New source for gas defects found: formerly thought to be produced through the reaction between aluminum and moisture in sand molds, aluminum induced gas defects in iron casting are in fact a result of precipitation of aluminum nitride during cooling.

Aluminum-induced gas defects in iron castings have plagued the metalcasting industry for many years. The source of aluminum generally is contaminated iron scrap or inadvertently high aluminum levels in ferroalloy inoculants. The main approach for the prevention of the defect is to screen charge materials and ferroalloys for aluminum.

There is now an important metalcasting application which requires the purposeful addition of aluminum. It has been found that additions of aluminum to cupola furnaces can eliminate the high loss of alloy that is inherent to the melting process.

Aluminum also can bring about a large increase in iron temperature. In this application, most of the added aluminum is oxidized; however, some unreacted aluminum is retained by the iron at levels that can produce aluminum-induced gas defects. This problem has discouraged almost all use of aluminum for this application

There have been many efforts to describe the mechanism for the formation of aluminum-induced gas defects. The unsatisfactory nature of the existing theories led to a new theory that better explains the characteristics of the formation of the defect and offers additional routes for the prevention of the problem.

Defect Definition

Aluminum-induced defects are typically graphite-lined, spheroid bubbles, 0.02-0.12 in. (0.5-3 mm) in diameter, located slightly below the cope section of a casting. The incidence of the defect is aggravated by high levels of moisture in the molding sand and high iron temperatures. On the other hand, high seacoal (highly volatile bituminous coal) levels in the molding sand seem to lessen the problem.

There are two unique characteristics of aluminum-induced pinholes. First, no metal other than aluminum has been associated with the defect. Second, the pinholes are associated with a characteristic range of aluminum concentrations (0.01-0.15%). Levels of aluminum outside this range do not produce the defect to the same degree. Any proposed mechanism for the formation of aluminum-induced gas defects must adequately explain these two characteristics.

Defect Detectives

The most widely accepted theory regarding the role of aluminum in pinhole formation attributed it to the reaction between aluminum dissolved in iron with moisture from the sand mold to produce aluminum oxide and hydrogen; hydrogen being absorbed by the liquid iron. During solidification, hydrogen bubbles form due to the relative solubility of nitrogen in austentite, which causes the gas to concentrate in the liquid phase. When the solubility limit is exceeded, bubbles form. The mechanism for the formation of hydrogen bubbles is supported by higher levels of dissolved hydrogen when aluminum is present in the iron.

Other theories have been advanced to account for the deficiencies in this dominant idea. The defect has in one case been attributed to the production of carbon monoxide and methane by reaction of aluminum sub-carbide with water.

Another proposal denied aluminum had any connection to gas defects. Rather, it suggests the observed defects were produced by the reaction of carbon dissolved in the iron with iron oxide-bearing slags. Although this reaction is responsible for many gas defects, it is not considered in the current discussion because there is such a large body of evidence that supports aluminum's connection to the formation of gas defects.

Defective Reasoning

None of the aluminum-related theories proposed thus far properly ac count for the two special characteristics of aluminum-induced pinholes described above. First, aluminum is not unique in its ability to increase hydrogen levels in cast iron in the presence of moisture in molding sand. Magnesium and manganese have been observed to produce hydrogen in this manner. Other common reactive metals present in iron, such as carbon, silicon and rare earths, also would be expected to increase hydrogen levels in cast iron; however, they have never been connected with the problem.

Second, none of the theories adequately account for the rather narrow range of aluminum concentrations (0.01-0.15%) that produce the defect.

Still another weakness of the theories that connect hydrogen with the aluminum defect is that seacoal in molding sand prevents the defect even though it increases, rather than decreases, the hydrogen level in the aluminum-bearing cast iron.

The inconsistencies noted here were recognized by previous researchers. As a result, those who adhered to the hydrogen hypothesis considered that other undetermined factors also were at work to induce the hydrogen gas defect.

The New Defect Detective

The critical aspect of the new theory is that it is not aluminum that produces the defect but rather aluminum nitride. The nitride is formed from soluble aluminum and nitrogen during solidification.

The mechanism was suggested by earlier work that found similar gas defects in continuously-cast steel were produced by the presence of titanium nitride. It suggested that a parallel reaction between aluminum nitride and those slags should be examined to determine if it could account for aluminum-induced gas defects in cast iron. The reaction produces nitrogen gas, as well as aluminum oxide and iron. Unlike the case involving titanium nitride where there was no water available for reaction, aluminum nitride also can react with water from the sand mold to produce both nitrogen and hydrogen.

The current hypothesis considers the following sequence for the formation of aluminum nitride. The solubility of aluminum nitride in liquid iron decreases with decreasing temperature. In the range of aluminum concentrations normally encountered in cast irons, aluminum nitride does not precipitate at iron pouring temperatures. However, aluminum nitride does precipitate as iron cools in the mold. The aluminum nitride particles that form (Fig. 1) subsequently react with iron oxide or water to create the gas bubble defect.

