ThermoCal coupled problem, NPLE and no_TQ problem

issues about thermodynamics and their coupling to MICRESS
Post Reply
nokkikku
Posts: 25
Joined: Wed Nov 25, 2009 2:07 pm
anti_bot: 333

ThermoCal coupled problem, NPLE and no_TQ problem

Post by nokkikku » Wed Feb 09, 2011 3:09 pm

Dear MICRESS team,

I am making a ThermoCalc coupled problem these days on the gamma_alpha transformation of Fe-C-Mn-Mo-Si-Cr-Al at 680°C and have the following questions:

1. In the dri. file, section phase diagram, there is an option 'database[local/global],

# Input of the phase diagram of phase 1 and phase 2:
# --------------------------------------------------
# Which phase diagram is to be used?
# Options: database [local|global] linear


can you explain me more about this?

2. I carried out a simulation under the option 'NPLE' as followed.

# Data for phase interaction 1 / 2:
# ---------------------------------
# Simulation of interaction between phase 1 and 2?
# Options: phase_interaction no_phase_interaction identical phases nb.
# [standard|particle_pinning[_temperature]|solute_drag|redistribution_control]
phase_interaction redistribution_control
# 'DeltaG' options: default
# avg ... [] max ... [J/cm**3] smooth ... [degrees]
avg 1. max 50.
# I.e.: avg +1.00 smooth +45.0 max +5.00000E+01
# Type of surface energy definition between phases 1 and 2?

# Options: constant temp_dependent
constant
# Surface energy between phases 1 and 2? [J/cm**2]
4.00000E-05
# Type of mobility definition between phases 1 and 2?
# Options: constant temp_dependent dg_dependent
constant
# Kinetic coefficient mu between phases 1 and 2? [cm**4/(Js)]
8.00000E-06


