M/G vs. Z: comparing refractory coatings on shell sand systems.

Inside This Story

* Directional heating of shell sand composites can burden metalcasters by forcing dimensional changes.

* Refractory coatings help assist in reducing such thermal expansion defects.

* This article details the thermal distortion, change in mass and impact strength properties of refractory-coated shell sand systems at aluminum and cast iron fill temperatures.

Every year, the metalcasting industry spends millions of dollars on refractory coatings for chemically bonded sand systems. Such investments are made because refractory coatings aid in surface finish improvements, reduce thermal expansion defects, such as veining, and unbonded sand defects, such as erosion. This is essential to maintain the integrity of mold and core shapes during pouring.

Because directional heating of sand composites (mold and core media) generates anisotropic thermal gradients in the materials, when a shaped sand composite comes in contact with molten metal, the heat transferred causes thermo-chemical reactions that force dimensional changes in the composite. At any given temperature these dimensional changes or thermal distortions are attributable to simultaneous changes in both the sand and the binder.

Thus, refractory coatings help prevent these sand deformations from occurring. However, when placed at the mold-metal interface, not all refractories act similarly. Investigations, there lore, were conducted to discover differences between refractory coating types applied to shell molds and cores at aluminum and cast iron fill temperatures. Refractory-coated and uncoated shell sand systems were evaluated for thermal distortion, mass change percentage and impact strength.

This article examines the effects of refractory coatings on shell sand under particular conditions and how well they assist in maintaining the sand's properties.

Make Preparations

In this investigation, shell specimens were made from a typical, acid-catalyzed, novolac-type resin, a type of low molecular weight polymers with a well-defined range of melt points and viscosities. This resin is generally regarded as thermally stable.

Basic properties for the sand are as follows. This resin is stable up to 518F (270C), after which thermal degradation begins. It also is reported that the thermo-chemical degradation of novolacs in shell sand is a thermal-oxidative process regardless of whether the pyrolysis reaction occurred in an oxidative or inert atmosphere. Between 572F (300C) and 1,112F (600C), the rate of degradation will increase with the evolution of gaseous components. Beyond 1,112F (600C), breakdown of the phenolic structure is evident.

Using washed and dried round grain silica sand (Table 1), the shell sand disc specimens were prepared employing common coremaking techniques.

The specimens were then refractory-coated with either a mica/graphite (M/G) or zircon (Z) refractory coating. Discs were hand-dipped into the refractory coating, placed horizontally (refractory-coated surface up) in a forced air oven and dried at 125F (52C) for 1 hr. The refractory coating dry deposits of the sample discs were determined using a dial thickness gauge that was zeroed on the refractory-coated surface, which is typically 0.006 in. thick, and a small section of the coating was carefully removed. The difference between the original surface and the substrate was then measured to the nearest 0.003 in.

Both M/G and Z specimens were tested at both aluminum (M/[G.sub.AL], [Z.sub.AL]) and cast iron (M/G.sub.FE], [Z.sub.FE]) temperatures. The M/G coating has been known to be thermally insulating, while the Z coating has been touted as being thermally conductive. There also were uncoated (C) and refractory-coated (CM/ G and CZ) control specimens. In addition, uncoated control specimens were tested at both aluminum ([C.sub.AL]) and cast iron ([C.sub.FE]) temperatures.

Using the TDT

A predetermined load on the thermal distortion tester (TDT) can be adjusted to simulate a specified force of molten metal acting on a mold (head pressure). However, for this investigation the load was kept constant during TDT at 0.73 lbs. (332 g) for 3 min. The test load used represents a 6-in. head height for cast iron and a 15-in. head height for aluminum.

The thermal distortion curves (TDC) for all systems tested showed undulations that indicate thermo-mechanical and thermo chemical changes in the binder system at elevated temperatures (Fig. 1). All curves had an initial expansion (upward movement of TDC) before plastic deformation (downward movement of TDC). Note that a TDC depicts an average for 10 specimens tested (each measured 50mm in diameter and 8mm thick).

