Reclamation and Reuse of Shell Ceramics in Investment Casting

The reuse of ceramics, specifically stucco and flour in investing casting, would reduce expenses in the purchase of raw materials and in the disposal of the spent shell. Some companies sell their spent shell material, but typically the only savings here are in the disposal side. Thus, reusing the material as a substitute for virgin materials would incur the greatest economic benefit.
Throughout the world zero waste or waste reduction regulations are increasingly in place. For many investment casters, the disposal of spent ceramic shell is one of their largest sources of waste. Eliminating or significantly reducing this waste would help meet these regulations.

With these points in mind, the reclamation and reuse of shell ceramics was investigated. This is AFS-sponsored research that is ongoing, and more work is planned for the future.

Bars of ceramic (6×3/4×3/8 inches [150x18x9 mm]) were mixed using recipes based on fused silica and alumino silicates common to investment casting. The bars were tested for green strength and hot strength at 1400F and 2012F (760C and 1100C). All heated specimens were then crushed manually and then broken down in a ball mill. The sieve size distribution of the milled material was compared to that of the virgin material. Then, the reclaimed material was remixed and retested, and the results were compared to the virgin mixes. The more promising mixes were run for several cycles.

The alumino-silicates were successfully reclaimed at the highest temperature while the fused silica materials were not successful at any temperature except for partial reclamation of flour at 2012F (760C) aluminum processing temperatures.

The first breakdown attempt was to ball mill the larger parts (~0.98 x 1.96 x 0.39 in [~25 mm x 50 mm x 10 mm]) directly. This was unsuccessful. So, in all subsequent breakdown processes the spent ceramic was reduced by manually hammering until the largest size was approximately 0.3 in. (8 mm). Then they were tumble-ball milled for five hours.

For comparison purposes, virgin flour and stucco were sieved. The results are shown in Figure 1. Comparisons of reclaimed particles were made against this virgin particle blend. In Figure 2, the distribution of one reclaimed mix is compared with the original particle distribution. One of the mixes was ball milled for five hours, but a sieving analysis was performed each hour to see the progressive breakdown of the ceramic and potential degradation of the ceramic particles. The results of this experiment can be seen in Figure 3. The ceramic breaks down slowly and there is little degradation of the original particles. Particle sizes of mesh 50, 70 and 100 were essentially zero in the original materials. The particles in these sieves were largely degraded stucco particles.

The breakdown of the shell materials can be done in such a way that the particle size distribution is similar to the original distribution. However, care must be taken as it is possible to overgrind the particles. In Figure 4, most batches that were broken down track the original blend (thick black line). These batches are represented by solid lines. The batches represented by the dashed lines were over-milled and resulted in significant degradation of the stucco particles. This occurred because the milling process was a batch process and these mixes had smaller amounts of material in the ball mill. Thus, while it is possible to recover the original particles of the shell material, care must be taken in the process.

At the end of the milling process, very few compound grains were within the stucco size range or smaller. Whether this is a material characteristic (that is, it tends to break apart easily) or it is in the nature of the specific secondary breakdown process (batch ball milling) is unclear.

The strength results for each batch are summarized in Table 1. These batches have been classified into V-100% virgin flour and stucco and R-Reclaim where the amount of reclamation was the maximum amount possible with the available materials. As can be seen in several batches, it was necessary to introduce virgin material to complete the mix. It is anticipated that due to losses in the breakdown process this will be necessary on a regular basis.

In Figure 5 the strength values are normalized by the type of mix (virgin or reclaim). The virgin mixes are considered to be the standard. Thus, the reclaimed mixes are compared to these. The values in this figure indicate that the average green strength from the reclaim mixes dipped by 10%. The hot strength greatly increased over 40%. One likely explanation is the increase of fines in the flour after milling. This would dilute the effect of the binder in the green strength but provide more sintering points for the hot strength.

In summary, alumino-silicates (mullite) of 60% alumina content may be feasible to reclaim. The particles, stucco and flour, maintain geometric integrity during breakdown and the reclaim mixes have comparable or superior properties to the virgin materials.

