Morphology of reversed austenite

solid-solid phase transformations, influence of stresses and strains
Post Reply
AZheng
Posts: 6
Joined: Thu Jun 14, 2018 2:47 pm
anti_bot: 333

Morphology of reversed austenite

Post by AZheng » Mon Sep 24, 2018 11:50 am

Hello,

I am trying to simulate the phase transition from ferrite to austenite.
The morphology of reversed austenite i simulated is shown below(Fig.1), but it seems not good. Compared with the real morphology(Fig.2), the tips should not be so sharp and the extrudations in the middle should not appear.
Is there any parameter i could adjust to solve these problems?
Many thanks in advance!

BR
Rui
Attachments
Fig.2.PNG
Fig.2.PNG (7.85 KiB) Viewed 4005 times
Fig.1.PNG
Fig.1.PNG (60.73 KiB) Viewed 4005 times

Bernd
Posts: 1247
Joined: Mon Jun 23, 2008 9:29 pm

Re: Morphology of reversed austenite

Post by Bernd » Mon Sep 24, 2018 2:56 pm

Dear Rui,

What your simulation shows is what can be expected if we assume diffusion-limited growth austenite in a pure FeC alloy. I here assume that the simulation was numerically correct, i.e. resolution is fine enough, and the interface mobility is chosen correctly (e.g. by using the "mob_corr" option). You essentially get a dendritic shape of the precipitate.

Within this assumption of simple diffusion-limited growth, the choice of the interface energy between austenite and ferrite will have a strong influence on morphology. Increasing this parameter will lead to less sharp tips, and the side branch in the middle will also be smaller or disappear. However, the shape will always remain somewhere between dendritic and cellular.

The microstructure will also change drastically if you assume kinetic undercooling (i.e. the interface mobility is smaller than what we assume in the diffusion-controlled case), if there is an elastic contribution, or if there are other (substitutional) elements present which diffuse slowly. In the latter case, the driving force for phase transformation is strongly reduced which prevents dendritic microstructures.

If you simulate the formation of austenite at ferrite/ferrite grain boundaries, morphology will again be different and furthermore depend on the interface energy and shape of the grain boundary (wetting effect).

Bernd

AZheng
Posts: 6
Joined: Thu Jun 14, 2018 2:47 pm
anti_bot: 333

Re: Morphology of reversed austenite

Post by AZheng » Thu Oct 04, 2018 1:49 pm

Dear Bernd,

Many thanks for your suggestion.
I tried to increase the interfacial energy between austenite and ferrite, but it seemed not work. The side branch is still obvious(like the figure below).
Also, is the existence of sharp tips due to the growth along grain boundaries? Is it possible to get more rounded tips?

The part of input file is shown for your reference.

Rui Zheng
Attachments
Figure.PNG
Figure.PNG (32.14 KiB) Viewed 3988 times
REV_1h_5_33_dri.txt
(15.27 KiB) Downloaded 167 times

Bernd
Posts: 1247
Joined: Mon Jun 23, 2008 9:29 pm

Re: Morphology of reversed austenite

Post by Bernd » Thu Oct 04, 2018 6:39 pm

Dear Rui,

From the attached input file I learned that your system is ternary Fe-Al-Mn, so that my assumption of Fe-C and possible effects of substitutional elements were wrong.

I also learned that your precipitate grows along a grain boundary. My idea to assume finite interface kinetics will increase kinetic undercooling and thus decrease the effects of the grain boundary. Then, you would get a morphology which gets shorter and more round - eventually ending with a circle if interface mobility is very small. However, what you want, namely an elongated shape with rounded tips, would probably not appear.

As I understand, what you need is wetting along the grain boundary with tips dominated by the wetting angle, and subsequent (or simultaneous) lateral growth. Conditions for that would be a low driving force (i.e. small overheating), otherwise you will always get a cellular or dendritic morphology. To achieve that, you could perhaps nucleate earlier upon heating (i.e. at lower temperature) or assume a driving force offset e.g. due to stress contributions (I don't now about your thermal boundary conditions yet...).

Another scenario which would enhance grain boundary wetting despite strong overheating would be to assume grain boundary diffusion...

Apart from that, I have few comments to your attached input file:

- you use faceted anisotropy for phase 2 which seems to have no effect. This is probably due to the high overheating situation and/or to the low anisotropy coefficient defined in the phase interaction data. Anyway, from a physical point of view, using faceted anisotropy for getting the desired shape is not a preferred approach.

