Search

While changes in the coordination number for liquid silicon upon supercooling, signaling an underlying liquid-liquid phase transition, have been predicted, x-ray and neutron measurements have produced conflicting reports. In particular some studies have found an increase in the first shell coordination as temperature decreases in the supercooled regime, while others have reported increases in the coordination number with decreasing temperature. Employing the technique of electrostatic levitation coupled with high energy x-ray diffraction (125 keV), and rapid data acquisition (100ms collection times) using an area detector, we have obtained high quality structural data more deeply into the supercooled regime than has been possible before. No change in coordination number is observed in this temperature region, calling into question previous experimental claims of structural evidence for the existence of a liquid-liquid phase transition.

Conventional techniques for measuring creep are limited to about 1700 C, so a new technique is required for higher temperatures. This technique is based on electrostatic levitation (ESL) of a spherical sample, which is rotated quickly enough to cause creep deformation by centrifugal acceleration. Creep of samples has been demonstrated at up to 2300 C in the ESL facility at NASA MSFC, while ESL itself has been applied at over 3000 C, and has no theoretical maximum temperature. The preliminary results and future directions of this NASA-funded research collaboration will be presented.

Determinations of the phase formation sequence, the crystal structures and the thermodynamic properties of materials at high temperatures are difficult because of contamination from the sample container and environment. Containerless processing techniques, such as electrostatic (ESL), electromagnetic (EML), aerodynamic, and acoustic levitation, are most suitable these studies. An adaptation of ESL for in-situ structural studies of a wide range of materials, including metals, semiconductors, insulators using high energy (125 keV) synchrotron x-rays is described here. This beamline ESL (BESL) allows the in-situ determination of the atomic structures of equilibrium solid and liquid phases, including undercooled liquids, as well as real-time studies of solid-solid and liquid-solid phase transformations. The use of image plate (MAR345) or GE-Angio detectors enables fast (30 ms - 1s) acquisition of complete diffraction patterns over a wide q-range (4 - 140/mm). The wide temperature range (300 - 2500 K), containerless processing under high vacuum (10(exp -7) - 10(exp -8) torr), and fast data acquisition, make BESL particularly suitable for phase diagram studies of high temperature materials. An additional, critically important, feature of BESL is the ability to also make simultaneous measurement of a host of thermo-physical properties, including the specific heat, enthalpy of transformation, solidus and liquidus temperatures, density, viscosity, and surface tension; all on the same sample and simultaneous with the structural measurements.

If crystallization can be avoided, liquids enter a metastable (undercooled) state below their equilibrium liquidus temperatures, T(sub l), finally 'freezing' into a glass below a characteristic temperature called the glass transition temperature, T(sub g). In rare cases, the undercooled liquid may undergo a liquid-liquid phase transition (liquid polymorphism) before entering the glassy state. This has been suggested from experimental studies of H2O and Si. Such phase transitions have been predicted in some stable liquids, ie. above T(sub l) at atmospheric pressure, for SiO2 and BeF2, but these have not been verified experimentally. They have been observed in liquids of P, Si and C, but only under high pressure. In this letter we present the first experimental evidence for a phase transition in a low viscosity metallic liquid that is driven by an approach to a constant entropy configuration state and correlated with a growing icosahedral order in the liquid. A maximum in the specific heat at constant pressure, similar to what is normally observed near T(sub g), is reported for undercooled liquids of quasicrystal-forming Ti-Zr-Ni alloys. A two-state excitation model that includes cooperativity by incorporating a temperature-dependent excitation energy, fits the specific heat data well, signaling a phase transition. An inflection in the liquid density with decreasing temperature instead of a discontinuity indicates that this is not a typical first order phase transition; it could be a weakly first order or higher order transition. While showing many similarities to a glass transition, this liquid-liquid phase transition occurs in a mobile liquid, making it novel.

