Increase Ceramic Shell Quality via the Infiltration Method.

PCC Structurals' method of ceramic shell production provides practical solutions to slurry aging, shell-metal interaction and limited shell mechanical properties.

The conventional investment casting ceramic shell process has enabled manufacturers to make useful shells for most applications. Nevertheless, this process has its shortcomings. Because of the very nature of the process, slurries go through aging, mechanical properties only can be designed to a limited degree and shell-metal reaction remains a concern.

This article provides a practical solution to the shortcomings of the conventional ceramic shell process. By integrating the infiltration method of ceramic shell production into shell building, a foundry can increase its shell process control and casting quality.

Problems with Tradition

Following is a look at the inherent problems in conventional shell production methods and some of the solutions used to overcome them.

Slurry Aging--In production, slurries age, lose their usefulness and must be disposed. There are at least two reasons why slurries age in a production environment. First, different components in a slurry may interact with each other and age over time. For example, it has been shown that in aqueous slurries, the soluble species originating from each component of a slurry may adsorb on the surface of other components. Surface adsorption may modify surface properties and cause aging over time. Similarly, binder particles may gradually agglomerate over a period of time if the repulsion energy between particles is low.

In slurries where silicon alkoxides are used as binder, an increase in pH (from the dissolution of components and/or water moisture adsorption from the environment) reduces the slurry lifetime. Aging behavior may differ depending on the slurry's amount of dissolution of components.

Besides aging due to component interactions, investment casting slurries age because inorganic binders such as colloidal silica are chemically reactive and permanently bond as the result of siloxane formation upon surface contact. While in the slurry, the possibility of direct surface-to-surface contact between colloidal silica particles is low due to the ionic cloud around each particle (diffused double layer). However, once these particles come to the slurry-air interface, their surfaces may contact and bonding can occur. Therefore, any stage of the process that promotes formation of a slurry-air interface could potentially cause permanent bond formation between colliding particles.

In some cases, attempts have been made to salvage aged slurries by recycling different components. However, it is difficult to separate inorganic binder from refractories in the slurry, and the process is costly, often with unsatisfactory results. On the other hand, organic components of an aged slurry can be pyrolyzed at relatively low temperatures and removed.

For the most part, because an inorganic binder is present in the slurry and liquid-air interface is continuously formed, slurry aging is inevitable in the conventional ceramic shell process.

Metal-Shell Reaction--Another problem with the conventional ceramic problem with the conventional ceramic shell process is its inability to produce silica-free shells. Silica binder, due to its low- and high-temperature binding efficiency, is a major contributor to the success of the ceramic shell process. However, silica is normally the lowest refractory component of high-temperature shells and its presence can cause a reaction with the metal, jeopardizing the quality of the cast part.

There have been many attempts to substitute non-silica inorganic binders in slurries. Many of these attempts, although successful in the laboratory, have proven impractical in production due primarily because of the rapid aging of non-silica binders. Another method in reducing the reaction between high-temperature alloys and the shell has been to reduce the amount of silica binder in the shell. Reduction of the amount of silica binder causes the loss of shell green strength (see next section). However, high-temperature strength may be compensated for by the addition of some sub-micron high refractory flours into the slurry.

Mechanical Properties--The third important deficiency of the conventional ceramic shell process is its limitation in the design of the mechanical properties of the shell. Without high enough mechanical properties of the shell, quality and price of a cast part can drastically suffer. This is especially true as the casting specifications become more stringent, and parts become larger and more complicated.

The first mechanical property of interest in an investment casting shell is green strength. Sufficient green strength is needed to allow construction of a shell around a wax pattern and to provide sufficient strength for a shell to survive cracking during wax removal. In an ideal autoclave during wax removal, tension exerted on the shell could be minimized. In reality, however, a shell may experience a large amount of tension during autoclave. If tension on a shell exceeds its strength, the shell may crack.

The general trend to increase the strength of a shell system is to increase the concentration of silica binder. After a critical binder concentration, however, shell strength decreases.

