EBSD · Start here

EBSD terms explained: what every number on your map means

Every point on an EBSD map holds a pattern, a phase, an orientation and a few quality numbers. This page follows one point from its pattern to the finished map. Each term gets one small thing to try. Links go to the longer pages when you want more.

What do you want to know?

How does the software turn a pattern into an orientation?

Hough transform. Finds each band as a straight line, then matches the angles between bands to a crystal.
Dictionary indexing. Compares your pattern with a library of simulated patterns and keeps the best match.
Spherical indexing. Paints the pattern on a ball and compares it with a simulated ball at every possible turn at once.

Method

Step

Try it: press Play all steps for each method. All three end at the same grain; they get there in different ways.

In real life: Hough indexing is fast and is the default in most programs. Dictionary and spherical indexing use the whole pattern, so they keep working on noisy patterns and look-alike phases, but they take more computer time. Their answers are usually refined afterwards.

Go deeper: how Hough indexing works, step by step · the four methods compared on one map · when the answer is wrong

Can I trust this point? IQ, CI and MAD

IQ / band contrast. How sharp the picture is: how strongly the bands stand out. EDAX says IQ, Oxford says band contrast, Bruker says pattern quality.
CI (confidence index). How clear the vote was: (votes for the winner minus votes for the runner-up) divided by all possible votes. 0 to 1. EDAX.
MAD / fit. How well the answer fits: the mean angle between the bands found and the bands the answer predicts. Oxford and Bruker say MAD, EDAX says fit.

Start from

Blur (damaged or strained surface)

Noise (short exposure)

Second grain under the beam

Votes for a look-alike answer

Band lines off the bands

IQ: the picture-

CI: the vote-

MAD: the fit-

In real life: the three numbers answer three different questions, so each can be good while another is bad. IQ drops with a damaged surface, strain or a boundary. A low CI means the vote was close, not that the answer is wrong; points under about 0.1 are often treated as unreliable. MAD stays low when a wrong answer fits the bands well, so look at all three.

Go deeper: where CI and MAD come from · reading IQ as data · why a bad polish lowers IQ

What phase is it? Phase maps and the candidate list

Phase. One crystal structure with its cell size, such as bcc ferrite or fcc austenite. EBSD sees the structure through the bands, not the chemistry.
Phase map. Each point coloured by the phase that fitted best. The software only chooses from the phases you load.

Phases loaded in the software

Ferrite (bcc) Austenite (fcc) Martensite (bct) Nickel (fcc)

How stretched martensite is (c/a)

Colour the map by

Try it: click a grain on the map. Then untick Austenite, or tick Nickel, and watch the map.

In real life: two phases with the same structure (fcc austenite and fcc nickel) give the same bands, so EBSD cannot tell them apart; EDS can. A phase that is not loaded is never found: its points are left empty or given to a look-alike phase. Martensite in most steels is so close to bcc that Hough indexing calls it ferrite.

Go deeper: phases that EBSD gets wrong · using IQ to find martensite

Which way does the crystal point? Euler angles, axis-angle and quaternions

Euler angles (φ1, Φ, φ2). Three turns in a fixed order that take the sample axes onto the crystal axes. Most programs use Bunge's order: about Z, then the new X, then the new Z.
Axis-angle. The same turn written as one rotation: an axis and an angle.
Quaternion. Four numbers: cos(half the angle), then the axis times sin(half the angle). Programs use them inside because turns are easy to compare and average.

φ1: first turn, about Z

Φ: second turn, about the new X

φ2: third turn, about the new Z

Euler angles (φ1, Φ, φ2)-

Axis-angle: axis [u v w] and angle-

Quaternion (q0, q1, q2, q3)-

In real life: a cube looks the same after 24 different turns, so one grain can be written with 24 sets of Euler angles. That is why two programs can print different numbers for the same grain. Programs also differ in their conventions (Bunge or Roe angles, how the sample axes are named), which can mirror a map or pole figure.

Go deeper: reference frames and conventions · the cube over the sample, coloured

How is an orientation turned into a colour? IPF maps and Euler maps

IPF-X, IPF-Y, IPF-Z map. Colours each point by which crystal direction lies along one sample axis (X, Y or Z), using the red-green-blue triangle. Z is usually the surface normal.
Euler map. Turns the three Euler angles straight into red, green and blue. Every angle counts, but the colours jump.

Colour the map by

Tilt the clicked grain about X

Euler angles of this grain-

Same grain in IPF-X, Y, Z-

Try it: click a grain, then drag the tilt slowly. In the strips on the right, the IPF colour drifts; the Euler colour jumps.

In real life: IPF maps are the easiest to read, but each one uses only one sample direction, so two different grains can share a colour. Check IPF-X or IPF-Y as well, or look at a pole figure. Euler maps use all three angles but cannot be read by eye.

