You might mean “Does martensite in carbon steels describe a phase or a microstructure in the scope of material science?”

Well, it is complex.

Just like Bainite, Martensite can be formed without carbon precipitation. Actually, even better, martensite can be formed without carbon or interstitial atoms. Phase is defined in thermodynamics as states of matter that have uniform physical properties. For me, martensite is a phase and has uniform physical properties, not only in steel but in titanium alloys, if the observation window is small enough.

Martensite can be well treated/modeled as a phase in materials engineering and materials physics. Because martensite can be well described macroscopically, microscopically, or statistically by certain independent parameters.

“If it is a phase, why we cannot find it in databases?”

Because it is metastable and it is diverse. Due to the undefined heat treatment history, martensite differs from each other irrelevant to the chemical compositions, mechanical and crystallographic properties.

I would say we knew far from enough about martensite.

Rigorously, solid phases differ in the lattice type and/or stechiometric formulas (for compositions). In this sense, steels have only gamma (austenite) and alpha phase, the latter appearing as SUBTYPES of ferrite, martensite or bainite. Besides, there may be Fe3C (cementite) and other dispersed (carbo-ntride, sulphide, ...) phases, of course. On the other hand, when distinct interfaces i.e.essential difference (dislocation density, solute carbon amount, etc) between neighboring MICROSTRUCTURAL units can be objectively recognized by appropriate local methods, an extended (non-rigorius) understanding of phase concept "helps" one to focus on certain types of transformed crystals in the complex transformed structure. Conditional character of the latter approach become clear, in particular, in tempering. Indeed, the diffecence between as quenced martensite and bainite gradually vanishes in tempering, where both the dislocation density and carbon (carbide) distribution become more and morer similar, whereas BCC /BCT/ lattice type remains invariant.

The definitions of martensite and phase are possibly subjective, so this answer is not definitive, but is intended to generate some discussion and clear my own understanding.

In undergraduate metallurgy we were taught that a phase is a physically distinct, chemically homogeneous state of matter. I find this definition is quite close to the IUPAC definition: https://goldbook.iupac.org/html/P/P04528.html

(I suppose 'chemically homogeneous' is broad enough to include compositional gradients, but the basic stoichiometry should be retained)

By this definition, martensite is definitely a phase. Bainite would fail because it is a mixture of ferrite and carbide, i.e. of two phases. The same is true for pearlite. So bainite and pearlite would be microstructures, but martensite would be a phase. It may be worth noting that the difference comes from the physical phase boundary, not from metastability or dislocation accumulation (which would change with tempering).

Coming to databases, the Crystallography Open Database (COD) lists plenty of martensites, one example of which is:

http://www.crystallography.net/cod/1524199.html

I'm not sure how accepted COD is vis-a-vis ICSD, but I felt its worth a mention.

As for martensitic transformations in other materials, I think the common feature is that they are all diffusionless transformations. So the nomenclature of 'martensite' is a matter of convenience, since it affords a common definition.

so...it sound like: martensite describes more some kind of microstructure than a "classic" phase, right?

In a current investigation I observe that the typical lath-like microstructure "somehow" remains but a huge amount of dislocations disappear? Is this then still martensite, or does it make more sense to call it already ferrite?

Hi Aashrant, martensite only appears in the title of the publication but not as phase name, at least for your example from COD.

If two phases are "practically" identical, it is usually one and the same phase. I only know from meteorites that taenite and kamacite are the same mineralogical names for austenite and ferrite, but in that case I hitherto assumed this is a peculiarity of the meteorite guys (like for cementite and cohenite....also the same but extraterrestrial).

Isn't this a nice example to overthink the classical definition of the term phase, i.e. not only the crystal structure and the chemistry (at similar environmental conditions like pressure) are essential?

Sorry, that may not have been a good example. Instead, CIF: 1532986 might be better:

Article The Crystal Structure and Magnetic Properties of Fe3Pt Marte...

With reference to the microstructure posted, I would go by the crystal structure rather than the dislocation density; if it is still BCT I'd call it martensite, but if it reverts to BCC then ferrite. One objection to the use of dislocation density-based classification is that we then have to go by a qualitative difference - how high is high? Then again, factors like prior processing (in addition to quenching) could change the dislocation density without changing the crystal structure.

