Variables affecting aluminum casting shakeout of coldbox cores.

Understanding the process variables in coldbox coremaking can help reduce aluminum shakeout problems for improved productivity.

Aluminum foundries face special problems for core removal that are not experienced by their ferrous counterparts. Aluminum's lower pouring temperatures result in lower core sand temperatures and less resin binder breakdown from thermal decomposition. Because of this reduced breakdown, the cores retain higher strength after casting and can be difficult to remove with mechanical vibration at shakeout. Additional time and/or labor may be needed to completely remove cores from narrow passages, increasing casting costs. Further, thin-wall castings sensitive to denting or deformation during shakeout and high sand-to-metal ratios can be particularly troublesome.

Earlier foundry testing indicated that the shakeout characteristics of resinbonded cores in aluminum castings were related to the mechanical strength of the core. As tensile strength increased, shakeout properties decreased. One common approach to improving shakeout was to reduce core resin levels as much as possible. While doing so reduced core strength and other mechanical properties, it improved shakeout characteristics, and increased core breakage, warpage and other coremaking problems. There also was a practical limit - around 0.7% resin for silica sand and as low as 0.5% for zircon sand - where lower resin volumes were not sufficient to uniformly coat the sand grains and provide consistent bonding. Problems with metering equipment and calibration of sand mixers also made it difficult to maintain resin levels near the lower limits.

As with most foundry problems, core removal from aluminum castings is dependent on a large number of variables. No "magic bullet" exists that can solve shakeout problems without presenting a new array of other problems. The variables that affect shakeout also tend to affect other coremaking or foundry processes. Attempting to optimize shakeout characteristics can result in a series of tradeoffs in which improved shakeout may result in core breakage, casting defects, loss of productivity or other problems. To intelligently make changes to their operations, foundrymen must understand the variables that affect shakeout, the relative strengths of the variables and their effects on other important properties. This article focuses on the phenolic urethane coldbox (PUCB) core process, but many of the findings apply to other types of core binder systems.

Coremaking & Core Properties

To understand how shakeout or core breakdown occurs, it is necessary to understand how cores are made and how their strength and thermal characteristics develop. Phenolic urethane coldbox cores typically use a mixture of 0.5-2.0% resin with sand. Since sand makes up more than 98% of the core, the sand composition, size and shape strongly influence core properties. Silica sand is the most widely used, but zircon, chromite and other specialty sands also are extensively used. The chemical composition of the sand typically controls bulk characteristics like density, heat capacity, heat transfer and thermal expansion.

Phenolic urethane coldbox resins are based on a three-part system: phenolic resin, polymeric isocyanate and vaporized amine catalyst. The core sand is coated with both the phenolic and polymeric isocyanate resins, and the mixture is blown into a corebox with dry air. The core is then hardened by passing the vaporized amine catalyst gas through the core to harden the resin. The resin on the surface of the sand grains forms resin "bridges" between the individual sand grains.

Figure 1 shows a scanning electron microscope (SEM) photo of these resin bridges. The strength of the core is determined by the number of these resin bridges, their cohesive strength (resin to resin), and their adhesive strength (resin to sand). Resin percentage and resin formulation determine these factors for any given sand. Exposure to humidity also affects the strength of the resin bridges. Moisture can be absorbed by the polymeric isocyanate portion of the resin, which causes the resin to lose strength and the tensile strength to decline.

Core strength and mechanical properties are determined by the combination of sand and resin. In general, strength is directly proportional to resin percentage within a given range since this determines the number of resin bridges. Sand shape and distribution also are critical since they determine the number of point-to-point contacts between sand grains, with a multi-screen, rounded sand generally giving the highest strength. Resin formulation determines both the adhesive and cohesive properties of the sand/resin bond. The base resin can be modified to change the strength and thermal characteristics of the cured resin by substituting some of the base resin with other reactive components. These may thermally degrade at lower temperatures and speed resin breakdown to facilitate core removal. Solvent packages formulated to embrittle the resin bonds also can improve mechanical shakeout characteristics.

