Douglas M. Matson

Convection in Containerless Processing

Abstract: Different containerless processing techniques have different strengths and weaknesses. Applying more than one technique allows various parts of a problem to be solved separately. For two research projects, one on phase selection in steels and the other on nucleation and growth of quasicrystals, a combination of experiments using electrostatic levitation (ESL) and electromagnetic levitation (EML) is appropriate. In both experiments, convection is an important variable. The convective conditions achievable with each method are compared for two very different materials: a low‐viscosity, high‐temperature stainless steel, and a high‐viscosity, low‐temperature quasicrystal‐forming alloy. It is clear that the techniques are complementary when convection is a parameter to be explored in the experiments. For a number of reasons, including the sample size, temperature, and reactivity, direct measurement of the convective velocity is not feasible. Therefore, we must rely on computation techniques to estimate convection in these experiments. These models are an essential part of almost any microgravity investigation. The methods employed and results obtained for the projects levitation observation of dendrite evolution in steel ternary alloy rapid solidification (LODESTARS) and quasicrystalline undercooled alloys for space investigation (QUASI) are explained.

Magnetohydrodynamic Modeling and Experimental Validation of Convection Inside Electromagnetically Levitated Co-Cu Droplets

A magnetohydrodynamic model of internal convection of a molten Co-Cu droplet processed by the ground-based electromagnetic levitation (EML) was developed. For the calculation of the electromagnetic field generated by the copper coils, the simplified Maxwell’s equations were solved. The calculated Lorentz force per volume was used as a momentum source in the Navier–Stokes equations, which were solved by using a commercial computational fluid dynamics package. The RNG k-ε model was adopted for the prediction of turbulent flow. For the validation of the developed model, a Co16Cu84 sample was tested using the EML facility in the German Aerospace Center, Cologne, Germany. The sample was subjected to a full melt cycle, during which the surface of the sample was captured by a high-speed camera. With a sufficient undercooling, the liquid phase separation occurred and the Co-rich liquid phase particles could be observed as they were floating on the surface along streamlines. The convection velocity was estimated by the combination of the displacement of the Co-rich particles and the temporal resolution of the high-speed camera. Both the numerical and experimental results showed an excellent agreement in the convection velocity on the surface.

Contrasting electrostatic and electromagnetic levitation experimental results for transformation kinetics of steel alloys.

Abstract: The delay between conversion of metastable ferrite to stable austenite during ternary Fe‐Cr‐Ni alloy double recalescence is seen to differ by over an order of magnitude for tests conducted using electrostatic and electromagnetic levitation. Several possible reasons for this deviation are proposed. Thermodynamic calculations on evaporation rates indicate that potential composition shifts during testing are minimized by limiting test time and thermal history. Simulation indicates that deviation would be limited to a factor of 1.5 under worst‐case conditions. Possible effects due to differences in sample size are also eliminated since the metastable array, where stable phase nucleation must occur, is significantly smaller than the sample. Differences in internal convection are seen to be the most probable reason for the observed deviation.

Internal convective effects on the lifetime of the metastable phase undercooled Fe-Cr-Ni alloys

Differences have been observed between the lifetimes of the metastable phases of undercooled samples of Fe–12 wt% Cr–16 wt% Ni alloy which had been subjected to electromagnetic levitation (EML) and electrostatic levitation (ESL). Internal flow is induced within the samples by positioning forces in EML and much weaker Marangoni forces in ESL. The hypothesis being tested is that the flow within EML samples is strong enough to cause the growing metastable dendrites to deflect so that the secondary arms of adjacent dendrites collide, resulting in early nucleation of the stable phase. Simulations using a commercial computational fluid dynamics software package were performed to determine the time required for collision of the secondary arms to occur. There is quantitative agreement between the numerical time to collision and the experimental lifetime of the metastable phase. It has been determined that the induced convective flow in EML samples is strong enough to cause collision and is the most likely cause of the difference between the lifetimes of the metastable phases in ESL and EML samples.

Microgravity experiments on the effect of internal flow on solidification of Fe-Cr-Ni stainless steels.