# Concentration data
# ==================
# Number of dissolved constituents? (int)
6
# Type of concentration?
# Options: atom_percent (at%)
# weight_percent (wt%)
weight_percent
#
# Options: diff no_diff infinite infinite_restricted
# multi database_global database_local
# [+b] for grain-boundary diffusion
# ('multi' can be followed by a string of "n", "d", "g", or "l",
# to describe each contribution: respectively no diffusion,
# user-defined diffusion coefficient, and 'global' or 'local'
# value from database, the default is global values from database).
# How shall diffusion of component 1 in phase 0 be solved?
no_diff
# How shall diffusion of component 1 in phase 1 be solved?
database_global
# How shall diffusion of component 1 in phase 2 be solved?
database_global
# How shall diffusion of component 2 in phase 0 be solved?
no_diff
# How shall diffusion of component 2 in phase 1 be solved?
database_global
# How shall diffusion of component 2 in phase 2 be solved?
database_global
# How shall diffusion of component 3 in phase 0 be solved?
no_diff
# How shall diffusion of component 3 in phase 1 be solved?
database_global
# How shall diffusion of component 3 in phase 2 be solved?
database_global
# How shall diffusion of component 4 in phase 0 be solved?
no_diff
# How shall diffusion of component 4 in phase 1 be solved?
database_global
# How shall diffusion of component 4 in phase 2 be solved?
database_global
# How shall diffusion of component 5 in phase 0 be solved?
no_diff
# How shall diffusion of component 5 in phase 1 be solved?
database_global
# How shall diffusion of component 5 in phase 2 be solved?
database_global
# How shall diffusion of component 6 in phase 0 be solved?
no_diff
# How shall diffusion of component 6 in phase 1 be solved?
database_global
# How shall diffusion of component 6 in phase 2 be solved?
database_global
#
# Interval for updating diffusion coefficients data? [s]
1.0000
#
#
# Phase diagram - input data
# ==========================
#
# List of phases and components which are stoichiometric:
# phase and component(s) numbers
# List of concentration limits:
# <Limits>, phase number and component number
# End with 'no_more_stoichio' or 'no_stoichio'
no_stoichio
#
#
#
# Is a thermodynamic database to be used?
# Options: database no_database
database
#
# Name of Thermo-Calc *.GES5 file without extension?
/work/xxxxxx
# Interval for updating thermodynamic data [s] =
5.0000
# Input of the phase diagram of phase 1 and phase 2:
# --------------------------------------------------
# Which phase diagram is to be used?
# Options: database [local|global] linear linearTQ
database local
# Maximal allowed local temperature deviation [K]
-1.0000000000000000
# Please specify the redistribution behaviour of each component:
# Format: forward [backward]
# Options: nple para normal
# Component 1
normal normal
# Component 2
nple nple
# Component 3
nple nple
# Component 4
nple nple
# Component 5
nple nple
# Component 6
nple nple
# The database contains the following components:
# 1: AL
# 2: C
# 3: CR
# 4: FE
# 5: MN
# 6: MO
# 7: SI
# Specify relation between component indices Micress -> TC!
# The main component has in MICRESS the index 0
# Thermo-Calc index of (MICRESS) component 0?
4
# Thermo-Calc index of (MICRESS) component 1?
2
# Thermo-Calc index of (MICRESS) component 2?
5
# Thermo-Calc index of (MICRESS) component 3?
7
# Thermo-Calc index of (MICRESS) component 4?
3
# Thermo-Calc index of (MICRESS) component 5?
6
# Thermo-Calc index of (MICRESS) component 6?
1
# 0 -> FE
# 1 -> C
# 2 -> MN
# 3 -> SI
# 4 -> CR
# 5 -> MO
# 6 -> AL
# The database contains 5 phases:
# 1: LIQUID
# 2: BCC_A2
# 3: CEMENTITE
# 4: FCC_A1
# 5: FCC_A1#2
# Specify relation between phase indices Micress -> TC!
# The matrix phase has in MICRESS the index 0
# Thermo-Calc index of the (MICRESS) phase 1?
4
# Thermo-Calc index of the (MICRESS) phase 2?
2
# 1 -> FCC_A1
# 2 -> BCC_A2
#
# Molar volume of (MICRESS) phase 1 (FCC_A1)? [cm**3/mol]
7.1824
# Molar volume of (MICRESS) phase 2 (BCC_A2)? [cm**3/mol]
7.2757
# Temperature at which the initial equilibrium
# will be calculated? [K]
953.0000
#
#
# Initial concentrations
# ======================
# How shall initial concentrations be set?
# Options: input equilibrium from_file [phase number]
equilibrium 1
# Initial concentration of component 1 (C) in phase 1 (FCC_A1) ? [wt%]
8.00000E-02
# Initial concentration of component 2 (MN) in phase 1 (FCC_A1) ? [wt%]
1.4400
# Initial concentration of component 3 (SI) in phase 1 (FCC_A1) ? [wt%]
2.70000E-02
# Initial concentration of component 4 (CR) in phase 1 (FCC_A1) ? [wt%]
1.90000E-02
# Initial concentration of component 5 (MO) in phase 1 (FCC_A1) ? [wt%]
0.14600
# Initial concentration of component 6 (AL) in phase 1 (FCC_A1) ? [wt%]
5.00000E-02


I found out in the log file that Mo has higher concentration in alpha phase.

