from the enhanced nucleation on the well dispersed MgAl₂O₄ particles in the melt. The Fe tolerance of LM24 alloy can be effectively improved by combining Mn alloying and intensive melt shearing.

A phase-field model of non-isothermal solidification in binary alloys has been used to study the variation of growth velocity, dendrite tip radius and radius selection parameter, σ*, as a function various process and materials parameters. We find that σ* is non-constant and, over a wide parameter space, displays first a local minimum followed by a local maximum as the undercooling is increased. This behaviour is contrasted with a similar type of behaviour to that predicted by simple marginal stability models to occur in the radius of curvature for constant σ*.

The authors present an integrated numerical model for the simulation of laser spot welding of an aluminium alloy at meso-scale in 2D. This model deals with the melting of the parent materials which form the weld pool and the subsequent solidification of the liquid metal in the pool, during the welding process. The melting of the parent materials due to the applied heating power is an important phenomenon, which determines the conditions at the onset of solidification, such as the geometry of the weld pool and the distribution of the temperature field. An enthalpy method is employed to predict the melting during the heating phase of welding. A Gaussian distribution is used to model the heat input from the laser. Once the laser beam is switched off and the melting halts, solidification commences. The UCD front tracking model [1,2] for alloy solidification is applied to predict the advancement of the columnar dendritic front, and a volume-averaging formulation is used to simulate nucleation and growth of equiaxed dendrites. A mechanical blocking criterion is used to define dendrite coherency, and the columnar-to-equiaxed transition within the weld pool is predicted.

Cobalt-based alloys ribbons for the fabrication of cored wires require high elongation, e.g. a good metallurgical quality associated with a fine grain structure. The melt-spinning process allows fabricating the 200 μm thick ribbons in a single casting step. However, the properties of such ribbon suffer from the presence of casting defects, leading to early fractures in the material. The present study investigates the microstructure and defect formation in melt-spun Co-6.5Fe-2.6Mn-(0-1.6)Si alloys (in wt.%). These alloys exhibit uncommon intrinsic behaviour and properties such as high nucleation undercoolings (up to 295K in DTA experiments) or reduced solidification interval (10 < T₀ < 25 K). It is found that the main defect leading to fracture is micro-shrinkage due to the solutal dendrite growth at low nucleation undercooling. Low silicon additions (0.6 wt.%) are found to reduce the size and the occurrence of shrinkage and increase the elongation, whereas larger silicon additions (1.6 wt.%) are detrimental to the ribbon properties.

Optimisation of laser sintering of submicron metal powders has been studied in connection with unsteady heat transfer in a porous powder layer under conditions of rapid phase transformations. The heating rate, cooling rate and depth of the sintered layer are estimated after analysis of geometrical characteristics of the metallic powder. The computer simulation revealed that the control parameters of the process are the scanning velocity and effective optical penetration depth. The last parameter depends on porosity and morphology of the powder layer. The effects of the laser annealing power, frequency of laser pulses and beam radius have a smaller effect on the depth of the sintering layer. At porosities higher than 70%, the mechanism of heat transfer drastically changes and the approximation of continuum used frequently in analytical models for draft estimations becomes incorrect.

Smoothed particle hydrodynamics is employed, for the first time, to develop a numerical model for the melting and fluid flow during laser welding process. In this meshlessLagrangian method the gas-melt two phase flow, heat transfer, surface tension, and melting of solid parent material are considered. This model was used to study the evolution of temperature field and fluid flow in the case study of laser spot welding in 2D. The simulation results show a strong influence of the melting process on the flow of liquid metal and a clear influence of the Marangoni flow on the heat transfer is also found.

Formation of Al₅₀Ti₅₀, Al₅₀Ti₄₅Ni₅ and Al₅₀Ti₄₀Ni₁₀ amorphous alloys by ball milling and their mechanical crystallization upon further milling have been investigated. Ni addition improves the thermal stability and increases the strength of the powder compacts obtained by hot pressing.

Based on critical review for the available experimental phase diagram data of the Al-Cu-Fe-Mn, Al-Cu-Fe-Ni, Al-Cu-Fe-Si, Al-Fe-Mg-Si, Al-Fe-Mn-Si, and Al-Mg-Mn-Zn systems, a set of self-consistent thermodynamic parameters for these systems has been established using CALPHAD approach. In combination with the constituent binary, ternary, and quaternary systems, a thermodynamic database for the Al-Cu-Fe-Mg-Mn-Ni-Si-Zn system is developed. The calculated phase diagrams and invariant reactions agree well with the experimental data. The obtained database has been used to describe the solidification behaviour of Al alloys: Al365.1(91.95Al-0.46Fe-0.3Mg-0.32Mn-6.97Si, in wt.%) and Al365.2 (92.77Al-0.08Fe-0.35Mg-6.8Si, in wt.%) under both equilibrium and Gulliver-Scheil non-equilibrium conditions. The reliability of the present thermodynamic database is verified by the good agreement between calculation and measurement for both equilibrium and Gulliver–Scheil non-equilibrium solidification.

The two-dimensional equilibrated grain boundary grooves (GBG) of succinonitrile (SCN) was directly observed by using a temperature gradient stage with high temperature stability to measure the solid-liquid interfacial free energy and its anisotropic factor. Based on the numerical solution of Gibbs-Thomson equation and parameter optimization method, the liquid-solid interfacial free energy and its anisotropy of SCN were extracted from the digitised GBG shapes respectively. The results show that the average solid-liquid interfacial free energy is γ₀ = 9.66×10⁻³J·m⁻², and the average anisotropic factor is ₄ = 0.85%.

