Dendrites’ Clues for Castability

How dendrites form in a casting during solidification informs plant metallurgists of its final properties.

Geoffrey Sigworth, GKS  Engineering Services, Dunedin, Florida

(Click here to see the story as it appears in May's Modern Casting.)

This article is the second in a series on solidification in aluminum castings. While the series focuses on solidification principles in aluminum alloys, many can be applied to other metals, as well.

A book on snowflakes at the Carnegie Library in Pittsburgh depicts many photographs of individual snowflakes. In the introduction, the author claims each snowflake is unique, and no two crystals are alike. This claim may be true, in spite of the incredibly large number of snowflakes that form each winter. The variety of snowflakes shown in the book is mind boggling.

Something similar happens every time metal solidifies in the mold. The liquid-to-solid transformation involves the formation of many small, individual crystals of solid aluminum. This is a fascinating area, one which has received a great deal of study. A brief overview will be given in this article, touching on the aspects of solidification most important to the casting industry.

The first article in this series described the use of phase diagrams to see the sequence of phases forming during solidification, which influences final casting properties and can provide insight into castability issues. Even more about alloy properties can be gleaned from delving deeper into how the structure of a metal changes as it freezes. This knowledge is an important tool in choosing alloying combinations for your desired result.    

The solid aluminum crystals forming during solidification are like snowflakes. The metallurgists first observing these crystals thought they resembled trees and called them dendrites, after the Greek word for tree (δένδρον or déndron). Dendrites were first observed by polishing metal samples or by etching the polished surface. More recently, real-time X-ray studies have observed the in situ formation of dendrites in Al-Cu alloys. Because the aluminum crystal contains much less copper than the surrounding liquid, they appear lighter in X-ray images. Examples are shown in Fig. 1.

The formation of dendritic crystals is a curious phenomenon, and many scientists have studied them. The technical literature in this area is extensive; however, a relatively simple explanation will suffice to understand what is happening.

One important clue is that pure metals do not form dendrites. But when silicon or other elements are alloyed to aluminum, dendrites appear. From the Al-Si phase diagram, only 13% of the silicon in the liquid metal remains in the first solid. This means that the silicon atoms pile up in front of the growing solid crystals. The situation is shown schematically in Fig. 2. In keeping with the snowflake analogy, the growing aluminum grain is represented by a snow plow.

Consider the act of shoveling snow. When a shovel is pushed, the snow quickly piles up in front, so one can go no farther. Years ago, sidewalks in some cities were cleared of snow by a horse-drawn plow. The plow would use a “V”-shaped blade the width of the sidewalk which easily cut through the snow, pushing it to the sides of the walkway. This is shown schematically in Fig. 3. Dendrites act much like this “V”-shaped plow. In other words, growing aluminum crystals adapt a dendritic shape as a response to the alloy composition.

Growing solid crystals adapt a planar or a non-planar (dendritic) shape depending on the interaction of two factors.

  • The growth rate of the crystal. This is usually defined as the velocity of motion of the solid/liquid interface, in microns per second (R), and is controlled by the thermal gradient in front of the crystal (G).
  • The rate at which the “piled up” solute elements can be removed, by diffusion, from the solidifying front.

The shape of the solidifying aluminum depends on the amount and type of solute dissolved in the alloy. The grain size also is influenced by the presence of growth-restricting solutes, like Si and Cu. This may be seen by comparing the grains of different Al-Cu alloys in Fig. 4. These alloys were solidified at an average cooling rate of 1.8F (1C) per second. All four figures are shown at the same magnification. Compare this to the crystals in Fig. 1, which are new and just forming. The arms on the branches of the dendrites are fine, much like needle-shaped leaves on a Christmas tree. Also, the dendrites are growing freely into liquid metal. They are still largely unimpeded by neighboring grains.

At some point, however, the “trunks” of the dendrites come in contact with neighboring grains. (This type of contact is called “dendrite coherency”.) After this time, any further solidification (and growth of dendrites) can occur only by thickening of the leaves and branches on the dendrite. As a result, the dendrites in the final casting are thicker. The spacing between arms also becomes larger.

It has long been known that the spacing of arms of the dendrite in the casting depends on the solidification time. One of the first detailed studies was published in 1963 by Alcoa researchers who related dendrite cell size to the solidification time.

Many of the early papers reported cell size in their studies. However, it is now known that a better measure is the secondary dendrite arm spacing (SDAS). The easiest way to measure SDAS is to use the linear intercept method. This is illustrated in Fig. 5 for a modified Al-7% Si alloy. Lines are drawn on a micrograph where well defined dendrite arms can be observed, and the average spacing between the centers of adjoining arms is measured. Typically, a number of measurements are made and the results averaged.

The SDAS can be used to determine the local solidification time at any point in a casting. The results of many commercial and laboratory measurements on Al-Cu alloys have been reviewed. Results from castings made from 356 and 319 alloys are also shown in Fig. 6, where measurements of SDAS are plotted versus the local solidification time (as measured by thermocouples in the casting).

It can be seen that, for a given freezing rate, the copper-containing 319 alloy has a somewhat smaller SDAS than the 356 alloy. The correlation for most other foundry alloys would probably lie somewhere between these two curves. The ability to measure SDAS, and the correlations shown in Fig. 6, represents a useful tool. It can help in learning about the thermal history of a sample from an “unknown” casting (e.g., a competitor’s product) or from one’s own castings. It may not always be convenient to place thermocouples in the mold, but the solidification time at various points in the casting can be estimated from the SDAS.

The dendritic structure is often visible if you look carefully into pores on the fracture surface of tensile bars. An example is shown in Fig. 7. The rounded ends of the secondary dendrite arms are sticking out from the left hand-side of this picture. The SDAS in the sample appears to be between 40 and 50 microns, which corresponds to a local solidification time of about two minutes (for an A356 alloy). 

This article was based on Paper 13-1224 which was presented at the 117th Metalcasting Congress.

Next month this series continues by examining how elements dissolved in an alloy segregate during solidification and determine a casting’s mechanical properties.