Ceramic Technology and Processing, King
.pdf
Ceramic Property Measurements 433
Cubic grains are isotropic and appear dark. Grain boundaries in this section are a little birefringent, containing some of the monoclinic phase and probably some impurities.
Alignment of the grains in the microstructure is not visible except in thin sections observed in polarized light, providing that the phase is birefringent. As the stage is rotated, a grain becomes extinct. Extinction occurs when the birefringence becomes zero and the grain darkens. When all or most of the grains become extinct together, there is alignment in the microstructure. For more sensitive viewing, a gypsum plate is inserted into the optical path between the polarizers. This imparts a background red color that accentuates the differences in orientation. The gypsum plate is called first order red or the sensitive tint plate. Figure 11.33 is a thin section of fused cast alumina near the interface with a steel casting plate.
Figure 11.33: Fused Cast Alumina. Thin section, polarized light, first order red interference plate. The edge on the right was the interface against the steel mold. Scale bar 400 μm.
434 Ceramic Technology and Processing
Alignment is a consequence of crystallization which is dendritic under these conditions. Contiguous regions of the same color have the same crystallographic orientation. Bundles of blue dendrites are attached to the same crystal lattice, as are bundles of red or yellow dendrites attached to each of their separate crystal lattices. These bundles are interfaces that deflect crack propagation and impart toughness to the ceramic. As an additional point of interest, the dark bands paralleling the cast surface on the right are waves of impurities pushed out ahead of the advancing crystallization front. Processing can also orient grains in a ceramic structure.
Shear orients platy or elongate particles because the shape is often related to the crystal structure. Thin sections are one way of observing alignment.
Due to the photoelastic effect, stresses can also be seen in thin sections. In particular, stresses around pores in zirconia result in dark bands around the pore as a void acts as a stress riser. This is rare with other ceramics due to lower photoelastic constants or less stress in the structure.
Zirconia refractories, where thermal shock is severe, are formulated from coarse monoclinic zirconia grains bonded together with a partially MgO-stabilized matrix. Figure 11.34 is a thin section of this material viewed with transmitted polarized light, and with the sensitive tint plate inserted.
Figure 11.34: Magnesia Stabilized Zirconia Refractory, Thin section, polarized light, first order red interference plate. The structure shows twinned large monoclinic zirconia grains. Scale bar 400 μm.
Ceramic Property Measurements 435
The field is largely covered by two large monoclinic zirconia grains. The matrix can be seen around the two large grains. Due to the large volume change when tetragonal zirconia transforms to monoclinic, the grains are extensively twinned.
On the same thin section, another location shows a concentration of the matrix, seen in Figure 11.35.
Figure 11.35: Magnesia Stabilized Zirconia Refractory (same sample as previous figure). Thin Section, polarized light, first order red interference plate. Fine particles are not mixed well into the body. Scale bar 200 μm.
As there is a concentration of the bonding phase in this location, there must be a sparsity of bonds in others. There was a mixing problem that
436 Ceramic Technology and Processing
resulted in a refractory structure off specification. Mixing often leads to problems, as the fines with a much higher surface area tend to stick to themselves. There are two ways to obtain a better homogeneity. One can either increase shear to break the agglomerated fines or one can pre-wet the coarse grain with the binder solution while mixing sift in the fines. This is a bit tricky and requires practice. Either a polished or thin section provides the measure of how well mixing has been accomplished.
There is another location on the same thin section where another problem is seen, as shown in Figure 11.36.
Figure 11.36: Magnesia Stabilized Zirconia Refractory (same sample as previous two figures). Thin Section, polarized light, first order red interference plate. The bond has shrunk away from the large grain. Scale bar 100 μm.
Ceramic Property Measurements 437
Around the large grog particle is an almost-continuous void where, during sintering, the matrix pulled away from the grain as the matrix shrunk. During sintering, when there is little liquid present, the particles must be in physical contact in order to bond. Atoms cannot jump very far. Contact is broken either by springback as the pressing load is released or by the absence of an adequate number of fines in the composition. When there is an inadequate number of fines, the volume decreases as the matrix shifts, leaving voids around the grog pack. In order too avoid this, the grog pack must be a little loose so that it can be carried down with the shrinkage of the matrix. In this particular case, there are not enough fines to go around because they have agglomerated and are not functional in the shrinkage structure. Mixing is important.
Sample Preparation
This section provides some helpful tips rather than giving detailed instructions on how to make sections. Detailed procedures are specific to individual materials and cannot be generalized except in an overall manner. The best source of detailed information is from colleagues working with the same materials who have previously worked out the methods by trial and error or from knowledgeable suppliers.
Polished sections. The first step is to select the location and orientation of the place to make the section. This could be an edge where a reaction has taken place or an orientation to see if the grains are aligned. Smaller is better unless there is a reason to make a large section. It is much easier to polish a small cross-section than a large one; 1 by 1 cm is a convenient size.
Sawing. When starting with an irregular shape or a large sample, sawing is the next step. Use a continuous-metal-rim, thin, diamond blade. Sawing often results in deep damage to the surface when severe, making it advisable to use a fine diamond grit (about 100 grit or finer), a low blade surface speed (1000 inches/min.), and a slow feed rate, sometimes regulated by a weight
438 Ceramic Technology and Processing
pressing the blade against the sample. Use water as the coolant, sometimes with a cutting fluid additive. Two figures of diamond saws are included. Figure 11.37 is of a small, slow-cutting saw that is useful for small samples.
