About solute drag

ripening phenomena, dislocations, grainboundary topology
Post Reply
betleenkim
Posts: 41
Joined: Tue May 07, 2013 1:01 pm
anti_bot: 333

About solute drag

Post by betleenkim » Fri Nov 01, 2013 7:35 am

Dear Bernd

Long time no question about MICRESS ;)
Today i came back to ask about solute drag model in the MICRESS.

First of all, I want to ask about the meaning of each line in the dri file as given below.

...
# Options: constant temp_dependent dg_dependent
temp_dependent
# File for kinetic coefficient between phases 1 and 2?
Mobility_ag.txt
# critical velocity [micrometer/s]
6.00000E-07
# transition range deltaV [micrometer/s]
1.00000E-07
# Drag factor (real) < 1.
0.50000

...

In this case, I use temperature dependence of mobility data and give "solute drag" assumption to the boundary between phase 1 (Ferrite) and phase 2(Austenite). I want to know the meaning of last three lines such as "critical velocity", "transition range..." and "Darg factor".

Secondly, as known that the solute drag force is proportional to the solute amount in the boundaries which means that if the solute content is high in the vicinity of boundaries the velocity of the boundary will get slower. And i experimentally observed that the solute (in this case Mn) is strongly segregated at the interphase boundary of Ferrite and Austenite. In my simulation, i solidify the system first from 1760K to 746K and reheated to 1550K holding there for 1 hour. I directly use diffusion data from thermo_calc and "multi" option is used in the dri file for all alloying elements such as Fe, Al, Si, Mn and C. And i do not use the enhanced diffusion assumption such as boundary diffusion.
Partitioning of Mn.pdf
(199.19 KiB) Downloaded 432 times
The result gives me quite nice Mn partitioning in the Ferrite and Austenite as seen above. However, boundary segregation of Mn does not appear in the simulation result. I will attach one of the experimental and simulation result of Mn content.
Do you have any way to simulate boundary segregation that is closed to real situation? What about using enhanced diffusion assumption along grain boundary? To get the above simulation result, i put the initial Ferrite grain at the both bottom of domain and assumed "ssii" for the both of boundary condition. I found that there is only large segregation of Mn at the very middle part of domain where the region becomes austenite during cooling.

Thank you

Seawoong Lee

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: About solute drag

Post by Bernd » Mon Nov 04, 2013 9:21 am

Dear Seawoong,

the solute drag model in MICRESS is a simple approach to take into account the effect of solute or impurities which are not explicitly included in the simulation, i.e. there is no coupling to the concentration solver! The model just makes the interface mobility depending on the velocity as described here: The mobility is decreased in a sharp step below a critical interface velocity by a drag factor. A transition range in place of a sharp critical velocity in some cases helps making the model numerically more stable.

But I don't think that solute drag is what you are looking for! If the high manganese composition in the experiment is due to adsorption at the interface, MICRESS currently does not have a solution for that. If it is a segregation effect, it may be interesting to try the nple redistribution model, if the diffusion length of Mn is not high enough compared to the numerical interface thickness, because then the artificial solute trapping which is expected without special models would lead to a partial trapping, i.e. reduced partitioning.

Best wishes

Bernd

Post Reply