Starting Microstructure - heat treatment - dissolution of carbide

[daysianne]

Dear,
I am trying to run a solid-state transformation simulation for a martensitic steel (Fe–13.7Cr–0.32C–0.3Mn) during heating.
The initial microstructure consists of BCC matrix + M23C6 carbides, and during heating the system is expected to transform mainly into FCC with a small remaining fraction of carbides.
For the initial microstructure, I defined approximately: 93% BCC, 6% M23C6.
I calculated the equilibrium phase compositions using Thermo-Calc and used these values in MICRESS under `# Initial Concentration` for each phase.
Initially, I intended to simulate heating from room temperature up to 970°C. However, Thermo-Calc predicts three phases at room temperature: BCC_A2 (84.44%), BCC_A2#1 (CrBCC) (9.2%), and M23C6_D84 (6.27%).
Since introducing the additional BCC_A2#1 phase would significantly increase the simulation complexity, I decided to start the simulation from 500°C instead. But, one question, would it be possible starting from RT without addition of BCC_A2#1?
Although the phase fractions, compositions, and temperature defined under `# Initial Concentrations` appear consistent with the Thermo-Calc results, the following error occurs when starting the simulation:

`--> Force automatic start values`
`initialisation failed, error = 105 interface BCC_A2/M23C6_D84`

Additionally, I noticed that the phase compositions shown in the results differ from the values provided in the input file (comand #input). Is there any suggestions on how to proceed or what might cause this initialization issue?
I have attached the driving file and GSFile.
Many thanks in advance,
Daysianne

Part of driving file:
# Initial Concentrations
# ======================
# How shall initial concentrations be set?
# Options: input equilibrium from_file [phase number]
input
# Initial concentration of component 1 (C) in phase 0 (Liq) ? [wt%]
0
# Initial concentration of component 2 (CR) in phase 0 (Liq) ? [wt%]
0
# Initial concentration of component 3 (MN) in phase 0 (Liq) ? [wt%]
0
# Initial concentration of component 1 (C) in phase 1 (FCC) ? [wt%]
0
# Initial concentration of component 2 (CR) in phase 1 (FCC) ? [wt%]
0
# Initial concentration of component 3 (MN) in phase 1 (FCC) ? [wt%]
0
# Initial concentration of component 1 (C) in phase 2 (BCC) ? [wt%]
0.00000309037
# Initial concentration of component 2 (CR) in phase 2 (BCC) ? [wt%]
9.408
# Initial concentration of component 3 (MN) in phase 2 (BCC) ? [wt%]
0.164
# Initial concentration of component 1 (C) in phase 3 (M23C6) ? [wt%]
5.650
# Initial concentration of component 2 (CR) in phase 3 (M23C6) ? [wt%]
85.195
# Initial concentration of component 3 (MN) in phase 3 (M23C6) ? [wt%]
2.573
# Temperature at which the initial equilibrium
# will be calculated? [K]
773


Part of results:
Start Composition and Limits for quasi-equilibrium
--------------------------------------------------
FE in FCC_A1: 97.1340 at% (>0 - <100at% )
C in FCC_A1: 0.903681 at% (>0 - 50.0000at% )
CR in FCC_A1: 0.862224 at% (>0 - <100at% )
MN in FCC_A1: 1.10008 at% (>0 - <100at% )
FE in BCC_A2: 49.6747 at% (>0 - 100.000at% )
C in BCC_A2: 2.78318 at% (>0 - 75.0000at% )
CR in BCC_A2: 47.0135 at% (>0 - 100.000at% )
MN in BCC_A2: 0.528644 at% (>0 - 100.000at% )
FE in M23C6_D84: 43.3532 at% (>0 - 79.3103at% )
C in M23C6_D84: 20.6897 at% ( 20.6897 - 20.6897at% )
CR in M23C6_D84: 35.4393 at% (>0 - 79.3103at% )
MN in M23C6_D84: 0.517850 at% (>0 - 79.3103at% )


