Ironing out thin-wall casting defects: low-density alumina silicate ceramic has been effective in producing thin-wall iron castings. By constructing "gate extensions," it can now be used more efficiently.

The use of low-density alumina silicate ceramic (LDASC) in molds and cores reduces heat extraction from liquid metal and slows cooling and solidification rates, which can eliminate many of the defects that often occur in castings with thin walls (less than 3 mm. However, the implementation of this technology has been slow in being utilized.

Mold cavity areas faced with LDASC act like heavier sections than they are, which effectively slows their solidification. In the past, testing has been conducted by facing large areas of the mold or core with LDASC, meaning the thermal properties of entire mold or core surfaces have been modified, perhaps where modification is not needed or wanted. One potential solution is to be very selective in where to use the LDASC. By limiting its use to direct the flow of metal into specific areas, thin sections can be filled without affecting other casting sections and using only the minimum amount of LDASC.

The newest concept is to use LDASC inserts as "gate extensions." These inserts can be placed on the mold surface leading from the ingates across thin sections of a casting. The inserts cause only that area to behave like a heavier section, thus creating a thermal channel from the gate for the metal to flow through. Thermally, the insert makes a thinner section act like a heavier section, but the actual casting section remains thin and uniform.

Thin is In

The need for thin-wall castings continues to grow. Weight reductions for automotive castings are driven by the demand to improve performance, reduce gas consumption and reduce emissions. This has resulted in a continuing shift to more aluminum and magnesium castings in automobiles. However, the strength to weight ratios of ductile and compacted graphite irons make them attractive if thin-wall parts can be produced with good dimensional accuracy and controlled physical and microstructural properties.

The creation of thin-wall iron castings requires a delicate balance of proper metal chemistry, inoculation practice, gating design and molding and core materials, which may be why many metalcasting facilities have been slow to adopt these practices and move toward thinner castings. But this is not an insurmountable problem. The procedures simply reflect an extension of existing practices rather than a totally new manufacturing process. There is a continuing need to convince both casting designers and producers that thin-wall gray, ductile and compacted graphite iron castings are a viable option, and proven production processes are available.

Thermal Everywhere

Thin-wall castings present special challenges both in preventing misruns, cold-shuts and other related casting defects, and in providing acceptable physical and metallurgical properties. Previous work has shown that LDASC and sand blends can modify and control the thermal properties of the mold or core and produce thin-wall castings that were difficult or impossible with conventional materials, as it provides the option of engineering the thermal properties of the mold and core components to match the local section thickness of the casting.

LDASC can be used alone or mixed with sand blown into cores or molded using conventional binders and production equipment. LDASC has a density of about one quarter of silica sand and corresponding reduced specific heat and thermal conductivity. When used at levels of 20-100% with sand, the heat extraction characteristics are dramatically changed, making thin-wall castings with good metallurgical and physical properties possible. The change in thermal properties slows the cooling and solidification rates in ductile iron compared to sand molds, as shown in Fig. 1.

[FIGURE 1 OMITTED]

Slower cooling can positively affect thin-wall castings in several ways. Apparent fluidity will increase, and the metal will flow further or through thinner sections. Slower solidification will reduce the tendency for "chill" or carbide formations that will increase hardness and reduce ductility. Casting microstructures and the associated physical properties will improve.

How Thin Can it Go?

The ability to produce thin-wall iron castings with LDASC has been demonstrated in several ways. First, standard fluidity spiral castings were produced with various blends of LDASC and silica sand. These blends reduced the cooling and solidification rates in these castings, increasing metal flow. Where a standard sand mold produced a flow distance of approximately 28 in., a 40% sand, 60% LDASC (by volume) blend resulted in flow all the way to the end of the spiral, or about 58 in. (Fig. 2). Even greater flow would be possible with higher percentages of LDASC.

[FIGURE 2 OMITTED]

Thin-wall plate castings were also produced to show the effects of LDASC additions. The test casting was approximately 8 x 8 x 0.0625 in. (200 x 100 x 1.5 mm) with two ingates. The mold was produced in phenolic urethane nobake, first with sand and then faced with 100% LDASC on both cope and drag surfaces. The sand mold produced a misrun when poured with ductile iron at 2,550F (1,400C). The mold faced with LDASC filled completely with no indications of misruns or coldshuts (Fig. 3).

[FIGURE 3 OMITTED]

More complex thin-wall castings were produced to show that the LDASC material could be used either on the mold or core surfaces. A manifold test casting was used that was approximately 15 in. by 2 in. in diameter with a 0.079-in. wall (380 x 50 mm in diameter with a 2-mm wall). By gating into all three heavy sections in a sand mold, the casting could be 90-95% filled with ductile iron. But by using at least 25% LDASC in either the core or the mold facing, the casting filled completely using a single gate, nearest the sprue.

