Magnesium driving to permanent mold: the development of magnesium alloys for casting via semi-permanent mold will mean wonders to the automotive industry if successful.

Reduced weight.

This is a key term throughout the automotive industry as OEMs search for methods to design and manufacture vehicles with efficient fuel economies. Aluminum alloys have been embraced greatly for engine and chassis components, but as further weight reductions are needed, lighter materials, such as magnesium alloys, have gained part of the spotlight.

In the automotive realm, magnesium alloys currently are used for diecast automotive components, including seat frames, instrument panels and steering wheels. Yet, future applications, such as in engine components, may require geometries and properties that cannot be met by diecast magnesium. Purchasers of metalcastings often encounter the difficulty of long-term fatigue properties magnesium alloys have when placed in high-temperature applications. A number of firms currently are investigating means to make magnesium alloys more heat resistant, but these alloys are designed primarily for sand and diecasting processes.

Recent investigations have examined several casting processes, such as low-pressure permanent mold (LPPM) casting, to allow for magnesium in applications that currently are not feasible. With LPPM. components with higher mechanical strengths than die castings can be achieved due to the controlled solidification trod tranquil mold fill of the process. Additionally, LPPM cast components can incorporate internal cores so more complex geometries can be cast. and these components can be heat treated to increase properties. LPPM technology also can accelerate structural magnesium market growth, but a number of technical barriers currently limit the use of LPPM magnesium casting.

The Magnesium Low Pressure Development Project was created to seek ways to improve these technologies by converting an existing cast aluminum front cradle for a Chevrolet Corvette GMX245 to a cast magnesium alloy cradle with the intention of meeting the C6 Corvette LPPM aluminum (356.1T6 alloy) cradle performance specifications, attachment and dimensional requirements, and safety, performance and mass reduction qualities. The goal of the project was to produce a magnesium cradle that ideally could be assembled in the Chevrolet C6 Corvette with an LS6 V8 base engine.

This article discusses the first phase of the project, detailing the engineering, process development and first casting runs.

Lightening Up

In order to make an accurate conversion, the computer data from the aluminum cradle was documented for linear static stress and modal, thermal and stiffness data. The cradle then was redesigned as a magnesium component for LPPM casting based on these analyses, with the objective of obtaining a 35% weight reduction over the aluminum design. The benchmark measurements of stress levels in the cradle structure and the room-temperature mechanical properties were based on diecast magnesium alloy AM50 (Table 1). The overall dimensions of the cradle were 36.2 in. x 23.6 in. x 7.8 in. (920 mmx 600 mm x 200 mm)--at front frame mounts.

The existing aluminum cradle used an open-ribbed design in the front and rear cross support beams (Fig. 1). Initial engineering tests showed the magnesium cradle would require additional ribbing support to meet the required stiffness properties on these cross support beams. Rather than adding weight with thicker or additional reinforcing ribs, engineers redesigned these cross support beams as a hollow cored box design (Fig. 2). The cradle included mounting points for attachment to the engine, and the magnesium design would need to be completely compatible with these and other attachment points for installation on the vehicle (Fig. 3).

[FIGURES 1-3 OMITTED]

The box section approach was necessary to meet the stiffness requirements of the aluminum cradle with the lower elastic modulus of magnesium. The stiffness requirement could be met with a box section at a lower total weight than what would be required in an open channel with ribbing. The nominal wall thickness on the hollow front cross support beam was 0.19-0.23 in. (5-6 mm), and the rear wall thickness was 0.35-0.39 in. (9-10 mm).

Through the design flexibility of LPPM, hollow cross support beams could be cast with a coldbox sand core (Fig. 4), something that could not be done through diecasting.

[FIGURE 4 OMITTED]

A finite element stress analysis for ambient temperature operating conditions was run on the magnesium design. To accurately simulate the road loading, the spring clamps were included in the model to represent the maximum von Mises stress encountered for each load case. The left and right engine mount stiffness values were 0.6 kN//mm and 0.8 kN/ mm below target. But these stiffness values could increase when using a high-temperature magnesium alloy, which may be needed because of the higher operating temperatures observed in the current design.