Abstract:  A new hypothesis has been developed to explain the effect of internal fluid flow on the lifetime of a metastable phase in solidifying Fe‐Cr‐Ni alloys. The hypothesis shows excellent agreement with available experimental results, but microgravity experiments are required for complete validation. Certain Fe‐Cr‐Ni stainless steel alloys solidify from an undercooled melt by a two‐step process in which the metastable ferrite phase forms first followed by the stable austenite phase. Recent experiments using containerless processing techniques have shown that the lifetime of the metastable phase is strongly influenced by flow within the molten sample. Simulations using a commercial computational fluid dynamics (CFD) package, FIDAP, were performed to determine the time required for collision of dendrites and compared to experimental delay time. If the convective velocities are strong enough to bend the primary arms, then the secondary arms of adjacent dendrites can touch. The points of collision form low‐angle boundaries and result in high‐energy sites that can serve as nuclei for the transformation to the stable phase. It has been determined that the convective velocities in electrostatic levitation (ESL) are not strong enough to cause collision. However, in ground‐based electromagnetic levitation (EML), the convective velocities are strong enough to cause the dendrites to deflect so that the secondary arms of adjacent dendrites collide. There is quantitative agreement between the numerically determined time to collision and the experimentally observed delay time in EML. The strong internal velocity due to convection within the EML samples is the reason for the observed difference in delay times between ESL and EML. Microgravity testing is essential because the significant change in nucleation behavior occurs between the ranges accessible by ground‐based ESL and EML. Testing in microgravity using EML will permit a large range of internal convective velocities including those that are inaccessible in 1 g.

Numerical Prediction of the Accessible Convection Range for an Electromagnetically Levitated Fe 50 Co 50 Droplet in Space

From December 2014, a series of space experiments will be performed to investigate the influence of the convection on the multiphase solidification phenomena of metallic alloys. For the success of the mission, it is of critical importance to predict the convection in molten samples under given test parameters. In this research, the convection induced in the molten Fe50Co50 alloy was predicted numerically. The magnetohydrodynamic model for the ground-based electromagnetic levitator developed in the previous research was extended to the space application. The same modeling strategies were applied to the electromagnetic levitator in space. Using the numerical model, the convection under various test conditions was predicted: The flow pattern was characterized as a function of the heating current. The maximum convection velocity at various temperatures was estimated with the increasing heating current. Finally, the range of accessible convection velocity was predicted as a function of the critical undercooling of the sample, the minimum positioner control voltage, and the undercooling of the sample. The results are expected to provide critical information for the design of the space experiments and the interpretation of the results.

Measurement of Density of Fe-Co Alloys Using Electrostatic Levitation

Abstract The density of a series of five different iron-cobalt alloys was measured using electrostatic levitator processing. Each sample was processed through multiple thermal cycles and the liquid density was measured, while the superheated sample was cooled down to its undercooled state. The volume of the sample was estimated by analyzing captured high-speed video data of the projected shape of the sample. The mass change during the melt cycle was also tracked using Langmuir’s equation of mass evaporation. The density was then calculated as a function of temperature based on these measurements of volume and mass. Density values obtained showed higher precision than existing data from the literature obtained using a variety of different techniques, although the accuracy was consistent.

Expanded Polystyrene Lost Foam Casting—Modeling Bead Steaming Operations

The retention of pattern voids observed in the production of expandable polystyrene patterns for lost foam casting can be traced to conditions developed during mold filling and subsequent steaming. Void formation and closure, or healing, was observed using high-speed video imaging through a clear acrylic sheet cut to match one-half of a test pattern mold. Two processing conditions, i.e., the initial bead packing density and the velocity of steam as it passes between beads, were shown to significantly impact the ability of a void to heal during steaming. A model is proposed to predict conditions where voids will heal based on three criteria that relate to a limitation of the processsing window, the void size, and the ability of the bead to swell.

Phase selection in the mushy-zone: LODESTARS and ELFSTONE projects

In a collaboration sponsored by ESA and NASA, international partners have developed a work plan to successfully address key issues relating to understanding the role of convection on alloy phase selection for commercially important structural alloys using the MSL-EML facility aboard the International Space Station. The approach is two-pronged. First, ground and space-based experiments will develop a baseline database to anchor subsequent modelling predictions. Tasks include sample preparation and verification, ground-based transformation evaluation, space-based experiments, and thermophysical property evaluation to support modelling activities. Second, modelling and theoretical analysis tasks will lead to a new understanding of the role of convection in phase selection for this class of materials. These models will allow prediction and control of microstructural evolution during solidification processing. Tasks include modelling of macroconvection induced by the EM levitation field, modelling of microconvection within the dendrite array, nucleation modelling, and modelling of the transformation kinetics specific to each alloy system. This paper outlines how two NASA-sponsored projects relate to the goals of the international collaboration.

Role of sample size in the nucleation kinetics of phase transformations in steel alloys

Certain commercial steel alloys exhibit a two-step phase transformation process during solidification when substantial undercooling of the liquid allows access to the metastable phase. This two-step transformation leads to a desirable microstructure under certain conditions. Electrostatic Levitation (ESL) and Electromagnetic Levitation (EML) are two methods of containerless processing used to study how nucleation and growth kinetics influence the transformation delay between phases. Because the two facilities show substantially different delay results, the test environment differences have been analyzed to determine the root cause of this deviation. In particular, the difference in sample size between ESL and EML is examined and modeling shows that this difference is not the controlling factor in determining transformation delay behavior.