Dry Forming

Die Pressing

Die pressing is the predominant powder-forming process with metals. In contrast, wet forming is the predominant powder-forming process with ceramics. With metals, more pressing pressure produces more consolidation, whereas with ceramics, more pressure usually causes problems such as delamination and nonuniform firing shrinkage. This is why die pressing of ceramics involves only low to moderate pressures and is generally used only with thin simple shapes, such as tiles.
In its simplest form, uniaxial die pressing involves relatively dry powders, containing a minimum amount of liquid and some binders. The powders are placed in a uniaxial hollow die with two opposing pistons or pressing plates (Fig. 9). The die cross section is commonly a simple shape: circular, square, or rectangular. Pressures are typically in the 20–50 MPa range for ceramics and range widely for metals (Table 8, Fig. 9). Pressure is applied to the two opposing pistons to consolidate the powder into a green body. Die pressing metal powders is a simple process in most cases because the plastic deformation of the particles enables excellent powder flow, cold welding, and in some cases almost complete consolidation by cold welding.

For ceramics, consolidation by uniaxial die pressing is much more problematic. Theoretical particle packing models predict that the highest packing density possible for mono-sized non-deformable spheres is 74 %, in a hexagonal close-packed array. The reality is that ceramic powders ordinarily attain a green density of only about 50 % from die pressing. Higher green densities can be achieved by gap-grading (mixing coarse medium and fine powders in appropriate ratios), vibratory pressing, and the careful use of lubricating binders. However, in practice, these strategies are rarely used except for a few specialty advanced ceramics.
The main advantages of uniaxial die pressing are no need for drying, simplicity, and ease of automation. With a fully automatic press, and multiple cavity molds, this process can produce hundreds of components per minute.
Unlike metal powders which deform plastically under pressure, the main disadvantage of uniaxial die-pressing ceramics is the nonuniform particle packing due to pressure gradients generated in the powder compact during pressing. The thicker the pressed part, and the more complex its shape, the more significant this problem becomes. Nonuniform particle packing translates into nonuniform firing shrinkage, which at best causes distortion, and at worst cracking. Cracking during sintering can involve obvious destructive macrocracks causing immediate rejection or invisible microcracks which may cause early catastrophic failure some time later during service life.
The main application of uniaxial die pressing for ceramics is in fabricating ceramic tiles. The major application of uniaxial die-pressed conventional ceramic tiles is porcelain, terra cotta, stoneware, and heavy clayware tiles for floor, wall, and roofing claddings, specialty curtain-walls for high-rise buildings, and die-pressed construction bricks. The major application of uniaxial die-pressed advanced ceramic tiles of alumina, silicon carbide, and boron carbide is ballistic armor and wear-resistant tiles for the mining industry. Industrial cermets such as tungsten carbide and titanium carbide are also commonly formed by uniaxial die pressing.

Ceramic tiles for the building industry have traditionally been no larger than about 300 x 300 mm in size. In recent years sizes up to 600 x 600 mm have become commonplace. In the 1990s, commercial tiles up to 1,000 x  1,000 mm in size came onto the market, from Germany (Buchtal), and China (Eagle Brand). The problems with increasing tile size are threefold:

• Extremely large press capacities are required. Typically ceramic tiles are pressed at pressures in the 20–50 MPa range, both ceramics for the building industry and advanced ceramics such as alumina, silicon carbide, and boron carbide. Thus, a 5,000-t press is needed for 1,000 x 1,000mm tiles, for example. Such a press costs many millions of dollars to purchase; is large, heavy, and very expensive to transport and install into a standard factory; and has a huge footprint on a production floor.
• Difficulty in supporting the tile during sintering. The larger the tile, the larger the absolute firing shrinkage. For a 15 % firing shrinkage, this equates to 150 mm linear dimensional change in a 1,000 x 1,000 mm tile. The tile support in the kiln needs to allow significant movement during shrinkage without any chemical interaction taking place between support and tile at the 1,100o C+ sintering temperatures. Various strategies are utilized such as ceramic balls, ceramic granules, and other proprietary substrate systems.
• Transportation and handling. The larger the tile, the higher the cost and the higher the risk of breakage during shipping and installation.

