Inclusions in aluminum foundry alloys.
By contrast, the adverse effects of inclusions in foundry products were not widely recognized until the last decade. This awareness was prompted by increased performance expectations such as higher strength-to-weight ratios, lower cycle fatigue and improved machined surface finish. Bearingized surfaces in automotive cast aluminum master cylinders, for example, may require melt cleanliness levels consistent with certain wrought alloy products. The effects of inclusions on hydrogen porosity distribution can also influence product performance. The heterogeneous nucleation of hydrogen porosity by inclusions is illustrated in Fig. 1.
Metal cleanliness has improved significantly because of higher performance expectations, sensitivity to quality and the emergence of filtration devices. Some common inclusion types in foundry alloys and prevalent inclusion sampling techniques will be covered. inclusion Types
Several general classes of inclusions exist in foundry products. Inclusions have different compositions, sources and morphology. A few of these sources are the following:
In Situ Oxides-Oxides, the most common inclusion, form either by direct melt oxidation or by the oxidation of certain elements during alloying. Oxide inclusions can also be introduced by furnace charges, but high-quality remelt ingot and careful melting practices generally preclude this as a major source.
Gamma or alpha alumina AI,O,) inclusions are frequently found in magnesium-free alloys. Gamma alumina of a lacy film morphology often forms initially and calcines after prolonged melt holding periods to the alpha phase (corundum). Alumina formation follows the parabolic oxidation law, where the initial rate of oxidation is high.
Magnesia (Mgo) inclusions are found in alloys containing magnesium. Several morphological forms exist that include colonies of 1-5 micron dispersoids, films
lusters. Dispersoid colonies typically result from oxidation of localized concentrations of magnesium that form during alloying malpractice. Failure to quickly submerge and disperse magnesium ingot generates this condition. Oxide films and clusters are generally formed by direct melt oxidation. Another oxide that is formed by melt oxidation is the magnesium aluminate spinel MgAI,O,). Spinels usually form a thick film or cluster and are created by high melt surface temperatures from direct flame impingement. Spinel formation can be accelerated by existing magnesia inclusions. Increased oxidation rates occur when the protective oxide layer breaks down during spinel formation. Spinel formation is reduced with additions of up to 0. 1 % beryllium.
Salts-Inclusions can also be formed by entrained salts. These salts form when chlorine or argon/nitrogenchlorine mixtures are used to remove hydrogen from melts. If the melt contains magnesium, a molten salt, magnesium chloride MgCl,) is formed in the reaction:
Since MgCI2 melts at 1307F, it can be present as a solid or as an immiscible liquid, depending on melt temperature. Sodium (added for modification) and calcium may also form solid salt particles that are difficult to remove by filtration.
Carbides-Carbide formation can produce another type of inclusion. Aluminum carbide Al4C,) may be formed during the aluminum smelting process and frequently results from poor quality secondary ingot or pig. These carbides are generally innocuous due to their small size (1-10 microns). Aluminum carbide can also be formed from using certain solid degassing tablets that consist of hexachloroethane C2CI,) via the reaction:
3C,cl, + 7A/ 4 3AIC13 +A14C3
Aluminum carbide particles can be as large as 50 microns, and most of them float to the surface along with aluminum chloride vapor formed in the reaction.
Intermetallic Compounds-Large titanium aluminide T1A2,) particles can cause inclusion defects. These large particles > 1 0 microns) are due to poor quality grain refining materials. Titanium aluminides, which are useful as grain refining nuclei, generally are smaller than 5 microns and do not constitute an inclusion problem.
High-iron diecasting alloys may exhibit chromium-manganese-iron-intermetallic "sludge" inclusions from low holding furnace temperatures. These sludges have been found to accumulate in channel induction furnaces.1
Exogenous/Refractory-Spalling of high-silica refractories may cause occluded particles in the melt that are not a result of melt oxidation. Under certain circumstances, this silica S1O2) can be chemically reduced by aluminum to form alumina. The breakdown of the refractory is accelerated by molten salt fluxes. Inclusions are illustrated in Fig. 2 and summarized in Table 1. Inclusion Sampling
Many techniques are used to measure inclusions. Monitoring of inclusions in the molten metal phase is favored because it provides a good indicator of casting quality. Unfortunately, standard methods for characterizing dilute micron-sized particles in aqueous systems are not amenable to molten metal, thus leading to the development of specialized techniques.1
One of the oldest methods of assessing melt cleanliness is the reduced pressure solidification (or vacuum density) test. Inclusion content is inferred from porosity distribution variations caused by the heterogeneous nucleation of hydrogen porosity, as in Fig. 1. This test is interpretive and useful only to approximate cleanliness within the same alloy system.
The break test is another interpretive technique where a molten sample is carefully drawn, solidified and fractured. The fracture surface is examined optically for inclusions. Neither the break nor vacuum density tests significantly concentrate the inclusions.
A more ideal direct melt cleanliness assessment technique that simply demonstrates gauge capability, has adequate sampling volume and has real time capability remains elusive. However, several clever methods for assessing molten metal quality have been developed and can be grouped by their operation modes (filtration, extraction/ centrifuging, ultrasonic, counting and nuclear).
