Setting for redistribution control & nucleation undercooling

solid-solid phase transformations, influence of stresses and strains
Post Reply
pmecozzi
Posts: 23
Joined: Mon Apr 15, 2013 3:20 pm
anti_bot: 333

Setting for redistribution control & nucleation undercooling

Post by pmecozzi » Thu Jul 08, 2021 11:34 am

Dear Bernd,
I am currently studying the effect of substitutional alloy elements (Ni, Cr and Mn) inhomogeneity on cementite precipitation during cooling of hyper-eutectoid steel.
In the first simulation (see file CarbidePrecipi_at%_cooling_0.4_faceted 8a_ParaTQ) I set the redistribution behaviour of Cr, Ni, and Mn for the austenite/cementite interaction as “ParaTQ”.
The starting temperature of the simulation was 923 K. The cooling rate was equal to 0.4 K/s.
I set the minimum nucleation undercooling for cementite equal to 175 K.
As expected, the nucleus distribution in the domain is dependent on the local value of the substitutional alloy elements. Figure 1 shows the frac3, Cr, Mn and Ni concentrations.
Figure 1.png
Figure 1.png (32.07 KiB) Viewed 1192 times
The undercooling values for nuclei formed at different times and temperatures during cooling and the overall nuclei undercooling distribution are reported in Figure 2 and Figure 3 respectively.
Figure 2.png
Figure 2.png (110.85 KiB) Viewed 1192 times
Fiure 3.png
Fiure 3.png (175 KiB) Viewed 1192 times
I repeated the simulation changing only the redistribution behaviour of Cr, Ni, and Mn from “ParaTQ” to “Normal”.
Also, in this case, the nucleus distribution in the domain is dependent on the local value of the substitutional alloy elements.
Since the para-equilibrium (PE) condition reduces the value of Acm. (Acm[ortho]=1120 K and Acm[PE]=988 K, calculated by Thermo-calc), I was expecting a higher value of the nucleus undercooling in this simulation (with “Normal” instead of “paraTQ” in the setting of the element redistribution) compared to the previous one. This is not the case as shown in Figure 4 and Figure 5.
Figure 4.png
Figure 4.png (80.93 KiB) Viewed 1192 times
Figure 5.png
Figure 5.png (100.16 KiB) Viewed 1192 times
I also compared the driving force for the austenite to ferrite transformation and I saw that it is higher for the “paraTQ” simulation than for the simulation with “Normal” setting.

Could you please explain the reason for these results?

Best Regards,
Pina

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

Re: Setting for redistribution control & nucleation undercooling

Post by Bernd » Thu Jul 08, 2021 4:54 pm

Hi Pina,

just to be sure that I understood correctly:
  • there is only austenite and cementite present in the simulation, so that nucleation of cementite is only depending on the local temperature and composition?
  • In your diagrams you aggregate all observed nucleation events, without consideration of the local temperature at the time of nucleation?
If the above is correct, then I see two major reasons why interpretation of your results may be difficult:
  1. The value of the undercooling is very sensitive to the dS+-value, which is calculated as the derivative of ΔG with temperature at fixed composition of the austenite phase. Extrapolation over a large range >175K therefore my be very inaccurate (this is definitively a case which reminds us that we should implement the option to specify critical driving force instead of undercooling as nucleation criterion...).
  2. You start at a temperature of 973K where cementite should be already stable in both cases. You aggregate all nucleation events without specifically considering the local temperature at this time. Thus, you primarily observe the width of spread of undercooling values (see histogram) which may reflect more the spread of local concentrations than the different ACm temperatures
I think you rather should try to redo the simulations starting at a higher temperature (above 1120K) and set the critical undercooling to 0K. Then, you should observer the onset of nucleation in both cases which definitively should be shifted to lower temperatures for the paratq case.

Bernd

pmecozzi
Posts: 23
Joined: Mon Apr 15, 2013 3:20 pm
anti_bot: 333

Re: Setting for redistribution control & nucleation undercooling

Post by pmecozzi » Fri Jul 09, 2021 3:50 pm

Dear Bernd,

Many thanks for the replay.
Let me first answer your questions.
• there is only austenite and cementite present in the simulation, so that nucleation of cementite is only depending on the local temperature and composition?
Yes, I consider only cementite as the product phase and its formation depend on the local composition (Mn, Cr, Ni and C) and the temperature during cooling.
• In your diagrams you aggregate all observed nucleation events, without consideration of the local temperature at the time of nucleation?
Figure 3 and Figure 5 collect all the nucleation events in the temperature range where they appear (∼920-840 K).
In Figure 2 and Figure 4 I plot the nuclei undercooling as a function of the time (and then temperature) of their appearance.


I agree with you that it is better to start from a higher temperature with ΔT=0K and, indeed, my first set of simulations were done with this condition.
Unfortunately, with this condition, the low value of the undercooling of the formed nuclei does not allow them to grow. Therefore I decided to increase the critical undercooling to 175 K
Nevertheless, following your suggestion, I repeated the simulations starting from 1123 K and ΔT= 0 K with “ParaTQ” and “Normal” settings to investigate the effect of the element redistribution on the nucleation undercooling.
As expected, the nuclei do not grow but, surprisingly, I got the same values of the nucleation undercooling in both cases (see the two log files)! As done for the previous simulations, I attach the histogram and the undercooling vs time for the new ones.
Figure 6.png
Figure 6.png (10.38 KiB) Viewed 1185 times
Figure 7.png
Figure 7.png (10.4 KiB) Viewed 1185 times
Best Regards,

Pina

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

Re: Setting for redistribution control & nucleation undercooling

Post by Bernd » Fri Jul 09, 2021 7:10 pm

Hi Pina,

Sorry for confusing you, you are right! In the current version 7.0 of MICRESS nucleation undercooling is always calculated relative to equilibrium...

Depending on the exact physical picture which one may have with "nucleation", this may even be consistent. However, it also is confusing because it is difficult to know the extra undercooling which is connected to the specific redistribution behaviour. Maybe, we should take this into account when calculating nucleation undercooling...

Bernd

Post Reply