In production environments, green strength also will vary due to slurry aging and the adsorption of moisture. The magnitude of stresses exerted on a shell during the wax removal stage can exceed 500 psi. The maximum obtainable green strength in the conventional ceramic shell process is 500-1500 psi. However, actual green strength for shell in a production environment is below these values. Therefore, shell strength and exerted stress values may overlap. This overlap makes investment casting shells prone to cracking.

High-temperature mechanical stability is another property of interest in investment casting ceramic shells. Shell deformation at high temperatures will force casting rework and, in some cases, can scrap the part. Severe shell deformation also can cause shell failure and metal leakage.

The primary cause for creep in investment casting shells is the presence of a glassy phase in the microstructure. A typical microstructure of a shell consists of flour grains, amorphous silica binder, and some possible interfacial reaction phase between the flour and the binder. In order to reduce creep and increase the refractory strength of shells, high refractory polycrystalline flours are normally used. These flours are creep-resistant and do not demonstrate noticeable creep under pressures and temperatures used in investment casting. Silica binder, however, is an amorphous material. Uncrystallized silica binder will creep at high temperatures. One method to reduce creep in shells is to devitrify silica binder.

General practice in the investment casting industry is to prefire shells to 1832-2192F (1000-1200C) before pouring. In a typical ceramic shell, amorphous silica binder does not totally crystallize at this temperature range. As a result, investment casting shells are prone to high-temperature creep. One exception may be an all-silica shell system. In this case, a considerable degree of crystallization can be achieved during the prefiring cycle at around 2192F. Consequently, an all-silica shell system can be designed to be creep-resistant.

Shell Removal--In investment casting, it is highly desirable to reduce the strength of shells after the metal is poured so that the cleaning stage can be done with ease, however, it is possible only in exceptional cases. For example, if the shell material is made of mostly fused silica flour, a large fraction of the flour and the binder can be made to crystallize. As fused silica shell cools down below 112F (600C) crystallized silica goes through a martensitic phase transformation. The induced internal stress of phase transformation causes the shell to crack and fall off by itself.

For the majority of other shell systems, silica binder does not crystallize to a high degree. Ideally, in a shell system, it is desirable to reduce creep and at the same time reduce shell removal strength. In practice, however, the trend works in a reverse order. Based on what has been discussed, if the creep is minimized by maximizing the shell strength, then shell knockout becomes difficult. On the other hand, if one tries to make the shell knockout easier by reducing the shell strength, then creep goes up.

INFILTRATION METHOD

The infiltration method of investment casting ceramic shell production is a simple modification of conventional shell building. In this process, a binder system is infiltrated into the shell body at a specific stage during the investing process. Depending on the type of infiltrating binder system, different properties can be obtained.

In the infiltration process, slurries similar to conventional building are used, however, all or part of the inorganic binder may be eliminated. In order to construct a shell, the slurry must contain some type of binding agent in order to keep the slurry layer around the pattern wax upon drying. To eliminate the inorganic binder from the slurry, it is necessary to incorporate some type of organic binder. Once a layer is dried, the shell layer can be infiltrated with an additional binder. This binder may be a silica binder or any different organic or inorganic binder. The shell may be soaked in a binder container or the binder may be poured or sprayed on the shell. It is not necessary to infiltrate a shell after each layer. One may choose to infiltrate a shell after construction of a number of layers or after the shell is fully constructed.

During infiltration, the wicking (pulling in) force will suck the binder into the porous ceramic layer. Since the porous shell contains air, as the binder is wicked in, air becomes compressed behind the infiltrating binder and resists or slows down infiltration. Therefore, it is necessary to either eliminate the air from the porosity or allow enough time for infiltration so that the binder can sufficiently infiltrate. Figure 1 illustrates the kinetics of infiltration and its effect on time.

Data demonstrates that it takes a few hours before the binder can completely infiltrate a 0.5-in, thick shell, because air compresses behind the infiltrating front, and it takes time before the air can relocate. In an experiment conducted in a vacuum chamber, a shell was completely infiltrated in 5 min.

