It would help if you would explain the abbreviations which are perhaps not clear to everybody. Moreover, already their names could help you where the differences are? Do you know, what KAM, GAM and GOS in your software are doing? Which software you are using?

My question is now reshaped to "In EBSD analysis, how to use KAM, GAM and GOS to estimate extend of deformation in grain?" I know what the software is doing but do not know what place which of these parameter is suitable for estimating what deformation?   

Look, until now EBSD only measures the local orientation. Nothing less but also nothing more. The idea to use these local orientations for other applications later arised since first it became obvious that one can link the local position AND the orientation to extract additional but still not evaluated information like the definition of grains as "volumes" of identical orientations. Especially for deformed materials this is still a quite problematic task since the condition "same orientation" is very soft. And also the conoditions for grains in a non-deformed material this is not easy to answer and default values are a simple number and have no real practical background. More tricky is the investigation of misorientations within a grain since there are two major problems: a) the reference frame, and b) what do we actually see? The first part is relatively easy. The orientation of a selected orientation, the orientation of an "ideal" orientation", the average orientation of a grain, the average orientation of a "kernal", the orientation of the former parent grain, the average orientation of a triple junction, or anything else. The number of possible reference orientations might be endless and ends up in a mathematical reanalysis of your data. The second one is more challenging since it refers to some physical or chemical meaning. Of course, you can insert some models like geometrically necessary dislocation in order to describe the visible rotation but this concept is a very simple one, and might match for some measurements the observed orientation deviations. Unfortunately, nature is commonly much more complicated than a simply extrapolation. I don't doubt that for specifically prepared samples containing disclocations of a well-defined character one can generate correlations between KAM, GAM or GOS, but you need to consider all parameters which may influence these specific results. GOS, KAM and GAM have so many interferences starting from sample preparation, over focusing the beam and operation of the SEM until the intrinsic factors of the microstructure that I don't expect a simple dependence. Their difference is only the reference orientation.

In hcp materials you can even use the IGMA (in-grain misorientation axes) to evaluate which slip system is most likely to introduce the misorientation of the grain. Published by Chun...

What do you mean by "extent of deformation"? is it strain you want to measure, stress ...? Using KAM you might be able to get an idea about geometrically necessary dislocations which are responsible for heterogeneous deformation. Measuring elastic strain is normally more complicated and needs calculations or external software.

But even KAM will not give you precise values since as Gert mentioned there are many factors involving. It is more of a qualitative measurement rather than a precise quantitative measurement....

Mr. Ashish Saxena, to be able to relate your "deformation" to your spatially changing orientation data (EBSD), I suggest that you establish what you are looking for theoretically first before trying to extract the correlations from experiments. I suggest that you start by a theoretical square grain (n x n pixels of same g(phi1, PHI, phi2)). Apply a deformation ( I will leave the definition for you based on the same definition on your question). Estimate the new evolved local orientations in the hypothetical square crystal ( new gi : i=1 to n2). Use the definition of KAM, GAM, and GOS (TSL documents) to calculate them for the deformed hypothetical square grain after the applied "Deformation". Build the correlation between your deformation ( scalar, vector, matrix ?? based on how you define it) and the calculated KAM, GAM, and GOS. Once you established that,  you can implement on your EBSD data by exporting Grain1 and Grain2 files from TSL which will give you grain ID with various orientation information within each grain. There are so many variables and assumptions in your question and I believe the efficient way to get help is to formulate it mathematically before you try to implement it on experimental data. Good luck!

I presume you mean plastic deformation which develops dislocation microstructures and local misorientations -which can then be measured in some way by these KAM etc methods. However I do not think there is any direct relation between KAM, GOS etc and plastic strain. Dislocation structures do not only depend on amount of plastic strain, they vary with material (through stacking fault energy) with temperature, with strain rate and even with strain mode (tension, rolling) and grain orientation. Ideally the local plastic strain tensor is required but there is little hope of getting this. You could aim to obtain a simple, empirical and approximate, measure of strain by doing some calibration EBSD experiments with your material deformed under your conditions. Good luck!

