Electron backscatter diffraction measures one orientation at a time. The beam sits on a point of a polished surface, the pattern that comes back is indexed against a list of candidate phases, and the result is a phase and an orientation for that point. Step the beam across a grid and you have a map: hundreds of thousands of orientations, each with a position.
That combination is what makes EBSD different from a diffraction experiment. A pole figure tells you which orientations are present in a volume; an EBSD map tells you which orientation is present where. Every question about microstructure - grain size, neighbourhood, gradients inside a grain, how a boundary is oriented - needs that "where", and this chapter is about getting it and trusting it.
plottingConvention.default('y↑→x');
mtexdata forsterite silent
% colour every Forsterite measurement by its orientation
plot(ebsd('Forsterite'),ebsd('Forsterite').orientations,'micronbar','off')
The colour is not a measurement. It is a colour key, a choice of how to turn a three-parameter orientation into three numbers you can see, and different keys make the same map look quite different. Nothing here can be read quantitatively off the colours alone.
What the data actually is
A ebsd variable is a list, not an image. Each entry carries a position, a phase and an orientation, and the entries happen to sit on a grid because that is how they were measured. This matters more than it sounds: filtering the list to one phase leaves gaps rather than blanks, and any operation that needs neighbours has to reconstruct the grid first.
Two kinds of entry are not ordinary measurements. Points whose pattern could not be indexed are given the notIndexed phase - they are recorded, not missing, and they matter, because a band of them along a boundary is information about the boundary rather than an inconvenience. Points outside the scanned region are not in the list at all.
The first thing to check is the reference frame
An EBSD file contains two coordinate systems that need not agree: the one the stage moved in, and the one the Euler angles were written in. Vendors differ, software versions differ, and nothing in the file forces them to be consistent.
The failure mode is unkind. A map with the wrong convention still plots, still reconstructs into grains, and still produces pole figures - they are simply rotated or mirrored, and there is no way to tell from the numbers alone. Reference Frame is therefore not an advanced topic to come back to later. Read it once, on your own data, before you trust anything downstream.
Where to start
Import then Reference Frame, in that order and without skipping the second.
Plot covers displaying a map, and IPF Maps the colour keys behind the picture above. Orientation Plots leaves the map behind and shows the same orientations in pole figures instead. Sharp Color Keys and Advanced Plotting are for when the default colouring hides what you are looking for.
Select and Select by Index cut a map down - by phase, by region, by property, by condition. Most real analysis starts with one of these.
Grain Reconstruction is the step from measurements to objects and the gateway to Grains. ODF Estimation is the other way to summarise a map, as a density over orientations rather than as a set of regions.
Real data is imperfect, and several pages address that. Denoising and Filling Missing Data repair it; Pseudo Symmetry deals with phases whose patterns are nearly indistinguishable under more than one orientation, which is an indexing failure rather than a noise problem. Square and Hex Grids, Regridding and Interpolation, Maps and Images and Spatial Transforms concern the geometry of the grid itself, and TrueEBSD Distortion Correction corrects the distortion that comes from measuring a tilted sample.
Three pages measure orientation change within grains rather than between them: KAM compares each point with its neighbours, Mis2Mean / GROD compares it with its grain's mean, and Profiles follows a line across the map. Orientation Analysis collects the rest.
Simulation generates maps with known answers, which is the honest way to test a processing chain. HDF5 Interface and Export handle files.
Next
Grains and Grain Boundaries are what a map becomes once its regions are found. 3D - EBSD extends all of this to volumes. The orientations themselves are Orientations, and their distribution is ODF.