Page 1 of 1

Nucleation on selected grain boundaries

Posted: Wed Dec 11, 2019 11:12 pm
by billyzhangubc
Hi all,
I am simulating austenite formation during intercritical annealing from a ferrite matrix. Nucleation of austenite occurs predominantly at ferrite grain boundaries. Based on my experimental observations, certain ferrite boundaries will not get nuclei.

I wonder if there is a way to simulate this in MICRESS. So far, by tuning shield distance, I am able to get maybe 1-2 nuclei on some grain boundaries, but I cannot have one boundary segment that does not have any nuclei.

Thanks,
Billy

Re: Nucleation on selected grain boundaries

Posted: Thu Dec 12, 2019 8:34 pm
by Bernd
Hi Billy,

the question is whether there is a way to distinguish this boundary segment from others, e.g. by specific misorientation, composition etc. Or is it a region which could be selected by a box?

If there is nothing which is different compared to other interface regions, you could try to put some difference there. If it is an interface between two specific grain, for example, where you do not want to nucleate, you could use "add_to_grain" as a trick to change certain properties of the two grains (phase number, orientation, reX-energy), which can be addressed by the nucleation parameters, before nucleating, and change them back afterwards...

What is special with these interface segments in your case?

Bernd

Re: Nucleation on selected grain boundaries

Posted: Sat Dec 14, 2019 6:13 pm
by billyzhangubc
Hi Bernd, thanks for your reply.

So far, everything is based on experimental observations (microscopy). All these grain boundaries have misorientations >15 degrees and there shouldn't be significant difference in composition (my material is microalloyed steel so it will be difficult to experimentally quantify segregations as well).

I will try the "add_to_grain" option first. Is this under "data for further nucleation"?

Billy

Re: Nucleation on selected grain boundaries

Posted: Mon Dec 16, 2019 12:29 am
by Bernd
Hi Billy,

yes, it is under "data for further nucleation". You can find more information here.

Bernd

Re: Nucleation on selected grain boundaries

Posted: Sun Jul 05, 2020 11:58 pm
by billyzhangubc
Hi Bernd,
Sorry to bring this up again. With "add_to_grain", it possible to place nuclei on selected boundaries based on boundary misorientation angle (for example, only boundaries with <15 degrees misorientation angle)?

Thanks,
Billy

Re: Nucleation on selected grain boundaries

Posted: Mon Jul 06, 2020 7:03 pm
by Bernd
Hi Billy,

unfortunately, there is still no specific nucleation option in MICRESS which restricts nucleation to certain misorientation values...

If you would have said you want to have nucleation only on ferrite-ferrite boundaries with >15° misorientation, this would be what I think is to be expected from a physical point of view. Accordingly, you would have at least 2 option how this can be achieved (indirectly) in MICRESS:

1.) You define a misorientation relationship for ferrite-austenite such that there is a high enough interface mobiliy (or low enough interface energy) only for high angle boundaries to allow austenite growth. Then, you nucleate with parent relation of 0°, thus that only for high-angle boundaries of ferrite-ferrite there is a high-angle growth condition of austenite vs. ferrite.

2.) If you would set nuclei of austenite with an analytical curvature model, but with an undercooling which is not sufficient to overcome the critical curvature, then ferrite-ferrite high angle boundaries would better stabilize these nuclei because they have a higher interface energy which allows wetting. Thus, having a suitable relation of the chemical driving force and the critical radius, one could achieve to allow growth only for the nuclei on high angle ferrite-ferrite grain boundaries while those on the other boundaries would vanish.

If you really observe the opposite (i.e., preferred nucleation on low-angle ferrite-ferrite boundaries - maybe I understood wrongly), it would be astonishing and unusual (see for example here), and there must be a specific reason for that, like different composition or dislocation density. The first choice then always should be to understand the mechanism behind and check whether it can be taken into account in MICRESS (e.g. by simulating the process which produced your microstructure (gamma-alpha transformation, recrystallisation).

If this is not possible, and you still want to have MICRESS to set austenite nuclei only on the low-angle grain boundaries, you still can use add_to_grain as a "dirty trick" to reverse the interface energies of the ferrite-ferrite grain boundaries directly before austenite nucleation (by switching all ferrite grains to a fake ferrite phase with these ugly behavior of higher interface energy for low-angle misorientation), and revert it afterwards when your unusual nucleation has happened...

Bernd

Re: Nucleation on selected grain boundaries

Posted: Mon Jul 06, 2020 7:55 pm
by billyzhangubc
Hi Bernd,
Thanks for your explanation and I'm sorry for the confusion. I meant to say >15 degrees. I am trying to do some proof of concept studies to introduce a bainitic initial structure and put austenite nuclei on high angle grain boundaries. I'm glad to know that's possible and the rest is for me to introduce misorientations and use some sort of multiscale tessellation to introduce sub-boundaries.

Thanks,
Billy