Methods based on filtration are the oldest techniques. All are grab sampling methods that pass a discrete sample (several kg) of molten aluminum through small-pore filter media that concentrates the inclusions in a filter cake. The cake is examined for inclusions by metallographic and constitutional analysis that may include optical, electron microprobe and energy dispersive techniques. Although filtration methods are primarily useful for qualitative inclusion sampling, Doutre et al.5 and Eckert et al.6 have reported improvements that provide semiquantitative melt cleanliness information through mathematical interpretation. All techniques based on filtration are categorically limited by sampling volume.
The extraction/centrifuging sampling method, which was pioneered by Siemensen, is also an off-line grab sampling technique. A 100 g sample is obtained from the bath and solidified under centrifugal force. Resulting body forces cause dense inclusions to migrate and concentrate in the sample's perimeter. The sample is examined in a manner similar to the filtration method. Siemensen's method reportedly can exhibit most inclusions larger than 0.5 microns. He also described solvent extraction techniques that attempt to quantify melt cleanliness levels.1 Ultrasonic inclusion sampling techniques developed within the last decade sonic provide real time, high-sampling volume and quantitative melt cleanliness information. All are based on the acoustic interaction of suspended inclusions within 3-25 Mhz that is shown by a probe immersed in the melt. Mansfield9,10 primarily measured the weakening of acoustic energy by the suspended inclusions, while Eckert" examined energy directly reflected from the particles. Because inclusion size sensitivity is directly related to the sound frequency, frequency and beam field intensity limit the sensitivity of ultrasonic techniques to about 8 microns.
A sampling technique has been developed by Alcan based on the Coulter counter principle in which metal is forced to flow through a small (0.25 mm) orifice. The electrical resistance changes caused by the inclusions are measured. The actual inclusion concentration is obtained by calculating inclusion particle diameters and the actual particle count. Detectable inclusion size is reportedly 10 microns.
The final popular method of inclusion sampling is a grab sampling technique based on nuclear principles and is known as the fast neutron activation analysis (FNAA). A directionally solidified sample (to control shrinkage) is obtained from the melt and machined into a small cylinder. The cylinder is irradiated, and the rate at which radioactive decay occurs is measured and related to the oxygen content of the sample. If the empirical formula for the oxides is known, a calculation of the mass fraction oxides can be made. Although FNAA is useful for research purposes, its use is limited by sample turnaround time, expense, analytical correction factors and an inability to differentiate types of oxides. An alternative method to determine oxygen content from the same samples uses a carburizing process.
Acknowledgments: The author would
like to thank Dr. R. Mutharasan of Drexel
University and Drs. D. Apelian and S.
Shivkumar of Worcester Polytechnic Institute
for input, and S.A. McKillop of Alcoa
for the metallographic preparation.
1 "Metallography and Microstructures,"
Metals Handbook, American Society for
Metals, vol 9 (1985).
2. Doutre, D.A., Doctoral Thesis, McGill
University, Montreal, Canada (1984).
3. Irwin, D.W., Sampling to Detect Inclusions
in Molten Aluminum," Aluminum
Melt Refining and Alloying -Theory and
Practice Symposium, Melbourne, pp51 - 58 (1989).
4. Guthrie, R., In-Line inclusion Monitor - ing," ibid, pp T1-T21.
5. Doutre, D., Gariepy, B., Martin, J.P. and
Dube, G., "Aluminum Cleanliness Monitoring:
Methods and Applications in Process
Development and Quality Control,
Light Metals, pp 1 1 79-1196 (1985).
6. Eckert, C.E., Mutharasan, R., Apelian,
D., Miller, R.E., "An Experimental Technique
for Determining Specific Cake Resistance
Values in the Cake Mode Filtration
of Aluminum Alloys," ibid, pp 1225 - 1248.
7. Siemensen, C.J., Strano, G., "Analysis of
Inclusions in Aluminum by Dissolution of
the Samples in Hydrochloric/Nitric Acid, "
Fresenius Z Anal. Chem., vol 308 (198 1).
8. Siemensen, C.J., Thesis, Dr. Techn., NTH,
Trondheim, Norway (1 982).
9. Mansfield, T.L., U.S. Patent 4,282,755. 10. Mansfield, T.L., "Molten Aluminum Quality
Measured with Reynolds4 MSystem,"
Light Metals, pp 1305-1328 (1984). 11. Eckert, C.E., U.S. Patents 4,563,895 and
4,662,215. 12. Guthrie, R. and Doutre, D.A., "On-Line
Measurements of Inclusions in Liquid
Metals, " Refining and Alloying of Liquid,
Aluminum and Ferro Alloys, pp 145-164
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||AFS Molten Metal Practices Committee Report|
|Date:||Apr 1, 1991|
|Previous Article:||Francis & Nygren: continuing a tradition of foundry 'firsts'.|
|Next Article:||Millionth engine marks milestone for Honda's U.S. production unit.|