Time-resolved absorption imaging using synchrotron radiation X-rays allows us to observe solidification of metallic alloys of interest. This paper presents peritectic solidification in Sn-Cd alloy and Fe-C alloys. In unidirectional solidification of Sn-Cd alloy, the formation of a banded structure, in which two phases were alternatively piled up in the growth direction, was clearly observed. Sequence of nucleation or fluctuation of triple junction (primary phase / secondary phase / liquid) resulted in the banded structure. Ease of nucleation for both phases contributed to the banded structure formation. In carbon steel (Fe-0.45mass%C), the transformation from δ phase to γ phase was observed. At lower cooling rates, γ phase was produced in semisolid state of δ phase and liquid, indicating the peritectic reaction occurred during solidification. In contrast, δ phase transformed into γ phase when solidification nearly completed at temperatures 100K below the liquidus temperature. Namely, the transformation seemed to be massive. The observation showed that two different transformation modes operated in Fe-C alloy.

Visualizations of the solidification process were obtained by means of X-ray radioscopy within a Hele-Shaw cell filled with Ga-25wt%In alloy. Thermo-solutal convection in the solidifying melt gives rise to the development of vertical segregation channels ("chimneys"). The probability of chimney formation depends sensitively on variations of the horizontal temperature distribution. A forced melt flow perpendicular to the growth direction accelerates the growth of the secondary dendrite arms on the upstream side and suppresses the development of secondary arms on the downstream side. The primary dendrite arm spacing is increased, whereas the secondary arm spacing remains unaffected. How-induced modifications of the local composition were observed within the mushy zone, which may contribute to the formation of spacious segregation pattern.

Understanding phenomena occurring at the scale of the crystals during the deformation of semi-solid alloys is important for the development of physically-based rheological models. A range of deformation mechanisms have been proposed including agglomeration and disagglomeration, viscoplastic deformation of the solid skeleton, and granular phenomena such as jamming and dilatancy. This paper overviews in-situ experiments that directly image crystal-scale deformation mechanisms in equiaxed Al alloys at solid fractions shortly after the crystals have impinged to form a loose crystal network. Direct evidence is presented for granular deformation mechanisms including shear-induced dilation in both equiaxed-dendritic and globular microstructures. Modelling approaches suitable for capturing this behaviour are then discussed.

Dynamical microstructure formation and selection during solidification processing, which has a major influence on the properties in the use of elaborated materials, occur during the growth process. In situ observation of the solid-liquid interface morphology evolution is thus necessary. On earth, convection effects dominate in bulk samples and may strongly interact with microstructure dynamics and alter pattern characterization. Series of solidification experiments with 3D cylindrical sample geometry were conducted in succinonitrile (SCN) –0.24 wt%camphor (model transparent system), in microgravity environment in the Directional Solidification Insert of the DECLIC facility of CNES (French space agency) on the International Space Station (ISS). Microgravity enabled homogeneous values of control parameters over the whole interface allowing the obtaining of homogeneous patterns suitable to get quantitative benchmark data. First analyses of the characteristics of the pattern (spacing, order, etc.) and of its dynamics in microgravity will be presented.

The irregular growth dynamics of the so-called locked (tilted) lamellar eutectic grains that are observed in directional solidification of nonfaceted/nonfaceted eutectic alloys, is attributable to a strong surface tension anisotropy of the interphase boundaries, which enters into the local-equilibrium (Young-Herring) condition at the trijunctions of the solid-liquid interfaces. Based on real-time observations of locked eutectic growth in thin samples, we propose that the lamellar tilt angle is selected by the system in such a way that the Hoffmann-Calm surface tension force ( vector) of the interphase boundaries is approximatively perpendicular to the solidification front.

The phenomena involved during equiaxed growth are dynamic, so that in situ and real-time investigation by X-ray imaging is compulsory to fully analyse the microstructure formation. The experiments on Al - 10 wt% Cu alloy of this paper are carried out at the European Synchrotron Radiation Facility (ESRF) in Grenoble (France). Equiaxed growth was achieved in nearly isothermal conditions and continuously monitored from the very early stages of solidification to an asymptotic state. First, measurements of dendrite arms velocity for a same grain showed slight differences in the early stages of the growth. This effect is attributed to a gravity-related "self - poisoning" of the grain. Then, the propagation of primary dendrite arms was analysed and two successive growth regimes were observed. First, due to the relative distance with neighbour grains, each grain could be considered as isolated (i.e. growing freely) and tip growth rate gradually increased. In a subsequent phase, tip growth rate slowly decreased towards zero, due to the proximity of neighbouring grains. Using an image analysis technique, we were able to measure the solute profiles in the liquid phase between interacting arms. These measurements confirmed that solutal impingement is responsible for stopping the grain growth.

We propose a method to determine the active particle distribution of nucleation undercooling in a refined alloy. The experimental data used in this work are inferred from solidification experiments on a refined Al-3.5 wt% Ni alloy performed with X-ray radiography at the European Synchrotron Radiation Facility. These in situ and real time observations allow the accurate and direct determination of the grain origin (heterogeneous nucleation on particles or fragmentation), of the density and of the equiaxed front growth rate. The LGK classical dendrite growth model is used to evaluate the front undercooling (ΔT_C) corresponding to the measured equiaxed front growth rate. Then, the corresponding cumulative distribution of active refining particles is determined. From this cumulative distribution, we derive the corresponding Gaussian and log-normal laws to obtain the nucleation undercooling distribution of active particles. Results are discussed and compared to available measurements in the literature. The standard particle distribution parameters (density of nuclei, mean nucleation undercooling and standard deviation) are determined. We plan to use the determined nucleation undercooling particle distribution in a stochastic CAFE model for the grain structure without preliminary adjustment of the nucleation undercooling.