Journal of the Virtual Explorer

Any type of sample can be analysed by electron backscatter diffraction (EBSD) so long as the sample is crystalline and hard enough to polish well enough to produce a topographically flat surface. After field mapping has been completed and orientated hand specimens have been collected or laboratory experimental samples have been produced, the most appropriate reference frame needs to be chosen to study the sample. A common geological reference frame for deformed rocks is the kinematic reference frame where samples are cut perpendicular to foliation and parallel to lineation. This is also referred to as the XZ reference frame. For laboratory experimental rock deformation samples it is more applicable to work in a stress-related reference frame, such as using the compression direction as the important external parameter. A common reference frame for metallurgical research is related to the rolling, normal and transverse directions that correspond to how the metal was processed.

It is vital to choose the most applicable reference frame for the study before preparing EBSD samples. It is important to keep linear features, whenever possible, parallel to either the long or short axis of the sample (depending upon shape) as this makes data interpretation easier later on. The overall size and shapes of samples that can be studied using EBSD depend upon the size and shape of the SEM EBSD holder, but most standard holders will normally take standard thin sections (30 µm thick slices of rock glued onto glass slides that are 4.5 x 2.5 cm) and blocks (normally rectangular or 1 inch diameter spheres), although oversized holders allow larger samples to be studied. EBSD is a surface technique, with the signal originating in the top 50nm of the sample (Lloyd, 1987a) and any damage to the crystal lattice in this surface layer will decrease the quality of the electron backscatter pattern (EBSP). EBSD can be performed on broken surfaces (this is common where samples do not polish well) but most laboratories work on specially polished sample surfaces.

After samples have been cut and made into standard polished blocks or thin sections, they will have normally have been polished using 3µm, 1µm and either 0.5µm or 0.25µm diamond paste. The surface of the sample will still exhibit topography at this point, any surface topography, scratches or attached particles will cause shadowing due to the high tilt angle (70°) and will reduce the quality of the EBSP. This final damage layer must be removed by an advanced polishing technique such as chemical-mechanical polishing (geological specimens), electro-polishing (metals), chemical etching (metals) or ion beam milling (metal and geological specimens). For geological specimens the most common technique used is chemical-mechanical polishing using either a colloidal silica suspension (0.02-0.06 µm) (Flynn and Powell, 1979; Lloyd, 1987b) or 0.05µm alumina suspension. The important differences between the two products are the grain shape and pH. Colloidal silica is more rounded in shape and has a pH of ~10, whereas the alumina suspension has a platy shape and a pH of ~7. Different brands of each product will have slightly different properties. For instance Buehler produces two colloidal silica suspensions, namely (1) MasterMet which is a 0.06μm amorphous silica colloidal suspension with a pH of ~10; and (2) MasterMet 2 which is a 0.02μm non-crystallizing amorphous silica colloidal suspension with a pH of ~10.5. Which suspension to use depends upon the bulk mineralogy of the thin section and the aims of the research. If part of the aim of the research is to quantify geochemical variations, as well as textural variations, Al is often one of the important constituent elements studied and if the sample is polished with an alumina suspension some of the polishing particles can become embedded into cracks, holes and in the sample surface and will create errors in any geochemical data collected. This chemical-mechanical polishing process results in very slow abrasion rates of 1-2 µm per hour or ~ 1 atomic monolayer per second (Lloyd, 1987a).

There are many different types of polishing machines and polishing pads available. Polishing pads can be made from cloth that is stretched out over a disc on the polishing machine to pads with an adhesive back that stick onto the base disc. There are a few main firms that produce quality equipment and consumables such as Buehler, Struers and Kemet. For EBSD purposes vibratory polishers, such as the VibroMet2 produced by Buehler give the best results. When selecting a suitable final stage polishing cloth or pad, choose items that are described as chemically resistant, durable pad ideal for rough and final polishing of hard steels, ceramics and composites for use with chemical-mechanical polishing (CMP) or for use with activated suspensions.

When chemical-mechanically polishing, the sample requires systematic checking to prevent possible problems such as ‘lifting’ (thin sections) or plucking of grains (thin sections and blocks). ‘Lifting’ in thin sections is a term used to describe when the rock slice begins to peel away from the glass slide and the polishing fluid gets in between the rock slice and glass slide. This can be a common occurrence if using thin sections which were produced a long time ago, as the adhesives used to prepare the thin sections may have degraded and were not as high quality as those used in thin section preparation today. If lifting does occur, cease polishing immediately since ‘lifted’ sample sections cannot be studied using EBSD and the sample will have to be used as is. As a rough guideline, quartz dominated samples usually require polishing for one hour on a vibrating polisher using colloidal silica fluid. The softer the mineral the less time is required for final stage polishing.