A Novel Prime Coat Slurry Mineral for Aluminum Investment Castings
Steven Ashlock
Director of Research and Technology
Ceramic Engineer
Kyanite Mining Corporation
Austin Scheer
Ceramic Engineer
Kyanite Mining Corporation
Kevin Rudolph
Metallurgical Engineer
O’Fallon Casting
ABSTRACT
The selection of a face coat refractory material to be used in investment casting is crucial when evaluating “as-cast” quality. The face coat material must be refractory (high temperature resistance), as it needs to able to withstand the “super heat” of the molten metal to avoid defects caused by mold-metal reactions. Ideally, the material would be non-wetting so that the molten metal never has a chance to penetrate the porosity of the prime coat. The prime coat slurries must have adequate flow-behavior (rheology) to ensure the casting maintains the intricate detail of the pattern. Zircon is nearly always used in prime coat slurries and prime coat stuccos. This is primarily because zircon is a stable refractory mineral and can withstand the superheats used to cast stainless alloys. However, in applications where the alloy temperature is much lower than iron or nickel-base superalloys, high temperature refractoriness may not be needed; e.g. investment casting aluminum.
In aluminum castings, the backup materials may differ from steel, but the prime coat material is often the same: zircon. Zircon is commonly used even though aluminum alloys melt at significantly lower temperatures than their steel counterparts. Molten aluminum is corrosive and readily reacts with most refractories through the reductive action of aluminum. This corrosiveness causes mold-metal reactions and leads to surface defects, increasing finishing labor. Zircon does a good job resisting this corrosion but comes at a significant cost and is over-kill for aluminum casting applications. Kyanite Mining Corporation has recently developed a new prime coat slurry mineral alternative that is both refractory enough to withstand the temperature of aluminum superheat and exhibits non-wetting characteristics with molten aluminum, leaving a smooth as-cast surface. This paper will discuss the viability of this new prime coat slurry for aluminum investment castings through laboratory testing and a foundry case study.
INTRODUCTION
Several years ago, the Department of Energy (DOE) funded a project that looked to improve the energy efficiency of handling molten metals, with a focus on aluminum alloys.1 Refractories were one of the main areas of focus. The reason for this focus is heat loss due to degradation of the refractories results in increased energy consumption. The degradation from aluminum attack was tested as far back as the 1950’s when the penetration of aluminum metal on silica refractories was tested by Brondyke.2 One way to reduce degradation due to corrosion is to reduce the penetration of metal into the refractory. For molten aluminum, the solution has often been to add non-wetting agents such as BaSO4 or CaF2.3,4
In the DOE study, samples of spent castable refractories from an aluminum alloy melting furnace were analyzed. One of the findings was a large amount of kyanite crystals in the metal penetrated zones. These kyanite crystals were unaffected by the molten aluminum while other aggregates had been degraded. Further studies5,6 have also shown that kyanite is more resistant to aluminum attack than major refractory aggregates such as mullite, silica, and bauxite. An example of this is shown in Figure 1.
Figure 1: A kyanite crystal (labeled Ky) shown as unaffected by molten aluminum attack while the mullite aggregate (blue) was degraded.
Resistance to molten metal attack is important in many refractory applications, including prime coats in investment casting shells. Inertness to the attack is key in minimizing mold metal reactions. These reactions can oxidize the surface of the casting which lead to, at best, extra finishing time and cost, and, at worse, scrapping the casting. Zircon has long been the key ingredient in the prime layer of the shell due to its chemical inertness. It has a very high melting temperature and generally does not react with any of the most cast iron and steel alloys. Zircon also finds a home in prime slurries for bronze, copper, and aluminum castings. Yet the question must be raised: Is zircon really needed in lower temperature castings such as aluminum? Concerns over availability, raw material cost, and disposal costs due to radiation are all factors to consider when choosing to use zircon in an application that simply isn’t very hot compared to steel castings.
The natural resistance to aluminum attack makes kyanite an intriguing option for a prime coat material. Kyanite has never been used as a prime slurry as far as this author knows. This is likely due to concerns about kyanite’s famous expansion. Kyanite converts to mullite at 2550°F (1400°C) when held for one hour.7 This conversion causes the crystal structure to rearrange, resulting in a 17% expansion (by volume). This would prohibit the use of kyanite as a prime coat material for alloys with high melting temperatures as the shell would crack due to the kyanite expansion. This would not be the case at lower melting temperatures of aluminum alloys as they are poured well below the conversion temperature of kyanite.