New short-range order data are presented for equilibrium and undercooled liquids of Ti and Ni. These were obtained from in-situ synchrotron x-ray diffraction measurements of electrostatically-levitated droplets. While the short-range order of liquid Ni is icosahedral, consistent with Frank's hypothesis, significantly distorted icosahedral order is observed in liquid Ti. This is the first experimental observation of distorted icosahedral short-range order in any liquid. although this has been predicted by theoretical studies on atomic clusters.

From extensive ground based work on the phase diagram and undercooling studies of Ti-Zr-Ni alloys, have clearly identified the composition of three different phases with progressively increasing polytetrahedral order such as, (Ti/Zr), the C14 Laves phase, and the i-phase, that nucleate directly from the undercooled liquid. The reduced undercooling decreases progressively with increasing polytetrahedral order in the solid, supporting Frank s hypothesis. A new facility for direct measurements of the structures and phase transitions in undercooled liquids (BESL) was developed and has provided direct proof of the primary nucleation of a metastable icosahedral phase in some Ti-Zr-Ni alloys. The first measurements of specific heat and viscosity in the undercooled liquid of this alloy system have been completed. Other than the importance of thermo-physical properties for modeling nucleation and growth processes in these materials, these studies have also revealed some interesting new results (such as a maximum of C(sup q, sub p) in the undercooled state). These ground-based results have clearly established the necessary background and the need for conducting benchmark nucleation experiments at the ISS on this alloy system.

The surface tension and viscosity of quasicrystal-forming Ti-Zr-Ni alloys were measured over a range of temperature, including both stable and undercooled liquids by an Electrostatic Levitation (ESL) technique. ESL is a containerless technique which allows processing of samples without contact, greatly reducing contamination and increasing access to the metastable undercooled liquid. The measured viscosity is typical of glass-forming alloys of similar composition to the quasicrystal-forming alloys studied here, while the surface tension shows an anomaly at deep undercoolings.

If crystallization can be avoided, liquids enter a metastable (undercooled) state below their equilibrium liquidus temperatures, TI, finally freezing into a glass below a characteristic temperature called the glass transition temperature, T,. In rare cases, the undercooled liquid may undergo a liquid-liquid phase transition (liquid polymorphism) before entering the glassy state. This has been suggested from experimental studies of HzO and Si4. Such phase transitions have been predicted in some stable liquids, i.e. above TI at atmospheric pressure, for Si02 and BeF;, but these have not been verified experimentally. They have been observed in liquids of P7, Sis and C9, but only under high pressure. All of these transitions are driven by an anomalous density change, i.e. change in local structure, with temperature or pressure. In this letter we present the first experimental evidence for a phase transition in a low viscosity liquid that is not driven by an anomalous density change, but by an approach to a constant configuration state. A maximum in the specific heat at constant pressure, similar to what is normally observed near T, is reported here for undercooled low viscosity liquids of quasicrystal- forming Ti-Zr-Ni alloys. that includes cooperativity, by incorporating a temperature dependent excitation energy fits the data well, signaling a phase transition.

Over fifty years ago, David Turnbull showed that the temperature of many metallic liquids could be decreased far below their equilibrium melting temperature before crystallization occurred. To explain those surprising results, Charles Frank hypothesized that the local structures of undercooled metallic liquids are different from those of crystal phases, containing a significant degree of icosahedral order that is incompatible with extended periodicity. Such structural differences must create a barrier to the formation crystal phases, explaining the observed undercooling behavior. If true, the nucleation from the liquid of phases with extended icosahedral order should be easier. Icosahedral order is often favored in small clusters, as observed recently in liquid-like clusters of pure Pb on the (111) surface of Si[3], for example. However, it has never been shown that an increasing preference for icosahedral phase formation can be directly linked with the development of icosahedral order in the undercooled liquid. Owing to the combination of very recent advances in levitation techniques and the availability of synchrotron x-ray and high flux neutron facilities, this is shown here.

