The directional solidification of a dilute binary alloy (Bi-1.0 at.%Sn)is investigated. Results are obtained at a gravity level of I pg. Computations are performed in two dimensions with a fixed, non-uniform grid. The simulation involves a solution of the species concentration equation (modified to account for solute rejection at the interface) and energy equation (modified to account for phase-change) for both the solid and liquid phases, in addition to the constitutive equations for describing convective flow in the melt. The effects of conductive heat transfer in the ampoule and in a capillary tube in the sample are included. To gauge the effects of including this growth capillary tube in the apparatus, simulations both with and without the capillary tube are presented and compared. Fully transient simulations have been performed; no simplifying steady-state approximations are used, however, the influence of solute on the melting temperature at the interface is not included. Both thermal and solutal convective cells are seen to form. Convective velocities are significantly damped inside the capillary, causing less segregation due to convection. As solidification proceeds beyond the capillary tube, longitudinal segregation arises as a result of the change in cross-sectional area of solidifying material. The magnitudes of the velocities in this cell increase significantly once the solid/liquid front passes beyond the end of the capillary tube; this causes a corresponding increase in the level of radial solute segregation in the solidified material.

It has been well understood that convective flows induced by g-jitter forces associated with spacecraft are responsible for defects formation and irregularity in product quality during melt growth of single crystals in microgravity. This research is concerned about numerical simulations and experimental measurements for the purpose of developing a fundamental understanding of the g-jitter induced fluid flows and their effects on solidification in microgravity with and without the presence of additional damping forces that are derived from the applied DC magnetic fields. The numerical models include both 2-D and 3-D transient fluid flow, heat transfer, mass transfer and solidification under the combined action of g-jitter and magnetic fields. Numerical simulations using both the 2-D and 3-D models are conducted for both idealized, synthesized and real g-jitter forces, and 2-D simulations are tested against the experimental measurements taken on the thermal oscillator. 2-D solidification models have also been developed and simulations are conducted. Results show that the numerical model predictions compare well with the measurements. Analysis of these results illustrates that an applied magnetic field can have a drastic influence on the convective flows induced by g-jitter and can be particularly useful to suppress the effects resulting from the spiking of g-jitter signatures, which are considered the most detrimental effects on quality of crystals grown in space. Work in progress includes developing 3-D numerical models for solidification phenomena with the presence of both g-jitter and magnetic fields and measurements of flow fluid and its effect on solidification in both transparent fluid and low melting point melts to verify numerical predictions.

The objective of the research is to quantitatively determine and understand the fundamental mechanisms that control the microstructural evolution during solidification of an assemblage of equiaxed dendritic crystals. A microgravity experiment will be conducted to obtain benchmark data on the transient growth and interaction of up to four equiaxed crystals of a pure and transparent metal analog (succinonitrile, SCN) under strictly diffusion dominated conditions. Of interest in the experiment are the transient evolution of the primary and secondary dendrite tip speeds, the dendrite morphology (i.e., tip radii, branch spacings, etc.) and solid fraction, the tip selection criterion, and the temperature field in the melt for a range of initial supercoolings and, thus, interaction strengths between the crystals. The experiment thus extends the microgravity measurements of Glicksman and coworkers for steady growth of a single dendrite [Isothermal Dendritic Growth Experiment (IDGE), first flown on USMP-2] to a case where growth transients are introduced due to thermal interactions between neighboring dendrites - a situation more close to actual casting conditions. Corresponding earth-based experiments will be conducted to ascertain the influence of melt convection. The experiments are supported by a variety of analytical models and numerical simulations. The data will primarily be used to develop and test theories of transient dendritic growth and the solidification of multiple interacting equiaxed crystals in a supercooled melt.

A space experiment, and supporting ground-based research, is proposed to study the microstructural evolution in free dendritic growth from a supercooled melt of the transparent model alloy succinonitrile-acetone (SCN-ACE). The research is relevant to equiaxed solidification of metal alloy castings. The microgravity experiment will establish a benchmark for testing of equiaxed dendritic growth theories, scaling laws, and models in the presence of purely diffusive, coupled heat and solute transport, without the complicating influences of melt convection. The specific objectives are to: determine the selection of the dendrite tip operating state, i.e. the growth velocity and tip radius, for free dendritic growth of succinonitrile-acetone alloys; determine the growth morphology and sidebranching behavior for freely grown alloy dendrites; determine the effects of the thermal/solutal interactions in the growth of an assemblage of equiaxed alloy crystals; determine the effects of melt convection on the free growth of alloy dendrites; measure the surface tension anisotropy strength of succinon itrile -acetone alloys establish a theoretical and modeling framework for the experiments. Microgravity experiments on equiaxed dendritic growth of alloy dendrites have not been performed in the past. The proposed experiment builds on the Isothermal Dendritic Growth Experiment (IDGE) of Glicksman and coworkers, which focused on the steady growth of a single crystal from pure supercooled melts (succinonitrile and pivalic acid). It also extends the Equiaxed Dendritic Solidification Experiment (EDSE) of the present investigators, which is concerned with the interactions and transients arising in the growth of an assemblage of equiaxed crystals (succinonitrile). However, these experiments with pure substances are not able to address the issues related to coupled heat and solute transport in growth of alloy dendrites.

