Aluminum casters discuss porosity, melt quality.

Conference speakers discussed enhanced ways to measure and predict aluminum melt porosity to ensure quality cast components.

As OEM's continue to look at aluminum for lighter and more complex components, the success of an aluminum caster is still based in the quality of metal it pours. At the 5th International AFS Conference on Molten Aluminum Processing, November 8-10, Orlando, Florida, the production of aluminum castings was under examination through 26 presentations that covered topics such as melt analysis, properties, and treatment and furnace control. In addition to the presentations, the 175 attendees also were treated to an exhibition at which 20 suppliers displayed the latest in aluminum casting technology.

Following is a look at two papers presented at the conference that focus on increasing the foundrymen's understanding of how and where porosity develops. As aluminum caster's work to meet the OEM's demands, one of their greatest challenges will be to produce complex, yet structurally sound castings. At the forefront of this challenge is porosity.

NUCLEATION AND POROSITY DURING SOLIDIFICATION 

In their study, "The Measurement of the Nucleation and Growth of Microporosity During the Solidification of an Aluminum-7% Silicon Casting Alloy," J.P. Anson, J.E. Gruzleski, and R.A.L. Drew, McGill Univ., Quebec, Canada, and M. Stucky and M. Richard, Centre Technique des Industries de la Fonderie, Quebec, Canada, used a quench during solidification technique to experimentally determine the conditions that influence pore nucleation and growth in a solidifying aluminum (Al)-7% silicon (Si) alloy.

According to the authors, microporosity in aluminum casting can lead to a reduction in tensile strength, fatigue resistance and a loss of pressure tightness. Its formation is attributed to two factors: shrinkage, coupled with a lack of interdendritic feeding during mushy zone solidification (shrinkage pores), and the evolution of hydrogen (H) gas bubbles due to a sudden decrease in H solubility during solidification (gas pores). Although research has been performed to determine the relationship between casting conditions and the amount of microporosity in a casting, the majority of the models capable of providing a qualitative description of the level of microporosity fail to give accurate values because the prediction of microporosity requires a detailed understanding of pore nucleation and growth in the melt. This study is incorporating all these factors to develop accurate values.

For the three main types of porosity in cast aluminum - gas pores, shrinkage pores and gas-shrinkage pores - nucleation and growth occurs as follows.

Gas Pores - Liquid Al reacts with water vapor in the atmosphere to produce Al-oxide and H gas. Gas porosity arises during solidification due to the difference in solubility of H gas in liquid and solid Al. The homogenous nucleation of gas pores necessitates the formation of microscopic bubbles of gas, one atom at a time, via diffusion in the bulk liquid. Since the resultant bubbles would be small (only several atoms across), the surface tension pressure, which is inversely related to the pore size, would be high. As a result, homogeneous nucleation is unlikely to occur.

Shrinkage Pores - If a casting is poorly fed during solidification, shrinkage will cause a hydrostatic stress in the liquid metal. This stress increases until a pore forms with the aid of a nucleus. The nucleation relieves the stress in the liquid, and the local conditions return. Any further shrinkage in the area around the pore will lead to pore growth and the area of influence can be large since hydrostatic tension far from the shrinkage pore can contribute to growth.

Gas. Shrinkage Pores - In most cases, gas evolution and shrinkage occur in the same volume of liquid metal at the same time. As a result, an interaction between these phenomena can be expected. Both the gas and shrinkage pressure aid in the nucleation and growth of the pore as the gas pressure pushes from the inside and the shrinkage pressure pulls from the outside.

Quench-During-Solidification Testing

The principle of the quench-during solidification testing is to freeze the casting microstructure during solidification by rapidly quenching it in water. The authors performed all tests using a non-modified, pre-refined, A356 aluminum alloy with a composition of 7.1% Si, 0.12% iron, 0.32% magnesium and 0.14% titanium.

The equipment for the quench tests consisted of a thin-walled steel mold (shaped like a funnel) over a bath of water. The mold was top-filled with 600 g of molten Al. The funnel spout contained the sample and the funnel basin acted as the riser. At the desired fraction solid, the water bath was raised until the spout of the funnel mold was completely submerged to rapidly chill the sample.

Eighty quench tests were performed to evaluate the formation of microporosity at five different gas levels - 0.11, 0.16, 0.19, 0.21 and 0.23 cu cm H.sub.2/100 g Al - and at a local solidification time of 215 sec (measured at 2 cm from the bottom of the mold).

H Threshold

During the formation of microporosity, according to the authors, nucleation and growth occur continuously over a wide time/temperature range. Figure 1 shows a plot of the final microporosity as a function of the initial H content. Little porosity exists until 0.17 cu cm H/100 g Al, after which the percent porosity increases linearly with the H content (the threshold H content). The threshold H content is higher than the 0.1 cu cm/100 g Al, the usually accepted concentration level below which negligible porosity occurs.

The different behavior found above and below the threshold implies that there are two mechanisms of pore formation. Below the threshold, the percent porosity is independent of the H level indicating that only shrinkage plays a role in pore growth. Above the threshold, however, the percent porosity increases linearly with respect to H content, indicating that the growth in pores is directly related to gas evolution.

In Fig. 1b, the pore density behaves differently. It is constant before the threshold with an increase at the threshold. Since there is a dramatic change in pore density at the threshold, the nucleation of the pores is shown to be partly dependent on the H content. Work by Tynelius et al has shown the number of pores is dependent on solidification times.

Critical Fraction Solid

When discussing the critical fraction solid for pore nucleation and growth, the authors said it is useful to divide the results into those found at low (below the threshold) and high (above the threshold) H contents.