New short-range order data are presented for equilibrium and undercooled liquids of Ti and Ni. These were obtained from in-situ synchrotron x-ray diffraction measurements of electrostatically-levitated droplets. While the short-range order of liquid Ni is icosahedral, consistent with Frank's hypothesis, significantly distorted icosahedral order is observed in liquid Ti. This is the first experimental observation of distorted icosahedral short-range order in any liquid, although this has been predicted by theoretical studies on atomic clusters.

Electrostatic Levitation (ESL) is an advanced technique for containerless processing of metals, ceramics, and semiconductors. Because no container is required, there is no contamination from reaction with a crucible, allowing processing of high temperature, highly reactive melts. The high vacuum processing environment further reduces possible contamination of the samples. Finally, there is no container to provide heterogeneous nucleation sites, so the undercooled range is also accessible for many materials. For these reasons, ESL provides a unique environment for measuring thermophysical properties of liquid materials. The properties that can be measured in ESL include density, surface tension, viscosity, electrical and thermal conductivity, specific heat, phase diagram, TTT- and CCT- curves, and other thermodynamic properties. In this paper, we present data on surface tension and viscosity, measured by the oscillating drop technique, and density, measured by an automated photographic technique, measured in the ESL at NASA Marshall Space Flight Center.

Over fifty years ago, David Turnbull showed that the temperature of many metallic liquids could be decreased far below their equilibrium melting temperature before crystallization occurred. To explain those surprising results, Charles Frank hypothesized that the local structures of undercooled metallic liquids are different from those of crystal phases, containing a significant degree of icosahedral order that is incompatible with extended periodicity. Such structural differences must create a barrier to the formation crystal phases, explaining the observed undercooling behavior. If true, the nucleation from the liquid of phases with extended icosahedral order should be easier. Icosahedral order is often favored in small clusters, as observed recently in liquid-like clusters of pure Pb on the (111) surface of Si(3), for example. However, it has never been shown that an increasing preference for icosahedral phase formation can be directly linked with the development of icosahedral order in the undercooled liquid. Owing to the combination of very recent advances in levitation techniques and the availability of synchrotron X-ray and high flux neutron facilities.

Results from electrostatic levitation studies using zirconium specimens will be presented. Analysis of the data permits the determination of nucleation kinetic parameters.

The first undercooling nucleation measurements of electrostatic-levitated droplets of TiZrNi alloys that form the icosahedral quasicrystal phase (i-phase) are presented. The reduced undercooling for crystallization decreases with an increasing polytetrahedral order of the primary solidifying phase, supporting the existence of a developing icosahedral short-range order in the undercooled liquid. X-ray diffraction measurements made at the Advance Photon Source on levitated liquid droplets of these alloys at their liquidus temperatures, however, show no evidence for increased icosahedral order. This suggests that significant ordering only occurs below the melting temperature.

The local atomic structures of undercooled liquid metals are presumed to be icosahedral; this order is incompatible with translational periodicity, constituting a barrier to the nucleation of the crystal phase. The extended atomic structure of the icosahedral quasicrystal (i-phase) is similar to that presumed in the undercooled liquid. Therefore, a comparison of the maximum undercooling in alloys that form the i-phase with those that form crystal phases provides a probe of the liquid structure.

The poster presentation provides an overview of the MSFC ESL facility. The technical capabilities and accomplishments of the facility are summarized. Future plans including the development of a pressurized processing chamber and additional high-energy X-ray beamline studies are outlined.

Zr58.5Nb2.8Cul5.6Nil2.8All0.3 is the first bulk glass forming liquid that does not contain beryllium to be vitrified by purely radiative cooling in the containerless electrostatic levitation process. The measured critical cooling rate is 1.75 K/s. The sluggish crystallization kinetics enable the determination of the time-temperature-transformation (TTT) diagram between the liquidus and the glass transition temperatures. At the nose of the TTT diagram, the shortest time to reach crystallization in an isothermal experiment is 32 seconds. In contrast to other bulk metallic glasses the scatter in the crystallization onset times are small at both high and low temperatures.

