Ceramic Technology and Processing, King

Effects of Processing on Properties 343

Figure 10.4: Fired Density contours as influenced by temperature and pressure, Alumina. Fine-grained alumina shows the maximum density towards the right of the diagram and, again, sloping off to the right.

Isopressed Alumina

Isopressed alumina of the same type shows a slightly modified behavior, as shown in Figure 10.5.

There is less interaction between temperature and pressure at low pressures. This decreased interaction is possibly due to the uniformlyapplied pressure that results in an improved green pack. The contour map is much more symmetrical in form than before. This is probably because the green density is much more uniform due to isopressing that hydrostatically presses the powder. Perhaps this effect can be considered as a fundamental advantage of isopressing over die pressing.

344 Ceramic Technology and Processing

Figure 10.5: Fired density contours as influenced by temperature and pressure, isopressed Alumina. Slope off to the right is not as steep as with die pressed alumina.

Shrinkage, Axial/Radial Die Pressed Alumina

There is little difference between axial and radial shrinkage until the pressure reaches a high value, as seen in Figure 10.6.

As pressure controls the green pack density, it also controls shrinkage when sintering to full density. Of particular interest in the contours is that the axial shrinkage, of say 21%, is achieved at a lower die pressure than that of the radial. Packing is anisotropic with a higher green density along the pressing direction.

Effects of Processing on Properties 345

Figure 10.6: Firing shrinkage as influenced by die pressure, Alumina. Axial shrinkage is less, due to better packing at high pressures.

A similar thing happens with coarse-grained ceramics, except that it does not show up as shrinkage. Rather, it manifests as a higher sonic velocity along the pressing direction, due to tighter packing.

There is another conclusion that can be drawn from the data. As pressure increases, so does the spread between axial and radial contours. Since particle packing is tighter along the axial direction, less pressure is needed to obtain an adequate green density for sintering. Friction within the compact retards lateral motion of the particles. This results in green density gradients with a lower overall density that require additional pressure to obtain full shrinkage.

346 Ceramic Technology and Processing

Check List, Effects of T/P

•Combination of T/P: locate maximum density.

•Statistically-designed experiments are preferred. There is an interaction between T/P.

•Too low a P does not sinter to a high density.

•High-forming pressure will sinter to a high density over a range of temperatures.

•Less reactive powders do not densify as readily.

•Isopressed powders are more isotropic in sintering behavior.

•Die-pressed parts shrink more in the diametrical direction than in the axial when the pressure is high enough.

•Packing density in the pressing has a large effect on sintering behavior.

4.0EFFECT OF TEMPERATURE ON PROPERTIES

Properties that are developed in the ceramic are the result of interactions between the starting powder properties, green structure, temperature, and time. Temperature of itself can affect the grain growth, gas desorption, melting, and phase changes such as those found in zirconia systems.

Gas Desorption

Bonding in many ceramics is very strong. When a crystal fractures, the surface bonds are broken and become extremely reactive. Surfaces are very strong absorbents. High-surface-area powders can absorb significant quantities of gases, which are difficult to remove. These types of powders have gained widespread use, making gas desorption an issue.

348 Ceramic Technology and Processing

There is a large burst of desorbed gases as the temperature is increased to 1010 °C. Less gas is evolved at 1260 °C, with a second small burst of desorbed gas when heated to 1470 °C. This last burst of gas is shown by the shading in the figure. The gas amount approximately equalled a monolayer desorbed from the ceramic surface. Thermal gravimetric analyses (TGA) have mistakenly fed the notion that gases are driven off at much lower temperatures such as 400 °C.

Bloating

The problem now becomes one of gases entrapped in the structure. This problem results from the inability to desorb gases prior to the ceramic sintering to an impervious state. When such a ceramic is reheated, it bloats. Figure 10.8 demonstrates bloating in alumina that was hot pressed to several temperatures.

Aluminas hot pressed at 1500 °C and 1600 °C both show decreased density when reheated to 1350-1400 °C. Hot-pressing at 1700 °C produces limited initial bloating but levels off above 1300 °C. It appears that the higher hot-pressing temperature allows the absorbed gases to escape by diffusion while the ceramic is still under pressure in the hot press.

It is important to note the following two factors. When the ceramic part is used at an elevated temperature, it will bloat if sintered at a lower initial temperature. It is likely that the grain boundary contains a monolayer of absorbed gas that is most likely water. In the case of alumina, the boundary bonding may be through OH-OH groups, which, if true, would alter the integrity of the structure. These considerations apply only when the sintering temperature is low enough to entrap absorbed gases, which is not always the case.

Effects of Processing on Properties 349

Figure 10.8: Bloating, Hot-Pressed Alumina. Alumina hot pressed to lower temperatures will bloat when reheated due to entrapped gases.

Decomposition

Temperature can also result in the decomposition of some ceramics, especially when one of the molecules has a higher vapor pressure than the other. Three cases are considered below.

