Data for further nucleation

dendritic solidification, eutectics, peritectics,....
Post Reply
CharMIC
Posts: 211
Joined: Mon Aug 21, 2017 5:01 pm
anti_bot: 333

Data for further nucleation

Post by CharMIC » Mon Sep 04, 2017 9:14 am

Hej
I have a question about "Data for further nucleation".
Lets assume a case as follows.
Consider a solidification scenario. Solidification will results two phases, namely 1 and 2. Phase 1 will appear first during the solidification. In the experiments phase 2 will only appear in the interdendritic regions due to the favourable concentration of segregated elements. So when I try to model this with MICRESS, if I understood correctly, I have to specify the following for the phase 2 seed types (I am only curious about the nucleation sites)
……..
# Type of 'position' of the seeds?
# Options: bulk region interface triple quadruple [restrictive]
interface
# Phase of new grains (integer) [unresolved|add_to_grain]?
2
# Reference phase (integer) [min. and max. fraction (real)]?
0
# Substrate phase [2nd phase in interface]?
# (set to 0 to disable the effect of substrate curvature)
1
…….
My question is,
If I do not know the nucleation site prior to the simulations (purpose of doing the simulation is to know where phase 2 will appear before doing experiments), do I need to specify the position as “bulk” without "restrictive" instead of “interface”.

Best Regards

Bernd
Posts: 1179
Joined: Mon Jun 23, 2008 9:29 pm

Re: Data for further nucleation

Post by Bernd » Mon Sep 04, 2017 11:22 am

Hey CharMIC,

in principle, your reasoning is correct. However, I want to point out the following:

- your approach still does not include all possible nucleation sites. You could also have nucleation of phase 2 inside phase 1, which would require a further nucleation type.

- without any experimental information (or good intuition) it is impossible to predict microstructure formation. Nucleation parameter are a priori unknown but can greatly determine the type of microstructure which will be formed (in simulation as well as in reality)

- even if you already know that phase 2 will nucleate in the segregated liquid between the dendrite arms of phase 1, it may still make a difference whether you use interface or bulk nucleation. In the first case, you will get eutectic-like growth while in the second case you rather will end up with blocky precipitates.

Bernd

CharMIC
Posts: 211
Joined: Mon Aug 21, 2017 5:01 pm
anti_bot: 333

Re: Data for further nucleation

Post by CharMIC » Mon Sep 04, 2017 1:21 pm

Hej Bernd,

Thanks for your reply.

I was not able to fully understood your 3rd point.

"....even if you already know that phase 2 will nucleate in the segregated liquid between the dendrite arms of phase 1, it may still make a difference whether you use interface or bulk nucleation. In the first case, you will get eutectic-like growth while in the second case you rather will end up with blocky precipitates."

If possible could you explain little bit more why you get such difference, eutectic like and blocky..

Chamara

Bernd
Posts: 1179
Joined: Mon Jun 23, 2008 9:29 pm

Re: Data for further nucleation

Post by Bernd » Mon Sep 04, 2017 2:26 pm

Hey Chamara,

the difference is just that with "bulk" nucleation it is possible to create seeds inside the interdendritic liquid which are still far enough away from the dendrite so that they can grow independently (divorced eutectic growth). These particles can reach a certain size and stay blocky (e.g. round) before they eventually touch the dendrite.
However, if you decide that they start growing already at the interface ("interface" nucleation type), they are forced to form a coupled eutectic microstructure from the beginning.
Of course, if there is only very little liquid between the dendrite arms, or spatial resolution is very poor, there may be no big difference...


Bernd

CharMIC
Posts: 211
Joined: Mon Aug 21, 2017 5:01 pm
anti_bot: 333

Re: Data for further nucleation

Post by CharMIC » Wed Nov 01, 2017 1:42 pm

Hej Bernd,

I have a question related to defining nucleation temperature ranges (min and max).

Depend upon the segregation of the elements, the min and max will be different for different solidification conditions. To avoid that I can put 0 and 2000 K for example for min and max. But this will cause loss in performance.

Instead, my thinking is looking at phase diagram and put the possible precipitation temperature for the phase nucleation. And later systematically change the min and max to see how it will affect the phase nucleation.

I would like to get your opinion on this so that I will not do some mistakes :)

BR
Chamara

Bernd
Posts: 1179
Joined: Mon Jun 23, 2008 9:29 pm

Re: Data for further nucleation

Post by Bernd » Wed Nov 01, 2017 5:24 pm

Dear Chamara,

this is exactly the way it should be done. However, the phase diagram gives you global equilibrium information which may be good for solid state precipitation or formation of primary phases. During typical solidification processes with formation of interdendritic precipitates, segregation leads to a different situation which is more close to what is predicted by the Scheil Model. Thus, for solidification, it is better to do a Thermo-Calc Scheil simulation to get the nucleation interval. In the first run, you should widen it a bit to be on the safe side. Afterwards, it can be narrowed to optimize for performance.

Bernd

CharMIC
Posts: 211
Joined: Mon Aug 21, 2017 5:01 pm
anti_bot: 333

Re: Data for further nucleation

Post by CharMIC » Thu Nov 02, 2017 8:57 am

Hej Bernd,

If I am interested in the precipitates that precipitates (eg. Laves phase in Ni superalloys) at the inter-dendritic regions, what type of position do I have to select? My guess is Bulk.

BR
Chamara

Bernd
Posts: 1179
Joined: Mon Jun 23, 2008 9:29 pm

Re: Data for further nucleation

Post by Bernd » Thu Nov 02, 2017 10:03 am

Dear Chamara,

which nucleation region fits best for your purpose depends on what happens in reality. However, we often do not know that. The two typical choices we have is "bulk" and "interface".

If we chose "bulk", we assume that there are seeding particles in the melt which serve as nucleation sites. Typically, this will lead to "blocky" precipitates because they have time to grow before touching the primary phase (at least if there is still enough liquid phase present).

If we chose "interface", we assume that nucleation appears at the interface. This will more likely lead to eutectic type growth because the secondary phase is growing together with the primary phase from the beginning.

If you don't know the morphology of the Laves phase from experiments, I would chose "interface" because it is more easily manageable. As long as there is liquid present, there is interface, and nucleation will be enabled.

If you want to use "bulk", it makes sense to use the "seed_density" model. However, to be sure that each entrapped liquid can form precipitates, you need to have a second nucleation type with "interface" with e.g. a higher critical undercooling so that "bulk" is favored.

Bernd

CharMIC
Posts: 211
Joined: Mon Aug 21, 2017 5:01 pm
anti_bot: 333

Re: Data for further nucleation

Post by CharMIC » Fri Nov 03, 2017 12:05 pm

Hej Bernd,

What did you meant by blocky? is it like block shape (square kind of) or larger irregular shape.

BR
Chamara

Bernd
Posts: 1179
Joined: Mon Jun 23, 2008 9:29 pm

Re: Data for further nucleation

Post by Bernd » Fri Nov 03, 2017 1:57 pm

Dear Chamara,

yes, of block shape. Whether round, squared, or whatever depends on anisotropy or shape of the rest liquid.

Bernd

Post Reply