Because of the close similarity between the infiltration method and the present conventional ceramic shell process, this method may be incorporated into the present production facilities. Using the infiltration method, many of the existing limitations can be eliminated from the conventional ceramic shell process. For example:

* slurry aging can be substantially reduced or the rejuvenation of aged slurries can become more practical;

* the need to use exclusively silica-based binder in shells is eliminated, allowing the use of a variety of other non-silicate binders that otherwise would not have been practical to use;

* greater flexibility in the design of mechanical properties allows foundries to attain higher green strengths, lower creep and lower shell knockout strengths.

Reduced Slurry Aging

The two main reasons for aging of slurries in the conventional ceramic shell process can be partially eliminated in the infiltration method. In this method, 100% of the inorganic binder and most of the organic binder can be removed from slurries. The removed binder can be infiltrated into the shell afterward. Since a major part of the binder system can be separated from the slurry, the aging process in slurries can be minimized. The infiltration method allows slurries to be made without any inorganic binder. Therefore, chances for interaction between the binder and the other components in the slurry is limited only to the short time period during the infiltration. Consequently, no substantial interaction will occur between the binder and the other components in the slurry. Moreover, since the binder can be excluded from the slurry, the formation of a slurry-air interface will not cause the binder to dry out. Both of these factors result in substantial enhancement of slurry lifetime. If the slurry does not in clude an inorganic binder, then an organic binder in an aged slurry may be pyrolyzed, and the flours can be reclaimed.

The following example shows how effectively the removal of silica binder from an aqueous slurry can reduce aging. An yttria slurry was made using an yttria flour with average particle diameter of 15 mm according to the first run in Table 1. This slurry has a pH of about 9.5. Data shows that at 9.5 pH, yttria slurry ages rapidly as in 5 days the viscosity increases more than 15 times.

In a second test, another yttria slurry was made according to the second run in Table 1. This slurry has a pH of about 9.5. Data shows this new yttria slurry does not age for 2 months.

The main difference between the above two slurries is the lack of colloidal silica binder in the latter slurry. It is shown that the main cause of aging in an yttria-silica slurry is the dissolution of yttria and, subsequently, readsorption of dissolved cationic species on the surface of negatively charged colloidal silica. Once colloidal silica is removed from the slurry, no such aging should occur. Besides colloidal silica, both the emulsion and surfactant also are negatively charged and can have a similar aging effect on the slurry as colloidal silica. In order to reduce aging, nonionic emulsion and surfactant were substituted for the emulsion and surfactant.

Non-Silica Binder

A major barrier in the utilization of non-silicate binders such as metal alkoxides or nanometer sized colloidal particles in investment casting has been the rapid aging of slurries. This barrier can be removed by using the infiltration method. In this method, binders can be infiltrated into the shell body in a separate step, thereby eliminating their presence in the slurry. The following example demonstrates how a silica-free facecoat can be made using the infiltration method.

A slurry was made according to the second run in Table 1. Test bar pattern waxes then were coated with this slurry. Next, the coatings were dried and infiltrated with colloidal yttria, colloidal zirconia or ammonium zirconium carbonate solution for 30 min. One set of samples also was left with no infiltration as a reference. After 4 hr of drying, three more layers were applied to each bar using a zircon slurry with 20 volume percent colloidal silica binder. The bars then were dried, dewaxed and fired at 2012F (1100C) for 3 hr and the facecoats were tested for chalkiness. The result was that the surface of the non-infiltration test bar was chalky, and the other three with infiltration were not chalky.

Enhancement of Mechanical Properties via Infiltration

The following discusses the additional flexibility in mechanical properties that can be achieved through the infiltration method.

Green Strength--Test bars of zircon slurry with 5 volume percent colloidal silica binder were constructed. Dry bars then were infiltrated with colloidal silica binder for 3 hr. Some of the infiltrated bars were infiltrated for a second time after being dried for 12 hr under 52% humidity. Bars were dried for 4 hr in an oven at 1 12F(45C),and then their green strength was measured.