One of the most useful innovations in recent years in ceramic powder forming is vibratory pressing. This typically involves a pneumatic press operating on compressed air, which pulses the pneumatic pressure during pressing, thereby vibrating the consolidating powder during the pressing process. This has two key benefits:

• Elimination of powder packing gradients.
• Significant reduction in required pressing pressure. Instead of 20–50 MPa which is usual for die-pressed ceramics, pressing can be achieved at pressures as low as 1–5 MPa. Thus, instead of a 5,000 t hydraulic press being required for 1,000 x 1,000mm tiles, they could be pressed with a 100 t vibratory press – a very much smaller, cheaper, and more portable unit.

Thus, vibratory pressing is most useful when pressing very large tiles or complex-shaped ceramic components.
When manufacturing advanced ceramics for specialty applications with serious consequences for failure, such as ballistic armor and wear-resistant linings for the mining industry, it is critical to ensure that the component contains no microcracks, which are most commonly caused by pressing nonuniformities. X-ray radiographic examination, once primarily the domain of metal fatigue inspection, is a common QA procedure for inspecting die-pressed advanced ceramics.

Cold Isostatic Pressing (CIP)

Cold isostatic pressing (CIP) involves uniform pressing from all directions. It is widely used for powder forming of ceramics. For example, three billion spark plug insulators are made every year by CIP. Metal powder CIP forming applications are much less common. Most powder metallurgy products are pressed by die pressing, and CIP is generally only used in the rare event that very complex shapes are required and powder injection molding and die casting are not available options.
CIP is usually done by placing the powder in a thin-walled deformable “bag” or “mold” and immersing in an oil chamber which is subsequently pressurized. Thus, consolidation occurs by immersing the powder (held in an impervious thinwalled mold) in a pressurized fluid. Typically pressures in CIP mass production are in the order of 70 MPa, although pressures up to 400 MPa are not uncommon. Molds are usually thin walled and made of elastomers such as latex, silicone, neoprene, nitrile, or polyurethane. Another manifestation of CIP is post-CIP in which a component is pre-pressed by uniaxial die pressing, and then the green part is either coated in elastomer or placed in an elastomer bag (vacuum de-aired) and then CIPed.
A CIP is basically a fluid-filled pressure vessel with a mechanism enabling rapid automated opening and closing, a high-pressure pump, and a system for gentle depressurization. The automated CIP has multiple racks or arms for supporting numerous molds in one production cycle.
While uniaxial die pressing is a widely used and effective means of producing PM metals and ceramic tiles, fabrication by pressing of any ceramic shape more complex than a simple tile usually requires CIP. The main benefit of CIP is that the nonuniform particle packing due to pressure gradients generated in the powder compact during pressing is much less significant than for uniaxial die pressing, although this needs to be seen from a viewpoint of relativity. For a simple thin tile, uniaxial die pressing should give a uniform result. For a complex shape such as a spark plug insulator, CIP is the only pressing process that is viable. Of course, wet-forming methods are also viable for complex shapes, but CIP has the advantage of being long established and well suited to automation, with large throughput rates. Multiple powder-filled molds can be loaded into the CIP for each pressure cycle.
For solid components, all that is required is a thin-walled elastomer mold. The elastomer mold wall needs to be thin so as to ensure maximum pressure transfer to the component. Moreover, the elastomer mold wall needs to be of uniform thickness so as to ensure uniform pressure transfer to the component. For hollow components, the elastomer mold provides the external wall dimensions and shape of the component, and a removable metal inner mandrel is used which provides the inside wall dimensions and shape of the component. The mandrel is usually made of hardened steel to ensure durability after multiple removals from ceramic powder components with their abrasive surfaces.

Some of the key advantages of CIP for ceramics include:

• Low mold cost which makes it ideal for complex parts with small production runs. For post-CIP, there is no CIP mold cost, simple elastomer bags, or coatings suffice. Of course, the uniaxial pre-pressing stage requires molds.
• There is no size limitation. Therefore, it is ideal for very large parts. Large refractory components heavier than 1 t have been produced by CIP.
• Unlike uniaxial die pressing, CIP produces uniform shrinkage in components with nonuniform cross sections, resulting in less cracking during sintering.
• Short cycle times. This is mainly because no drying or binder burnout is required.