Of course, to be fair this is somewhat true of BCC/BCT too- how much tetragonality (in terms of c/a maybe) classifies it as BCT? Still, this seems a little more consistent to me.

Its nice to hear about taenite/kamacite, so much learning happens here. :)

In a strict thermodynamic sense, a phase is defined as a “macroscopically homogenous body of matter”. This definition can be further broken down to: a phase is the region of microstructure which is "physically distinct, chemically homogeneous and mechanically separable".

Following the definition, martensite is simply a phase. Now when the as quenched martensite is subjected to tempering (in steels), the carbon supersaturation is released resulting in the decrease of tetragonality, recovery, etc. I am assuming the microstructure shown is of steel.

Therefore, I am bending to call the microstructure a mixture of martensite (since it is still distinct with respect to the conventional ferrite) and carbide. Most importantly, martensite forms through displacive mechanism while the ferrite is diffusional.

Perlite and Bainite are microstructures, because they are a combination of 2 distinct phases.

Martensite, I think, refers to the kind of transformation. In maraging steels, for example, the martensite is just like ferrite (cubic), since there is almost no carbon in these steels. In carbon steels, carbon is the responsible for the tetragonality of martensite (the martensitic transformation is so fast that carbon becomes "trapped"). Maybe that is the reason why the term “martensite” is not used in databases. We may have distinct martensites depending on the alloy composition.

You might mean “Does martensite in carbon steels describe a phase or a microstructure in the scope of material science?”

Well, it is complex.

Just like Bainite, Martensite can be formed without carbon precipitation. Actually, even better, martensite can be formed without carbon or interstitial atoms. Phase is defined in thermodynamics as states of matter that have uniform physical properties. For me, martensite is a phase and has uniform physical properties, not only in steel but in titanium alloys, if the observation window is small enough.

Martensite can be well treated/modeled as a phase in materials engineering and materials physics. Because martensite can be well described macroscopically, microscopically, or statistically by certain independent parameters.

“If it is a phase, why we cannot find it in databases?”

Because it is metastable and it is diverse. Due to the undefined heat treatment history, martensite differs from each other irrelevant to the chemical compositions, mechanical and crystallographic properties.

I would say we knew far from enough about martensite.

Dear Jian Feng, if you find martensite in steel but also in Ti-alloys than it is from my point of view no phase since it has another chemical composition (unequivocal). However, this is more related to my imagination. I assumed that a phase is not only chemically uniform but also unique, i.e. one cannot use the name austenite for Ni, only because it is also fcc.

From your explanation I would tend more to the name describing a microstructure, what do you think?

Dear Gert,

A good example is topologically close-packed phase’s family, they are not defined by chemical compositions. Another example is gamma double prime phase, it is very important in both Ni-based and Ti-Al based alloys. It is not unique in only one material.

We come back to martensite.

Martensite is not defined by chemical compositions or mechanical properties, but the transformation type and the product of the transformation. Martensite and its subcategories are clearly defined in certain condensed matter. When we call plate martensite in steels, or everyone knows what the meaning is. And so is martensite in titanium alloys. But if you say the name austenite for Ni, nobody can get it.

Dear Jian, TCPs and gamma' are structure types but no phases (as far as I understand their definition).

Martensite originates from the field of iron alloys. However now it is used to describe any product of martensitic transformation. You can find many systems that undergo martensitic (displacive) transformation. For example all shape memory alloys undergo martensitic transformations.

So in other words if BCC iron based solid solution is a product of diffusive transformation it is called ferrite (there are also different types of ferrite that are not listed as different phases) and when it is a product of dispacive (shear) transformation it is called martensite. In some cases it differs crystallographically in some cases not exactly or it is not possible to distinguish it from ferrite.

There are grades of steels called "dual phase" and microstructure of those consist of ferrite and martensite.

In your case the lath type microstructure will remain until recrystallization temperature is reached but other diffusion transformations occurred already like carbides precipitation (that are present on lath boundaries in your sample). I would call it "microstructure of tempered martensite".