How Breakdown Occurs

For core breakdown at shakeout to occur, there must be a failure of the resin bridges in the core sand. This can be either a cohesive failure (the resin bridge breaks) or an adhesive failure (the resin bridge becomes detached from the sand grain). The strength of the adhesive and cohesive bonds will determine the mechanical properties of the core, that may not simply be related to the tensile strength. While tensile strength is an important characteristic, core breakdown during mechanical shakeout appears to have two phases. As the casting is vibrated, there is an initial breaking or shattering of the core into smaller fragments. This is related to the good mechanical trans- mission of energy from the casting to ' the core while the core is still tightly in contact with the walls of the casting. However, after the first shattering occurs, the mechanical coupling of the core to the casting wall is lost, and the efficiency of the transmission of energy into the core decreases rapidly. Cores that break down easily seem to shatter into relatively small fragments within the first few seconds of shakeout.

After the first shattering of the core, the breakdown mechanism seems to be more abrasion than additional breakage of the fragments. The remainder of the core consists of a collection of fragments within the cavity that are too large simply to fall out of the opening to the core cavity. They continue to rub against each other, removing sand grain by grain, until the core fragments become small enough to fall out.

Tests were conducted to examine the mechanism of core breakdown in mechanical shakeout. Aluminum castings were produced with PUCB cores and allowed to cool to room temperature. One face of the casting was carefully cut away, and a clear plastic plate was clamped against the casting face and core. The casting then was placed in the vibratory shakeout and impacted. Within the first seconds, a thin layer of loose sand developed against the surface of the plastic and rapidly ran out through the opening in the casting. This "decoupled" the core from the walls of the casting and reduced energy transmission from the casting into the core. The core had shattered into numerous chunks, which vibrated against one another, but did not further fracture into smaller chunks. Rather, sand was abraded away by the rubbing action. The abrasion continued until all the fragments had worn down and fallen from the casting. Figure 2 shows the core shortly after shakeout was begun, including the initial fragmentation of the core.

Figure 3 shows typical weight loss curves for test castings. The first rapid weight loss is believed to be related to the initial fracturing while the slower, flat part of the curves is related to abrasion.

The thermal and mechanical breakdown of a core is related to the breakdown of the resin bridges. Thermal breakdown can occur by either pyrolysis (thermally induced chemical changes) or oxidation (reaction of resin with oxygen to give C[O.sub.2], water and other oxides).

In core applications, there may be limited oxygen available because of small, enclosed passageways and limited oxygen infiltration. In these circumstances, pyrolysis may he the predominant breakdown mechanism. The degree of thermal breakdown of the core resin is related to the resin formulation, temperature, time exposure and oxygen availability. A series of tests were conducted to quantify typical time/temperature exposure profiles for aluminum casting. The standard test casting used was a shakeout tree. The casting has a wall thickness of approximately 0.25 in. and was produced with a 1-in.thick core. A large boss was designed on one side to facilitate clamping in the vibratory hammer during shakeout. The sand-to-metal ratio for the casting is approximately 1:1.4 (silica sand). Thermocouples were placed at the core/metal interface, the center of the core was placed beneath the heavy boss and the corner of the core was placed farthest from the casting wall. Temperatures at the interface rose rapidly but also cooled rapidly. At the center and corner, temperatures rose to a maximum and appeared to meet the interface temperatures. Temperature at all three points then slowly dropped along a smooth curve.

Temperature profiles also were conducted with different core geometries and materials. First, the core was cut to 0.5 in. using the same mold cavity. This effectively doubled the casting wall and reduced the sand-to-metal ratio to 1:2.8. Next the core thickness was doubled while the wall remained the same. This increased the sand-to-metal ratio to 1.4:1. Cores made from silica, lake and zircon sand were evaluated in the standard mold. Finally, the temperatures in a standard core were measured in a cast iron semi-permanent mold. As expected, changing the sand-to-metal ratio of the castings produced significantly different core temperatures. Also, the profiles for silica vs. lake sand were nearly the same, as expected. However, the zircon core showed nearly the same temperatures. While the heat capacity is higher for the zircon sand, the conductivity also is higher. It evidently absorbed more heat, but transferred it faster into the core, so that the overall temperature pattern was nearly the same as that for silica sand.