Dear Ashish Saxena,

You can find some details in my papers that are available on ResearchGate. I used the  EBSD to estimate a fraction recrystallized  and deformed volume of material by a local misorientation approach (GOS-based) and a pattern degradation approach (IQ - analysis).

Kind regards,

Wojciech

https://www.researchgate.net/publication/272478927_Electron_Backscatter_Diffraction_Study_on_Microstructure_Texture_and_Strain_Evolution_in_Armco_Iron_Severely_Deformed_by_the_Differential_Speed_Rolling_Method

https://www.researchgate.net/publication/263765640_EBSD_and_X-ray_diffraction_study_on_the_recrystallization_of_cold_rolled_Ni3Al_based_intermetallic_alloy

https://www.researchgate.net/publication/262527213_Evolution_of_crystallographic_texture_and_strain_in_a_fine-grained_Ni3Al_%28Zr_B%29_intermetallic_alloy_during_cold_rolling

Article Electron Backscatter Diffraction Study on Microstructure, Te...

Article EBSD and X-ray diffraction study on the recrystallization of...

Article Evolution of crystallographic texture and strain in a fine-g...

WHAT you may get from  an invented measure of plastic strain is hopefully clear from the above discussion. As J.H. Driver hinted, I would also recommend use of reference specimens subject to certain strain magnitudes. 

More challenging is WHICH EBSD-based parameter is most rational to map ! 

Indeed, GAM or kAM are somewhat rough estimates, whereas the more refined GND density as restored from EBSD data involves several assumptions and apply we

well only to rather extended features under indentors, near second phase particles etc. meanwhile the desired measure should be free of assumptions, rigorously related to lattice defects, and sensitive to even subtle substructural objects. Such is div(Q) where Q is the local orientation VECTOR. If feels very low angle (les then 0.5 deg) dislocation boundaries in particular. See doi: 10.1016/j.msea.2007.06.005 in MSE-A 2008.

I will make just a short comment based on our experience: We used KAM parameter as a qualitative measure of strain build-up/relief during annealing and partially during fatigue tests. But to be able to judge something from our data we had to make the experiments in a "site specific" way, i.e. we observed the same site before and after we did something to the specimen. Making our experiments like this, we were able to judge the local increase or decrease of the level of strain (we rather use the vague term "lattice distortion").But we were really far from making any quantification other than sorting based on the extent of some change - the reason being explained above.

We chose KAM since it is less dependent on the assumption of your "grain tolerance angle". Using KAM, you have to be aware of its definition and namely its dependence on the step size of your EBSD data. Since you told you use TSL system, you can then make full use of the spacing of datapoints in a hexagonal grid  - the KAM kernels will be symmetric and of equal radius from the centre point.

Nice article about various quantities to judge strain based on EBSD data is this one: 10.1017/S1431927611000055

Very interesting discussions. 

Although, there is often a good correlation with the plastic strain and the misorientation. We found no correlation between strain and mean misorientation (albeit surface strain) for different grains deformed to a particular strain. We find a loose correlation with mean misorientation and maximum Schmid factor or grain size. Taylor factor, Taylor rotation, orientation spread have no correlation.

I suspect misorientation differences are determined mostly by local neighborhood effects (i.e. surrounding grains) which mean that different parts of the grain deform differently. I hope to try to eliminate this effect (CPFEM, averaging many grains etc) to quantify the other effects. 

Be good to know more if there are theories that exist between EBSD, plastic strain and plasticity models to understand this area.

Tom

IGMA

One interesting feature of deformation slip in hcp crystals is that basal slip should be visible as rotation around the axis, while prismatic slip should be seen in EBSD patterns as rotation around the -axis. 

For more detail see Chun's paper from 2010: "Distribution Characteristics of In-Grain Misorientation Axes in Cold-Rolled Commercially Pure Titanium and Their Correlation with Active Slip Modes" 

http://link.springer.com/article/10.1007%2Fs11661-010-0410-4#/page-1

Interesting article, thanks Arnas. 