High - Figure 2 presents three graphs of typical results for a high H content (0.21-0.22 cu. cm H.sub.2/100 g). In Fig. 2a, the pore density is plotted vs. the fraction solid at quench. A baseline region is evident on the curve where the number of pores is constant, and a second region, where the number of pores increases up to the maximum at 100% solid, also is seen. The boundary between these two regions is the point at which new pores begin to nucleate - the critical fraction solid for nucleation.

In Fig. 2b, the percent porosity is plotted against the fraction solid at quench. The same two regions as in 2a are evident, however, the boundary between these two regions is the critical fraction solid for growth. It is important to note that the two critical fraction values are not equal so the pores do not begin to grow at the same fraction solid as they nucleate.

Low - Figure 3 displays three graphs of typical results for low H content quench tests (0.11-0.12 cu. cm/100 g). The graphs of pore density (a) and percent porosity (b) are similar to those obtained at high H contents. Two distinct regions of behavior also are found - a baseline region in which there is no change in the pore density or the percent porosity and a region in which there is a sudden increase in each parameter at a distinct fraction solid. At low H contents the average pore size does not peak (as it does at high H levels) but rather increases until a maximum at a fraction solid of 100% (c). The difference in behavior is due to fact that the boundary fraction solids are equal at about 65%.

Conclusions

From their experiments, the authors have established some baseline information:

* a threshold H content exists below which only insignificant porosity forms;

* two modes of pore growth exist - one at high H contents and one at low;

* a minimum amount of porosity always exists (baseline porosity);

* pore growth that begins above the H threshold baseline is dependent on initial H content and occurs before significant new pore nucleation;

* new pore growth below the H threshold is independent of H content and occurs simultaneously with new pore nucleation;

* new pore nucleation above and below the threshold is independent of H content and occurs at the same fraction solid.

POROSITY PREDICTION VIA RPT AND SIMULATION

In their paper, "Using the Reduced Pressure Test and Solidification Modeling to Predict the Porosity Distribution in Aluminum A356 Castings," Franco Chiesa, Centre de Metallurgie du Quebec, and Pascal Desilets and Francois Morin, College de Trois-Rivieres, demonstrated how the result of a reduced pressure test (RPT) can be combined with solidification modeling to predict microporosity distribution in a sand or permanent mold casting.

According to the authors, at the heart of this marriage is a correlation (equation) that has been derived between the local microporosity in a primary A356 alloy and two thermal parameters-local solidification and solidus velocity. The equation is:

microporosity (% by volume) = C x ([d.sub.rpt]) x [[t.sub.sl].sup.m] x [[V.sub.s].sup.n]

In this equation, C is a constant that depends on the density of the reduced pressure sample ([d.sup.rpt]), [[t.sub.sl].sup.m] is the local solidification time and [[V.sub.s].sup.n] is the solidus velocity. The correlation is valid until the maximum level of porosity (obtained at a low gradient) is reached for the values of t and d. The graph in Fig. 4 illustrates the maximum porosity that results from a melt of a given gas level as a function of the solidification time. This assessment implies that all points in the casting are cooling at the same rate and feeding is not available from or given out to the surroundings.

This correlation and equation for the prediction of microporosity was applied to two industrial castings - an A356 sand-cast pressure cylinder and an A356 low pressure permanent mold cast wheel.

Casting Models

The 70-lb sand-cast pressure cylinder (15 in. diameter and 2 in. wall thickness) was specified to a radiographic quality grade C, which corresponds to a porosity of 1.5% for plates 1.5 in. thick. When the casting was evaluated via the correlation equation, the porosity distribution shown in Fig. 5 was obtained. It shows that maximum porosity will take place around the ring and above the mid-height of the casting. It also indicates that for a melt gas content of 0.13 ml/100 g, the maximum porosity level is 1.2%, which satisfies the grade C specification. The examination of radiographs of five prototypes confirmed the presence of a ring of higher microporosity in the upper part of the casting.

The modeling of the same casting was carried out for an alternate rigging comprising three risers instead of just one as in the original setup. The result, using the equation and correlation, was a reduced maximum porosity to 0.8%.

For the low pressure permanent mold wheel casting, the same porosity correlation equation as with the sand casting was used to predict the effect of mold coating and mold run-in on the occurrence of porosity. The melt gas level for the casting was at a RPT density of 2.61. The simulation was carried out for a filling time of 30 sec, a cooling time and mold open time of 1 min after ejection, and a total cycle time of 7 min.

Three cycles were run with the wheel casting, each showing a different porosity level. The porosity level is lower after 3 cycles by 0.4% or more. This is particularly true in the spoke area. To improve wheel quality, artificial cooling could be applied at two locations in the rim section of the wheel. In addition, porosity improvements can be seen when a surface finish coating is used instead of an all-purpose coating.

Testing using the equation and correlation also was carried out on sand and permanent mold cast test bars. In these tests, a correlation including solidification time and porosity level showed that the elongation could vary from 1-6% for an as-cast unmodified A356 alloy. In addition, the distribution of solidification times and porosity in the test bars allowed for a correlation of actual local mechanical properties inside a casting to the measured tensile properties of the separately cast test bars.

Conclusions

In the study, the authors determined how a criteria function can be used to quantitatively predict the local porosity in A356 castings poured in sand or permanent mold. It was demonstrated that riser location, mold coating and mold run-in can influence porosity distribution. The correlation equation also showed that elongation can vary from 1-6% for an as-cast, unmodified A356 alloy and that the distribution of solidification times and porosity in the standard test bars for sand and permanent mold correlate the actual local mechanical properties inside a casting to the measured tensile properties of the separately cast test bars.