Practical Uses for Foundries

Data from foundry tests can be used as a tool to assist foundry personnel in process changes. Results show both the effects and relative strength of a number of possible control factors for improving shakeout characteristics and their effects on other important coremaking and casting properties. By using all of the available tools, it may be possible to optimize shakeout while minimizing any detrimental effects. Through the data, it is possible to: select the factor levels to optimize a desired characteristic, understand the side effects on other properties, determine which factors have the strongest impact and predict responses for a given set of factor levels. This can remove much of the guesswork from a very complex problem.

There are economic and practical constraints, though, that limit changes that can be made. For instance, a foundry's coremaking process is generally built around a particular sand. It is not possible to change the sand to improve shakeout without disrupting an array of other processes. While the effects on tensile strength were examined, other properties and costs were not. Pouring temperature changes also may have other unexpected effects. The lower temperature could cause misruns while the higher temperature could cause excessive dross and gas pickup. Because different sands, casting types and processes are used from one foundry to the next, the data results might explain significant differences in shakeout. However, within a given foundry, the most important factors continue to be percentage resin, type of resin, additives and temperatures.

Another important part of the data is the identification of factors that do not affect shakeout. For instance, the pressure setting on the shakeout to control frequency and amplitude had little or no effect on total removal time. This might go against popular opinion, as the typical belief is that increasing shakeout intensity will speed shakeout. Excess intensity could cause unnecessary casting damage with little improvement in shakeout. Other factors having little effect are pouring temperature, time from pouring to shakeout and core age. It should be possible to vary these factors to facilitate operations without significantly affecting shakeout.

The mechanical shakeout properties of coldbox cores are affected by a large number of variables. Core composition (sand, resin percentage, resin type and additives) has the strongest effect on shakeout, but foundry process variables also have a strong influence. The interaction of these variables determines the actual results. An understanding of the relative strengths of effects allows the foundryman to balance the variables to optimize not only shakeout but other important casting, coremaking and foundry characteristics.

In addition to the traditional approaches to improving shakeout, like reducing resin percentage, recent advances in binder technology and sand additives provide additional tools for optimization. Often these changes can be made without sacrificing other important properties like core tensile strength or casting surface finish. It also may be easier to change these factors than sand type, pouring temperature or other variables that are key to the foundry process.

This article was adapted from a paper (99-115) presented at the 1999 AFS Casting Congress and is available from AFS Publications at 800/537-4237.

RELATED ARTICLE: Foundry Reveals Variable Effects on Shakeout

A design of experiments was developed to quantify the relative effects of the variables shown to affect shakeout performance. This used several levels for the important factors and also examined sand additives and a wide pouring temperature range. Factors included in the design of experiments were pouring temperature, shakeout temperature, shakeout control pressure, additive type, resin type, resin percentage and sand type. Selected pouring temperatures were 1350F (732c) and 1500F (815C) while the shakeout temperature was set at 350F (176C). This was believed to represent more accurately casting temperature immediately after removal from the mold. Two types of antiveining additives were included: Additive A, a proprietary iron oxide/carbohydrate compound and Additive B, a proprietary engineered sand additive. The additive levels were set at the "normal" levels (2% and 6%, respectively) and compared to an additive-free system. Three different resins were included: a resin formulated for iron, a new technology resin formulated especially for aluminum and a resin containing a high level of plasticizers. The sands included a silica, a lake and a Florida zircon sand. Both the silica and lake sand were of similar GFN, 50 and 52 respectively, but the zircon had a GFN of 110. The testing also included the effects of the variables on core tensile strength and on surface finish of the casting (Table 1).

