Molten aluminum contamination: gas, inclusions and dross.
In any battle, the first step to success is understanding the "enemy" and how it "thinks." Similarly, the only way to truly fortify your aluminum melt from factors that impair its quality is to understand the processes at hand.
Aluminum exhibits properties that make it unique among commercially important structural metals, particularly in its liquid state. The alloy has a low density and a high interfacial energy between both atmospheric gases and most oxides, including its own. At common processing temperatures, its oxide - alumina or [Al.sub.2][O.sub.3] - is solid and exists for extended periods of time as amorphous or crystalline (a or y phases) surface films.
These [Al.sub.2][O.sub.3] films can be a blessing or a curse. By coating surfaces of aluminum, films inhibit continued oxidation. However, as little as 5 ppm oxygen (O) in the form of entrained alumina films or other inclusions can significantly reduce properties by nucleating gas porosity during solidification or by acting as planar defects in the finished part.
Another property of aluminum is its high liquid solubility for hydrogen (H) compared to its solid solubility. In terms of weight, the solubility of H in solid, pure aluminum is small [0.05 ppm at 1220F (660C), for instance]. Comparatively, it is 0.85 ppm in the liquid state.
Thus, early measurement methods relied on measurement of volume of [H.sub.2] gas at STP released upon remelting a given weight of aluminum in a vacuum or carrier gas. This resulted in the arcane units of [cm.sup.3]/100 g (volume [H.sub.2]/weight aluminum), which have become an industry standard. Coincidentally, the units are nearly equal. (Note: the conversion factor for [H] is: ppm = 1.11 x [cm.sup.3]/100 g.)
Entry and exit of H in molten aluminum and its interaction with inclusions are now becoming well understood. Progressive aluminum foundries, recognizing the mechanisms that produce gas and inclusions, now use technically correct methods for transfer, cleaning and degassing.
Aluminum cast houses have used these methods for years to continuously cast plate, sheet and can stock, but technology transfer to the foundry has been slow. The role of inclusions is difficult to grasp due to the obvious opacity of aluminum, the lack of efficient inclusion measurement systems and the tendency for inclusions to segregate into nonuniform distributions.
Reactions between Molten Aluminum, Oxygen and Hydrogen
Figure 1 illustrates common phenomena involving O and H in molten aluminum and the resulting distribution of contaminants.
The surface of molten aluminum may be covered with films and dross, as shown in Inset A. Inset B illustrates the structure of dross. As shown in Inset C, inclusions tend to float in suspension in molten aluminum. Chemical reactions involving the diffusion of Al and H through the [Al.sub.2][O.sub.3] film are shown in Inset D. Suspended inclusions interact with dissolved hydrogen, modeled in Inset E. Finally, dissolution and desorption of hydrogen into and out of molten aluminum are shown in Inset F. Each phenomena is detailed below.
Alumina Films (Inset A)
Whenever Al is exposed to O - as when molten aluminum is poured - an amorphous [Al.sub.2][O.sub.3] film forms within milliseconds. [Al.sub.2][O.sub.3] films are always present because the free energy of formation for [Al.sub.2][O.sub.3] is so large that avoiding it would require an impossibly low pressure of 10-45 torr. While it is impossible to prevent films, inert cover gases can effectively retard the thickening of [Al.sub.2][O.sub.3] films. However, whenever O contacts Al, the atoms bond.
As mentioned above, surface films act as a protective coating on aluminum, although they are less stable at molten aluminum temperatures than at room temperature. Although [Al.sub.2][O.sub.3] films are initially amorphous, they will crystallize, first to a and later to y, given sufficient time and temperature. Once crystallized, films substantially lose their protectiveness. In addition, the diffusion rate of [Al+.sup.3] through the film is greater in crystalline vs. amorphous [Al.sub.2][O.sub.3]. (See text for Inset D.)
Crystallization is accelerated by three factors: the presence of suspended inclusions that are already crystalline; Mg concentration in the alloy, which introduces MgO and Mg[Al.sub.2][O.sub.4] into the film; and temperature, which is perhaps the most abused variable. When crystallites coarsen, they draw nearby [Al.sub.2][O.sub.3] and thus tear the surrounding film. The immediate reoxidation at each break in the film launches linear-rate film thickening.
