MICRESS Forum

Hi Chamara,

The classical method (which you should know from the Basic MICRESS Training ) is to use different seed types for bulk, interfaces and triple junctions, and then assign different critical undercooling (using the "seed_undercooling" model).

The only alternative I can imagine would be to spatially resolve the critical nucleus (e.g. seed particle). Then, while nucleating at the interface of the seed particle, a grain boundary would help the seed to overcome the critical curvature. However, I don't know how this could work in solid state, where nucleation probably occurs on defects or dislocations...

Bernd

Hi Bernd,

Actually I thought about the classical method you have mentioned. But was not sure if its the right path to take. I thought may be there might be other special models in MICRESS for this.

BR
chamara

Hi Bernd,

Consider the following case.

I want to set nucleation r distance to each other. The nucleation will be checked at dt intervals. Initial nucleation will be checked at t₀. When the nucleation is checked at t₀, the average distance of the seeds will be around r. However when the nucleation checked at t₀+dt, the new seeds will nucleate at an average distance of r to each other. However they will also nucleate close (< r) to the seeds that were precipitated at t₀.

If I want the seeds that nucleate at t₀+dt to nucleate at an average distance r to the previously nucleated seeds at t₀, how should I do it.

BR
Chamara

Hi Chamara,

For this purpose you would use a shield distance equal to r, and a shield time > dt. The optional parameter "nucleation distance", if set to a value < r would increase the number of checked points, but never override the shield distance, even at t0.

Bernd

Hi Bernd,

Thanks for your input. I did as you told. I put shield time 5000s and nucleation checking for 20s. And I set shield distance for 50nm and nucleation distance for 10nm. However, when the nucleation happens, distance between particles become around 10nm. see the image. It seems like it override the shield distance?

Hi Chamara,

Yes, you are right! I looked it up in the code, and nucleation in the same time step only recognizes the nucleation distance and not the shield distance. This is reasonable in the sense that the shield is some physical (thermal or solutal) effect which would not exist if nuclei really would form simultaneously.
Thus, as in reality exactly simultaneous nucleation is rare anyway, one should in most cases use the nucleation distance equal (=default) or bigger than the shield distance. A reasonable strategy is to increase the nucleation distance but at the same time decrease the checking interval dt such that a sufficiently high "checking density" is reached without increasing the calculation time too much.

Bernd

Hi,

Can you change the minimum nucleation temperature when you use the restart option for an interrupted simulation. Restart option has been used without any other options.

BR
Chamara

Hi,

Yes, there should be no problem!

Bernd

Hi Bernd,

when you switch the remaining liquid in a simulation to a solid phase (using nucleation at interface of the solid phase) is there a way to set the orientation of the newly created solid phase to take the same orientation as the parent solid phase orientation? I do not want to create and interface between the newly precipitated solid grain and the parent solid grain.

Hi Chamara,

The principle of switching rest liquids to solid is to nucleate grain 0 again (which normally is the liquid) using the option "add_to_grain". By this procedure, you have the possibility of redefining all properties of grain 0, not only the phase identity. Of course, you always can do that by "hand", i.e. specify directly the fix orientation value which you prefer. You can do that also automatically by using "parent_relation". Then you should nucleate on the interface between the rest liquid and your dendrite. The dendrite grain then should be the parent, therefore you chose liquid as reference phase. Of course, in this case, you cannot automatically identify the liquid grain as parent any more, and you thus need to directly specify it as grain 0:

#Phase of new grains (integer) ...
1 add_to_grain
#Which grain number the new grains shall be added to?
# ...
grain_number 0
#...

However, you always will keep the grain number 0 existing, you just redefine the properties. Thus. the interface of grain 0 with the dendrite grain will remain - it is just converted to a solid-solid interface between the same phase with misorientation 0!

Bernd