Structural and Microstructural Effects

The use of LDASC has significant control over the resulting microstructure and physical properties of thin-wall castings. This was first demonstrated with gray iron "chill wedges." These castings are typically used to show the chilling tendencies of gray iron with different chemistries or inoculation. However, they also can be used to show the effects of slower cooling with LDASC additions on casting microstructure, particularly on carbide formation.

The effects of LDASC additions on ductile iron microstructure were shown most dramatically in the thin-wall manifold casting in the 0.079-in. (2-mm) section. When both a 100% sand core and mold were used, the section contained massive carbides in fine pearlite. When a 50/50 (by volume) mixture of sand and LDASC was used in both the core and the mold, the section showed small graphite nodules in ferrite and pearlite, with some small retained carbides. However, when 100% LDASC was used in both the core and the mold, the 0.079-in. (2-mm) section contained large graphite nodules in a typical "bulls-eye" structure that might be expected in sections of 0.197 in. (5 mm) and above.

The physical properties of thin-wall castings are dependent on the cooling rates and resulting microstructures. Hardness often is used as an indicator of tensile strength in castings where it is impractical to cut and test actual sections of the metal. For the 0.079-in. (2-mm) manifold test casting, hardness values were recorded from the 0.079-in. (2-mm) section and plotted against the volume percentage of LDASC in contact with the casting. The resulting graph (Fig. 4) shows a dramatic reduction in hardness, which would indicate a corresponding reduction in unwanted carbides and an increase in ductility.

[FIGURE 4 OMITTED]

Researchers also have performed tensile tests on samples cut from thin-wall ductile iron castings produced with varying amounts of LDASC. They found a significant decrease in ultimate tensile strength (UTS) and yield strength and a corresponding increase in % elongation with increasing amounts of LDASC, as shown in Fig. 5.

[FIGURE 5 OMITTED]

Thermal Where?

The benefits of LDASC are not without cost. The LDASC material is more expensive than sand. Special mixing and handling methods may be needed. Techniques may require the use of additional cores or inserts, which increases mold complexity and cost.

To address those difficulties, LDASC gate extensions were developed for thin-wall castings. Tests were conducted with the LDASC material to produce carefully placed inserts within conventional sand molds. These extensions provide a means of controlling the temperature of metal as it is directed into very thin sections of a mold cavity, while limiting the amount of LDASC needed and leaving the microstructure of the rest of the casting unchanged. To develop modeling parameters for the application of this technology to new casting designs, several examples were produced and used.

Out of the Gate

An extreme test casting was chosen for trials. The 8 x 8 x 0.0625-in. (200 x 200 x 1.5-mm) casting could not be poured in either gray or ductile iron at normal pouring temperatures without misruns. The casting was impossible without the use of special techniques. Pouring trials in normal nobake sand molds produced castings that showed the expected flow pattern spreading from the ingates, but the metal cooled and solidified before the entire mold could fill (Fig. 6). Note that there was some variation from casting to casting, but that the fill pattern was generally consistent.

[FIGURE 6 OMITTED]

Gate extensions were created by placing 1 x 0.5-in. (25 x 12-mm) bars of 100% LDASC as inserts on the mold surface. Four configurations were used---extension from one gate (Test 1), extensions from both gates (Test 2), one cross piece between extensions (Test 3) and two cross pieces between extensions (Test 4). Initially, molds were made with extensions only in the drag, but later molds were made with extensions also in the cope (Fig. 7).

[FIGURE 7 OMITTED]

The molds were poured in Class 30 gray iron at about 2,500F (1,370C). The castings with the inserts in the drag only showed improved metal flow and fill distance, and the castings with the inserts in both the cope and drag showed even greater improvement. The gate extensions had acted as thicker sections and helped fill the impossible casting (Fig. 8).

[FIGURE 8 OMITTED]

The castings also were sectioned and polished to show the microstructures in and away from the gate extensions. At the center of the plate, between the runner extensions, the casting contained small Type A graphite with some Type D. Etching showed a pearlitic matrix with some small retained carbides. The edge of the plate showed the effects of very rapid cooling and solidification with fine Type D graphite with massive carbides in a pearlitic matrix. The section at the gate extension showed larger Type A graphite with some ferrite around the smaller graphite flakes in a pearlitic matrix. This confirmed the much slower cooling and solidification in the gate extension regions.

For More Information

"A Process for Thin-Wall Sand Castings," R.E. Showman and R.C. Aufderheide, 2003 AFS Transactions (03-145).

"Using 'Gate Extensions' to Produce Thin-Wall Castings," R.E. Showman, R.C. Aufderheide and N.P. Yeomans, 2006 AFS Transactions (06-068).

Ralph Showman, Ronald Aufderheide and Nigel Yeomans, Ashland Casting Solutions, Dublin, Ohio 

Ralph Showman is a senior staff scientist, Ronald Aufderheide is a senior product manager, and Nigel Yeomans is a project engineer for Ashland Casting Solutions, Dublin, Ohio.