- you define a pre-factor on diffusion data. However, this does not work like that and should provoke at least a warning on screen; "factor" is an extra line option, i.e. you should put it in extra lines behind the definition of the diffusion coefficients:
# How shall diffusion of component 1 in phase 1 be solved?
1 1 multi gg
1 1 factor 20

Further, in terse mode, "1 0 diagonal n" can be omitted, as it is default for all diffusion terms which are not explicitly specified.

- your updating scheme for thermodynamic data leads to a first relinearisation after 1800s. This is definitively not sufficient, because even if the simulation is isothermal, compositions are changing, especially shortly after nucleation. You should have at least about 100 updates during your simulation period.

Best wishes

Bernd

AZheng
Posts: 6
Joined: Thu Jun 14, 2018 2:47 pm
anti_bot: 333

Re: Morphology of reversed austenite

Post by AZheng » Fri Oct 12, 2018 5:14 pm

Hello Bernd,

Sorry for the missing information, attached is the full input file.
I have some questions about "Conditions for that would be a low driving force (i.e. small overheating), otherwise you will always get a cellular or dendritic morphology. To achieve that, you could perhaps nucleate earlier upon heating (i.e. at lower temperature) or assume a driving force offset".

Could you please explain more about the relation between morphology and nucleation time? And if there are some references available?

Best regards,
Rui
Attachments
REV_1h_5_32_dri.txt
(18.06 KiB) Downloaded 194 times

Bernd
Posts: 1247
Joined: Mon Jun 23, 2008 9:29 pm

Re: Morphology of reversed austenite

Post by Bernd » Fri Oct 12, 2018 7:46 pm

Hi Rui,

By changing nucleation parameters, you define under which conditions nucleation occurs, including at which undercooling - at least if temperature is not constant.
For example, if you would have a heating rate of 100K/s, and you only check only at an interval of 1s, then you could easily have an extra overheating of 100K just because of the low checking frequency...

However, in your case, as I can see now, there is no heating or cooling because you have isothermal conditions. In this case, undercooling or overheating is only determined by temperature (and local composition if you have concentration gradients). In this case, of course, nucleation parameters do not affect undercooling/overheating!

Anyway, what I wanted to say is that if driving forces are very high, you will hardly see any effects of grain boundaries. For seeing strong wetting of a grain boundary, the overall driving force should not be much bigger than curvature.

Bernd

AZheng
Posts: 6
Joined: Thu Jun 14, 2018 2:47 pm
anti_bot: 333

Re: Morphology of reversed austenite

Post by AZheng » Thu Nov 08, 2018 4:30 pm

Hi Bernd,

Thanks a lot for your explaination and it is helpful. While there is another question.
The austenite should have smooth interface (lateral growth) but the result shows dendrite-like morphology shown below. Is there any solution to get smooth interface?

Best regards,
Rui
Attachments
REV_2h_185_dri.txt
(17.47 KiB) Downloaded 137 times
a1.PNG
a1.PNG (226.17 KiB) Viewed 3601 times

Bernd
Posts: 1247
Joined: Mon Jun 23, 2008 9:29 pm

Re: Morphology of reversed austenite

Post by Bernd » Sun Nov 11, 2018 5:45 pm

Dear Rui,

As I can see from your input file, you reduced the interface energy of the 1/1 and 1/2 interfaces by ~4 orders of magnitude - that is why your interfaces get unstable!
Did you have a specific reason for that?

Bernd

AZheng
Posts: 6
Joined: Thu Jun 14, 2018 2:47 pm
anti_bot: 333

Re: Morphology of reversed austenite

Post by AZheng » Mon Nov 12, 2018 1:29 pm

Dear Bernd,

Yes..I know the interface energy is small but if not, the austenite will not become bigger and show dendritic shape. I have no idea about how to make austenite grow along inerface(lateral growth)....
Any suggestion will be appreciated.

Best regards,
Rui

Bernd
Posts: 1247
Joined: Mon Jun 23, 2008 9:29 pm

Re: Morphology of reversed austenite

Post by Bernd » Tue Nov 13, 2018 2:46 pm

Hi Rui,

What you say is not logic: The higher you chose the interface energy of the 1/2 interface, the more it will grow along the grain boundary (as long as 1/1 and 2/2 interface energies have a similar or slightly higher value) and be less dendritic.
The problem is that above a certain value of the interface energy 1/2 phase 2 will not grow at all because of a too high curvature undercooling. You should prevent that by increasing the critical radius for nucleation of phase 2 (to about 0.1 µm or even higher) and by choosing the interfacial energies at a suitable level.

Bernd

Post Reply