B" Alumina. This composition is a sodium-alumina composition that is used in a variety of fuel cells. The problem with its manufacture is that Na2O evaporates during firing. There are two ways to handle this problem: either pack the ceramic in a bed of the same material or encase the ceramic in a platinum enclosure.2

350 Ceramic Technology and Processing

MgO-ZrO2. As previously mentioned, MgO volatilizes from the surface during firing, leaving a monoclinic zirconia layer that is fluorescent under UV light. In one case, the surface of the part was crazed with cracks; this is unusual as a monoclinic surface should result in compressive rather than tensile forces. By sectioning and inspecting with UV, one could explain the crack layer as follows. The layer just below the monoclinic surface was enriched in MgO, making it cubic with a higher thermal expansion than the surface layer. At high temperatures molecular species are mobile, sometimes in unpredictable ways.

SiC/Si3N4. In the case of SiC, the higher vapor pressure atom is Si, which evaporates leaving a graphitic surface. In the case of Si3N4, N2 is the molecule with the higher vapor pressure. In both cases, the ceramic tends to change composition during firing. Again, there are two ways to address this problem: either bury the part in a bed of the same composition or suppress volatilization by increasing the vapor pressure of the more volatile species. For silicon carbide, the furnace interior is presiliconized. In the case of silicon nitride, an over pressure of N2 can be used. When burying the part, it is necessary to use a coarse-grained sand of a material that will sinter to a crumbly consistency. Otherwise, the part will be entombed and difficult to release.

Gas Absorption

The previous discussion showed some of the effects of absorbed gases on microstructure. One must now address the source of these gases.

Effects of Processing on Properties 351

Absorption from the Atmosphere

One obvious source is the atmosphere. Many a practicing ceramic engineer has personally experienced the whole fabrication process fail in the humid summer months due to absorption of atmospheric moisture. In one case, a calcined powder had to be processed within an hour for it to behave properly.3 Absorption was inevitable and was only a question of time. All ceramic surfaces absorb water vapor; the quantity of absorption is decreased with large particle sizes but the saturation is the same. This becomes critical when working with ceramics that readily form hydrates, such as MgO and CaO. While a dry box will slow down hydration, it will not prevent hydration from eventually occurring. However, if the processing is done quickly, hydration can become a manageable problem. Often unrecognized are the absorption processes that transpire during firing.

Absorption in a Kiln

Composition of the atmosphere in a gas-fired kiln consists of CO2, CO, H2O, O2, N2, and possibly some residual hydrocarbons. At elevated temperatures, various molecular species can become volatile and can subsequently be absorbed by the ceramic surfaces. Much of this behavior is predictable from tables on the standard free energy of formation.4 The following species are suspect: Na2O, K2O, B2O3,and, in some cases, SiO2. Water vapor or carbon monoxide can react with these molecules and cause them to become volatile. These reactions are temperature-dependent since the shape of the free energy curves can change with temperature.

As most kilns are lined with refractories containing at least some of these species, contamination inevitably occurs. Unfortunately, these are fluxes that react on the surface or in the microstructure of the grain boundaries as glassy phases. Glass on the grain boundaries decreases resistance to sag at elevated temperatures and decreases resistance to corrosion. Attack along the grain boundaries is not uncommon. Lab furnaces are often lined with highalumina fiber blocks. These kilns are much cleaner than those lined with fire brick. However, one must proceed with caution. When lab work is done in a clean kiln and production is done

352 Ceramic Technology and Processing

in a firebrick kiln, the lab test results may be misleading. These differences can be very large. For example, a ceramic for application in a molten metal had a hundred fold decrease in corrosion rate when fired in a clean environment.

In reducing atmospheres, silica can become volatilized as SiO, which will be absorbed on interior kiln surfaces and on the ware. This is not always harmful as in the case of firing SiC ceramics. As previously mentioned, the graphite must be silicionized to prevent decomposition of the silicon carbide surface. The volatile species reacts with graphite to form a SiC lining.

Gases in Graphite

Any light-colored ceramic will discolor to grey when fired in a graphite-lined kiln. Graphite itself has a very low vapor pressure at the temperature where oxides are usually sintered. For example, at 1627 °C (1900 °K), the vapor pressure of monatomic carbon is 3.47x10-12 atmospheres, which is not enough to account for the contamination.5 However, there are other gases that evolve from graphite when it is heated, including CO2 , CO, and N2. Some grades of graphite contain as much as 0.115% S. ATJ graphite has only 0.043% S, but the odor is still noticeable. Sulfur absorbed in ceramics is difficult to remove and can have harmful effects on properties such as strength.

Hot pressing in graphite molds places graphite in direct contact with the ceramic, and evolved gases are absorbed onto the ceramic surfaces. An alumina with a surface area of 10m2/g, hot-pressed in ATJ at 1700 °C and then abraded by sand blasting exhibited the abraded surface contour seen in Figure 10.9. Please note the 10x vertical exaggeration.

This is a cross-section view, with the graphite contact surface on the left. This edge is abrasion-resistant, possibly because it was coldworked by shear during hot pressing. The weakened zone extends inward for about 0.1", and then the interior levels off.

There is substantial weakening of the ceramic as indicated by the wear pattern. This is also seen by microhardness indentations where this zone, at its deepest, has twice the incidence of fractures around the