Some of the key disadvantages of CIP for ceramics include:

• Poor dimensional control. A common solution to this is to CIP oversized components, or blanks, followed by green machining. This requires the use of binders and carries the risk of cracks introduced from green machining.
• It is suitable for relatively simple shapes only. For highly complex shapes, powder injection molding is a better choice.
• Powders must have good flowability. This generally means ceramic powders must be spray dried, which adds an extra cost to the process. Mold vibration helps for mold filling, but if complex-shaped molds are involved, spray-dried powders are essential.

Thus, CIP is an appropriate choice for mass production of advanced ceramics of complex shape or extremely large ceramic components. The number one product is spark plug insulators. Other examples of commercial ceramic products  mass produced byCIP include powerline insulators, specialty refractory components, oxygen sensors, specialty wear parts for pumps, ball milling balls and grinding media, and tableware.
CIP is not to be confused with hot isostatic pressing (HIP) which is a densification process (section “Hot Isostatic Pressing” discussed later).

Wet Forming

Introduction to Wet Forming

Powder wet forming dates back thousands of years to the invention of pottery –plastic forming of clayware. The potter’s wheel was invented in ancient Mesopotamia more than 4,000 years ago. Early pottery was fired in bonfires, and kilns were invented much later. In the last century, many novel powder wet-forming methods have been reported. The most widely used methods are extrusion, plastic forming, slipcasting, and injection molding. Extrusion and plastic forming are considered low-technology powder-forming methods and are almost exclusively used for clayware. The potter’s wheel is the most common type of plastic forming commercially, and there are many automated methods that utilize the principle of the potter’s wheel, machining the rotating plastic clay body, for example, in making tableware. Extrusion of plastic powder slurries is widely used to make clayware building bricks and pipes. Extrusion is also used to make advanced ceramics such as alumina furnace tubes, hollow insulators, and thermocouple sheaths, using a highly optimized water/powder/deflocculant/binder mix.
One of the most important issues with powder wet forming is drying. Drying shrinkage can be more than 10 % in many cases, and even when large amounts of binder are used, this can cause cracking during drying, both visible cracks and invisible cracks which manifest only later, either after sintering or after some time in service. Preventing cracking requires attention to the following issues:

• Add binders to the slurry (a powder–liquid suspension is known as a slurry).
• Maximize solids loading to reduce shrinkage.
• The finer the particles, the greater the drying shrinkage (and conversely the better the sintering).
• A broader particle size distribution reduces shrinkage.
• Slower drying times reduce cracking.
• The larger the component, the larger the overall shrinkage, and therefore the greater the cracking risk. Thus, wet forming large ceramics is highly problematic.

Several high-technology drying solutions have recently begun to be introduced: controlled humidity drying chambers, drying in ethylene glycol, and supercritical drying (SCD) using CO2. All three of these techniques can greatly reduce cracking during drying; however, they are slow and not widely used commercially.
There are a number of high-technology wet-forming methods that warrant further discussion here, both from the viewpoint of their novelty and their potential to push the frontiers of powder forming into new possibilities.

Slipcasting

Slipcasting is a powder-based wet-forming method which has been used for a long time in the traditional ceramic industry for the manufacture of sanitary ware and tableware. This process has been also used for the manufacture of technical ceramics, e.g., alumina. This method involves a filtration process which separates powders from the solvent using a porous medium. In a typical example of this process, a suspension, usually a water-based suspension (slip), is poured into a porous plaster mold which, by its porosity, creates capillary forces and removes liquid from the suspension. A flow of solvent causes the particles to create a consolidated layer or a filter cake on top of the mold surface. The cast formation time is reduced if high solids loaded slurry is used. Although less concentrated ceramic slurries give better control of the thickness of the cast layer, the chance of segregation among particles increases, especially when composite slurries are used.
Once the required layer thickness is obtained, the process is stopped, either by removing the excess slurry or by letting the casting cakes meet each other in the center of the piece (solid slipcasting). Generally, the casting shrinks in the mold during air-drying enabling easy removal from the mold for further drying or sintering (Adcock and McDowall 1957; Tiller and Tsai 1986; Lewis 2000).