Utilizing the containerless electrostatic levitation facility at NASA/MSFC, we were able to undercool the Ni-59Nb-41 (atomic) alloy by 210 K which was 160 deg. farther than the results of previous flight experiments. Undercoolings were clustered around 200 K during the repeated melting-freezing cycles on a single sample. Prior to this work, a metastable liquid separation had been presumed to limit the undercooling of this alloy. However, microstructural observations have revealed that undercooling was limited by crystal nucleation.

Utilizing the containerless electrostatic levitation facility at NASA/MSFC, we were able to undercool the Ni59Nb41 (atomic) alloy by 210 K which was 160 degrees farther than the results of previous flight experiments. Undercoolings were clustered around 200 K during the repeated melting-freezing cycles on a single sample. Prior to this work, a metastable liquid separation had been presumed to limit the undercooling of this alloy. However, microstructural observations have revealed that undercooling was limited by crystal nucleation.

Previous tests of the classical nucleation theory as applied to liquid-liquid gap miscibility systems found a discrepancy between experiment and theory in the ability to undercool one of the liquids before the L1-L2 separation occurs. To model the initial separation process in a two-phase liquid mixture, different theoretical approaches, such as free-energy gradient and density gradient theories, have been put forth. If there is a large enough interaction between the critical liquid and the crucible, both models predict a wetting temperature (T(sub w)) above which the minority liquid perfectly wets and layers the crucible interface, but only on one side of the immiscibility dome. Materials with compositions on the other side of the dome will have simple surface adsorption by the minority liquid before bulk separation occurs when the coexistence (i.e., binoidal) line in reached. If the interaction between the critical liquid and the crucible were to decrease, T(sub w) would increase, eventually approaching the critical consolute temperature (T(sub cc)). If this situation occurs, then there could be large regions of the miscibility gap where non-perfect wetting conditions prevail resulting in droplets of L1 liquid at the surface having a non-zero contact angle. The resulting bulk structure will then depend on what happens on the surface and the subsequent processing conditions. In the past several decades, many experiments in space have been performed on liquid metal binary immiscible systems for the purpose of determining the effects that different crucibles may have on the wetting and separation process of the liquids. Potard performed experiments that showed different crucible materials could cause the majority phase to preferentially wet the container and thus produce a dispersed microstructure of the minority phase. Several other studies have been performed on immiscibles in a semi-container environment using an emulsion technique. Only one previous study was performed using completely containerless processing of immiscible metals and the results of that investigation are similar to some of the emulsion studies. In all the studies, surface wetting was attributed as the cause for the similar microstructures or the asymmetry in the ability to undercool the liquid below the binoidal on one side of the immiscibility dome. By removing the container completely from the separation process, it was proposed that the loss of the crucible/liquid interaction would produce a large shift in T(sub w) and thus change the wetting characteristics at the surface. By investigating various compositions across the miscibility gap, a change in the type and amount of liquid wetting at the surface of a containerless droplet should change the surface nucleating behavior of the droplet - whether it be the liquid-liquid wetting or the liquid-to-solid transition. Undercooling of the liquid into the metastable region should produce significant differences in the separation process and the microstructure upon solidification. In this study, we attempt to measure these transitions by monitoring the temperature of the sample by optical pyrometry. Microstructural analysis will be made to correlate with the degree of undercooling and the separation mechanisms involved.

Bulk Cu-Co alloys at compositions ranging from 10 to 80 wt pct Co were highly undercooled using a melt fluxing technique. The metastable liquid separation boundary has been directly determined from the measured temperature-time profiles. It was found that the critical point of the miscibility gap is slightly shifted towards the Co side and about 90 K below the liquidus. A droplet-shaped microstructure was observed for all solidified specimens (Cu- 10 to 80 wt pct Co), when the melts were undercooled into the metastable miscibility boundary.

