EBSD could reveal subgrains.

Jianguo Wang, from my experience it shows grain boundaries, while I need to see the subgrains. Should I use any specific EBSD mode?

Alright, SEM–based EBSD microscope via any software (e.g., Channel 5 [HKL technology]) allows to identify and quantify, as well the majority of sustructure components (i.e., LABs, HABs and TBs) after a specific etching procedure (related to each kind of examined material).

through heat treatment which depend on the chemical composition of austentic stainless steel

EBSD or TEM is capable. You may see the subgrain boundaries in SEM or OM, but it is difficult to distinguish them from normal grain boundary. If the misorientation angle of the subgrain boundary is below 2 deg., you cannot reveal it clearly by EBSD either. Then you should go to TEM.

EBSD is the easiest to cover a relatively large field of view.

EBSD data, a bit better processed as now in standard software can do this very nicely. Of course the question arises which precision do you expect.

Have a look at the attached link. It shows it for quartz but it should be comparable for stronger scattering materials, like steel.

Conference Paper EBSD: more reliable and precise

Oksana Klok, EBSD will reveal the local orientation at a pixel. To show your subgrain boundaries, I suggest that you record EBSD with the highest spatial resolution possible. Do not clean up the data. Plot the IPF maps which will show you the gradient in orientation within a grain. You can later plot extra boundaries with various orientations angles. I suggest assigning a different color for each range. You may start by 2-5 degrees (white), (6-10) yellow, then 10-15 deg (blue) and >15 deg (black). The selection of the ranges will depend on the amount of deformation you have. There are multiple ways to identify the boundaries automatically. I do not know which analysis software you use to recommend the suitable method. Good luck! 

@ A. Salem: I guess it was clear that O. Klock is interested in subgrain boundaries, i.e. smaller than 2 degrees. You can do this only roughly with EBSD because of the orientation-specific Hough indexing which is excellent for fast but not for precise band detection. Form this follows, that this is OK for texture measurements but has strong limitations regarding misorientation analyses focusing on small angular differences. Therefore, an investigation of these kind of misorientation is only in first approximation acceptable, but for more precise measurements you need a higher orientation precision technique. Do you agree?

Best greetings! G. Nolze 

@ G. Nolze, I agree that looking for disorientation less than 2 deg ( angular accuracy for commercial EBSD) will require different tool. The key is the definition is subgrains. With EBSD, that definition can be a user input. As a first "rough" analysis, I ask my students to exhaust all possibilities they have in hand before seeking different and more expensive data recording. Once there is data confirming that they need more resolution, then I support the investment of time and resources. Thank you for the note!

@ A. Salem: the definition of a grain using EBSD data is quite tricky and also very misunderstandable. The grain recognition compares the orientation of adjacent pixels only. As long as they are below a misorientation limit both pixels are assumed to belong to the same grain. On this way you can get accumulated misorientations in the scale of tens of degrees if adjacent pixels do not show a bigger misorientation angle. But it is still worse since a grain boundary is only then defined as grain boundary if all adjacent pixels fulfill this condition. Take 20 pixels along a ring or another line, and for sake of simplicity all have a misorientation angle of 1.5°, then the twentieth pixel has a misorientation of 30° compared to the first, but also adjacent pixel. The software will define all pixels as part od one grain since there is a connection between these pixels which fulfills the condition of being smaller than 2°, although the misorientation to the other adjacent (first) pixel is 30°. No grain boundary will be drawn! With GAM and KAM you can see this but anyway no grain boundaries are derived. Good software has tools which also show incomplete grain boundaries, i.e. these procedures simply show all misorientations within  a grain as long as they show a bigger misorientation than a certain angle. I am not sure which software you have and whether you can visualize these type of small-angle boundaries.

Nevertheless, the major problem is the inaccuracy of the band detection using Hough transform, cf. the link given above. 