Kyanite Mining Corporation (KMC) set out to prove the idea of using kyanite as a prime slurry mineral. As shown, initial lab testing was encouraging enough to warrant a foundry test which was conducted at O’Fallon Casting.
LAB TESTING
As this was the first time kyanite would be used in a prime coat slurry application, great care was taken to choose the optimal particle size distribution (PSD) of the ground product. Several PSDs were studied to examine slurry viscosity and shear behavior. A material with a d50 of around 21 µm was chosen to be the optimal PSD for this project. This ground material was used to create a slurry with the following properties:
Table 1: Properties for the slurry used in lab testing
This slurry was then used as the prime layer on the initial cup tests. Wax coated paper cups were filled with wax for rigidity including a hanger for drying the test cup. A second cup was stapled to the outside of these wax filled cups. (Figure 2) These cups were then dipped into the kyanite slurry. A 70x100 kyanite stucco, created to mimic zircon sand, was applied as the prime stucco. The kyanite slurry was used in the 6 backup layers and seal coat with Virginia Mullite 20x50 acting as the backup stucco. The staples were then pulled, and the wax filled cup removed. The paper on the inside of the dipped cup was peeled away leaving a ceramic shell in the form of the cup. A356 was melted at 1450°F (790°C) and poured into the cup. After cooling, the metal ingot was removed from the shell. No mold-metal reactions were observed.
Figure 2: Wax coated paper cups were used as a pattern to create a cup for the cup tests.
KMC wanted to compare the surface finish of these aluminum ingots to some made with a zircon prime. More cups were created using a zircon prime coat and zircon sand prime stucco with the same backup slurries and stuccos. There were no distinguishable surface finish differences between the ingots made using kyanite or zircon prime coats.
After the first proof of concept, KMC wanted to try a more complex shape. Metal molds the shape of the state of Virginia were filled with molten wax to create a wax pattern. (Figure 3) The same dipping sequence, wax burnout, and alloy used in the cup tests were repeated for Virginia patterns with both kyanite and zircon prime slurries. No noticeable surface differences were observed between the Virginia shaped castings made with zircon or kyanite prime slurry.
Figure 3: The state of Virginia was cast with A356 using kyanite as a prime coat slurry
The last lab test was designed to look at shell removal characteristics in corners and holes. Numerical birthday candles were dipped in a similar fashion as the cups from the first test. Dewaxing these shells with the torch proved to be problematic for these more complex shapes. To remove the wax, the shells were placed through an inspection port on one of the mullite calcination kilns for a pseudo flash fire dewax. This did crack some of the shells due to thermal shock, but most survived for testing. The numbers were cast with A356 at 1450°F (790°C). Knockoff and surface finish for these castings showed no issues from using kyanite as the prime slurry material.
O’FALLON CASTING TRIALS
1st Batch
The first foundry trial was designed with the slurry having properties similar to lab tests. The mixing occurred in a high shear mixing tank, and then pumped over to the tank that would be used for dipping. Most of the ingredients were added without difficulties. However, the additions for the larger tank (120 gallons) did not scale linearly as expected compared to the lab scale tank (3 gallons). This resulted in some difficulties as more flour was added; the slurry was not wetting the latter half of the flour additions quickly. To prevent the slurry from becoming too thick while mixing, the speed of the mixing tank was periodically increased as well as having to “burp” the slurry – briefly stopping the rotation of the tank to allow trapped air bubbles to escape off the surface. By the end of the additions the speed of the mixing tank needed to be increased dramatically.
After all ingredients were added, mixing occurred for about five hours, after which the slurry was pumped to the tank used for dipping. The slurry did not pump well and there was some buildup remaining on the edges of the mixing tank, indicating that more mixing time would have been beneficial. The day after mixing, the slurry was tested to confirm quality. These tests include viscosity, plate weight, and specific gravity. The slurry was maintained with daily viscosity checks using a No. 4 Zahn cup. If water was added, a second viscosity check was performed. The slurry was maintained for about three months. After one month more evaporation than expected was occurring, causing much more water to be needed each day than would be needed for O’Fallon Casting’s (OFC) typical prime slurry. Kyanite slurry did not gel, and was able to be recovered even though it became thick.
Before parts were dipped, a shear test was also completed to ensure an even coat. Parts for testing were chosen to include different alloys as well as different features such as thin walls, flat faces, or tube passages. A356, A357, C874, and metal matrix composite (MMC) alloys were tested. Due to the size of the tank, parts were dipped by hand, using an alumino-silicate stucco. After the prime layer, the parts received three intermediate layers, three backup layers, and a dip seal. The Kyanite slurry was only used for the prime layer.
