Hi Bernd,
I hope to grow new small grains at the edges and boundaries of large grains, which shows that grains are easier to nucleate and grow in the presence of a substrate than in pure liquid phase.
Ling
Shape of grain boundary
-
dingling312
- Posts: 17
- Joined: Mon Jun 11, 2018 10:11 am
- anti_bot: 333
Re: Shape of grain boundary
Hi Ling,
I cannot follow your argument: Nucleation of phase 1 on a 0/1 interface makes only sense in case of strong kinetic undercooling of the 0/1 interface, like e.g. in case of silicon growth during solidification of Si-containing aluminium alloys. Then, twinning can be observed if the silicon grain is not well-oriented to the temperature gradient and if there furthermore is a strong kinetic anisotropy. However, this is not the "normal" case.
If the nucleating phase is already present, epitaxial growth is typically favored because nucleation of a grain with different orientation requires formation of a critical nucleus and thus an additional undercooling. However, your argument would be correct for nucleation of a further phase, e.g. phase 2, on the 0/1 interface.
Bernd
I cannot follow your argument: Nucleation of phase 1 on a 0/1 interface makes only sense in case of strong kinetic undercooling of the 0/1 interface, like e.g. in case of silicon growth during solidification of Si-containing aluminium alloys. Then, twinning can be observed if the silicon grain is not well-oriented to the temperature gradient and if there furthermore is a strong kinetic anisotropy. However, this is not the "normal" case.
If the nucleating phase is already present, epitaxial growth is typically favored because nucleation of a grain with different orientation requires formation of a critical nucleus and thus an additional undercooling. However, your argument would be correct for nucleation of a further phase, e.g. phase 2, on the 0/1 interface.
Bernd
-
dingling312
- Posts: 17
- Joined: Mon Jun 11, 2018 10:11 am
- anti_bot: 333
Re: Shape of grain boundary
Hi,Bernd,
According to our experiments, adding a small amount of Fe into ti-6al-4V can effectively refine the grain size. According to my previous simulation, with the increase of Fe content, the segregation of Fe at the grain boundary is to inhibit the mobility of the grain boundary and promote the formation of dendrites. I used to think that the dendrite surface is easier to form grains than the flat interface because of local undercooling, which leads to grain refinement.You think there is not enough kinetic undercooling at the interface of phase 1 and 1/0, so I wonder if it's possible the segregation of Fe hinders the movement of grain boundaries, reduces the annexation of large grains to small grains, which refines the grains. If I want to solve this problem, do you think I should use nucleation model or grain growth model?
Ling
According to our experiments, adding a small amount of Fe into ti-6al-4V can effectively refine the grain size. According to my previous simulation, with the increase of Fe content, the segregation of Fe at the grain boundary is to inhibit the mobility of the grain boundary and promote the formation of dendrites. I used to think that the dendrite surface is easier to form grains than the flat interface because of local undercooling, which leads to grain refinement.You think there is not enough kinetic undercooling at the interface of phase 1 and 1/0, so I wonder if it's possible the segregation of Fe hinders the movement of grain boundaries, reduces the annexation of large grains to small grains, which refines the grains. If I want to solve this problem, do you think I should use nucleation model or grain growth model?
Ling
Re: Shape of grain boundary
Hi Ling,
The typical model for that in solidification science would be the following:
Starting from the melt, there typically are particles floating in the melt which can serve as nucleants for the primary phase. I would assume that this is also the case in smelting reactions where you have a lot of impurities. In casting processes, particles often are intentionally added as seeding agents.
The nucleation undercooling on such particles depend on their size, and we can assume a certain size distribution. When the melt cools down, nucleation starts on the biggest particles. The question whether smaller particles also get active as nucleants depends on how low temperature is dropping: If the first nuclei can grow rapidly, latent heat release prevents temperature from dropping, thus leading to few seeds or big grains. If they cannot grow such fast due to strong segregation (or a high growth restriction factor), temperature is dropping such that additional nucleants are activated, and the grain size is reduced.
In MICRESS this can easily be simulated by using latent heat and the seed density model. A simple example for that is AlCu_Equiaxed_dri.txt This approach should allow you to show that adding Fe will automatically lead to smaller grains. The only problem is that you need to assume a certain grain size distribution which in most cases is unknown.
Bernd
The typical model for that in solidification science would be the following:
Starting from the melt, there typically are particles floating in the melt which can serve as nucleants for the primary phase. I would assume that this is also the case in smelting reactions where you have a lot of impurities. In casting processes, particles often are intentionally added as seeding agents.
The nucleation undercooling on such particles depend on their size, and we can assume a certain size distribution. When the melt cools down, nucleation starts on the biggest particles. The question whether smaller particles also get active as nucleants depends on how low temperature is dropping: If the first nuclei can grow rapidly, latent heat release prevents temperature from dropping, thus leading to few seeds or big grains. If they cannot grow such fast due to strong segregation (or a high growth restriction factor), temperature is dropping such that additional nucleants are activated, and the grain size is reduced.
In MICRESS this can easily be simulated by using latent heat and the seed density model. A simple example for that is AlCu_Equiaxed_dri.txt This approach should allow you to show that adding Fe will automatically lead to smaller grains. The only problem is that you need to assume a certain grain size distribution which in most cases is unknown.
Bernd