[FIGURE 1 OMITTED]

More specifically, for systems M/[G.sub.AL] and [Z.sub.AL] tested at aluminum temperature 1,400F (760C), there was continuous expansion for the first 65 sec. followed by plastic deformation for the next 60 sec. before the curves leveled out. The uncoated control system, [C.sub.AL], had a similar TDC, but this system showed further plastic deformation to ward the end of the test. M/[G.sub.AL] and [Z.sub.AL] tested at aluminum temperature showed no significant difference in thermal distortion. When comparing those refractory-coated systems to the uncoated control systems ([C.sub.AL]) tested at the same temperature, the [C.sub.AL] specimens had greater thermal distortion than the refractory-coated systems.

Furthermore, systems tested at cast iron temperature 2,350F (1,288C) showed expansion for the first 70 sec. There were two expansion phases (see curves [C.sub.FE], M/[G.sub.FE], and [Z.sub.FE] in Fig. 1) followed by plastic deformation over the rest of the test. The TDC for refractory-coated specimens temporarily leveled out after an additional 50 sec. but continued with plastic deformation for the remainder of the test period. After the initial expansion, the [C.sub.FE] system exhibited only plastic deformation.

Like the aluminum tests, there was no significant difference in thermal distortion between the refractory-coated systems tested at cast iron temperature (M/[G.sub.FE] and [Z.sub.FE]). However, the [C.sub.FE] showed the greatest thermal distortion range and was significantly different from all other systems tested (Table 2). Thus, the refractory coatings helped reduce such distortion. Yet, test temperatures caused significant difference in thermal distortion between the cast iron test temperature and the aluminum test temperature.

Even though the samples were cured prior to testing, some residual reactivity was seen by the undulation of the TDC. If the application of heat caused further cross-linking reaction in the specimen, it would generate gases as novolac-curing reactions typically do, causing some distortion in the specimen. Due to the inherent thermal stability of the resin, it was expected that there should have been minimal increases in the region of thermal stability and more in the ranges where decomposition occurred.

After thermal exposure, the test specimens remained intact allowing for the investigation of additional data, including impact strength, visual observation for cracks, mass change measurement that relates to pyrolysis of binder bridges and the amount of loose, unbonded sand generated at the mold metal interface.

Finding Mass Changes

Each specimen was weighed before and after TDT to calculate the percent change of mass. Next, the specimens were visually examined for signs of thermally induced cracking of the surface, loss of sand where contact was made with the hot surface and any other discolorations. If the core/mold media broke down, this may have been indicative of the tendency to produce cuts and washes, erosion/inclusion-type detects. In interpreting this data, it was critical to identify the components causing the change in mass. All percent change in mass values represented the percentage of weight lost.

The uncoated control systems, [C.sub.AL] and [C.sub.FE], had significant mass losses when compared to the refractory-coated systems (Table 2), (Fig. 2). Further, [C.sub.AL] had less loss due to lower thermal stress. Mass change was not significant among the refractory-coated systems regardless of refractory coating type, which would indicate that refractory coatings offered some thermal resistance to mass change.

Observations from the heat-affected zone on the surface of tested specimens revealed that the uncoated control systems had visible sand losses and crack propagation. For all uncoated control systems tested, the hot surface/specimen interface showed black to white to brown discolorations due to various levels of binder degradation (Fig. 3). In addition, sand binder losses were evident at the hot surface/specimen interface where binder bridges pyrolyzed and sand grains broke loose; this was apparent in [C.sub.AL] and especially [C.sub.FE].

[FIGURE 3 OMITTED]

Expansion cracks were macroscopically evident on the uncoated control systems and to a much lesser extent on refractory-coated systems. However, the crack propagation was most pronounced in [C.sub.FE]. For the refractory-coated specimens, there were only faint cracks on M/G and Z coatings regardless of test temperature. This demonstrated that refractory coatings can help in the prevention of expansion-(veining) and erosion(inclusion) type defects.

Strong on Impact

The overall durability of the specimens was then investigated using an impact testing machine to measure the strength of the sand specimens before and after the TDT. Impact strengths before TDT related to handling of the core/mold material after core/mold production and prior to pouring. The impact strengths after TDT testing relate to shakeout/ collapsibility characteristics.