The copper-niobium phase diagram has been under active debate; thus, a corroboratory experimental study is needed. In this investigation, the melts of Cu-Nb alloys at compositions ranging from 5 lo 86 wt% Nb were processed in different environments and solidified at relatively low rates of 50-75 C/s to determine liquidus temperatures and to study solidification behavior. For all samples processed under very clean conditions, only Nb dendrites in a Cu matrix were observed; while in the presents of oxygen impurities, the alloys containing 5-35 wt% Nb exhibited microstructure of Nb-rich spheroids and Nb dendrites in the Cu matrix. The results obtained from clean conditions are in fair agreement with the Cu-Nb phase diagram having an S-shaped, near-horizontal appearances of the liquidus. The formation of Nb-rich droplets at slow cooling rates is discussed in terms of a stable liquid miscibility gap induced by oxygen.

Container and containerless processing was employed to determine liquidus temperatures and to examine microstructural development in the Cu-Nb system. The Cu-Nb phase diagram of an S-shaped, near-horizontal liquidus, has been confirmed by both the temperature-time profiles and resultant microstructures with only Nb dendrites in a Cu matrix, which were obtained from crucible experiments under clean conditions. However, the microstructural pathways of Cu-Nb alloys are particularly sensitive to processing variables. By the addition of oxygen impurities or rapid solidification, droplet-shaped morphology was observed for some compositions, implying occurrence of a liquid-phase separation. The effects of impurities and cooling rates are analyzed in connection with a stable and metastable liquid miscibility gap, respectively.

A systematic study on the liquid separation in supercooled Cu-Co and Cu-Fe alloys was performed using a melt fluxing which permits high supercooling to be achieved. Moreover, this method renders it possible to directly measure binodal temperatures and establish metastable liquid miscibility gap (LMG). All phase-separated samples at compositions ranging from 10 to 80 wt pct Co or to 83 wt pct Fe were found to exhibit droplet-shaped morphologies, in spite of various droplet distributions. Uniformly dispersed microstructures were obtained as the minority component was less than 20 vol.%; while beyond this percentage, serious coarsening was brought about. Calculations of the miscibility gap in the Cu-Co system and Stokes movement velocity of Co and Fe droplets in Cu matrix were made to analyze the experimental results.

The copper-niobium phase diagram has been under active debate; thus, a corroboratory experimental study is needed. In this investigation, the melts of Cu-Nb alloys at compositions ranging from 5 to 86 wt pct Nb were processed in different environments and solidified at relatively low cooling rates of 50 to 75 C/s to determine liquidus temperatures and to study solidification behavior. For all samples processed under very clean conditions, only Nb dendrites in a Cu matrix were observed; while in the presence of oxygen impurities the alloys containing 5 to 35 wt pct Nb exhibited microstructure of Nb-rich spheroids and Nb dendrites in the Cu matrix. The results obtained from clean conditions are in fair agreement with the Cu-Nb phase diagram having an S-shaped, near-horizontal appearance of the liquidus. The formation of Nb- rich droplets at slow cooling rates is discussed in terms of a stable liquid miscibility gap induced by oxygen.

Microprocessor-controlled susceptometer measures cryogenic temperature dependence of ac magnetic susceptibility of small cylindrical samples.

An enclosure provides a containerless environment in which a sample specimen is positioned. The specimen is heated in the containerless environment, and the specimen melt is dropped through the tube in which it cools by radiation. The tube is alternatively backfilled with an inert gas whereby the specimen melt cools by both radiation and convection during its free fall. During the free fall, the sample is in a containerless, low-gravity environment which enhances supercooling in the sample and prevents sedimentation and thermal convection influences. The sample continues to supercool until nucleation occurs which is detected by silicon photovoltaic detectors. The sample solidifies after nucleation and becomes completely solid before entering the detachable catcher. The amount of supercooling of the specimen can be measured by knowing the cooling ratio and determining the time for nucleation to occur.