Table 1. Design of Experiments Layout

Factors                   Level 1   Level 2   Level 3

A. Pouring Temperature      1350F     1500F        NA
B. Shakeout Temperature        RT      200F      350F
C. Shakeout Pressure       50 psi    40 psi    30 psi
D. Additive Type                A      none         B
E. Resin Type                  Al      Iron      Plaz
F. Resin Percentage          0.7%      0.9%      1.1%
G. Sand Type               Zircon      Lake    Silica

Trial #        A       B      C        D      E        F          G

1           1350      RT     50        A     Al     0.7%     Zircon
2           1350      RT     40     none   iRON     0.9%       Lake
3           1350      RT     30        B   Plaz     1.1%     Silica
4           1350     200     50        A   Iron     0.9%     Silica
5            130     200     40     none   Plaz     1.1%     Zircon
6           1350     200     30        B     Al     0.7%       Lake
7           1350     350     50     none     Al     1.1%       Lake
8           1350     350     40        B   Iron     0.7%     Silica
9           1350     350     30        A   Plaz     0.9%     Zircon
10          1500      RT     50        B   Plaz     0.9%       Lake
11          1500      RT     40        A     Al     1.1%     Silica
12          1500      RT     30     none   Iron     0.7%     Zircon
13          1500     200     50     none   Plaz     0.7%     Silica
14          1500     200     40        B     Al     0.9%     Zircon
15          1500     200     30        A   Iron     1.1%       Lake
16          1500     350     50        B   Iron     1.1%     Zircon
17          1500     350     40        A   Plaz     0.7%       Lake
18          1500     350     30     none     Al     0.9%     Silica

Sand type and resin percentage once again had the greatest overall effects. Immediate and 24-hr tensile strength, percentage removal at 15 and 30 sec, total removal time and surface finish were all significantly impacted by the sand type. The lake sand gave the lowest tensile strengths, the best shakeout and the worst surface finish. The silica sand gave the highest tensiles, reduced shakeout and moderate surface finish. The zircon sand produced somewhat reduced tensile strength, shakeout similar to silica and the best surface finish. These effects may have been partially the result of the difference in GFN rather than the chemical composition of the sand. The resin percentage had the expected effect of increasing tensile strength and reducing shakeout. Tensile strength was directly proportional to resin level, and shakeout characteristics were inversely proportional. As resin percentage increased, tensile strength increased and shakeout characteristics declined. There also was a noticeable impact on surface finish, with the highest resin level producing the best surface.

Other factors had more mixed results. Both the iron and aluminum resins had nearly equal and improved tensile strength compared to the plasticized resin. The new aluminum resin system significantly outperformed the iron and plasticized systems for shakeout with greater percentage removal and lower shakeout times. The iron system was good for tensile strength, but poorer for shakeout and surface finish. The additives had some interesting effects. The iron oxide additive at 1.5% drastically reduced tensile strengths, had relatively little effect on the percentage removal at 15 and 30 sec, but improved total removal times. The proprietary additive at had relatively little effect on strength but improved both shakeout measurements. It was unexpected to see the improvement in shakeout without an accompanying loss in core strength. Neither additive affected the surface finish.

The foundry process variables generally had smaller effects than the core variables. The hammer control pressure had a slight positive effect on shakeout, but not to the extent that was expected. Pouring temperature did improve shakeout somewhat, but not to levels that are statistically significant.

There was a clear relationship between resin percentage, tensile strength and shakeout characteristics. However, many of the other factors do not exhibit the same influence. For instance, Additive A resulted in lower tensile strength but without a corresponding improvement in shakeout. Additive B had only a minimal effect on strength but improved shakeout. Sand types also affected strength and shakeout in different ways. It is apparent that these nonlinear factors are introducing enough variation to the systems that the known relationship of tensile strength to shakeout is masked. It also means that there may be ways of improving shakeout without significantly sacrificing tensile or other important characteristics.