The main advantage of slipcasting is that it is a low-cost process with the capability to produce complex geometries with high homogeneity. Furthermore, the mold material is relatively cheap. However, the durability of the molds is an issue which limits the application of this technology for mass production, requiring many molds and large plant area.
Pressure slipcasting or pressure casting solves some of the problems with the slipcasting process. In this process, instead of using plaster molds, molds of polymeric materials are used which have a high porosity consisting of larger pores that do not give the same capillary forces but require an externally applied pressure to drive the filtration process. The higher the applied pressure (<40 bar), the faster is the casting cycle. In conventional slipcasting, the capillary forces correspond to a pressure of 1–2 bar. Using pressure can help to dry the parts immediately and cast pieces can be demolded, and a new casting cycle can be started. The durability of plastic filters is much better than plaster and can help to achieve parts with better tolerances.
Centrifugal slipcasting is a similar process in which ceramic particles in a suspension are forced toward a mold cavity which can be nonporous, and the top solvent is poured off after the casting. This process can improve the packing density of the filtered cake and improve the quality and reliability of the final product (Steinlage et al. 1996; Huisman et al. 1994).

Tapecasting

Tapecasting, as the name suggests, is used to form powder slurries into sheets that have a large surface area and a thin cross section. These tapes are widely used to produce electronic components and packages (Howatt et al. 1947; Mistler 1990). Tapecasting arose in the 1940s as a method for fabricating thin layers of dielectric materials for capacitors and piezoelectrics. Unlike slipcasting, tapecasting slurries are usually nonaqueous. The slurry is fed through a device known as a “doctor blade” which is either stationary over a moving substrate or moves over a stationary substrate. The slurry is screeded by the blade so that it spreads out as a thin layer of slurry due to the relative movement of doctor blade substrate, then dries rapidly by evaporation. Thickness is controlled by adjusting the gap between the doctor blade and the substrate. Tapecasting involves the following stages:

• Ball milling the slurry
• De-airing the slurry
• Filtering (to remove lumps)
• Tapecasting using the doctor blade
• Drying

Some of the key parameters involved in optimizing tapecasting include:

• Particle size and size distribution of the powder
• Chemical composition of the solvent (suspending liquid medium)
• Composition and concentration of the deflocculants, so as to maximize solids loading
• Composition and concentration of the binders, which enhance the green strength of the cast tape
• Composition and concentration of the plasticizers, which enhance the flexibility of the cast tape
• Solids loading
• Doctor blade clearance
• Relative velocity of doctor blade and substrate
• Drying cycle

There have been many patents filed on tapecasting since the original one in 1952 (Howatt 1952). Tapecast alumina is now used for many cutting-edge microelectronic and biomedical technologies, for example, fabricating advanced multi-array microelectrodes for specialized applications such as the bionic eye (Guenther et al. 2013).