--> Force automatic start values

--> Force automatic start values
initialisation failed, error = 105 interface BCC_A2/M23C6_D84
Initial concentration in the phases:
------------------------------------
FCC_A1, FE: 85.68000 wt%
FCC_A1, C: 0.3200000 wt%
FCC_A1, CR: 13.70000 wt%
FCC_A1, MN: 0.3000000 wt%
BCC_A2, FE: 100.0000 wt%
BCC_A2, C: ************* wt%
BCC_A2, CR: ************* wt%
BCC_A2, MN: ************* wt%
M23C6_D84, FE: 100.0000 wt%
M23C6_D84, C: ************* wt%
M23C6_D84, CR: ************* wt%
M23C6_D84, MN: ************* wt%

Routine init calls routine initConcFeld
Routine init calls routine initMol
Thermo-Calc error 1146 MICRESS error 24 Thermo-Calc
WARNING IN toolGetMolVolTQ!
Error in first equilibrium! BCC_A2

[Bernd]

Dear daysianne,

The problem seems to be associated with the .GES5-file which you are using. My guess is that this .GES5-file has been created using a very new database and/or Thermo-Calc version which is not compatible to your older MICRESS Version. Which versions did you use for making the .GES5-file? And which is your exact MICRESS version?
If you have the chance, please try an older database version or/and a newer MICRESS version...

Bernd

[daysianne]

Dear Bernd,
Many thanks for your attention. I have generated the .GES5-file with old thermodynamic and mobility database, and now that error (=105 ) disappears.
Answering your question, I am using Micress version 7.202 and ThermoCalc version 2026a.
Data base used for the first .GES5-file that showed error = 105: TCFE15 and MOBFE9.
Data base used for the new .GES5-file without error =105 : TCFE13 MOBFE8.

My question now is regarding to the initial concentration, although I have added the initial concentration as “input”, I don’t know if it is using it or from the database or my input. Demixing error appears.

Driving file:
# Initial Concentrations
# ======================
# How shall initial concentrations be set?
# Options: input equilibrium from_file [phase number]
input
# Initial concentration of component 1 (C) in phase 0 (Liq) ? [wt%]
0
# Initial concentration of component 2 (CR) in phase 0 (Liq) ? [wt%]
0
# Initial concentration of component 3 (MN) in phase 0 (Liq) ? [wt%]
0
# Initial concentration of component 1 (C) in phase 1 (FCC) ? [wt%]
0
# Initial concentration of component 2 (CR) in phase 1 (FCC) ? [wt%]
0
# Initial concentration of component 3 (MN) in phase 1 (FCC) ? [wt%]
0
# Initial concentration of component 1 (C) in phase 2 (BCC) ? [wt%]
0.00000309037
# Initial concentration of component 2 (CR) in phase 2 (BCC) ? [wt%]
9.408
# Initial concentration of component 3 (MN) in phase 2 (BCC) ? [wt%]
0.164
# Initial concentration of component 1 (C) in phase 3 (M23C6) ? [wt%]
5.650
# Initial concentration of component 2 (CR) in phase 3 (M23C6) ? [wt%]
85.195
# Initial concentration of component 3 (MN) in phase 3 (M23C6) ? [wt%]
2.573
# Temperature at which the initial equilibrium
# will be calculated? [K]
773
#

Results:
Routine init calls routine initTQ

Start Composition and Limits for quasi-equilibrium
--------------------------------------------------
FE in FCC_A1: 97.1340 at% (>0 - <100at% )
C in FCC_A1: 0.903681 at% (>0 - 50.0000at% )
CR in FCC_A1: 0.862224 at% (>0 - <100at% )
MN in FCC_A1: 1.10008 at% (>0 - <100at% )
FE in BCC_A2: 49.6747 at% (>0 - 100.000at% )
C in BCC_A2: 2.78318 at% (>0 - 75.0000at% )
CR in BCC_A2: 47.0135 at% (>0 - 100.000at% )
MN in BCC_A2: 0.528644 at% (>0 - 100.000at% )
FE in M23C6_D84: 43.3532 at% (>0 - 79.3103at% )
C in M23C6_D84: 20.6897 at% ( 20.6897 - 20.6897at% )
CR in M23C6_D84: 35.4393 at% (>0 - 79.3103at% )
MN in M23C6_D84: 0.517850 at% (>0 - 79.3103at% )

