There are a number of ways to reduce or even eliminate charging on insulators:
Ensure that there is no surface topography
Ensure that the polish is very good
Only turn the beam on once the sample has been tilted to 70°
Coat with gold before the final polishing stage – this will fill cracks and voids with a conducting network
Work in variable pressure / low vacuum mode (if your SEM has this capability): ideally in the pressure range of 10-50 Pa. Any higher and the signal of the diffraction pattern will be too weak
Work at higher speed so that the beam does not stay in one area for a long time
Work at lower probe currents and/or accelerating voltages
Prepare a conductive tract from the sample to the stub/holder – using some conductive paint or metallic tape
But sometimes you will have to apply a conductive coat...
The key point for Electron Backscatter Diffraction (EBSD) coating is to keep the coat very thin, typically in the range of 2-5nm, if it is too thick, the signal to noise ratio will decrease significantly and result in very poor EBSPs. If the coating is too thin, there won’t be sufficient conductive material to dissipate the charge. Ideally the coat should be carbon either sputtered or evaporated onto the sample, but it is possible to use other coating materials such as gold or tungsten. In cases where the coat is a little too thick, it may be possible to obtain good EBSPS by increasing the accelerating voltage to penetrate through the coating.
Typical effects observed when samples are charging:
Variations over time in signal intensity, typically seen in SE images but can also be visible in the EBSD data
Charging can cause beam deflections which can appear as sample drift, resulting in a mismatch between initial electron images and EBSD mapping data, including change of grain shapes and size
Charging can lead to discontinuities in EBSD maps caused by either large beam deflections or sudden changes in build up charge near the mapping area
Electron image from uncoated Zirconia sample. To the left is shown the secondary electron image and to the right the orientation contrast image based on the FSD. Charging effects are clearly visible in the SE image and less so in the FSD image.