# The linearisation parameters of the phases FCC_A1/BCC_A2 are:
# -------------------------------------------------------------
953.00000 ! T0 [K]
-57.080636 ! dG [J/cm**3]
0.59874035 ! dSf+ [J/cm**3K]
0.48840043 ! dSf- [J/cm**3K]
669.23144 ! dH [J/cm3]
8.00079239E-02 ! c0(C)/FCC_A1
1.01709775E-03 ! c0(C)/BCC_A2
1.4401103 ! c0(MN)/FCC_A1
0.35137437 ! c0(MN)/BCC_A2
2.69996069E-02 ! c0(SI)/FCC_A1
3.08752214E-02 ! c0(SI)/BCC_A2
1.90006792E-02 ! c0(CR)/FCC_A1
1.22949850E-02 ! c0(CR)/BCC_A2
0.14599155 ! c0(MO)/FCC_A1
0.22935210 ! c0(MO)/BCC_A2
4.99987475E-02 ! c0(AL)/FCC_A1
6.23530968E-02 ! c0(AL)/BCC_A2
-83.888545 ! m(C)/FCC_A1
-8428.2963 ! m(C)/BCC_A2
-14.278846 ! m(MN)/FCC_A1
-68.439379 ! m(MN)/BCC_A2
8.1175936 ! m(SI)/FCC_A1
9.5659570 ! m(SI)/BCC_A2
-6.1153430 ! m(CR)/FCC_A1
-10.371514 ! m(CR)/BCC_A2
6.4601977 ! m(MO)/FCC_A1
5.5256093 ! m(MO)/BCC_A2
5.7543647 ! m(AL)/FCC_A1
10.850346 ! m(AL)/BCC_A2
-6.83658505E-04 ! dcdT(C)/FCC_A1
8.79722774E-06 ! dcdT(C)/BCC_A2
-7.76264727E-03 ! dcdT(MN)/FCC_A1
1.94283755E-03 ! dcdT(MN)/BCC_A2
-2.57004902E-05 ! dcdT(SI)/FCC_A1
2.24090767E-05 ! dcdT(SI)/BCC_A2
-4.02497141E-05 ! dcdT(CR)/FCC_A1
2.24873205E-05 ! dcdT(CR)/BCC_A2
-3.89580810E-05 ! dcdT(MO)/FCC_A1
-1.58395495E-05 ! dcdT(MO)/BCC_A2
-6.17543895E-05 ! dcdT(AL)/FCC_A1
1.37775617E-04 ! dcdT(AL)/BCC_A2
# Minimum undercooling for stable growth, seed type 1: 6.680692 K [r=0.2000000 mic.]
# Minimum undercooling for stable growth, seed type 2: 6.680692 K [r=0.2000000 mic.]
# Minimum undercooling for stable growth, seed type 3: 6.680692 K [r=0.2000000 mic.]


What is the reason for that? Does it mean that my simulation went wrong?

3. Since NPLE is not derivable from ThermoCalc, could you please explain me a bit more what is the principle of calculation?

4. Since sometimes the simulation with TQ is resource consuming, I tried to transfer the available Thermodynamics data (ex.from ThermoCalc) to a none-database, linearized phase diagram section in no_TQ calculation. I understand this is easier in case of paraequilibrium because we need the phase diagram slope only respect to carbon. In such the case, where can we take the dSf+ amd dSf- from ThermoCalc? Which one - or + do we need to use? However, to my understanding, paraequilibrium at 680°C is still too extreme.

5. But is there any possibility to connect the LE and paraequilibrium thermodynamics data derived from ThermoCalc and modified into NPLE case and put into the phase diagram section in no_TQ dri. file?

With big thanks and best regards

Bernd
Posts: 1088
Joined: Mon Jun 23, 2008 9:29 pm

Re: ThermoCal coupled problem, NPLE and no_TQ problem

Post by Bernd » Thu Feb 10, 2011 6:48 pm

Hi nokkikku,

Let me answer your questions one by one:

1.) Even if you decided to use a thermodynamic database using the "database" keyword and giving the path to the .GES5 file, you are asked for all phase interactions (which are switched on and which are not between same phases) whether you want to use thermodynamic data from this database or a linearised description for this interaction:

...
# Is a thermodynamic database to be used?
# Options: database no_database
database
#
# Name of Thermo-Calc *.GES5 file without extension?
C:\MICRESS\MICRESS_5_501\Examples\GES_Files_INTEL\Al_Cu
# Interval for updating thermodynamic data [s] =
1.00
# Input of the phase diagram of phase 0 and phase 1:
# --------------------------------------------------
# Which phase diagram is to be used?
# Options: database [local|global] linear linearTQ
database
# Maximal allowed local temperature deviation [K]
-1.0000000000000000
...

This allows you to combine the usage of a thermodynamic database with linearised descriptions if e.g. one specific phase is not in the database or you want to construct a "pseudo-phase" like pearlite. You can specify such linearized descriptions either in the normal "linear" or the "linearTQ" format (see here).
If you choose "database" you have an additional option which allows you to reduce the calculation effort by replacing all the linearisation data of the individual grid points of a grain interface by one "global" linearisation. More information you can find here.

