Ternary system

solid-solid phase transformations, influence of stresses and strains
Sudhee
Posts: 23
Joined: Tue Mar 28, 2017 7:33 pm
anti_bot: 333

Ternary system

Post by Sudhee » Wed May 31, 2017 2:21 pm

Dear Bernd,
I have a question regarding selection of initial microstructure for a ternary system during a heat treatment.
Step 1: I created an initial microstructure with 40% austenite and 60% ferrite (by quenching from 100% austenite) with a binary system (FeC).
Step 2: I am using this final microstructure as an initial microstructure for the next step (example: isothermal holding at 400°C) also for the FeC system. This worked completely fine.
Is it okay if I use the final microstructure in 'step 1' (FeC system) as an initial microstructure (step 2) for a ternary system (FeCMn) instead of FeC system. My idea is to just take the microstructure from step 1 and use it for different systems. Does this work??

Thank you,
Sudhindra.

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

Re: Ternary system

Post by Bernd » Wed May 31, 2017 2:57 pm

Dear Sudhindra,

from a technical point of view, the answer of your question depends on the way how you use the result of step 1 for the next simulation step. If you want to use the .rest file for reading the initial microstructure, it is not possible because you cannot add a further element by this way.
An alternative method is to read the initial microstructure from ASCII or vtk files which you have created using DP_MICRESS. In this case, you can easily prepare an additional "fake" composition input file (vtk or ASCII) for Mn by editing the composition file for carbon, replacing all values by e.g. a constant value.

I think, from a conceptional point of view, adding Mn or other elements with small constant composition in this way makes sense if there is no considerable segregation of these elements at this stage. However, you should not assume that the initial microstructure is in equilibrium if it was in equilibrium at the end of step 1. And you should think whether step 1 would have been completely different if you had included the element(s) already there...

Bernd

Sudhee
Posts: 23
Joined: Tue Mar 28, 2017 7:33 pm
anti_bot: 333

Re: Ternary system

Post by Sudhee » Wed May 31, 2017 3:31 pm

Dear Bernd,
I want to use the microstructure alone. So, I tried to create an ASCII file from the Korn file of step 1 and used as input for the 2nd step. I am giving constant initial concentration for both C and Mn in the next step. My question is, if adding another alloying element 'X' in the 2nd step, in this way, has some effect on the diffusivity of the elements (C and X) or is it same if I add 'X' in step 1.
I am raising this question because I see, in the 2nd step, after holding for some time the concentrations at the interface are going into negative values (I don't understand why!!).

I am having some troubles to create a microstructure in step 1 by including Mn and also the simulations take longer times (half day, at least)!!.

Best regards,
Sudhindra.

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

Re: Ternary system

Post by Bernd » Wed May 31, 2017 5:07 pm

Dear Sudhindra,

I would not expect any problems with diffusion coefficients - anyway you can look them up in the .TabD and .diff (in case you use mobility database) files and see if there is anything strange.
When you say, you start with constant compositions, I guess you set them for one phase and put the other to equilibrium? If you would set the same composition in both phases, you would be out of equilibrium (what could be reasonable for Mn).
How do you define thermodynamic data? Do you use database or linearized diagram?
Do you use any redistribution control (nple, para, mobCorr)?

It would be helpful if you could paste your input file here.

Bernd

Sudhee
Posts: 23
Joined: Tue Mar 28, 2017 7:33 pm
anti_bot: 333

Re: Ternary system

Post by Sudhee » Wed May 31, 2017 6:03 pm

Dear Bernd,
I am using a ThermoCalc coupling to define phase diagram (FeCMn). I did define the concentration of each component in each phase individually (as I treat Acicular ferrite as Martensite). I defined redistribution behaviour of C - normal, Mn - nple. I also observed that using the database for diffusion coefficients did not really help. This file worked fine when I defined the diffusion coefficients based on literature (but for shorter holding times).
Please find the attached file. Kindly, let me know if you can find any issue within the program.

Thank you,
Sudhindra.
Attachments
Part_TQ_400°C_gg_cells_FeCMn_in_in.txt
(26.12 KiB) Downloaded 286 times

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

Re: Ternary system

Post by Bernd » Wed May 31, 2017 7:03 pm

Dear Sudhindra,

the input file itself seems to be fine in general. What I am wondering is down to which temperature you have calculated step 1. 400°C is a very low temperature where diffusion is very slow. Cementite and also other types of carbides should be expected to start forming about 300K higher. So, you should be aware that your initial microstructure at 400°C is extremely far from equilibrium, and very fine resolution would be required to simulate precipitation at this temperature. Furthermore it is probably impossible in reality to cool down to this temperature without having formed cementite before.
Thus, my guess is that you should better include cementite/pearlite formation into simulation step 1.

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

Re: Ternary system

Post by Bernd » Fri Jun 02, 2017 7:25 pm

Dear Sudhindra,

thanks for sending additional information per PM. The unexpected high "diffusion length" of Mn is just a ghost image form carbon diffusion: Diffusion changes the C content in at% (balanced by Fe), however in wt% it will change all composition, even if it is only slightly.
Essentially, you should not expect any Mn diffusion at 400°C, no matter whether you use nple or para-equilibrium.

Negative compositions can be due to a too small relinearisation frequency, especially when starting so far from equilibrium. This parameter would be my first try.

Do you get any error messages in screen output?

The Thermo-Calc output from the script to make the .ges5 file looks fine.

Bernd

Sudhee
Posts: 23
Joined: Tue Mar 28, 2017 7:33 pm
anti_bot: 333

Re: Ternary system

Post by Sudhee » Sat Jun 03, 2017 8:53 am

Dear Bernd,
Thank you for the reply.
Regarding diffusion of Mn:
When I run the simulation for the FeMn system. Mn diffuses as it is expected to (nothing @ 400°C and very slow at 600°C).
So, for solving the 'diffusion length' problem of Mn in a ternary system, is it better to define the element concentrations in 'at.%' instead of wt.%. Please correct me if I understood you wrong. Or any other method?

Regarding negative compositions:
Yes, I do see some error message like 'trying hard phases 2 1 level: 3 zp= 502638 error= 3'
I am trying to run the simulation with longer relinearisation intervals. Will let ou know once I have the results.

Thank you,
Sudhindra.

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

Re: Ternary system

Post by Bernd » Sat Jun 03, 2017 11:51 pm

Hi Sudhindra,

2 small remarks:
Regarding diffusion of Mn:
correct, but it is no error, just a problem of the representation of the results...
Regarding negative compositions:
you should use smaller, not longer relinearisation intervals!

Bernd

Sudhee
Posts: 23
Joined: Tue Mar 28, 2017 7:33 pm
anti_bot: 333

Re: Ternary system

Post by Sudhee » Tue Jun 06, 2017 4:11 pm

Dear Bernd,
Defining the element concentration in 'at%' helped to solve the problem :D

Regarding the negative values, it is working fine at higher temperatures (500-600 °C, 0.5secs relinearization interval). At 400°C I tried to use a relinearization interval of 0.05 seconds and still I get negative concentrations at some interfaces!!! I am still trying to find out how to solve this. Meanwhile, if you could try to find a reason for this it would be great :).
Thank you so much.

Best regards,
Sudhindra.

Post Reply