Searching for an Alloy

Because portions of the cradle are near the engine exhaust system (Fig. 5), the magnesium design was analyzed to calculate the probable temperatures these locations experience under operating conditions. Values for emissivity and film coefficients for pure magnesium indicated an expected temperature of 237-257F (114-125C), which could be too high for magnesium alloy AM50 to withstand without a significant reduction in mechanical properties. At tile maximum predicted component operating temperature (257F/125C), the typical yield stress of AM50 alloy is 90 MPa, which is lower than cradle requirements in the operating temperature. Thus, a high-temperature alloy was considered for the application.

[FIGURE 5 OMITTED]

New magnesium alloys that provide higher properties at elevated temperatures have been developed for diecasting and sand casting applications, but there are no available high-temperature alloys specifically designed for permanent mold casting. Using magnesium alloy AZ91 as a benchmark, several new commercial magnesium alloys were poured into sand molds to compare fluidity, hot tearing, shrinkage and other castability issues and to determine the most suitable high-temperature alloys for the cradle. Trials were run using low-pressure engine block sand molds to provide a more complex geometry. Engine blocks were poured from each of the alloys, and all exhibited complete fill with the same level of casting defects. Magnesium alloy MRI 202S was chosen for the LPPM cradle because of its castability and probable mechanical properties (Table 2).

On Trial

The initial casting trials were conducted in MRI 202S alloy (a sand casting alloy), but to conserve this alloy, much of the subsequent process development work was done in AZ91E. A three-runner gating system was used, and the mold was coated with a commercial aluminum permanent mold coating with 10% NaF addition. Although this coating demonstrated improved durability, the coating continued to show wear and occasional casting detects related to the reaction of the metal with the coating, a typical problem with mold coatings in magnesium permanent molding. A 10% calcium fluoride addition (which is less soluble in water) was used in subsequent casting trials and showed a reduction in coating reactivity. The cradle mold was fully vented from the start of production in a manner similar to what would be used in a comparable aluminum component.

The trials were conducted under a maximum atmospheres pressure of 0.75 psi, and a protective cover gas was used to prevent magnesium's strong oxidation tendencies. Castings were poured at 1,320-1,400F (715-760C). The mold temperatures were measured at 475F (246C) for the upper die and 700F (371C) for the lower die. The difference was due to the heat of the metal as it entered the mold through the gating system, radiant heat from the intermediate tube and the attached intermediate tube heating system.

Better Than Expected

Investigations prior to the trials indicated that magnesium would experience poor fluidity under low pressure, causing difficulties with filling the cradle mold. This would force the need for higher LPPM pressures than aluminum alloys in similar applications. However, differences in response to filling pressure were noted, and less pressure was required for magnesium alloys than would typically be used for aluminum alloys (Fig. 6). The fluidity of the molten magnesium under pressure exceeded expectations, and the mold cavity filled completely on the first pour. The mold was reworked for better closure on the parting lines to minimize flashing on future runs. The metal feeder box and front and rear runner systems were eliminated on subsequent casting trials, and the casting was adequately filled from the single center sprue and runner.

[FIGURE 6 OMITTED]

Although a protective cover gas was used on the melt, purging the mold cavity itself with protective gas had little apparent effect on quality, and mold purging was eliminated in the later trials.

Several trial castings continued to show defects possibly related to the turbulence in the gating system. Multiple metal flow patterns converged from the front and rear runner systems that were no longer being used to fill the casting. Molten magnesium appeared to be filling these gates and reentering the mold cavity, causing an interrupted flow to the casting and a variable fill rate. The ingates that connected the cradle cavity to the front and rear runner systems were closed off to prevent metal from filling these runner systems during future casting trials. The trials also demonstrated the need for additional venting to aid in the removal of mold and core gas.

The initial cradles showed hot tearing in isolated locations. Pure metals and alloys of eutectic compositions are not prone to hot tearing, while alloys of intermediate compositions are. This depends largely on alloy composition, part design, feeding system and casting parameters. Hot tearing defects arise partly due to the inability of the casting to shrink freely during cooling and can be more severe with the constraint of a permanent mold. The high-temperature magnesium alloy was grain refined using Mg-Zr master alloy. The proper melt treatment of these alloys is critical to their use as permanent mold alloys.

