Every metal object we see around us has specific material properties such as yield strength, fracture toughness, and electrical resistivity.  Many of these properties are dependent on the crystalline structure of the metal on a microscopic level.  For example two pieces of identical alloy composition may have vastly different yield strengths because one of them may have very large crystal grains, while the other may have very fine crystal grains.  One of them may have been heat-treated or work hardened leading to substantially better strength or hardness.  Thus the manufacturing processes with which a metal is created is very important to the life and usefulness of that metal.

Among other things, the Thermal Analysis of Materials Processing Laboratory (TAMPL) at Tufts University studies a particular facet of the creation of metals, solidification.  Using both an electromagnetic levitator (EML) and an electrostatic levitator (ESL), we are studying the process of rapid solidification. 

Rapid solidification occurs when a liquid is brought to a temperature below its freezing point while still a liquid.  In this state, the metal is called “undercooled” or “supercooled”.  Because this state is very unstable thermodynamically (the metal really wants to be a solid, not a liquid), as soon as one tiny piece of solid is formed inside the liquid, the entire piece of metal solidifies very quickly.  This tiny piece of solid is called a nucleus, and its formation is called nucleation.  A nucleus can form by liquid atoms sticking to either some surface on its container, or some impurity in the liquid; this is called heterogeneous nucleation – a nucleation that occurs on something that is not part of the alloy itself.  A nucleus can also form by liquid atoms sticking to each other, or homogeneous nucleation.  Because it is much easier for heterogeneous nucleation to occur than homogeneous, we must minimize the possibility of heterogeneous nucleation in order to reach the undercooled state in our tests.  The biggest place for atoms to stick in a molten metal is the walls of the container in which it is sitting, so this should be the first thing we remove.  In order to remove the container from the process, we melt and solidify our samples in either the EML or the ESL, which allow us to simply suspend the sample in an either a magnetic or electric field, not touching anything. 





We also use only high purity metals to create our alloys to minimize the amount of impurities inside the sample.  In the lab, we frequently have undercoolings as great as 150° C.  At such undercoolings, the solid crystals grow through the undercooled liquid at speeds of 3-10 meters per second.  Because our samples are small (EML samples are ~7mm, ESL samples are ~2mm), the entire sample solidifies in less than a millisecond.  The only way to view these events is using a very high-speed camera, 40,000 to 100,000 frames per second. 

There are certain kinds of alloys which exhibit a phenomenon known as “double recalescence” when they are undercooled below a certain Temperature.  Because solidification occurs so rapidly in these scenarios, the energy of the liquid that is lost to the creation of atomic bonds (latent heat of fusion) is suddenly released into the sample and the surrounding environment.  The result is a sudden rise in temperature that is visible to the eye as a flash of light.  This is called a recalescence.  A double recalescence is caused by a second phase transformation, from one type of solid to another type of solid.  This means that these alloys actually solidify twice!  First they transform from the liquid phase to a metastable solid phase, then after a short incubation time, they transform to a different solid phase, this time a stable solid phase.  The result of this whole process is that the finished material has a uniform, very small crystal grain structure, which means that it has favorable material properties.  The temperature plot below shows a typical double recalescence event.  The sample heats up (1), the sample melts (2), the sample undercools (3), the first recalescence (4), the second recalescence (5), and the sample cools as a solid (6).


Unfortunately, there are many factors involved in this event.  Because the growth velocities of the two phase transformations are different functions of undercooling, there are regimes where the first is faster than the second and vice-versa.  Therefore, in order to have the entire material experience the double recalescence and to have the favorable microstructure, a third factor comes into play, the incubation time of the metastable phase. 

We know that the incubation time of the metastable phase, commonly referred to as the delay time, is a strong function of at least the alloy composition and undercooling.  Past research has shown that it is also a function of something else, and that is the focus of our current research. 

When comparing delay time data taken in the EML with that taken in the ESL, it is clear that there is roughly an order of magnitude difference between them. 


We have considered 3 possible factors so far, based on the differences in test environments: first, the effect of alloy change due to preferential evaporation of Chromium from the melt, second, the effect of difference of sample size between EML and ESL samples, and third, the presence of magneto-hydrodynamic (MHD) convection in EML and its absence in ESL.  While we have shown that the first two cannot be the driving factors in the difference in delay times, it is possible and appears likely that the MHD convection may be an important factor. 

This research is sponsored by NASA grants NAG8-1685, NNM04AA31G, and the NASA GSRP grant NAGT5-50445.  The ESL facility is located at the NASA Marshall Spaceflight Center in Huntsville, Alabama; the EML facility is located at Tufts University. 

|