Go deeper: IPF-X, Y and Z on one map, and grains that share a colour

How different are two grains? Misorientation and disorientation

Misorientation. The turn that takes one grain onto the other: an axis and an angle.
Disorientation. The smallest of the equivalent misorientations. A cube has 24 symmetry turns, so there are 24 angles to check; the smallest wins. Boundary maps use this one. Its largest possible value for cubic crystals is 62.8°.

Turn grain B about

by this angle

Misorientation, as stored-

Disorientation (smallest of 24)-

Try it: set 85° about [001]. The cubes look 5° apart, and the disorientation says 5°. Then try 60° about [111]: that is a Σ3 twin.

In real life: boundaries are drawn from the disorientation. A common rule: 2 to 15° is a low-angle boundary, above 15° a high-angle boundary, and below about 2° is too close to the noise to trust.

Go deeper: boundary thresholds, Σ3 twins and the random-grain histogram

Is this grain bent? KAM, GROD, GOS, GAM and GND

KAM (kernel average misorientation). For each point: the average turn to its nearest neighbours, skipping neighbours above a cutoff (often 2 or 5°) so boundaries do not count. Older Oxford software calls it local misorientation.
GROD (grain reference orientation deviation). For each point: how far it has turned from its grain's average orientation.
GOS (grain orientation spread). One number per grain: the average GROD of all its points.
GAM (grain average misorientation). One number per grain: the average KAM of all its points.
GND density. Worked out from KAM: geometrically necessary dislocations per square metre, about α × KAM (in radians) ÷ (b × step). α is 2 to 4; this page uses 2.

The grain

Colour by

Step size

KAM, this point-

GROD, this point-

GND, this point-

GOS, whole grain-

GAM, whole grain-

Cutoff for KAM5°

Try it: pick Flat, with one bad point. GROD flags only the bad point; KAM also lights up its four neighbours. Then pick Bent and switch the step size: GND changes even though the bending did not.

In real life: KAM and GND show where the lattice bends from point to point, so they pick out deformation near boundaries and particles. GROD shows how far each point has drifted from its grain's average. GOS is often used to split recrystallised grains (GOS under about 1 to 2°) from deformed ones. Always report the step size, the kernel and the cutoff with a KAM or GND map.

Go deeper: why KAM depends on step size, and the noise floor

Do the grains line up? Texture, pole figures and the ODF

Texture. Grains sharing similar orientations instead of pointing every way.
Pole figure. A map of where one crystal direction (here the three cube axes) points in each grain, seen from above the sample.
ODF (orientation distribution function). How crowded each orientation is, counted in all three angles. 1 means as crowded as a random sample (multiples of random, m.r.d.).

How much the grains agree

Strongest spot in the ODF-

Grains within 15° of the shared orientation-

Try it: drag from 0 to 100. The cubes line up, the dots in the pole figure gather into spots, and the dots in the box gather into one cloud.

In real life: rolled sheet, wire and printed parts usually have texture, and it changes how stiff and strong they are in each direction. A pole figure shows one crystal direction at a time; the ODF holds all three angles, so it is the full answer. Peaks of a few to a few tens of m.r.d. are common in rolled metals; the value depends on how the ODF is smoothed.

Go deeper: build a texture grain by grain, and read ODF sections

Which grain slips first? The Schmid factor

Schmid factor. m = cos φ × cos λ. φ is the angle between the load and the slip-plane normal, λ the angle between the load and the slip direction. The largest possible value is 0.5.

φ: tilt of the slip plane

Turn the slip direction within the plane

φ (load to normal)-

λ (load to slip direction)-

Schmid factor m-

In real life: for a chosen load direction, EBSD software works out m for every slip system of each grain (12 for fcc metals) and maps the largest. Grains with m near 0.5 usually start to slip first, as long as their neighbours let them.

Go deeper: strengthening and slip · how grain orientation changes hardness

One grain, five maps: how do the measures differ?

The same bent grain and its neighbours, drawn five ways. Move the pointer (or a finger) over any map: the cross follows on all five and each map gives its number for that point.

IPF-ZWhich way it points (Euler angles). -

KAMBend from point to point. -

GRODDrift from the grain average. -

GOSSpread of the whole grain. -

IQHow sharp the pattern is. -

Questions people ask about EBSD terms

What is the difference between KAM and GROD?

KAM compares each point with its nearest neighbours, so it shows local bending. GROD compares each point with the average orientation of its whole grain, so it shows how far the point has drifted overall. A steadily bent grain has a low, even KAM but a GROD that grows toward its edges. One bad point shows in GROD at that point only, but in KAM it also raises its neighbours, because they are compared with it.

What is the difference between misorientation and disorientation?

Misorientation is any turn that takes one grain onto the other. Because a crystal looks the same after its symmetry turns, there are many equivalent ones; for cubic crystals they come down to 24 different angles. The disorientation is the one with the smallest angle. Boundary maps and misorientation histograms use the disorientation.