Preexisting inclusions facilitate crystallization of the amorphous film by acting as nuclei. The presence of Mg in the alloy enables MgO to form. Since MgO has a lower free energy of formation than [Al.sub.2][O.sub.3], it tends to form preferentially, particularly in alloys containing more than 0.5% Mg by weight. When MgO enters the [Al.sub.2][O.sub.3] film, Mg[Al.sub.2][O.sub.4] (magnesium-aluminum spinel) forms, which exhibits faster crystallization kinetics than [Al.sub.2][O.sub.3]. Above 1400F (760C), the film loses its protectiveness rapidly due to rupturing of the film, as discussed above, and accelerates crystallization and diffusion kinetics. Both crystallization and diffusion are enhanced by the higher temperature, crystalline vs. amorphous structure and the presence of short-circuit diffusion paths, such as grain boundaries.
Dross (Inset B)
Dross is composed mainly of crumpled films that encapsulate a significant amount of unoxidized aluminum. In addition to films in various stages of crystallization, dross will also contain inclusions scavenged from the liquid bath, flux constituents, ash, sludge particles and oxidation products from the contained metallic aluminum.
The most common origin of dross is turbulent metal handling, which breaks films and causes new films to form. Examples of such metal handling are dipping of ladles, stirring, pouring, and even degassing - as bubbles of sparging gas break the surface and momentarily expose unprotected aluminum to the atmosphere. The common thread is temporary creation of oxidizing surfaces later have less surface area. The solid films maintain their two-dimensional character even as they fold into dross.
Dross can be roughly classified into two groups - wet and dry.
Wet dross is predominantly metallic aluminum alloy enveloped by oxide films. New dross is typically wet and lies in intimate contact with the molten metal bath. Such contact allows free communication between the liquid bath and oxides in the dross. Wet dross can easily bleed inclusions into molten aluminum because the interfacial energy barrier is already overcome. AS will be discussed, [Al.sub.2][O.sub.3] and other metal oxides have a greater specific gravity than molten aluminum so they should naturally tend to sink.
Dry dross, in contrast, has the appearance of a nonmetal, though it contains as much as 50% metallic aluminum, and is partitioned by surface tension from molten aluminum. This is the preferred dross condition because the contained oxides can't readily re-enter the molten bath. Fluxes are usually the only means of draining enough metallic aluminum from dross to make it dry.
Suspended Inclusions (Inset C)
Although oxide inclusions have a greater density than molten aluminum, they easily float in suspension. This is due to minimal density difference between inclusions and molten aluminum, and high surface to volume ratio of the inclusions.
The specific gravity of molten aluminum at casting temperatures is about 2.3, while that of [Al.sub.2][O.sub.3] varies from 3.5-3.9. Other common inclusions such as MgO and Mg[Al.sub.2][O.sub.4] have similar specific gravities.
Moreover, inclusions tend to be small, irregularly shaped, or both, particularly in the case of entrained films. Films suspended in molten aluminum tend to be crumpled and float like "newspapers blown around on a windy day." Crumpled films, the primary constituent in dross, have an irregular appearance on a metallographic section [ILLUSTRATION FOR FIGURE 2 OMITTED].
Diffusion and Reactions with Oxygen and Moisture (Inset D)
A common misconception is that [O.sub.2] penetrates the alumina film and combines with Al. Actually, Al is the species (as [Al+.sup.3]) that migrates through the film, oxidizing on the outer surface of both molten and solid aluminum. This occurs because the [Al+.sup.3] cation has an atomic radius of 0.51 [Angstrom] and the 0-2 anion, 1.32 [Angstrom]. Obviously, the smaller atom diffuses more rapidly through the film.
The requirement of penetration through an ever-thickening film is the source of the parabolic growth rate of films on aluminum. Initially rapid film thickening slows dramatically as the diffusion distance through the film increases. On the other hand, microfissuring during film crystallization (discussed above) removes this diffusion barrier, accelerating penetration by [Al+.sup.3].
Two oxidation reactions can occur when molten aluminum contacts air:
2[Al.sub.(s)] + 3/2 [O.sub.2(g)] [right arrow] [Al.sub.2][O.sub.3(s)]
2[Al.sub.(1)] + 3[H.sub.2][O.sub.(g)] [right arrow] [Al.sub.2][O.sub.3(s)] + 6[H.sub.(g)]
These reactions are irreversible in an engineering sense due to the large negative free energy change and the segregation of the reaction products. The nascent (newly liberated) H that is produced can dissolve into the film (and subsequently the molten aluminum) or combine into [H.sub.2].