Bars that were infiltrated once with colloidal silica binder showed green strength of 1100 psi. The total amount of silica in the shell was calculated at 11 volume percent. Shells made with 11 volume percent silica using the conventional method showed green strength of only 495 psi. After the second infiltration, silica content went up to 16 volume percent and green strength became 1650 psi. Shells made with 16 volume percent silica using the conventional method showed green strength of only 650 psi.

Reasons for the increase in strength by the infiltration method can be speculated based on the explanation of a maximum in a strength curve. As the concentration of the binder in the slurry increases, the green strength goes up. However, above a critical concentration, flours become unable to touch surfaces when the shell dries. In addition, some part of the binder will dry inside the cavity between particles and does not contribute to bond formation between particles. On the other hand, during infiltration, flour particles are already in contact and bonded together. Once the additional binder is infiltrated inside the porous shell, binder liquid is drawn to the neck region between particles due to capillary action. As the binder dries, it dries at the neck region and further increases bonding between particles.

High - Temperature Creep--Infiltration produces stronger shells for an identical binder level. As the shell strength increases, total stress at high temperatures is supported by higher strength and creep decreases. Figure 2 shows the comparison between fired strength and creep of two zircon shells with identical silica levels. One shell was constructed using the conventional ceramic shell process and contained 16 volume percent silica. The other shell was made by infiltrating colloidal silica binder into a shell with 5 volume percent silica. The final silica level in the infiltrated shell was 16 volume percent as well. Data shows that the fired strength of infiltrated shell increases about 25%, and creep decreases over 46%.

Another way to reduce creep is to increase the viscosity of the glassy phase in the shell. One way to increase viscosity of non-crystallized silica in the shell is to disperse a fine crystalline phase into it. This idea can be accomplished by precipitating fine dispersion of a crystalline material inside the glassy phase.

Zircon test bars with 20 volume percent silica were infiltrated with zirconium ammonium carbonate solution. After infiltration, samples were creep tested. Data shows that creep was reduced by 50%. At the same time, strength remained similar to the value achieved for the uninfiltrated bars. Zirconium ammonium carbonate easily decomposes under high temperatures and forms zirconium oxide. It also can react with the neighboring silicon atoms to form fine zircon precipitates. In either case, forming a zirconium compound inside the silica matrix will retard the mobility of the amorphous phase.

Ease of Shell Removal--One way to improve shell knockout is to reduce the shell-fired strength. This can be achieved by transforming amorphous silica binder to crystalline form before the shell cools to room temperature. However, in the absence of an ability to crystallize silica, simply reducing the shell strength results in a higher creep.

In the infiltration method, the shell has a low green strength to start and then the shell is infiltrated with a nonsilicate inorganic binder. If the binder type is chosen such that the binder precipitates as a crystalline dispersion, it will normally act as a dispersion hardening agent without increasing the shell strength noticeably. These precipitates reduce creep without sacrificing the ease of shell knockout. Since this method allows selective infiltration, one can easily tailor design creep vs. shell knockout strength as required.

In one example, test bars were prepared from a zircon slurry with 20 volume percent colloidal silica binder. In this slurry, instead of -325 mesh flour, a coarser -250 mesh zircon flour was used. Bars then were infiltrated with zirconium ammonium carbonate solution, fired and creep tested. Figure 2 shows that creep in these bars was similar to uninfiltrated bars made with the -325 mesh flour, however, green strength was reduced by 50%. In addition, creep was reduced by increasing the viscosity of the glassysilica phase. Viscosity increase was accomplished by precipitating zirconium compounds inside the shell through infiltrating the shell with a zirconium ammonium carbonate solution.

             Two Yttria Flour Shell Compositions Comparing the
                        Effects of Collodial Silica
First Run                 Second Run
    Material     Weight %    Material     Weight %
Deionized water      5    Deionized water   5.5
    Emulsion       1.9       Emulsion       3.2
   Surfacant       0.2       Surfacant      0.2
Colloidal silica   7.8
  Yttria flour      85     Yttria flour      91
    Defoamer       0.1       Defoamer       0.1