The disc-shaped specimen was supported on its edge on a specimen holder on the impact testing machine. It was then subjected to impact energy by drop ping a uniform load with a 2 mm thick rounded edge blade across its diameter.

The uncoated control systems had no significant difference in strength when compared to the refractory coated systems (Table 2), (Fig. 3). Further, the impact strength between the two uncoated control systems was not significantly different, nor was it between the refractory-coated systems regardless of refractory coating type. The same can be said for the control specimens. This indicated that the addition of refractory coating offered negligible impact resistance and that strength was derived primarily from the shell sand system.

Keep the Coat On

Overall, with the exception of percent change in mass, there was no difference between the M/G and Z coatings. Additionally, there were only faint cracks on M/G and Z coatings regardless of test temperature. The faint cracks were undoubtedly the result of expansion/contraction differentials between the refractory coating and sand composite. The positive effects of applying refractory coatings to the shell system were shown in the reduction of mass loss and surface cracking. Higher mass loss and surface cracking were prominent on the uncoated control systems. For these uncoated systems, cast iron temperature caused greater mass loss and expansion cracks compared to that found at aluminum temperature.

Although these investigations demonstrated the capabilities of laboratory test methods, the effects of cast metal chemistry on refractory coatings and shell systems were not considered. There also are numerous other chemical binder systems from which additional data could be gathered to learn more about thermal properties. Further work could be perforated at different loads and different temperatures simulating other alloys and pressures representative of larger or smaller castings.

However, these test deserve actual casting trials to fulfill their validity.

This article was adapted from a paper (04-122) presented at the 2004 Metalcasting Congress.

For More Information

"Baume: Complete Coating Control?," AFS Molding Div. Mold-Metal Interface Reactions Committee (4-F), MODERN CASTING, October 2003, p. 28-30

About the Authors

Sam N. Ramrattan is the Foundry Educational Foundation Key Professor at Western Michigan Univ., Kalamazoo, Mich. O. Brian Guyer is a Lab Manager at HA International, Toledo, Ohio. Kim M. Fisher is a Lab Supervisor at HA International. Suet Fong Cheah is a graduate student at Western Michigan Univ.

Table 1. Properties of the Shell Resin Investigated

Source   AFS/gfn   Shape     Screens    % Resin

 III.      90      Round        3          3

Roundness/Sphericity     pH      Acid demand
     (Krumbein)                     (pH-7)

      0.8/0.8          Neutral        <1

          FE        AL

Z        0.3%       0.2%
MG       0.4%       0.2%
C        1.2%       0.5%

Table 2. Physical and Thermo-Mechanical Properties of the
Refractory-Coated and Uncoated Shell Sand Specimens

   Systems      Test Temp   Thermal Distortion Range (in.)
     (F)                           '@ 332g for 3 min.

      C            75                    N/A
    CM/G                                 N/A
     CZ                                  N/A
 [C.sub.AL]       1,400             0.003 (0.0004)
M/[G.sub.AL]                        0.003 (0.0003)
 [Z.sub.AL]                         0.003 (0.0005)
  C.sub.FE]       2,350             0.005 (0.0009)
M/[G.sub.FE]                        0.004 (0.0001)
 [Z.sub.FE]                         0.004 (0.0002)

   Systems       % Change     Impact Strength
     (F)          in Mass        (in.-lbs)

      C             N/A            5 (2)
    CM/G            N/A            6 (3)
     CZ             N/A            6 (2)
 [C.sub.AL]     0.45 (0.21)        5 (2)
M/[G.sub.AL]    0.24 (0.21)        5 (2)
 [Z.sub.AL]     0.2 (0.21)         5 (2)
  C.sub.FE]     1.28 (0.17)        5 (2)
M/[G.sub.FE]    0.4 (0.19)         5 (2)
 [Z.sub.FE]     0.36 (0.13)        6 (2)

Note: Values Represent Averages and the Standard Deviations
Are Shown in Parentheses)