2.) NPLE is not a characteristic of a thermodynamic equilibrium (like para), but of a special redistribution condition, when slow diffusing components are overrun by the phase transformation front due to a high driving force although they are redistributed, i.e. local equilibrium is established. That means, on a local scale you cannot expect zero partitioning although it is observed on a global scale!
The initial equilibrium which MICRESS calculates is obtained starting from the gamma composition, i.e. the fraction of alpha is assumed to be 0 for the initial equilibrium. Therefore it is quite logic that you get the composition for Mo which is in equilibrium to 0.146 % in fcc (and which is higher).
This (or some similar) high value of the Mo composition will also be used for calculation of the driving force during the phase transformation under NPLE. It should be visible in the .c5Pha2 output, but will never appear in the .conc5 output because it exists only locally inside the interface!
So, to me nothing seems wrong...

3.) Unfortunately, there is still no publication on the NPLE and PARA model in MICRESS, although one is on the way.
Shortly speaking, the model applies modified phase fractions for redistribution of the elements to assure that the element pile-up is present all over the interface region and the driving force is calculated correctly.

4.) The dSf+ and dSf- (as well as dcdT values) are needed if you use the "linearTQ" option, which has the same format as the local linearisations obtained by MICRESS via TQ from the database. The advantage is that you can copy them directly from the .log file of a previous TC-coupled simulation. In case of "linear" you need only one value for dS which is easier if you use literature data but less convenient for using data from previous MICRESS simulations. If you need it nevertheless, just take the average of dSf+ and dSf- for dS.
If the main reason for you to switch to linearised phase diagrams is because TC coupling takes too much time, you should consider the "database global" option which is much faster (see 1.)), so that it is not necessary to use a "no_TQ" driving file.

5.) I hope for NPLE I gave already the answer: Start TC-coupled simulation with MICRESS using NPLE, then copy the linearisation output in the .log or .TabLin file into the MICRESS driving file while using the option "linearTQ".
For para-equilibrium the answer is more complex, because with the current version 5.5, the option "paraTQ" is still not running perfectly. But with the new version which we plan to release until end of march, it will be possible, completely analogous to nple. In this case, para-equilibrium is calculated on a thermodynamic level.

Bernd

nokkikku
Posts: 25
Joined: Wed Nov 25, 2009 2:07 pm
anti_bot: 333

Re: ThermoCal coupled problem, NPLE and no_TQ problem

Post by nokkikku » Wed Dec 14, 2011 12:37 pm

Dear Bernd,

recently I got to know how to calculate the phase boundary under NPLE mode (I hope this is not too late.). So I want to resume our discussion.

Attached you will see the gamma/alpha phase boundary calculated by NPLE approach compared with LE approach. Both were calculated with ThermoCalc in Hillert's paper.

So my question is, when NPLE solute redistribution is selected in MICRESS (coupled with TCC), it means the phase boundary and the Gibbs free energy will be re-calculated according to the NPLE approach (like in the attachment), right?

Thanks
nokkikku
Attachments
NPLE boundary.jpg
NPLE phase boundary calculated with TCC by Hillert
NPLE boundary.jpg (29.6 KiB) Viewed 1881 times

Bernd
Posts: 1088
Joined: Mon Jun 23, 2008 9:29 pm

Re: ThermoCal coupled problem, NPLE and no_TQ problem

Post by Bernd » Wed Dec 14, 2011 8:44 pm

Dear nokkikku,

I do not know what this diagram you show exactly means, or what it can be used for - can you give me more information? Does this diagram just show the stability regions of the phases or is it really a "phase diagram" which could be used in linearized form in MICRESS? Perhaps, you could give a reference to Hillert's paper.

Bernd

nokkikku
Posts: 25
Joined: Wed Nov 25, 2009 2:07 pm
anti_bot: 333

Re: ThermoCal coupled problem, NPLE and no_TQ problem

Post by nokkikku » Mon Jan 02, 2012 11:25 am

Dear Bernd,

Happy new year 2012.

Sorry that I misled you with my improper picture and explanation. I am also late for replying this. I don't know why I don't get the warning when my post is replied?

Here I have a new attached picture. You can see that if we calculate the Ae3 temperature (although here it was written Ac3 instead) under NPLE, It is lower than that from orthoequilibrium (or local equilibrium). (Therefore the ferrite fraction calculated from the lever rule for low carbon steel such as C of 0.1% will be less than that calculated from orthoequilibrium.)

