Freshly formed Martensite has been described as a body centered tetragonal structures under high microstresses,  and greatly super  saturated with carbon or containing very disperse carbides.Reheating of martensite would be expected  to permit a reversion to  the stable bcc lattice. On the basis of age-hardening scenario , precipitation  of carbide particles followed by growth of agglomeration of these carbides , as permit ed  by temperature and time. Upon tempering or drawing at successively higher temperatures for a constant time of  1 hr, the whole process, for carbon steel, can be divided up into four stages: (150 to 230 oC), (230 to 400 oC), (400 to o650C)  and 

 (650 to 723 oC)  At this stage continued growth of carbides brings them to a size may be resolved  at  X500, although their form better to be seen at X1000. They appear as globular particles  in a continuous ferrite matrix, which now becomes visible as a separate phase for the fist time. The structure is called , and it naturally soft, from C20-5, tough, and gummy to machine.  Note: Spheroidal carbide may also be obtained by heating pearlite in the same temperature range, since carbides tend to assume to he shape having least surface  area  or lowest energy. A long time ia required to spherodized pearlites, abd the minimum  size  is limited by the size of the lamellae from which they developed.

The title of your post/question is ill posed, but the body of your post reveals more clearly what issues you are trying to clarify.

Anyway, if someone comes here looking just at the title of your question, and is wanting basic information on that question, the following might help.

Annealing of steel conventionally involves reheating steel into the austenitic phase field, then furnace cooling. This slow  cooling generally allows diffusional transformation to more equilibrium-like structures, typically a mixture of ferrite, and pearlite, (a lamellar mixture of ferrite and cementite). Cementite is of course not a true equilibrium phase in the iron-carbon system, the equilibrium phase would be graphite, cementite is actually a metastable phase. Some common alloying elements can stabilise graphite, e.g. Si and Ni.

The non-equlibrium phases termed martensite and baintes can only form from austenite, not from each other.

However, what you are really asking is quite a longstanding diificult question that has been challenging metallographers, ever since these phases were originally described, i.e. differentiating between, martensite, tempered martensite, and the various bainitic forms, both as freshly formed, and after a tempering treatments.

A former colleague of mine, Dr G. Thewlis, worked on some aspects of this problem and tried to come up with a classification scheme for such microstructures (which IIRC was a modification of a pre-existing WI scheme):-

http://www.maneyonline.com/doi/pdfplus/10.1179/026708304225010325

Another colleague, Dr D. Hanlon, was involved in an RFCS project acronym PHAST which had similar goals re phase identification/quantification.

You can get the report from the EU bookshop.

http://www.google.co.uk/url?url=http://bookshop.europa.eu/mt/quantitative-phase-analysis-of-multi-phase-steels-phast-pbKINA22387/downloads/KI-NA-22387-EN-S/KINA22387ENS_002.pdf%3Bpgid%3DIq1Ekni0.1lSR0OOK4MycO9B0000FJ6sKlIN%3Bsid%3DhDuKbiGAi3KKTHXi2N8UyUOlJvj2PiZTgug%3D%3FFileName%3DKINA22387ENS_002.pdf%26SKU%3DKINA22387ENS_PDF%26CatalogueNumber%3DKI-NA-22387-EN-S&rct=j&frm=1&q=&esrc=s&sa=U&ei=tHHYVN6EMJDraK2SgvgO&ved=0CCMQFjAD&usg=AFQjCNFipBLUZA1rvWD0NPxAGhyHCpJ7XQ

You also mention EBSD, as this is often the technique of choice for phase identification and classification these days, and in fact there has been at least one European RFCS project trying to arrive at schemes for phase identification in steels via EBSD - Micro Quant, again we collaborated with other major companies, in an EU part funded RFCS project.

M. P. Aarnts, R. A. Rijkenberg, F. A. Twisk, D. Wilcox, M. J. Zuijderwijk, A. Arlazarov, D. Barbier, L. Germain, M. Gouné, A. Hazotte, D. Huin, B. Krebs, L. Barbé, J. Hagström, J. Komenda, S. Zajac, W. Bleck, M. Fischer, F. Gerdemann, F. Friedel, V. Marx, C. Commenda, E. Füreder, R. Hackl; Microstructural quantification of multi-phase steels (MICRO-QUANT) Contract No RFSR-CT-2006-00017 (2009) Full report can be downloaded from https://bookshop.europa.eu/en/home/

The bottom line is that these phases are difficult to distinguish, especially when tempering treatements potentially have been applied, despite the advent of the newer techniques. In an optical metallography setting the physical appearance of the phases varies depending on the carbon an alloy contenet of the steel.

So I understand your real question, but there still is not an easy answer to it, even with the techniques available today, including the computed reconstruction methods (for the parent austenite phase). Familiarity and long experience with particular steel grades, in its various states, coupled with hardness data does help.

As Orgutani mentions, on reheating pearlitic structures to sub-eutectoid temperatures, the lamellar pearlitic structure with its high surface/interfacial area can start to spheriodise, giving globular carbides in a ferrite matrix, which has a lower overall free energy. One can consult classic papers if one wants to understand why such lamellar structure form during cooling in the first place from the parent phases.

Dear John !

Thanks a lot for your detailed message and particularly for references. Indeed, I poorly formulated the title, anyway tempering (in alpha-state) should be mentioned rather than annealing !

As to challenging situation with identification of coexisting martensites and bainites, I agree that only EBSD is not enough, particularly when tempering is involved.

Currently we try to supplement EBSD "defectness" (GAM or KAM) with independent crystallographically sensitive characteristics and hopefully are close to success.

You will be the first to get preprint, if any.

Thanks a lot again for your deep insight in my question !!

Alexander