Powder Injection Molding

Powder injection molding is best suited for the high-volume near-net-shape production of small metal or ceramic parts starting from powders. The use of metal or ceramic powders enables a wide variety of ferrous and nonferrous alloys in powder form to be used as well as many advanced ceramics in powder form. The material properties (e.g., strength, hardness, wear resistance, corrosion resistance, temperature resistance) can be close to those of wrought metals or ceramics made by other processes. Since no melting of the functional material occurs in the PIM process (unlike metal casting processes), the total heat input of the process is much lower compared with casting processes. High-temperature alloys can be used without any adverse effect on tool life. Metals commonly used for PIM parts include low-alloy steels, high-speed steels, stainless steels, cobalt alloys, tungsten alloys, nickel alloys, and titanium alloys (Piotter et al. 2007, 2008; German et al. 1991; Blackburn and Wilson 2008). Ceramics such as alumina, zirconia, silicon carbide, silicon nitride, and aluminum nitride have been successfully demonstrated and commercially produced by the PIM process, a prominent example being ceramic turbocharger rotors, now mass produced by PIM on a large scale. Composites of ceramics and metals can also be fabricated using this process (Ye et al. 2008).
Powder injection molding parts are found in numerous industry sectors, including automotive, aerospace, oil and gas, consumer products, medical/dental, and telecommunications.
Depending on the type of material, powder injection molding is sometimes called metal injection molding (MIM) or ceramic injection molding (CIM). As shown in Fig. 23, the process uses fine powders of ceramics or metals (typically <20 μm dia.) that are heated and mixed with thermoplastics polymer binders, waxes, or other organics to form a feedstock (Supati et al. 2000). Some additives are added to keep the dispersion of particles and plasticity of the component. After cooling, the mix is granulated into the form of pellets with a nominal size of 1–5 mm and fed into an injection molding machine. The granulated mixture is heated to a temperature below 250o C to form a flowable paste and injected into a closed mold to form the “green” compact. The equipment which is used for PIM is very similar to plastic injection molding. However, in PIM the material used for the mold and extruder is more critical because of the risk of erosion from the abrasive binder-powder mix. Unlike conventional powder metallurgy processes where the organic content is ~1 wt%, the binder content of the green PIM compact may be as high as 40 wt%. The green part is cooled and removed from the mold. Chemical and/or thermal methods are used to dissolve and remove the majority of the binder. The remainder is removed from the porous compact by heating at an intermediate temperature, then the temperature is elevated to the sintering temperature. The PIM process can be used for mass production of parts with much greater complexity than for press-and-sinter powder metallurgy parts, but smaller in size.

Design rules for plastic injection molding are still applicable when the parts are to be manufactured by PIM. However, there are some exceptions or additions, such as the following (www.custompartnet.com):

• Wall thickness limitation: Wall thickness has to be minimized and be kept uniform in order to facilitate binder removal and minimize warpage or distortion. In normal PIM processing, a thickness below 7 mm would help to make crack-free and uniform structures. However, very thin-walled sections need a support to avoid deformation from the thermal cycle.
• Draft: Although using draft in mold design for plastic injection molding is essential, many PIM parts do not require any draft. The polymer binder used in the powder material releases more easily from the mold than most injection molded polymers. Also, PIM parts are ejected before they fully cool and shrink around the mold features because the metal powder in the mixture takes longer to cool.
• Sintering support: During sintering, PIM parts must be properly supported or they may distort as they shrink. By designing parts with flat surfaces on the same plane, standard flat support trays can be used. Otherwise, more expensive custom-made supports may be required.

Direct Coagulation Casting

Direct coagulation casting (DCC) is a near-net-shape fabrication method for making ceramic components from concentrated aqueous powder suspensions (Balzer and Gauckler 2003; Baader et al. 1996; Prabhakaran et al. 2009). In DCC, the powder suspensions are cast into a mold and they set through an in situ coagulation or destabilization of the suspensions by producing acid, base, or electrolyte from water soluble precursor molecules present in the suspension medium. Setting via generation of ammonia (base) and electrolyte like ammonium bicarbonate and ammonium carbonate from urea by urease catalyzed in situ hydrolysis is a common approach, well studied for most of the ceramic powder systems, such as alumina, zirconia, silicon carbide, and silicon nitride (Prabhakaran et al. 2009). By a pH shift near the isoelectric point, the ammonia can coagulate the slurry prepared in acidic pH. Ammonium bicarbonate and ammonium carbonate coagulate the powder suspensions by compressing the electrical double layer thereby promoting coagulation. In situ generation of acid from acid anhydrides, esters, lactones, hydroxyl aluminum acetate, and MgO has also been reported for destabilization of powder suspensions.

Although the preparation of the slurry for this process is very similar to gelcasting, the mechanism for gelation is different. There are some limitations corresponding to this method, such as premature coagulation of the slurry, low green density, cracking of the parts via drying, or sintering. This process is still in the research stages and has not been commercially taken up for manufacturing of ceramics.

Gelcasting

The gelcasting process dates back to 1990s when Omatete and Janney from Oak Ridge National Laboratory, USA, came up with a new slurry-based manufacturing process for making ceramic components in which conventional wet processing techniques were combined with polymer chemistry (Janney and Omatete 1989; Omatete et al. 1997; Yang et al. 2011) (Fig. 24).

