MICRESS Forum

Hi Bernd,

I am trying to simulate the reverse Austenite formation via Micress. Since the Martensitic transformation is NOT equilibrium and is not a diffusion controlled transformation, it's not possible to simulate the whole process including martensite formation and then tempering to create the reverse austenite in boundaries. (right?)
MY question ist how can I define different diffusion rate for the C atoms inside the phases and into the boundary? As far as I see, this is only possible to define for the phases, not for the boundaries. What comes to my mind is; I can define the diffusion coeficients and surface tension energy between phases along with defining a potential nucleation position at boundary for the new phase in which move of the carbon atoms to boundary and creating the new phase would be easier than move of carbons into the phases, so instead of growth of existing phases, I probably will see creation of new nuclei of austenite at boundary. Does it make sense? What do you feel about not using this procedure, and using keyword +b instead? From your viewpoint which one fits better?
Moreover, is there any phase standing for Martensite in the Thermocalc database, so that I could generate my desired Ges file?

Regards,
Omid.

Dear Omid,

It is clear that martensite is not an equilibrium phase, and this is also a reason why it is not included in thermodynamic databases. In certain cases and with respect to certain properties it may be appropriate to consider martensite as an "undercooled" bcc, i.e. to define it as bcc with a strong driving force offset and altered diffusivities. Nevertheless it seems useless to describe the martensitic transformation in terms of a moving boundary, because that is definitively wrong.
Unfortunately, I do not know much about the reverse transformation to austenite, and I have no idea what are the observations and the mechanism of what you call "reverse austenite formation". Can you explain?
From a technical point of view, using the +b option allows for decreasing or increasing the activation energy for diffusion inside an interface relative to the bulk of the phase, and thus to increase or decrease diffusivity. However, I cannot tell you at present whether this could be helpful for your simulations or not...

Bernd

Dear Omid,

According to the paper you sent me per PM (L. Yuan et al. / Acta Materialia 60 (2012) 2790–2804), I understand that for the reverse austenite formation martensite is considered thermodynamically as ferrite which is supersaturated in carbon due to its history. They calculate the driving force for its growth as if were ferrite.
Thus, you just could treat reversed austenite in MICRESS as a second austenite phase which nucleates either at interfaces between martensite grains or between martensite and retained austenite. You should not define any phase interaction between retained austenite and martensite (or set mobility to zero), otherwise this phase also would grow inside the martensite.
What they say about adsorption of carbon at interfaces does not seem reasonable to me. At least it would not stabilize formation of reverse austenite, because carbon activity cannot be increased by equilibrium adsorption. Adsorption cannot be treated in MICRESS up to now.

Bernd