Understanding Microstructural Evolution
For instance, most silicon nitride materials have an intergranular glassy phase of approximately 1 nm in thickness after sintering, which can have a negative impact on creep resistance and high temperature strength. These properties can usually be improved by crystallizing the glassy phase. High temperature properties of silicon nitride can also be improved by achieving direct bonding between the grains or the grains and grain boundary phase. If alumina is present, it can disrupt this direct bonding, as well as lower the creep resistance. Other work has shown that the toughening behavior of self-reinforced silicon nitride ceramics with identical microstructures can be increased by reducing the Al and O content of the epitaxial SiAlON layer formed on the reinforcing grains.
A complete understanding of grain boundary behavior is therefore needed, both on the experimental and theoretical level, and progress is being made in both areas. For instance, molecular dynamics simulations are being used to understand the bonding behavior of interface regions on the electronic and atomic level. From these simulations, properties such as thermal conductivity can be calculated for materials like nanocrystalline silicon carbide. Optimizing mechanical properties of materials like partially stabilized zirconia is also possible since simulations can reveal how to control grain boundary and surface properties. Being able to control the microstructure can lead to the design of materials with novel functions as well.
Experimental studies are also needed to understand grain growth and microstructural development. This often requires such sophisticated tools as high resolution transmission electron microscopy and energy dispersive spectroscopy. JFCC researchers have applied these and other methods for such studies. For instance, crystal phase transformation of zirconia under stress in situ has been observed using micro Raman spectroscopy. Electron back scatter diffraction has helped to establish the relationship between the crack propagation path and grain boundary characteristics. A new method for evaluating internal stress distribution also has been developed based on low-incident angle X-ray diffraction. Using such tools, JFCC has developed Mg-doped LiCoO2 cathode materials, flexible silicon nitride ceramics, solid-electrolyte fuel cells, and other novel materials for a wide range of applications.
Other experimental results were also discussed at the workshop. In a study of alpha-Al2O3, microstructural analysis indicated that abnormal grain growth starts if a certain level of impurities segregates at grain boundaries. Other research has looked at the effect of additives on grain boundaries of alumina. Grain boundaries that have facets result in abnormal grain growth; when no facets are present, normal grain growth occurs. The type of grain growth depends on the composition and amount of oxide impurities such as SiO2, CaO and MgO. Grain boundaries with facets are produced when SiO2 and CaO are added in small amounts leading to abnormal grain growth; when MgO is also added at much higher amounts, grain growth reverts back to normal. Another study showed that a change in the ionic bonding state of grain boundaries due to dopant cations’ segregation can improve creep resistance in polycrystalline alumina.
Additives used to promote sintering can also produce phase transformations. For instance, MgAl2O3-ZrO2 (MAZ) and Y2O3-Al2O3 (YA) systems are used to densify silicon nitride but produce different phase transformation behavior. Both additives produce an alpha to beta phase transformation that begins at 1550∞C and is completed around 1700∞C. In general, an increase in MAZ inhibits the transformation, although a very small amount can achieve the same effect, whereas an increase in YA has no effect. Likewise, the grain size decreased with an increase in MAZ, but showed little change with the addition of YA. The difference in behavior is attributed to two different controlling mechanisms: diffusion control for MAZ and reaction control for YA.
Electrical properties of SrTiO3 are also dependent on grain boundary segregation of dopant atoms and impurities. Doping with Fe/Ti=0.4 atom % produces amorphous films of 0.8-nm thickness at the grain boundaries. Such films can change the current-voltage characteristics of the material from linear to nonlinear. Other work has shown that both impurities at the grain boundaries and changes in boundary structure can change the current-voltage behavior in a similar way.
As grain boundary behavior becomes better understood, ceramic engineers will be able to design materials on the microstructural and even nanometer level. For instance, researchers have already determined that controlling intergranular brittleness in lead zirconate titanate ceramics will be required for developing high performance materials for future micromachines. Unique materials with multiple functions (both mechanical and electrical) will also be possible.