In this technique, a ceramic powder and a monomer solution are mixed together to form a slurry. A high solids loading is possible with a minimum amount of polymer. The slurry should be highly loaded with the ceramic powder and maintain high flowability for the casting process. Therefore, a suitable slurry formulation is necessary. Mechanical milling is essential to achieve high solids loading. The slurry is degassed under a low pressure and poured into a nonporous mold (unlike slipcasting which uses a porous mold) after addition of a chemical initiator. Over time, a polymer gel network is formed in situ to the shape of the mold. On polymerization, all ceramic particles are immobilized because of the formation of a cross-linked gel network. After removal from the mold, the part is dried, the polymer is burned out, and the sample is sintered to a high density. Gelcast green bodies also have a high green strength, which allows them to be machined easily before sintering (Janney et al. 1998).
Gelcasting overcomes many limitations associated with other ceramic processing methods, such as injection molding (e.g., long debinding time, flaw generation, or distortion via binder removal) and slipcasting (e.g., slow casting rates, insufficient strength for green machining), and is suitable for the complex shape fabrication by offering short processing times, high yields, high green strength, and low-cost machining. Unlike the powder injection molding process in which a high amount of binder (15–20 wt%) is used and long debinding steps are required, the polymer content in the dried gelcast samples is 2–5 wt%. This amount of binder can be easily removed by pyrolysis without a significant effect on distortion or cracking of the parts (Yang et al. 2011; Janney et al. 1998).
One of the advantages of the gelcasting process is the variety of materials usable for molds. Wood, plastic, wax, and metals such as aluminum and stainless steel have been used for gelcasting molds. However, some plastic and metallic materials tend to react or interfere with the gelation process which must be avoided. Demolding agents need to be used to ensure fast and easy part removal (Yang et al. 2011; Janney et al. 1998).
Gelcasting has been used for many years for manufacturing of ceramic components. Gelcasting metal parts has also been successfully demonstrated. However, since metals can be more easily processed by casting from a melt, forming, or conventional powder metallurgy and machining, gelcasting of metals has not received as much attention as ceramics (Omatete et al. 1997; Yang et al. 2011; Janney et al. 1998; Huang et al. 2007; Stampfl et al. 2002).
Gelcasting has been combined with other methods such as tapecasting, injection molding, centrifugal casting, and stereolithography (Stampfl et al. 2002). Gel-tapecasting is considered the main manufacturing method for large-scale ceramic sheet production with thickness between 0.1 and 1.0 mm (Yu et al. 2004). By combining injection molding as a fully automated method and gelcasting, the advantages of the two processes can be merged to improve productivity and reliability of ceramic parts. The new process is called colloidal injection molding of ceramics (CIMC) (Huang et al. 2004). In this process, pressure is used to induce a gel reaction, which can avoid the effects of the temperature gradient on the gelcasting process. The efficiency and reliability of the product were greatly improved due to more homogeneous solidification which reduced the internal stresses from the colloidal-forming process and prevented the formation of microcracks in the green body.
In another method, gelcasting is combined with centrifugal casting. In this method, the slurry is transferred to a centrifuge machine with an embedded mold, after addition of the initiator and catalyst. The centrifugal force not only can form the ceramic slurry into the mold cavity but also can eliminate air bubbles and accelerate the gelation process. Therefore, no degassing is required and the reliability of parts has been shown to be improved. Moreover, very fine features of the mold can be filled by slurry by the action of centrifugal force. However, the centrifugal force applied for this process is much lower than for conventional centrifugal slipcasting processes, as the role of the force is to shape the slurry and not to cause powder segregation (Maleksaeedi et al. 2010).
In a novel concept, gelcasting is merged with 3D stereolithography, as a solid free form fabrication method. In this case, dense ceramics directly from a CAD file have become a reality (Zhou et al. 2010). The gel formula should be adapted for a photopolymerization process (Chartier et al. 2002).
Some limitations with the gelcasting process are as follows (Omatete et al. 1997; Yang et al. 2011):