This reports on the MEPHISTO-4 experiment on the Space Shuttle Columbia, STS-87, November 19-December 5, 1997. Involved were NASA; the University of Florida at Gainesville; groups from France that developed and built the furnace; the National Institute of Standards and Technology; The University of New South Wales, Australia; and Purdue University. This was a solidification study in which three long rods of Bismuth- 1 at.% Tin were directionally solidified. The goals were to solidify in an environment free of natural convection; to determine the relationship among solidification growth velocity, growth mode, and temperature; and determine the diffusivity of Sn in Bi. The flight samples grew with a planar solid/liquid interface at velocities less than 3.4 gm/s, and cellular growth was present at velocities greater than 6.7 um/s; grain orientation influenced the planar to cellular transition. The temperature gradient in the liquid was 204 K/cm. The s/l interface was flat with slight concavity. Diffusion-dominated conditions were present during MEPHISTO-4. The Seebeck technique was used to determine the s/I interface temperature during growth, however, to date, analysis of the Seebeck results has not yielded a reliable measurement of the interface temperature. The partition coefficient for Bi alloyed with Sn was measured, k = 0.029.

An experimental study of the horizontal Bridgman growth of pure succinonitrile (SCN) and of a succinonitrile-1.0 mol% acetone alloy (SCN-1.0 mol.% ACE) has been performed. Experiments involving both a stationary thermal field (no-growth case) and a translating thermal field (growth case) were conducted. Growth rates of 2 and 40 micrometers/sec were investigated. For the pure SCN experiments, the velocity field in the melt was estimated using video images of seed particles in the melt. Observations of the seed particles indicate that a primary longitudinal convective cell is formed. The maximum velocity of two different particles which traveled along similar paths was the same and equal to 1.49 +/- 0.01 mm/s. The general accuracy of velocity measurements is estimated to be +/-0.08 mm/s, though the data shows consistency to within +/- 0.02 mm/s. The shape of the solid/liquid interface was also quantitatively determined. The solid/liquid interface was stable (non-dendritic and non-cellular) but not flat: rather it was significantly distorted by the influence of connection in the melt and, for the growth case, by the moving temperature boundary conditions along the ampoule. It was found that the interface shape and position were highly dependent on the alignment of the ampoule in the apparatus. Consequently, the ampoule was carefully aligned for all experiments. The values for front location agree with those determined in previous experiments. For the alloy experiments, the solid/liquid interface was determined to be unstable at growth rates greater than 2.8 micrometers/sec, but stable for the cases of no-growth and growth at 2 micrometers/sec. When compared to the shape of the pure SCN interface, the alloy interface forms closer to the cold zone, indicating that the melting temperature decreased due to the alloying element. Extensive temperature measurements were performed on the outside of the ampoule containing pure SCN. The resulting thermal profiles are presented in detail in the results section. It is intended that the interface shape, thermal boundary condition, and velocity data presented in this paper be used to test numerical simulations.

This report summarizes the results of the In situ Monitoring of Crystal Growth Using MEPHISTO (Material por l'Etude des Phenomenes Interessant de la Solidification sur Terre et en Orbite) experiment on USMP-4. The report includes microstructural and compositional data obtained during the first year of the post flight analysis, as well as numerical simulation of the flight experiment. Additional analyses are being continued and will be reported in the near future. The experiments utilized MEPHISTO hardware to study the solidification and melting behavior of bismuth alloyed with 1 at% tin. The experiments involved repeated melting and solidification of three samples, each approximately 90 cm long and 6mm in diameter. Half of each sample also included a 2 mm. diameter growth capillary, to assist in the formation of single grain inside. One sample provided the Seebeck voltage generated during melting and freezing processes. Another one provided temperature data and Peltier pulsed demarcation of the interface shape for post flight analysis. The third sample provided resistance and velocity measurements, as well as additional thermal data. The third sample was also quenched at the end of the mission to preserve the interface composition for post flight determination. A total of more than 45cm of directionally solidified alloy were directionally solidified at the end of the flight for post mission structural and compositional characterization. Metallurgical analysis of the samples has shown that the interfacial kinetics play a key role in controlling the morphological stability of faceted alloys. Substantial differences were observed in the Seebeck signal between the ground-based experiments and the space-based experiments. The temperature gradient in the liquid for the ground-based experiments was also significantly lower than the temperature gradient in the liquid for the space-based experiments. Both of these observations indicate significant influence of liquid convection for the ground-based experiments.