Due to its small size and mass, H diffuses rapidly in the film as H+ cations (actually, protons). However, three groups of H atoms escape into the atmosphere and dissipate: those that never enter the [Al.sub.2][O.sub.3] film; those that enter the film, but return to the surface; and H atoms that desorb from the molten aluminum bath and find their way through the surface film. The key point is that the dissolution and desorption of H are dynamic processes as H constantly enters and leaves the molten aluminum bath.
Inclusions and Hydrogen (Insets E and F)
Inclusions suspended in the aluminum bath interact with dissolved hydrogen. The high interfacial energy between inclusions and molten aluminum causes reentrant corners to collect H from the melt as [H.sub.2] in a reversible reaction:
2[H.sub.(soln)] [tautomer] [H.sub.2(g)]
How Do Heavier-than-Aluminum Inclusions Become Buoyant?
For years, foundry operators have suspected that [H.sub.2] or some other gas is adsorbed onto inclusions. Inclusions have a buoyancy too great to be fully explained by low specific gravity differences and large surface area to volume ratios. Early evidence that inclusions nucleate gas porosity during solidification stemmed from increases in casting soundness after metal filtration. In addition, inclusions often rise to the surface instead of sink. A logical conclusion is that inclusions become buoyant as a result of adsorbed gases.
Inclusions in Induction Melted Aluminum - GM Powertrain melts large quantities of machining chips at its foundries in Bedford, Indiana, and Massena, New York. Initially, this practice produces a relatively dirty mixture of molten aluminum and films that were originally on each chip. A 1991 filtering experiment at Massena consisted of pouring 30 lb (13.5 kg) through 3 in. (75 mm) diameter by 0.875 in. (22 mm) thick reticulated foam filters. The object was to evaluate the capture efficiency of various filter pore sizes with induction-melted chips.
In the first trial, operators ran the 15,000 lb (6820 kg) furnace at full power for 2 min, turned the power off, dipped aluminum within 30 sec, and poured through a 25 ppi filter. In the second trial, the procedure was changed slightly - the aluminum was poured through a 15 ppi filter, just 2 min after furnace power was shut off. This procedure change was made with the expectation that suspended inclusions would sink, thus providing cleaner metal near the surface.
Surprisingly, the filter clogged up after only about 5 lb (2.3 kg) was poured. Evidently, the [Al.sub.2][O.sub.3] films (exemplified in [ILLUSTRATION FOR FIGURE 3 OMITTED]) left behind by the melted chips were made buoyant by adsorbed gases and segregated toward the surface instead of sinking.
Inclusions Suspended Just Below the Surface - An early concern in the green sand aluminum casting program at GM Powertrain Saginaw Metal Casting Operations (GMSMCO) was the probable generation of excessive dross and inclusions by 240 dips/hr of ladles holding 200 lb (91 kg) each. Operators simulated dipwell contamination in a 200-lb crucible furnace by repeatedly dipping metal out and periodically testing for gas content. The reduced pressure test (RPT) sampling procedure included dipping the sample cup just below a freshly skimmed surface.
The results, shown in Fig. 4, were a gradual increase in [H] as measured by the AlScan technique, compared with irrational changes in specific gravity of the RPT samples. Suspecting suspended near-surface inclusions, the test was rerun, taking RPT samples from metal about 6 in. (150 mm) beneath the surface. The result, also plotted in Fig. 4, more closely follows the [H] curve. GMSMCO has since instituted deep dipping into its RPT sampling procedure.
Analogy to Solidification - As stated previously, inclusions are recognized as nucleating sites for gas porosity during solidification. This is due to a higher interfacial energy between aluminum and inclusions than between molten aluminum and [H.sub.2] gas. Hydrogen can be drawn from the solidifying melt particularly easily where two inclusion surfaces form a re-entrant corner, thus nucleating a bubble. This is favored thermodynamically since the aluminum oxide interfaces are replaced by aluminum-[H.sub.2] interfaces.
Interestingly, the principles described above do not require solidification, only re-entrant corners formed by inclusions. Hence, it is only reasonable to expect that one source of the buoyancy in inclusions is adsorbed [H.sub.2], precipitated from the melt. Additionally, this would explain how adsorbed air or even sparging gas can attach and buoy inclusions.
This article was excerpted from a proceedings paper presented at the 1995 AFS 4th International Conference on Molten Aluminum Processing.
|Printer friendly Cite/link Email Feedback|
|Comment:||Molten aluminum contamination: gas, inclusions and dross.|
|Author:||Crepeau, Paul N.|
|Date:||Jul 1, 1997|
|Previous Article:||A method for extending your cupola campaign.|
|Next Article:||Lost wax to lost foam: reflections on past, present and future.|