Two samples were cut from the cradle casting at the location of the hot tearing. These samples were analyzed, and hot tears (Fig. 7) and severe cracking were discovered. After a chemical analysis of the samples comparing the average composition of the ingot (Table 3), it was found that the zirconium content in the samples was less than in the ingots and also less than the standard limit of 0.4% for sand castings of zirconium-containing alloys. To prevent this, the metalcasters must perform zirconium grain refinement at 1,418-1,427F (770-775C), maintain a zirconium content of more than 0.4% in the final casting and eject the casting from the mold as early as possible to minimize restraint on coating.

[FIGURE 7 OMITTED]

Driving Forward?

Since this first set of investigations, several key factors have been discovered for consideration in future casting trials. Because hot tearing continued to occur despite several grain refinement efforts, additional studies are examining different high-temperature alloys to determine which are best suited for permanent molds. Magnesium alloys designed for sand casting have caused some difficulty because they are not designed for casting in a metal die.

Development of a non-reactive mold coating has been a technical barrier that has inhibited the permanent mold casting of magnesium. Investigations with the mold coating formulas, however, have been finalized. The use of a fluorinated compound has proven its ability to reduce risk of reactions between the molten magnesium and the mold coating and provide suitable mold coating life.

The project also addressed melt-protection strategies. The cover gas used in this project was a new formula that provided an environmentally safe alternative to the traditional use of SF6 as a cover gas.

With the implementation of the new coating and cover gas, the first casting trials proved LPPM magnesium casting can fill the geometries of intricate permanent mold dies. Although magnesium LPPM casting holds potential in the future, further studies are examining methods to attain the maximum properties, and investigations in the project's second phase will continue to study the relation of pressure to fluidity in magnesium alloys.

This article was adapted from a paper (05-217) presented at CastExpo '05.

For More Information

"Shrinkage Behavior of AM50, AM60 and AZ91D," J.F. Wallace, D. Schwam, D. Druyor, Y. Zhu, American Foundry Society Transactions (05-028).

////

D. Weiss, Eck Industries, Manitowoc, Wisconsin

S. T. Robison, AFS, Schaumburg, Illinois

David Weiss is the vice president of engineering at Eck industries inc., Manitowoc, Wis.

Steve Robison is the senior technical director for the American Foundry Society, Schaumburg, ill.

Table 1. AM50 Room Temperature Mechanical
Properties Used to Design the Magnesium Cradle

Typical Properties                   Value

Ultimate Tensile Strength (MPa)       210
0.2% Yield Strength (MPa)             125
Elastic Modulus (GPa)                  45

Table courtesy of Dead Sea Magnesium Ltd.

Table 2. Mechanical Properties of MRI 202S-T6

Tensile Yield Strength (MPa)
  * At 68F (20C)                      150
  * At 302F (150C)                    145

Ultimate Yield Strength (MPa)
  * At 68F (20C)                      250
  * At 302F (150C)                    220

Elongation (%)
  * At 68F (20C)                        7
  * At 302E (150C)                     15

Compression Yield Strength (MPa)
  * At 68F (20C)                      145
  * At 302F (150C)                    140

Fatigue Strength (MPa)
  * At 68F (20C)                       85

Table courtesy of Dead Sea Magnesium Ltd.

Table 3. Chemical Compositions of Ingot and Magnesium
Cradle Samples

              Zn       Zr        Nd        Y       Ca

Ingot        0.32     0.46      3.1      0.28     0.04
Sample 1     0.31     0.20      2.6      0.20     0.01
Sample 2     0.36     0.17      3.0      0.20     0.01

              Fe       Ni        Cr       Si

Ingot       0.0028    0.001    0.0009    0.008
Sample 1    0.0005    0.001    0.0005    0.006
Sample 2    0.0005    0.001    0.0006    0.004

Table courtesy of Dead Sea Magnesium Ltd.