• Toxicity of the gelcast materials
• Inadequate strength of the parts after gelation and via demolding which can cause severe distortion or cracking if the part is not well supported
• Difficulty to achieve high solids content of gelcast slurry for some specialty ceramics
• High susceptibility for cracking during the drying stage
• Not easily automated, preventing gelcasting being utilized in the industrialization of high-performance ceramics

Thixotropic Casting

Thixotropic casting is a ceramic wet-forming process derivative of solid slipcasting that utilizes a casting slip that is fluid under vibration and plastic in the absence of vibration (Ruys and Sorrell 1999; Chesters 1973; Norton 1974). Thixotropic casting is the powder-forming method that gives the highest green densities of all known powder-forming methods. Green densities up to 82 % are achievable. Solids loadings of up to 78 vol.% have been reported for thixotropic casting (Legrand and Da Costa 1989). Most powder-forming methods give green densities in the 40–60 % range. Theoretical particle packing models state that the highest packing density possible for mono-sized spheres is 74 %, in a hexagonal close-packed array. The reality is that ceramic powders ordinarily attain a green density of only about 50 % from die pressing and slipcasting. Even lower from wet-forming processes involving large amounts of binder removal, such as injection molding. No known method other than thixotropic casting can give green densities above 70 %.
Thixotropic casting is unsuitable for powders with particle sizes below one micron, for example, clays or submicron advanced ceramic powders. The high surface area of colloidal-based slurries limits the maximum solids loading attainable (Sushumna et al. 1991).
Secondly, thixotropic casting gives by far the lowest drying shrinkage of any known wet-forming method. Typically, it is 1 %, although it can be less than 0.5 %. This has proven very effective when manufacturing large ceramic components by thixotropic casting. A one-meter-sized ceramic component, such as a large refractory or a large floor tile, requires a 5,000 t press for manufacture by conventional uniaxial pressing. No wet-forming method other than thixotropic casting can be used to make such a large ceramic as the drying shrinkages are in the order of 100 mm or more. However, a 1-m-sized ceramic component can be manufactured crack-free by thixotropic casting, as 0.5 % drying shrinkage corresponds to just 5 mm.
Thixotropic casting is a simple process. A highly concentrated powder–liquid slurry is prepared, typically in the 70–80 vol.% solids range, and vibrated into a mold. The thixotropic casting process involves the following stages:

• Powder and aqueous deflocculant solution are batched and mixed into a uniform stiff doughy consistency.
• The molds are mounted on a vibrating table.
• Vibration is switched on and the thixotropic casting slurry is slowly vibrated into the mold.
• Vibration is switched off as soon as the mold is filled, so as to prevent segregation.
• The component is left to air-dry in its mold, followed by oven drying and finally densification.

The high solids loadings, and by association high green density and low drying shrinkage, are achieved by three means:

• Firstly and most importantly, the slurry is so concentrated that it will only flow under vibration. A thixotropic fluid has a lower viscosity when subjected to shear. This is the principle of thixotropic casting. Therefore, a thixotropic casting slurry typically shows no fluid characteristics in the absence of vibration. Its consistency depends on the particle size of the slurry and varies from a stiff “paste” to a damp “mortar.”
• The use of gap-grading to maximize the efficiency of particle packing. This can be achieved either by using a powder of broad particle size distribution or by blending coarse and fine powders in optimized ratios.
• The use of chemical deflocculants at optimal concentrations to maximize the thixotropy of the slurry and to maximize the fluidity for a given solids loading.

Thixotropic casting is primarily used in the refractories industry, where it is ideal because of the widespread use of “grog” (coarse particulates) in refractory mixtures, and the large size of many refractories, which necessitates the minimization of drying shrinkage. It has also seen use in boutique applications such as specialty bioceramic components, ceramic composites, functionally graded materials, and very large ceramic tiles (Ehsani et al. 1995; Ruys et al. 1994; Kerdic et al. 1996; Ehsani et al. 1996; Ruys and Sorrell 1992, 1994). Thixotropic casting molds can be porous (e.g., plaster or timber) or nonporous (e.g., metal or plastic).