Experiments were carried out to study the morphological stability of Bi- 1 atomic % Sn alloys using the MEPHISTO directional solidification apparatus aboard Space Shuttle Columbia (STS-87, launched Nov. 19, 1997) and in ground-based studies. The Seebeck signal and temperature measurements indicate that convection was significant for ground-based studies. In the space-based experiments, interface breakdown was observed at growth velocities of 6.7, 27, and 40 microns/sec, but not at 1.8 and 3.3 microns/sec.

The shuttle flight experiment Situ Monitoring of Crystal Growth Using was accomplished during STS-87 as part of the fourth flight of the United States Microgravity Payload (USMP-4), which was flown from November 19 to December 5, 1997. The data returned from that flight are just now beginning to yield quantitative results. This project is an international collaboration: the furnace system known as MEPHISTO was built in France by CNES (French National Space Agency) and CEA (French Atomic Energy Commission); the principal investigator, Prof. Reza Abbaschian, is from the University of Florida at Gainesville; and numerical and analytical modeling support includes collaborators from the University of New South Wales, Australia, the University of Wisconsin at Milwaukee, the National Institute of Standards and Technology, and the NASA Lewis Research Center. MEPHISTO is a French acronym that translates into English as Materials for the Study of Interesting Phenomena of Solidification on Earth and in Orbit. Since this was the fourth flight of the MEPHISTO furnace, the experiment is referred to as MEPHISTO-4. MEPHISTO-4 was a directional solidification experiment that studied the liquid-to-solid transformation of bismuth alloyed with tin. Directional solidification is a freezing technique common to the processing of the electronic materials used in integrated circuits and detectors, such as silicon and germanium. When liquids are frozen on Earth, they must be cooled. The cooling causes stirring because of density variations in the liquid. This stirring, known as natural convection, influences the quality of the resulting solid. During freezing, regions of high and low concentrations of tin are created. This introduces another important phenomenon: diffusion, or the movement by molecular action of matter from regions of high concentration to regions of lower concentration. In MEPHISTO-4, it is tin that diffuses from the high-concentration region in front of the solid-liquid interface to more distant low-concentration regions.

A study of directional solidification of a weak binary alloy (specifically, Bi - 1 at% Sn) based on the fixed grid single domain approach is being undertaken. The enthalpy method is used to solve for the temperature field over the computational domain including both the solid and liquid phases; latent heat evolution is treated with the aid of an effective specific heat coefficient. A source term accounting for the release of solute into the liquid during solidification has been incorporated into the solute transport equation. The vorticity-stream function formulation is used to describe thermosolutal convection in the liquid region. In this paper we present a numerical simulation of g-jitter: the small, rapid fluctuations in gravitational acceleration which may be experienced in an orbiting space vehicle. A background gravity of 1 micro-g has been assumed, and new results for the effects of orientation angle of the periodic disturbances over a range of amplitudes and frequencies on solute field and segregation have been presented.

An experimental and numerical study of the horizontal Bridgman growth of pure succinonitrile (SCN) and of a succinonitrile- 1.0 mol.% acetone alloy (SCN- 1.0 mol.% ACE) has been performed. Experiments at growth rates of 0, 2 and 40 micron/s were investigated. The solid/liquid interface was stable (non-dendritic and non-cellular); however, it was not flat. Rather, it was significantly distorted by the influence of convection in the melt and, for the growth cases, by the moving temperature boundary conditions along the ampoule. For the alloy, the interface was.determined to be unstable at growth rates greater than 2.8 micron/s, but stable for the no-growth and 2 micron/s growth cases. When compared to the pure SCN interface, the alloy interface forms closer to the cold zone, indicating that the melting temperature has been suppressed by the addition of the alloying element. Two-dimensional computer simulations were performed for the no-growth case for both the pure and alloy materials. These simulations indicate that a primary longitudinal convective cell is formed in the melt. The maximum magnitude of velocity was calculated to be 1.515 mm/s for pure SCN and 1.724 mm/s for the alloy. The interface shape predicted by the computer simulation agrees well with the experimentally determined shape for the pure SCN case. In ongoing work, numerical simulations of the process during growth conditions are being performed.