Molten Al-In leaked unexpectedly out of the crucible of a proposed MEA materials processing in space experiment. The molten metals use a spring loaded plunger to eliminate most free surfaces. The critical criteria necessary to initiate flow and the rate of fluid flow into the crucible/plunger annulus is calculated. Experimental in situ X-radiographs are interpreted according to the calculations. A note on possible effects of capillary flow if wetting occurs between crucible/plunger and liquids is included.

A 32 meter evacuated drop tube has been used to investigate the solidification of Nb-Ge alloys after deep undercooling. Samples have been supercooled as much as 500 K below the liquidus by using free-fall conditions to eliminate crucible induced nucleation. Final microstructures are dependent on the quenching rates at the bottom of the drop tube with a striking extension of the beta phase solubility limit at the higher quenching rates.

Experiments using containerless undercooling and low-gravity solidification of Nb(1-x)Ge(x) alloys for x=0.13-0.27 have been carried out in a 32-m drop-tube apparatus to study the feasibility of forming metastable Nb3Ge in bulk form. It is found that bulk samples (2-3 mm diam) of Nb-Ge alloys with 18-22% Ge can be undercooled by large amounts (300-500 K) and solidified in a containerless environment. Subsequent quenching of the solidified samples in oil helps preserve the metastable A-15 phase by removing the latent heat of fusion and quickly cooling the samples to a stable temperature below 1000 C. Even at an undercooling of less than 100 K, the superconducting transition temperature of the material is enhanced over the cast material by about 1 K.

A susceptibility apparatus to measure superconducting properties of samples made in the MSFC Drop Tube was used to measure the transition temperature (Tc) and susceptibilities of Nb and Nb Ge Alloys prepared in bulk spherical (2-4 mm diameter) form using a 32 m drop tube in which containerless low gravity solidification could take place. Results indicate that a drop tube processing environment was beneficial for increasing the Tc of the superconducting phase of the material over that of arc melted material. The increase in Tc is found to be related to the amount of solidification of the total sample that took place before reaching the bottom of the drop tube. In phase and quadrature phase measurements of the specimen's susceptibility indicated that some improvement in homogeneity takes place in drop tube processing. These phase measurements also indicated little or no shielding of a lower Tc phase by a higher Tc filamentary structure.

A containerless low-gravity environment, produced within a 32 m drop tube apparatus, has been used to undercool and solidify metals, alloys or glasses by eliminating crucible induced nucleation processes. Niobium droplets with diameters in the range of 2 to 5 mm have been undercooled by 525 K which corresponds to the maximum undercooling reported by Turnbull and others on fine dispersions of low melting point metals. Solidification at large undercooling resulted in single crystalline spheres with the formation of interdendritic shrinkage channels on the sample surface rather than interior shrinkage cavities. The grain refinement as observed for Ni samples undercooled and solidified in fused silica crucibles does not occur in free-falling drops of Nb. A calculated solidification speed of undercooled Nb is compared to Ni. A solidification speed of 320 m/s is found for the Nb drops. This solidification speed is greater than or comparable to the solidification speeds calculated in splat cooled samples. Thus, a drop tube apparatus can be useful in the preparation and study of high temperature metastable compounds or alloys in bulk form.

Drop tube makes possible preparation of exotic materials. The 100 foot tube is oriented precisely vertical to prevent free-falling drop from hitting tube walls. Inert-gas supply, evacuation pumps, viewing ports, and flexibility in choice of melt technique allow precise control and monitoring of solidification.

Equipment indicates that system can rapidly and accurately process metallurgical and materials-processing data for wide range of applications. Advantages include increase in contract between areas on image, ability to analyze images via operator-written programs, and space available for storing images.

The application of a specific digital computer system (known as the Image Data Processing System) to the analysis of three NASA-sponsored metallurgical experiments is discussed in some detail. The basic hardware and software components of the Image Data Processing System are presented. Many figures are presented in the discussion of each experimental analysis in an attempt to show the accuracy and speed that the Image Data Processing System affords in analyzing photographic images dealing with metallurgy, and in particular with material processing.