@ G. Nozle, I totally agree with your summary of some of the current widely used tools. The reasons you mentioned and the inability of many of those software tools to handle large EBSD datasets ( 10 million and higher) were the motivation to build our own software tools that we just commercialized as a cloud computational tools to handle EBSD datasets of 50-100 million orientations. Ref (TiZone and Ti-texture on www.ICMRL.com). We are slowly adding more apps every couple of weeks based on customer demands. The issues you raised above are inherent for using spatial domain to identify boundaries.  To address this, we have developed an app (for Ti alloys for now)  to capture primary alpha particle fragmentation into grains/subgrains using our own algorithm that will be lunched for our subscribers to use soon. If you have an idea of a unique algorithm and you like to deploy on our cloud (i.e Microstructructure Informatics Cloud= MiCloud), please let me know. We have a procedures to work with you to share with the world. Thank you for echoing our frustration with the limitations of some of the standard EBSD analysis tools of the shelve. Regards

Ayman

@ A. Salem: Do you know Mtex? It  is a free-available Matlab toolbox for EBSD as well as XRD data. 

http://mtex-toolbox.github.io/

@ G. Noize. Mtex is a great free tool to conduct many of standard EBSD data analysis. We have used it in many occasions. Our industrial needs outgrown MTEX. Therefore, we had to write our own. In particular, we had a patented algorithm for calculating the ODF using the generalized spherical harmonics (GSH) for 100s of millions of EBSD orientations very efficiently which enabled us to develop many other industrial tools for inserting EBSD data into simulation and modeling software such as crystal plasticity finite element analysis. Thank you for the note about MTEX!   

Using TEM, EBSD.

Please see the following papers and the Ph.D. thesis by R.K. Bhargava :- (i) "Microstructure and transient creep in an austenitic stainless steel", O. Ajajaa & A..J. Ardell, Philosophical Magazine A  Volume 39, Issue 1, 1979, pp. 65-73. (ii) "POWER-LAW BREAKDOWN AND THE DISLOCATION MICROSTRUCTURE IN TYPE 304 STAINLESS STEEL", M.E. KASSNER, MATERIALS LETTERS, Vol. 2, 1984, pp. 451-454. (iii) PhD Thesis: R.K. Bhargava, "The influence of substructure on the elevated temperature deformation of an austenitic alloy", Univ. Cincinnati, 1975 (the work is on AISI 304 stainless steel). (iv) Book "Fundamentals of Creep in Metals and Alloys", By Michael E. Kassner, In Chapter 2, Section 2.8, page 28.

(v) Review: "Grain and subgrain characterisation by electron backscatter diffraction", F.J. Humphreys, Journal of Materials Science, August 2001, Volume 36, Issue 16, pp 3833-3854.                 (vi) "Subgrain Structures Characterized by Electron Backscatter Diffraction (EBSD)", S. Bunkholt, K. Marthinsen, E. Nes, Materials Science Forum, Vols. 794-796, pp. 3-8, Jun. 2014. (vii) Book  "Introduction to Texture Analysis: Macrotexture, Microtexture and Orientation Mapping", By Valerie Randle, Olaf Engler, CRC Press, 2000. 

Also, please see papers published by  D. Caillard and J. L. Martin in Acta Metallurgica (1982, 1983, 1984) and in Materials Science and Engineering, 81 (1986) 349-354, D. CAILLARD (but this paper in MSE is on Aluminium) as well as papers by R. Lagneborg group (Swedish Institute for Metal Research, Stockholm, Sweden).

For Prof. Alan J Ardell on RG and and his work/papers on dislocation link length distribution during elevated temperature deformation:-https://www.researchgate.net/profile/Alan_Ardell/publications?sorting=newest&page=3

Best wishes.

Dear Oksana!

Just weak substructures ( tool instability). Thus,  while waiting for more efficient tools, we has to make most use of smart post-processing raw data. See for instance doi: 10.1016/j.msea.2007.06.005