As with the mullite, the fused silica was heated to 2,012F (1,100C) and broken down with a hammer. Then it was milled in the ball mill and the sieve size of the progressive breakdown was taken each hour. The stucco particles were not preserved, as compared to the distribution of the original blend (Fig. 6.). As the larger particles (larger than mesh 6) are broken down they do not become the size of stucco particles, but rather appear to shatter into smaller particles. Based on this, a batch of virgin fused silica stucco was milled following the same procedures to see what degradation occurred. The results in Figure 7 show the degradation of particles occurs much more slowly than that seen in Figure 6.

After seeing these results, specimens then were produced and heated to aluminum processing temperatures (1,400F/760C) to avoid the fast cristobalite formation temperature range (above 1,652F/900C). The stucco degradation also occurs at these processing temperatures, not as aggressively as at the higher temperatures, but sufficiently to significantly change the particle size distribution of the stucco fraction if reintroduced to the process.

Fused Silica-Strength
The strength results for each batch are summarized in Table 2. These batches have been classified into V-100% virgin flour and stucco and RF-Reclaim 100% flour, and RF50-Reclaim 50% flour. No stucco reclamation was possible. Only materials processed at a given temperature were used in the reclaimed mixes destined for those same temperatures.

The hot strengths decrease significantly for the 2,012F (1,100C) mixes at any level of flour reclamation. The green strength also sees a decreasing pattern, but there is no statistical difference between virgin and 50% reclaim, while there is a statistical difference between 50% and 100% reclaim. As the amount of reclaimed flour is increased the result worsens.

Where material processed at 1,400F (760C) is reclaimed, the results are more promising. The green Modulus of Rupture remains stable and there is no statistical difference between reclamation quantities. The hot MOR swings up and then drops as the reclaimed flour content is increased. Why the swing in hot strength occurs is unclear. The 50% level of reclamation was not statistically different from either the 100% virgin or 100% reclaim formulations. However, there was a statistical difference between the 100% virgin and the 100% reclaimed hot strengths. These results are much better than the results for the 2,012F (1,100C).

In summary, fused silica systems have limited reclaiming potential. First, the stucco is not reclaimable, and at either steel or copper base processing temperatures the flour is not reclaimable. At aluminum processing temperatures, it may be possible to reclaim flour on a limited basis.

Industrial Equipment Appropriate for Use in Industrial Scale Reclamation
In the study, several types of equipment were investigated and conversations with representatives of this equipment were held to identify equipment that could be suitable for investment casting reclamation. Generally, most suppliers would approach the problem by crushing the shell material to about ½-in. (12-mm) particles and then milling using self-abrasion or media such as ball milling. To minimize iron contamination, two methods were mentioned. The first is to use ceramic-lined and grinding media. The second consisted of using magnets to remove the iron. There are important considerations for foundries that use water blasting, since the spent ceramic is wet and in smaller pieces, for this application the ceramic would need to be dried but would likely not need a pre-crushing step. For segregation by sieve size, generally mechanical sieves would be used for stucco sizes. For flour, both mechanical sieves and air separators were recommended.

Multiple companies have crushing, milling, and size separation experience.

An alternate method reviewed would consist of using acoustic energy to break the ceramic into particles. In this system, it may be possible to place the cast tree in a tank prior to shell removal, apply acoustic energy and in the process, shakeout the castings, break the ceramic and possibly segregate the ceramic by grain size. There is the possibility that this may also help with core removal. Only one supplier of this technology was found, and testing would be necessary.

Quality Control Photographs of Stucco and Flour
The virgin and reclaimed stucco and flour were separated by sieve size and photographed with the goal of developing quality control reference standards. Then these standards could be used with visual examination with the use of 10X to 100X pocket microscopes. The image collection was unsuccessful because the quality of the images obtained was too poor to use in this capacity. However, the surface and shape of the grains in stucco could be assessed with optical microscopes by changing the focus, but because of the depth of field could not be effectively photographed. Thus, it could be possible to save reference samples to use as side by side comparison standards for quality control purposes. ■

This article is based on paper 17-104 that was presented at the 2017 Metalcasting Congress.