The solidification of a dilute bismuth-tin alloy under Bridgman crystal growth conditions is investigated in support of NASA's MEPHISTO space shuttle flight experiment. Computations are performed in two-dimensions with a uniform grid. The simulation includes the species-concentration, temperature and flow fields, as well as conduction in the ampoule. Fully transient simulations have been performed; no simplifying steady state approximations are used. Results are obtained under microgravity conditions for pure bismuth, and Bismuth-0.1 at.% Sn and Bi-1.0 at.% Sn alloys. The concentration dependence of the melting temperature is neglected; the solid/liquid interface temperature is assumed to be the melting temperature of pure bismuth for all cases studied. For the Bi-1.0 at.% Sn case the results indicate that a secondary convective cell, driven by solutal gradients, forms near the interface. The magnitude of the velocities in this cell increases with time; this causes increasing solute segregation at the liquid/solid interface.

Recent progress in modeling of transport phenomena during dendritic alloy solidification is reviewed. Starting from the basic theorems of volume averaging, a general multiphase modeling framework is outlined. This framework allows for the incorporation of a variety of microscale phenomena in the macroscopic transport equations. For the case of diffusion dominated solidification, a simplified set of model equations is examined in detail and validated through comparisons with numerous experimental data for both columnar and equiaxed dendritic growth. This provides a critical assessment of the various model assumptions. Models that include melt flow and solid phase transport are also discussed, although their validation is still at an early stage. Several numerical results are presented that illustrate some of the profound effects of convective transport on the final compositional and structural characteristics of a solidified part. Important issues that deserve continuing attention are identified.

The MEPHISTO space experiments are collaborative United States and French investigations aimed at understanding the fundamentals of crystal growth. Microgravity experiments were conducted aboard the USMP-1 and -2 missions on STS-52 and 62 in October 1992 and March 1994 respectively. MEPHISTO is a French designed and built Bridgman type furnace which uses the Seebeck technique to monitor the solid/liquid interface temperature and Peltier pulsing to mark the location and shape of the solid/liquid interface. In this paper the Bridgman growth of Sn-Bi and Bi-Sn under terrestrial and microgravity conditions is modeled using the finite element code, FIDAP*. The numerical model considers fully coupled heat and mass transport, fluid motion and solid/liquid phase changes in the crystal growth process. The primary goals of this work are: to provide a quantitative study of the thermal buoyancy-induced convection in the melt for the two flight experiments; to compare the vertical and horizontal growth configurations and systematically evaluate the effects of various gravity levels on the solute segregation. Numerical results of the vertical and horizontal Bridgman growth configurations are presented.

An experimental study was conducted of the crystal growth of succinonitrile during solidification, melting, and no-growth conditions using a horizontal Bridgman furnace and square glass ampoule. For use as input boundary conditions to numerical codes, thermal profiles on the outside of the ampoule at five locations around its periphery were measured along the ampoule's length. Temperatures inside the ampoule were also measured. The shapes of the s/l interface in various two dimensional planes were quantitatively determined. Though interfaces were nondendritic and noncellular, they were not flat, but were highly curved and symmetric in only one unique longitudinal y-z plane (at x=O). The shapes of the interface were dominated by the primary longitudinal flow cell characteristic of shallow cavity flow in horizontal Bridgman; this flow cell was driven by the imposed furnace temperature gradient and caused a 'radical' thermal gradient such that the upper half of the ampoule was hotter than the bottom half. We believe that due to the strong convection, the release of latent heat does not significantly influence the thermal conditions near the interface. We hope that the interface shape and thermal data presented in this paper can be used to optimize crystal growth processes and validate numerical models.

A technique resulting in large undercoolings in bulk samples (23g) of lead-tin alloys was developed. Samples of Pb-12.5 wt percent Sn, Pb-61 wt percent Sn, and Pb-77 wt percent Sn were processed with undercoolings ranging from 4 to 34 K and with cooling rates varying between 0.04 and 4 K/sec. The nucleation behavior of the Pb-Sn system was found to be nonreciprocal. The solid Sn phase effectively nucleated the Pb phase of the eutectic; however, large undercoolings developed in Sn-rich eutectic liquid in the presence of the solid Pb phase. This phenomenon is believed to be mainly the result of differences in interfacial energies between solid Sn-eutectic liquid, and solid Pb-eutectic liquid rather than lattice misfit between Pb and Sn. Large amounts of segregation developed in the highly undercooled eutectic ingots. This macrosegregation was found to increase as undercooling increases. Macrosegregation in these undercooled eutectic alloys was found to be primarily due to a sink/float mechanism and the nucleation behavior of the alloy. Lead-rich dendrites are the primary phase in the undercooled eutectic system. These dendrites grow rapidly into the undercooled bath and soon break apart due to recalescence and Sn enrichment of the liquid. These fragmented Pb dendrites are then free to settle to the bottom portion of the ingot causing the macrosegregation observed in this study. A eutectic Pb-Sn alloy undercooled 20 K and cooled at 4 K/sec had a composition of about Pb-72 wt percent Sn at the top and 55 percent Sn at the bottom.