Introduction
The excellent mechanical and electrical properties of engineering ceramics make them popular in a wide variety of applications. In this example the high dielectric strength of alumina promotes its use as an electrical insulator. These dielectric properties are influenced by the porosity of the material, which needs to be kept to a minimum after sintering to ensure optimum performance. The properties are engineered by controlling the grain size of the raw material and by the use of various additives.
Therefore, characterisation of the ceramic microstructure is vital in understanding and controlling its properties and Electron Backscattered Diffraction (EBSD) is the ideal technique for this. Typically the properties of interest include:
- Grain size and shape
- Grain boundary characterisation
- Porosity
- Texture
- Phase identification (when used in conjunction with EDS)
This data is typically presented in the form of maps and pole figures.
Sample preparation of ceramics can be more challenging than for metallic materials, but standard preparation techniques can be successful in producing sample surfaces suitable for EBSD.
Care must be taken to minimise the 'pull out' of grains during the initial stages of sample preparation [1]. Pull out can result from intergranular fracture caused by high grinding loads on the sample. Regions where grains have been 'pulled out' can be mistaken for genuine porosity which would give erroneous results. Careful sample preparation, and monitoring of the same area of the microstructure after each preparation stage, can minimise 'pull out' and hence produce more representative microstructures for EBSD analysis.
Another issue to be aware of is specimen charging. This can cause problems both in degradation of the EBSD pattern or in sample drift. This can be overcome by coating the sample with a thin layer (2–5 nm) of carbon or gold.
This application note illustrates how EBSD can be used to characterise the microstructure of an alumina electrical insulator.
Experimental Method
The sample was prepared using standard metallographic techniques, with an additional final polish using a mixture of colloidal silica and hydrogen peroxide. The sample was examined in a FEGSEM equipped with a Nordlys EBSD camera and the data analysed using OI CHANNEL5 software. EBSD acquisition parameters are listed in Table 1.
The sample was not coated, and charging was minimal. Any sample drift was overcome using the CHANNEL5 drift correction tool.
| Parameter | Value |
| Grid dimensions | 446 x 335 |
| Step size | 0.3 µm |
| Number of points | 149,410 |
| Mapping speed | 20 points / sec |
Table 1. EBSD acquisition parameters
Results
Figure 1(a) is a Differential Interference Contrast (DIC) image of the alumina ceramic microstructure. In addition to the alumina grains, there is also an intergranular glassy phase and porosity. Figure 1(b) is a typical EBSD pattern obtained from this sample and illustrates that high quality patterns can be obtained from ceramic materials provided sufficient care is taken with sample preparation.

Figure 1. (a) DIC image of the alumina and (b) typical alumina EBSD pattern.


Figure 2. (a) Band contrast map with the glassy phase and porosity showing as black regions and (b) IPF plus grain boundary map. The lack of a predominant colour indicates no dominant texture is present.
Figure 2(a) is a band contrast map of the alumina where porosity is shown in black. This map has been 'cleaned up' using the noise reduction facility in Tango. In cases such as this where there are large regions of non-indexed points i.e. porosity, care should be taken to ensure that existing grains do not 'grow' into these regions during the noise reduction procedure, as this leads to an unrepresentative microstructure. Note, however, that if problems arise it is always possible to restore the original data and start the noise reduction process again.
Figure 2(b) is a grain boundary plus Inverse Pole Figure (IPF) map of the same area. In this map non-indexed regions are displayed in white. The absence of a predominant colour indicates that there is no dominant texture. Analysis of the data in the map showed that porosity (or amorphous phase) makes up 6.6% of the total area and, in addition, the mean grain size is 3.4 µm with a maximum of 16.4 µm.


Figure 3: (a) Grain size map; (b) legend from (a) showing the colour scheme used and the grain size distribution, and (c) contoured {0001}, {11-20} and {10-10} pole figures.
Figure 3(a) shows a grain size component map with the corresponding histogram and colour key shown in Figure 3(b). In this map the grains are coloured according to size based on the equivalent circle diameter (other options are available in Tango), i.e. the diameter of a circle having the same area as the grain. One grain (coloured red) has a larger diameter than the majority which are less than 5 µm. Figure 3(c) shows the contoured {0001}, {11‑20} and {10‑10} pole figures from the area of the sample mapped in Figure 3(a).
These confirm that there is no dominant texture present in the sample.
Conclusions
Careful sample preparation enables good quality EBSD data to be collected from an alumina insulator. This data provides valuable information on the microstructure of the material – in particular grain size and distribution, texture and porosity.
Data of this type can be valuable in the production of these materials as it is fundamental in controlling or defining the properties of the insulator.
References
- G Vander Voort, Metallography Principle and Practice, ASM International, Metals Park, Ohio, USA, 1984.