Figure 11.37: Small Diamond Cut Off Saw. Useful for light laboratory work; it produces a smooth cut. (Courtesy of Buehler)
Cutting is slow, but there are other things that can be done in the waiting period. As the cut is easy and slow with a fine diamond blade, there is relatively little damage done to the sample surface. By contrast, a rough cut will shatter the surface to a considerable depth, requiring a lot of lapping to remove it. A continuing problem is holding an irregular shape in the saw vise. If necessary, the sample can be mounted in a polymer cast to make a regular shape suitable for gripping in the vise. Another alternative is to
Ceramic Property Measurements 439
rough-saw a regular shape and then make the section cut. Figure 11.38 is of a larger saw for more heavy duty work that is still of a lab scale.
Figure 11.38: Laboratory Sized Diamond Cut Off Saw. Used for general laboratory work. (Courtesy of Buehler)
Impregnation. When the sample is porous, it is impregnated with a polymer. Epoxy is most commonly used. Refer back to Figure 11.17 of an apparatus used for saturating a porous sample for specific gravity determinations. The same apparatus is used for infiltrating the sample with the epoxy polymer. The procedure is the same. A vacuum is pulled, the epoxy is let in, and the chamber is pressurized to about 60 psi. By using pressure, a higher viscosity epoxy can be used. The epoxy will have a higher molecular weight, less shrinkage, and better physical properties. The resin will spatter as it is admitted into the chamber, however that will not be a problem as long as the tube is far enough below the top of the beaker.
440 Ceramic Technology and Processing
After use, the apparatus must be cleaned. The valve is a stopcocktype dismountable by a snap ring, making it easy to clean off the residual resin with a solvent such as acetone. Chaining the snap-ring pliers to the apparatus ensures that they will always be available.
Use a casting resin to mount the sample, as hot molding phenolic overheats the epoxy.
Another procedure for infiltration is to use a glass frit with a low melting point. Melt it on the surface over a burner. Lead borosilicate glasses melt as low as 450-550 °C, low enough to prevent reaction with almost all ceramics. Infiltration is shallow as no pressure assistance is used, making it necessary to limit the amount of surface that can be ground off. Frit is advantageous for having a higher modulus than a polymer, providing greater support for the sample at the edges of the porosity. When the porosity is very fine, frit infiltration is the preferred technique and may be the only option for making the section.
Premolding Treatment. Scratches are usually caused by pieces chipping off the edges of the section and being dragged across the surface. A few scratches do not make any difference, but too many scratches obscures much of the surface. Grinding a small bevel on the top edges of the sample greatly reduces the tendency for the edge to chip. Lapping off the excess epoxy or frit prior to mounting helps the sample to lie flat during mounting.
Mounting. If the sample is to be polished by hand, there is no need for mounting. When the sample is to be polished in an apparatus, a mount must be made for it to geometrically fit into the polisher. Polishing equipment is the preferred method, especially when there is more than one sample. Offhand polishing is quick and dirty and often used for a quick look. Samples can be mounted in a phenolic resin by hot molding or cast in a mold with a polyester resin. Some polymers have a strong exotherm and will char and bubble during curing. Large castings are susceptible to overheating. Procedures for mounting in a phenolic resin, as given by the vendor, do not allow enough time for the resin to crosslink. Do not follow the manufacturing directions in this case, but give the molding additional time (+5 min.) to crosslink at the molding temperature. Otherwise, since the
Ceramic Property Measurements 441
phenolic is soluble in organic solvents, it will produce a mess when the polishing oil is cleaned off with a solvent.
Lapping. Sawing leaves a rough surface, with damage extending into the interior of the sample. This must be lapped off with abrasives using water as a cutting fluid. Figure 11.39 shows a set of laps mounted in a polishing bench.
Figure 11.39: Polishing Lapidary Table. Can be used for wet grinding and polishing if kept clean. (Courtesy of Buehler)
The photo shows a setup for polishing. But, the bench can be used for both lapping and polishing. When the ceramic sample is hard, it is necessary to use diamond laps where the grit is held on the surface with
442 Ceramic Technology and Processing
nickel plating. These laps are thin and are conveniently held in place with a magnetic disc. This makes it easy to change grit sizes. Grit sizes start for many applications at 60 grit and then go sequentially to about 100 grit and finally 220 grit diamond. Finer sizes do not work well with this system as the lap surface seems to glaze over and ceases to cut. All lap surfaces should have voids. For loose abrasives, these voids are a spiral groove machined into the lap surface. The nickel-plated laps are best when perforated with an array of small holes. These are seen in a sketch of the interface between the sample and the lap surface, as depicted in Figure 11.40.
Figure 11.40: Nickel bonded Diamond Lap with Perforations in Surface. Fastens with a magnetic disc for easy removal.
Voids on the lap surface help to break the "suction" between the sample and the lap surface and supply lubricant (water) to the interface. The sketch shows air bubbles that are drawn out into films on the interface. Air, as a gas, expands when pressure is lowered. This expansion allows the sample to glide freely over the lap surface. A solid lap surface "grabs" the sample, seemingly pulling it from one's hands. If the lap runs dry, the sample overheats, and there is no means to carry away the swarf. Laps made from