What do IPF-X, IPF-Y and IPF-Z mean?

Each one colours a point by the crystal direction that lies along one sample axis: X, Y or Z. IPF-Z (the surface normal) is the one most often shown. The three maps of one sample look different, and you need more than one to pin down an orientation.

What CI value can I trust?

A common rule is to treat points with CI below about 0.1 as unreliable. One study (De Vincentis and Field, 2021) found about 90% of solutions correct at CI of 0.1 or more when at least 8 bands were used. CI only measures how clear the vote was, so look at IQ and the fit too.

Are KAM and local misorientation the same thing?

They are the same idea: an average misorientation between a point and its close neighbours. EDAX, Oxford AZtecCrystal and Bruker call it KAM; older Oxford software (Channel 5) called it local misorientation. Kernel size and cutoff can differ, so state them when you report a value.

What GOS cutoff marks a recrystallised grain?

Many papers count grains with GOS below 1 to 2° as recrystallised. Some use 3° for steels or up to 5° for dynamic recrystallisation. The right cutoff depends on your noise level, so check the GOS of a grain you know is undeformed first.

Why do two programs give different Euler angles for the same grain?

A cube looks the same after 24 turns, so the same grain has 24 valid sets of Euler angles; each program picks one by its own rule. Programs can also use different sample axes or Euler conventions. Compare disorientations, not raw Euler angles.

Is GND density measured by EBSD?

Not directly. It is worked out from the bending between neighbouring points (KAM or the full lattice curvature), divided by the Burgers vector and the step size. It counts only the dislocations needed to bend the lattice, and it inherits the noise of KAM, so a smaller step can raise it even when nothing changed.

Where does this connect?

References

Show the 13 references

H. J. Bunge. Texture Analysis in Materials Science: Mathematical Methods. Butterworths, London, 1982.
D. Rowenhorst, A. D. Rollett, G. S. Rohrer, M. Groeber, M. Jackson, P. J. Konijnenberg and M. De Graef. Consistent representations of and conversions between 3D rotations. Modelling and Simulation in Materials Science and Engineering 23 (2015) 083501. doi:10.1088/0965-0393/23/8/083501.
J. K. Mackenzie. Second paper on statistics associated with the random disorientation of cubes. Biometrika 45 (1958) 229-240. doi:10.1093/biomet/45.1-2.229.
D. P. Field. Recent advances in the application of orientation imaging. Ultramicroscopy 67 (1997) 1-9. doi:10.1016/S0304-3991(96)00104-0.
N. S. De Vincentis and D. P. Field. Factors affecting confidence index in EBSD analysis. Ultramicroscopy 225 (2021) 113269.
S. I. Wright and M. M. Nowell. EBSD image quality mapping. Microscopy and Microanalysis 12 (2006) 72-84. doi:10.1017/S1431927606060090.
S. I. Wright, M. M. Nowell and D. P. Field. A review of strain analysis using electron backscatter diffraction. Microscopy and Microanalysis 17 (2011) 316-329. doi:10.1017/S1431927611000055.
M. Calcagnotto, D. Ponge, E. Demir and D. Raabe. Orientation gradients and geometrically necessary dislocations in ultrafine grained dual-phase steels studied by 2D and 3D EBSD. Materials Science and Engineering A 527 (2010) 2738-2746. doi:10.1016/j.msea.2010.01.004.
L. P. Kubin and A. Mortensen. Geometrically necessary dislocations and strain-gradient plasticity: a few critical issues. Scripta Materialia 48 (2003) 119-125. doi:10.1016/S1359-6462(02)00335-4.
F. J. Humphreys. Review: grain and subgrain characterisation by electron backscatter diffraction. Journal of Materials Science 36 (2001) 3833-3854. doi:10.1023/A:1017973432592.
Oxford Instruments. EBSD Explained: From Data Collection to Analysis. 2015.
MTEX documentation: Grain orientation parameters (GOS, GROD, GAM) and Schmid factor. mtex-toolbox.github.io.
W. C. Lenthe, S. Singh and M. De Graef. A spherical harmonic transform approach to the indexing of electron back-scattered diffraction patterns. Ultramicroscopy 207 (2019) 112841.

Cite this page: Tripathy, Manisha. “EBSD Terms Explained: IPF, KAM, GROD, Euler Angles, CI and IQ.” untethered atom, 2026, https://untetheredatom.com/ebsd/ebsd-terms-explained.

BibTeX

@misc{tripathy2026ebsdterms,
  author = {Tripathy, Manisha},
  title  = {EBSD Terms Explained: IPF, KAM, GROD, Euler Angles, CI and IQ},
  year   = {2026},
  howpublished = {\url{https://untetheredatom.com/ebsd/ebsd-terms-explained}},
  note   = {Interactive web page}
}