Dear Bernd:
Hi, I want to simulate pearlite growth with MICRess, which is a lamellar structure.
As you know, the thickness of cementite may be pretty small according to Experimental data.
Thus the small size may yield large curvature undercooling, which may depart from the reality to 50K or more.
One possible way is the alter its interfacial energy. However, will it also affect the kinetics of pearlite transformation? I mean, the process is accelerated if we use a small interfacial energy.
For my simulation, a smaller interfacial energy is better or a larger one?
Also, do you have an idea of how to eliminate to undercooling induced by curvature?
Best wishes!
Yang
question of the stability and curvature induced undercooling
Re: question of the stability and curvature induced undercoo
Dear Yang,
Of course, also in reality the curvature undercooling of the cementite lamellae has an important effect on the transformation kinetics as it reduces the driving force which is available! The important point for the simulation is to get the correct curvature undercooling, and this means that you need the correct value of the thickness of the cementite lamellae and the correct values for the interface energies!
Unfortunately, interface energies are typically unknown. Thus you need to calibrate them such that experimental behaviour is recovered. If in the simulation pearlite cannot grow, while you see it is growing in experiments under the same conditions (and and with the same lamellar spacing), then you can assume that the interface energies which you specified are too high (anyway I think that 1E-4 J/cm2 is too high...). If you know from experiments up to which temperature pearlite is still growing (and we assume the thermodynamic database to be correct) you could calibrate interface energies to get a realistic value.
There is an additional effect of the triple point angle which may play a certain role. You can check it out using "multi_obstacle" potential. It will only take effect when the interface energies of the boundaries which build up the triple junction are not identical!
Bernd
Of course, also in reality the curvature undercooling of the cementite lamellae has an important effect on the transformation kinetics as it reduces the driving force which is available! The important point for the simulation is to get the correct curvature undercooling, and this means that you need the correct value of the thickness of the cementite lamellae and the correct values for the interface energies!
Unfortunately, interface energies are typically unknown. Thus you need to calibrate them such that experimental behaviour is recovered. If in the simulation pearlite cannot grow, while you see it is growing in experiments under the same conditions (and and with the same lamellar spacing), then you can assume that the interface energies which you specified are too high (anyway I think that 1E-4 J/cm2 is too high...). If you know from experiments up to which temperature pearlite is still growing (and we assume the thermodynamic database to be correct) you could calibrate interface energies to get a realistic value.
There is an additional effect of the triple point angle which may play a certain role. You can check it out using "multi_obstacle" potential. It will only take effect when the interface energies of the boundaries which build up the triple junction are not identical!
Bernd