--> Force automatic start values

Time t = 0.0000000 s
CPU-time: 5 s
Current phase-field solver time step = 1.46E-04 s
Average conc. of comp. C = 0.2824459 wt%
Average conc. of comp. CR = 13.1965848 wt%
Average conc. of comp. MN = 0.2844257 wt%
Temperature at the bottom = 773.00 K
Temperature gradient = 0.00000 K/cm
Fraction of phase LIQUID: 0.00000
Fraction of phase FCC_A1: 0.00000
Fraction of phase BCC_A2: 0.93778
Fraction of phase M23C6_D84: 0.06222

Updating of diffusion data from database...

Intermediate output for t = 0.50000 s
CPU-time: 12 s
Current phase-field solver time step = 3.05E-03 s
Average conc. of comp. C = 0.2824459, Variation = +0.0000000 wt%
Average conc. of comp. CR = 13.1965848, Variation = +0.0000000 wt%
Average conc. of comp. MN = 0.2844257, Variation = +0.0000000 wt%
Temperature at the bottom = 775.00 K
Temperature gradient = 0.00000 K/cm
complete relinearisation!
Warning: Demixing in interface BCC_A2/M23C6_D84, component CR


Kind regards.

[Bernd]

Dear daysianne,

The output of the "Start Composition and Limits for quasi-equilibrium" which you show have nothing to do with your initial compositions. They show the start values which will be applied by the "set_default_major_composition" command of Thermo-Calc to those phases where no composition condition is set, and thus are only a property of the .GES-file itself. To check the initial compositions which have been set in MICRESS you should use the .conc*, .c*pha*, or .TabC output.

The "demixing" warning (and potential errors linked to this problem) can be avoided either by using the "diagonal" switch in "Numerical Parameters" / "Concentration Solver", or you set all element in phase M23C6_D84 to "stoichiometric (in the same place). "diagonal" is recommended when using high-alloyed materials, or when this "demixing" warning is appearing. Setting all elements of a phase to "stoichiometric" is recommended for intermetallic phases, which have only a small solubilit of the matrix component. My recommendation in your case is to do both:

# Concentration solver
# --------------------
# Factor for diffusion time stepping? (0.0 < factor < 1.0)
0.9500000000000
# Optional parameters for thermodynamic definitions
# Finish input with 'end_of_options'.
# {stoichiometric | stoich_solubility | stoich_enhanced} <phase> <component list>
# steep_limit <value>
# limits <phase> <component> <min (at%)> <max (at%)>
# penalty <phase> <phase 2> <component>
# interstitial <phase> <component list>
# sum_limits <phase> <component list>
# diagonal [<component list>]
# interaction <component 1> <component 2>
# {criterion_higher | criterion_lower} <phase> <phase 2> <component>
# {ordered | disordered} <phase> <sublattice 1> <sublattice 2>
# switch_diff_data <phase> <reference phase>
# {diff_comp_sets | vol_comp_sets | enth_comp_sets} <phase list>
# temp_limit_TQ <lowest (K)> <highest (K)>
# temp_limit_phase_TQ <phase> <lowest (K)> <highest (K)>
# independent_sublattice <phase> <submatrix component> <component list>
# database_consistent
# pressure <value (MPa)>
# start_value <phase> <phase 2>
# init_fraction <phase> <phase 2> <fraction>
# tic_tq_segments | tic_z_segments
diagonal
stoichiometric 3 1-3
no_more_options
#

Furthermore, I would recommend not to set the initial composition of fcc to zero (even if it is not present at the beginning), and rather use some more realistic values. The user-defined initial compositions are also used in case of numerical trouble as start values for iteration, and 0.0 is a very bad start value for a compostion...

Bernd