Dear Grzegorz, for me as crystallographer it really sound irritating :-). There is only one phase if the crystal structure is (practically) identical. Well, the difference between theory and practice. Comparable with single crystal superalloys which cannot be single crystalline because of a) changing chemistry, b) i millions of dendrites with slightly varying orientations, and c) the existence of two phases. Nevertheless, they are technically single crystalline. The same happens with martensite (to me). But thanks you for your explanation. I think, I can live with that., even if the tempered martensite is so perfect that it fits much better to a recrystallized ferrite.

Dear Gert,

Definitely, martensite (found in many materials, not only in steels) is a completely different phase.

Best regards.

Prof. Jose SAN JUAN

@ Jose SAN JUAN: No offend, you can call it phase if you like, but until now all definitions proposed are identical to ferrite. At least I do not realize anything which is different (chemical composition, crystal structure, ). In contrast, we found out that martensite does not have to be BCT, it does not have to contain carbon, it even does not have to contain Fe since it nowadays stands for a special kind of transformation forming a very characteristic microstructure called martensite. Therefore, I was asking this question. Actually, there is no reason to define a new "phase" martensite...except of the situation that it is separately noted in phase diagram. We also do not call a deformed ferrite by another name in order to discriminate it from a recrystallized. Also the Widmanstätten structure is not described by another phase but refers to ferrite (maybe it is somewhere martensitic?). Only because nowadays martensite is mainly used as phase name does not mean that it is correctly used. And I see this opinion confirmed by the missing phase description in crystallographic databases which are hopefully generated from specialists. As far as I know, also the former Metal structure database (from NIST, since about 10 years included in the ICSD) did not contain martensite as phase.

I have to admit that it is not essential whether it is a phase or not. Everybody can call it how does he / she like, but it only shows that there is some kind of discrepancy about terms or their use, and we simple use them as we "believe". I know this from the terms structure and lattice which are mixed as well, although their difference is quite clear.

Dear colleagues,

Lets talk about steels. Did someone measure the XRD pattern of a martensite structure? I think yes. What did you see there? Depend on the C content one can see patterns of ferrite+cementite+ledeburite etc. These are phases. Did someone detect some distinct peak pattern of a "martensite"? If yes = it is a phase. If No = please read further. That is why it cannot be considered as a "new" phase. Let someone correct me. I am not so familiar with steels, but rather with Zr alloys I would also add the following. After hydrogenation and cooling from beta region of the Zr-H phase diagram you can see something like Widmanstätten type of structure (sometimes it looks the same as martensitic structure in steels). After hydrogenation and cooling in alpha region - unchanged initial grain structure with additional hydrides. Absolute different appearance it has the same phases on XRD pattern, namely alpha Zr and two phases of Zr hydrides. It allows me to conclude that if no one was able to detect some distinct XRD peaks for a "new phase" - it means no "new phase" present. Unltil someone finds new peaks for the "Martinsite" it is a martensitic structure.

Ready to discuss further.

Best wishes,

Anton

When I entered graduate school I asked the head of the laboratory this question in the early days of my studies. This surprised me a lot; no base contains a martensite structure. The head of the laboratory thought for a long time about the answer and said the following - the martensite structure is so famous that scientists decided not to add it to the database :-)

Unfortunately this is not a joke but the truth.

We possibly do not like this but science is often only an opinion..or a belief..until the next belief :-).

In my opinion, the only unambiuguos cryterion is the lattice type, if any.

In solid steels the phase transitions in the true sense are mutual

transformations of ALPHA (bcc/bct) and GAMMA (fcc) phases; let alone

cementite and pearlite. Thus, ferrite, bainite and martensite should be

better treated as subtypes -- structural species -- of ALPHA (bcc//bct) phase.

Yet, I think the discussion is somewhat theological because materials science

(as well as real life in general) is not at all a subject to definite wording

or dogmas. Even if there exist alternative terms, the matter is whether or not

you properly understand and effectively use them in simulating natural phenomena.

Yes, it is far from elegance and certainty of mathematics, however it is not

resonable to spend time and enegry in fighting "wind mills".