Once shell build was completed, the shells were dewaxed and fired. No cracking occurred during these steps. A356 and A357 parts were poured with a conventional ladle pour, while the MMC and C874 parts used counter gravity casting. When pouring, none of the parts ran into issues. Ease of knockout and surface finish were evaluated when cleaning up the parts. The A356 and A357 parts did not see a difference in surface finish, or in water wash according to operators. The MMC parts had slight issues with metal penetration in some tight corners. This did not result in any scrapped parts, but did increase the work time needed for water wash. Although the parts cast in C874 did not have any problems with water wash, the sprue stick and some runners showed increased signs of oxidation compared to a similar sprue with OFC prime layer. This also did not lead to any scrapped parts, with the surface confirmed to be conforming at fluorescent penetrant inspection (FPI). Aluminum parts were poured with a metal temperature of 1300°F and mold temperatures ranging from 800-1000°F. The C874 sprue saw the highest temperatures, with the mold temperature at 1450°F and the metal at 1900°F.
Figure 4: The surface finish of a sample part cast using a kyanite prime slurry was similar to castings using the standard OFC prime.
Figure 5: Metal penetration on the MMC alloy required more time during water wash to remove the shell.
2nd Batch
The primary goal with the second trial batch was to further prove the success of the prime layer while also making it easier to mix. This time the recipe had been altered to more closely match the properties used in O’Fallon Casting’s prime slurry. This resulted in a recipe with lower percent solids, lower percent silica, and lower viscosity. The second batch was smaller than the first to better match the dip tank size, making 90 gallons instead of 120. The mixing of the slurry went significantly easier on the second batch. There was a large improvement on how fast the flour wet and the mixing tank did not have to be set nearly as high. Biocide was added as well in order to more closely match the OFC prime slurry.
The slurry was pumped over to the same tank as the first trial after five hours, like the first batch. The pumping was not an issue with the same wait period due to the smaller batch size. The same tests were used to verify and maintain the slurry as well. When attempting to first dip parts, there was an issue with the slurry sticking to the wax patterns. When dipping in one side of the tank, however, the slurry would stick fine. It appeared that the slurry had somehow separated and did not properly mix. The slurry appeared to be more mixed when stirred just before dipping, so, the slurry was maintained with water additions to combat evaporation. After two weeks the shear tests came out good with an even coat indicating the problem was that the slurry was not fully mixed. Once shear test was complete, parts were dipped like first batch, with no issues. The parts finished shell build, dewax, and burnout as expected. Parts in this trial were A356 and MMC alloys. The results of the pour and cleaning were the same as the first trial: all A356 parts poured well and cleaned well with a good surface finish. Meanwhile the MMC parts showed a small amount of metal penetration in tight corners and needed extra time to remove the leftover shell. The sprues that were dipped while the slurry was separated showed a slightly rougher surface finish in some areas the shell did not stick but were otherwise the same as the well-dipped parts.
CONCLUSION
The surface finish and wash ability are comparable to the normal OFC process; the only parts that suffered a slightly longer wash time was the MMC alloy parts. Mixing the slurry proved to be significantly easier with the second batch, showing that the issues from the first batch could be attributed to batch size. So, after testing parts across multiple alloys and complexities, the Kyanite slurry can be considered a success for aluminum and copper alloys.
REFERENCES
X. Liu, et al., “Multifunctional Metallic and Refractory Materials for Energy Efficient Handling of Molten Materials”, United States Department of Energy, 2008.
K.J. Brondyke, “Effect of Molten Aluminum on Alumina-Silica Refractories”, Journal of the American Ceramic Society, 36[5] 171-174 (1953).
J. Aguilar-Santillan, “Wetting of Al2O3 by Molten Aluminum: The Influence of BaSo4 Additions”. Journal of Nanomaterials, (2008).
S. Afshar and C. Allaire, “Furnaces: Improving Low Cement Castables by Non-Wetting Additives”, Journal of Materials, 53 24-27 (2001).
M. Karakus, W. L. Headrick, and E. Feiner, “Aluminum Melting Furnace Post-Mortem”, pp. 135-156, in Monolithics: Advances, Installation and “Boom”, 41st Symposium on Refractories, The American Ceramic Society, March 30-31st, 2005, St. Louis.
W. Headrick, M Karakus, and D. Jain, “Qualification of Micronized Kyanite and Silica Fume” MS&T October, 2006, Cincinnati.
S. Ashlock, “A Property Comparison of Commercially Available Sillimanite Minerals,” EUROGRESS, 1-10 (2017).