If it's the intra-grain deformation you're after, I'm not convinced you can get that information using misorientation metrics. We'd all love it if it worked, but I've seen three papers trying to validate that, and unfortunately all the signs are suggesting that it doesn't work. I don't think there is any theoretical basis for it working at this size scale, while there are both arguments and experimental data against.

The upside is that we managed to show a pretty decent correlation between local KAM and local strain hardening (note: not local deformation). The correlation is not quantitative. In particular, low KAM does not mean that there is little or no hardening. A proper interpretation is that the whole grain has hardened, and the regions with high KAM are likely to have hardened slightly more.

Article: 10.1016/j.msea.2016.07.123

For plastic strain estimation I'd recommend to average your KAM/GAM/GOS over at least an area corresponding to ~10 grains, preferably more, especially if you know that the material is highly strained. For this I would recommend GOS since it seems to be insensitive to EBSD step length, orientation measurement scatter and average grain size in the sampled area. KAM on the other hand is highly sensitive to these factors.

Where can I find additional information about GAM in the software channel 5?

Dear Ashis,

GOS concerns a little more global disorientation i.e., for a given measurement point and all others from one "EBSD grain", GAM and KAM concern only neighboring measurement points. KAM - applies to all neighbors , GAM only those lying in the same "EBSD grain". These parameters can be easily used to compare different microstructures as at work: :A. UNIWERSAŁ, et all, Materials Characterization  2016 vol. 118, s. 575–583. However, the network of measurement points and the EBSD grain definitionmust be matched to the problem you want to attack

Yours

Mirek

I recommend Gert's answer

Best Regards

Can anybody suggest me why we get low KAM value i.e. less than 1 degree and which is reducing with deformation? What can be the reason for that? Suggest me some experimental papers. Please

KAM shows the misorientation of one data point from its neighbouring data points within a Kernel. GAM average out these point-to-point misorientations within a grain and color the grains by the average misorientation value. On the other hand, GOS displays the total misorientation in a grain from one end to other end and color the grains accordingly. If you are interested in the strain localization along the grain boundary, low-angle boundary, or any other defects, then KAM is the best choice. Other two are grain-wise data and do not contain much of local information within a grain. However, the strain partitioning (in terms of extent of deformation) in two different grains belonging to different phases can suitably be showed by GAM or GOS. GOS is also used to quantify the percentage of recrystallized grains in the microstructure.

Thank you Mr. Arka Mandal for your answer. But my question is different. I did not ask about KAM, GOS and GAM analysis. I asked why KAM is reducing with deformation at room temperature but not each and every time. Sometime, it is increasing or decreasing as well with strain. And I have highly recrystallized sample. So, if you can suggest me regarding this then please go ahead.

I obtained a histogram from GAM. I do not know what is the meaning of "Number" in the Y-axis?!!!!

I attached a picture.

Hi Arka Mandal

Thank you.

There is a problem. We cannot observe these numbers of of grain in the our image!!!

Hi Arka Mandal,

Yes, I am using the same step size throughout experiment. The material I am using is single phase FCC material and it is highly recrystallized having ~ 500 micron grain size.

Still thank you so much for your answer. I will see for the points suggested by you.

Hi Sandhya Verma,

Since your grains are huge in size, please check the reported KAM values are based on area covering a sufficient number of grains which is statistically significant. Compare the areas showing high KAM values with the phase map to check if there is any preferential strain-induced transformation. It is also possible that your scanned area in a higher-strained sample does not contain these transformed hard microstructural constituents but that in a lower-strained sample does. This would also reflect in your result. However, it is hard to identify the real cause without looking at the data.

Thank you so much for your reply. It is difficult for me to discuss like this. Can we have discussion separately? Can you please give me your email id or You can ping me at email id: [email protected]

Please do if possible for you.

Thanks for your response. Tomorrow I will drop you an email. Thanks again.