[FIGURE 1 OMITTED]

The amount of nitrogen that precipitates as aluminum nitride can be determined from the thermodynamically derived data in Fig. 2.

[FIGURE 2 OMITTED]

The Range of the Defect

The data in Fig. 2 provide the solubility of nitrogen at 2,100F and 2;552F (1,150C and 1,400C) as a function of the aluminum content of cast iron (3.5% carbon, 2.5% silicon). The 2,552F temperature represents the pouring temperature of a casting. The 2,100F temperature represents the eutectic temperature. The horizontal line at each temperature represents the solubility of iron under conditions where no aluminum nitride is formed (57 ppm at 2,552F and 27 ppm at 2,100F). The angled lines represent the concentration of nitrogen in the regions where aluminum nitride is formed. The nitrogen (ppm) that precipitates as aluminum nitride is given by the difference between the concentrations of nitrogen at 2,100F and 2,552F for any given level of aluminum. As shown, no aluminum nitride particles can form at aluminum concentrations below 0.003% (region I). Between 0.003 and 0.07% aluminum (region II), aluminum nitride forms only during solidification. As shown, the amount of aluminum nitride formed increases with increasing aluminum. When aluminum increases beyond 0.07% (region III), aluminum nitride begins to precipitate at the pouring temperature, reducing the nitrogen in solution. Hence, lesser amounts of aluminum nitride are produced during solidification.

Above 0.003% aluminum, the difference in the soluble nitrogen levels at the two temperatures in Fig. 2 gives the amount of nitrogen precipitated as aluminum nitride. These data are represented by the heavy black line in Fig. 3. As seen, the range predicted by thermodynamics corresponds very closely to the concentration range where most aluminum-induced pinholes are detected. If hydrogen is produced in addition to nitrogen, it would quadruple the total gas volume indicated in Fig. 2. Aluminum nitride also can be produced in the region enclosed by the thin and heavy black lines. It represents a region where insufficient aluminum is present to react with all the nitrogen.

[FIGURE 3 OMITTED]

Fig. 3 is defined by two conditions. The low aluminum end is determined by a concentration below which aluminum nitride cannot form at the eutectic temperature. The high aluminum end, which is not sharply defined, is established by the aluminum concentrations where aluminum nitride forms at casting temperatures. In this region, the low density of aluminum nitride particles that form will rise to the surface where they are either oxidized or absorbed by slag.

The Selective Defect

The predicted range for nitride formation depends on nitride stability. For nitrides more stable than aluminum, formation is shifted to lower concentrations. Conversely, the range for less stable nitrides will be shifted to higher concentrations. The stability of metal nitrides is progressively greater among the following metals: titanium, zirconium, aluminum, boron, silicon and chromium.

As indicated in Fig. 4, because of the stability of the nitrides, titanium nitride and boron nitride precipitate at lower and higher concentrations of the metal than aluminum nitride. It may appear as an anomaly that titanium is well recognized as a remedy for nitrogen defects while it is claimed here that aluminum creates nitrogen defects. But precipitation of titanium nitride occurs prior to casting, so the particles have the opportunity to separate from the iron prior to solidification. Boron also could produce nitrogen defects, but it is not generally present in castings at these levels.

[FIGURE 4 OMITTED]

External Factors

Defect formation requires the interaction of aluminum nitride with external factors such as moisture and seacoal in the molding sand or the presence of iron oxide-rich slags. Thus, the degree of gas-defect formation depends also on the amount of pouring ladle slag that enters the mold or the oxidation of iron by moisture in the molding sand; conversely, the generation of hydrogen from the decomposition of seacoal reduces oxidizing conditions in the mold, which in turn reduces the formation of the defect.

Defect Correction

Existing remedies for the prevention of aluminum-induced gas defects are eliminating green sand molds, adding seacoal to green sand, limiting the water content of green sand and screening charge materials for aluminum content. However, these remedies make aluminum addition impossible for some metalcasting operations.

Remedies now can focus on reducing nitrogen content, rather than aluminum. Facilities can use strong nitride-formers to reduce the nitrogen levels in the iron. Iron containing residual titanium levels greater than 0.03% should prevent pinhole formation by aluminum.

Less attractive ways to reduce the nitrogen content of the liquid iron include bubbling with an inert gas or causing a carbon boil.

This article was adapted from a paper (06072) presented at the 2006 Metalcasting Congress, Columbus, Ohio.

For More Information

"System Approach to Casting Defect Analyses and Reduction: Hydrogen Gas Defect in Iron Castings," & Kannan and J.E. Thixton, 2004 AIS Transactions (04-128).

Seymour Katz is president of S. Katz Associates, West Bloomfield, Mich.