So my question is, if we assign the NPLE solute redistribution in MICRESS with the TQ interface, will MICRESS link back to ThermoCalc to calculate the Gibbs free energy and so on from such the Ae3 temperature under NPLE (which is lower than the Ae3 under LE) as shown in the attached phase diagram?

Thanks and regards,
nokkikku
Attachments
Ac3D, Ac3 (eq para nple).jpg
Actually it must be Ae3 and Ae3d. The subscribt 'd' means under deformation.
Ac3D, Ac3 (eq para nple).jpg (38.58 KiB) Viewed 1866 times

Bernd
Posts: 1088
Joined: Mon Jun 23, 2008 9:29 pm

Re: ThermoCal coupled problem, NPLE and no_TQ problem

Post by Bernd » Mon Jan 02, 2012 1:36 pm

Dear nokkikku,

I also wish you a Happy New Year 2012!

NPLE is not a thermodynamic condition - you cannot tell Thermo-Calc to switch to NPLE! It is rather a special kinetic situation where the driving force for phase transformation is reduced because of the immobility of the substitutional elements.
Therefore, I would interpret this diagram not as a phase diagram but as a diagram which shows the existence of the two phases after a predefined kinetic process (the gamma-alpha transformation under NPLE/LE/PARA).

MICRESS should give the same results in a simulation of this process if "NPLE" is used, but not due to modified thermodynamics, but rather due to a modified driving force!

Best regards,

Bernd

PS: You will get notified of all new posts if you click "Subscribe Forum" at the bottom in the corresponding sub-forum while you are logged in!

zhubq
Posts: 84
Joined: Mon Jun 22, 2009 7:33 pm

Re: ThermoCal coupled problem, NPLE and no_TQ problem

Post by zhubq » Tue Oct 30, 2012 8:04 am

Hi Bernd,
I wanna add some questions here for my simulations.[ I used micress 4.08a. unfortunately cannot use newest version 6.003 in our institute.]

1) the linearisation parameters obtained in a TQ coupled simulation depend on the input of initial concentration of each phase even when the fraction of the phase is 0. [I thought it is calculated based on the overall concentration, but it seems not.]
2) the dS+ and dS- are quite different in my simulation (See below). Is it because I made some mistakes in my input?
1043.1500 ! T0 [K]
-5.7901929 ! dG [J/cm**3]
-0.10734394 ! dSf+ [J/cm**3K]
-0.30175914 ! dSf- [J/cm**3K]
-398.53986 ! dH [J/cm3]
7.83515318E-03 ! c0(C)/BCC_A2
0.29345708 ! c0(C)/FCC_A1
1.0892543 ! c0(MN)/BCC_A2
2.6347178 ! c0(MN)/FCC_A1
0.27729574 ! c0(CR)/BCC_A2
0.40335453 ! c0(CR)/FCC_A1
0.17602370 ! c0(SI)/BCC_A2
0.13777902 ! c0(SI)/FCC_A1

Bernd
Posts: 1088
Joined: Mon Jun 23, 2008 9:29 pm

Re: ThermoCal coupled problem, NPLE and no_TQ problem

Post by Bernd » Tue Oct 30, 2012 1:53 pm

Hi zhubq,

What you are showing is the initial linearisation data which is written into the .log file. These initial values can be calculated under different assumptions, but it is important to know that they lose importance after the first updating or when the interface is moving!

1.) If the user specifies initial concentrations for all phases, then we assume that these phases already exist, and initial equilibrium is calculated for the mixture composition at the center of the interface. On the other hand, if only the composition of the matrix phase is given, the initial equilibrium is calculated with the assumption that the fraction of the other phases are 0, i.e. they are constructed as to be in equilibrium with the matrix composition. This proceeding is a bit arbitrary, but I do not know a better way to do that!
Up to now, nobody told us about problems with this point. But if needed it would be possible to include an extra parameter (in the case of specifying all phase compositions) which determines the fraction of the matrix phase during initialisation.

2.) It is quite common that dS+ and dS- are quite different, especially for solid-solid reactions. If they have different sign it would be an indication of being inside or close to a miscibility gap! Your linearisation data are looking quite reasonable!

Bernd

PS: micress 4.08?, I hope you mean 5.408 :shock:

Post Reply