How do I prepare a sample for EBSD?

Run a disciplined conventional sequence with polycrystalline diamond down to 1 micron, then an extended chemo mechanical finish: 5 to 15 minutes of colloidal silica on a chemical resistant pad, or one to several hours on a vibratory polisher for difficult materials. EBSD reads only the top 10 to 50 nanometers, so removing all residual deformation in the final step is the entire point.

Why are my EBSD patterns not indexing?

Low band contrast and unindexed pixels across the surface almost always mean residual subsurface deformation or surface contamination. The fix is more chemo mechanical final polishing, not software cleanup, plus fanatical cleaning, since polishing residue, carbon contamination, and fingerprints sit exactly in the signal escape depth.

Can I use alumina instead of colloidal silica for EBSD prep?

Yes, when the sample chemistry rules out silica, such as carbonate bearing materials or alloys that stain at pH 10. Use fine deagglomerated alumina with a longer polishing time, ideally on a vibratory polisher, to approach the deformation free surface that silica's chemo mechanical action normally provides.

EBSD Sample Preparation: Deformation Free Polishing Guide

Beta Diamond colloidal silica polishing suspension bottle

By Sam Desai, Technical Team, Beta Diamond Products

Electron backscatter diffraction reads the crystal lattice in roughly the top 10 to 50 nanometers of your sample. Any deformation in that layer, and standard mechanical polishing leaves deformation far deeper than that, scrambles the diffraction patterns into noise. The route that works: a disciplined conventional sequence down to 1 micron diamond, followed by an extended chemo mechanical finish, typically 5 to 15 minutes of colloidal silica on a chemical resistant pad, or hours on a vibratory polisher for the most demanding materials. The final step is not a finishing touch. For EBSD it is the entire point.

Why EBSD is unforgiving in a way imaging is not

Ordinary SEM imaging looks at topography and composition; a little subsurface deformation is invisible to it. EBSD diffracts electrons off the lattice planes themselves, and the signal escapes from only tens of nanometers of depth. Mechanically polished metal carries a deformed skin, dislocations, smeared material, residual strain, typically tens to hundreds of nanometers thick even after a careful 1 micron finish. So a sample that produces gorgeous optical micrographs can return EBSD maps that are mostly black unindexed pixels. The pattern quality map does not lie: low band contrast across the surface means residual deformation, and the fix is always more and better final polishing, never more software cleanup.

The route, step by step

Run your normal sequence with extra discipline. Gentle sectioning under flood coolant, mounting, SiC grinding through 600 grit, then 9, optionally 6, then 3 and 1 micron diamond, polycrystalline throughout, because monocrystalline's deeper deformation is exactly what you are trying to avoid creating. Light pressure at every stage. Every shortcut taken here gets paid back with interest in the final step.

Then the chemo mechanical finish. Colloidal silica, 0.02 to 0.06 micron, on a soft chemical resistant pad, light force, and let it run: five minutes minimum, fifteen for harder or more strained materials. The mild alkaline chemistry dissolves the deformed skin while the soft spherical particles sweep it away, a removal mechanism mechanical abrasives cannot replicate. Flush the cloth with water for the final fifteen seconds and clean the sample immediately and thoroughly, ultrasonic in water then a solvent rinse, because dried silica residue ruins patterns as surely as deformation does. A low crystallizing silica earns its keep here, since EBSD runs are long and the cloth spends a long time loaded.

For the difficult cases, titanium alloys, zirconium, nickel superalloys, heavily deformed samples, geological materials, a vibratory polisher with colloidal silica for one to several hours produces the most consistently indexable surfaces in routine use. It is slow, but it is unattended slow, which is the good kind.

Alternatives exist for specific materials: electropolishing for many steels and nickel alloys, broad ion beam milling for ceramics and cross sections that mechanical methods torment. But for the broad run of metals, extended silica is the workhorse.

Details that separate indexed maps from black ones

Clean fanatically. Carbon contamination, polishing residue, and fingerprints all sit exactly in the escape depth of the signal. Handle finished samples by the mount only.

Mind charging. Non conductive mounts and samples charge under the beam; use conductive mounting resin, paint a grounding path, or apply the thinnest possible carbon coat, since every nanometer of coating attenuates patterns.

Mind silica pH on sensitive materials. Carbonates etch and some alloys stain at pH 10; fine deagglomerated alumina with a longer dwell, ideally on a vibratory polisher, is the fallback.

And budget honestly: the final step takes longer than every other polishing step combined, and trying to compress it is the single most common reason EBSD prep fails.

There is no single product bundle for this work, because the right combination varies by material, which is exactly why we stock the full route: polycrystalline suspensions in every micron size, low crystallizing colloidal silica, deagglomerated alumina, and the pads to run them on. Describe your material and we will spec the route, with free samples to prove it before you commit a research budget to it.