WHICH MODULE ? mu
MULTIPHASE OPTION ? def sy 'Fe,Cu' sou plus !
MULTIPHASE OPTION ? cl ab p(*) no p(FCC_A1) !
MULTIPHASE OPTION ? set temp 1273 !
MULTIPHASE OPTION ? set w 1 !
MULTIPHASE OPTION ? step w(2) 0 1 0.01 !
MULTIPHASE OPTION ? comp pri gra !
MULTIPHASE OPTION ? ord gibbs system !
MULTIPHASE OPTION ? plot go
If you cannot follow this example, you should have a look at the section First Use.
You should obtain the following graph:
Clearly, there are 2 minima in the Gibbs energy vs composition curve. An alloy of bulk composition 50% Cu should give a Fe-rich phase with < 5% Cu in solution , and a Cu-rich phase with < 5% Fe in solution.
MULTIPHASE OPTION ? set w(2) 0.5 !
MULTIPHASE OPTION ? co pri ea_st_re !
MTDATA predicts a single phase, even though with 50% Cu we are clearly in the miscibility gap. This is because MTDATA does not automatically detect miscibility gap. To authorise one miscibility gap in FCC_A1 (and allow MTDATA to search for 2 minima for this phase), you need to do:
MULTIPHASE OPTION ? class misc(FCC_A1) 1 !
MULTIPHASE OPTION ? com pri ea_st_re !
The results should now show 2 phases (with the same name) corresponding to the two end solutions described above.
This example may seem trivial, nevertheless unclassified miscibility gaps are behind 95% of the 'weird answers' sometimes produced by MTDATA.
In particular, most MX precipitates (NbC, NbN, TiC, TiN, etc) are modelled by FCC_A1 which is also austenite.
There are different ways to work around this problem. If you have
a maximum of two such phases (example, TiC and TiN in a steel
containing Ti,C,N), you can set 2 miscibility gaps in FCC_A1:
The other way is to include the phases you know may cause problem,
when you define your system, for example:
MULTIPHASE OPTION ? def sy 'Fe,Ti,N,C' sou plus sub_sgte !
MULTIPHASE OPTION ? cl ab p(*) no p(FCC_A1,CTi,NTi,BCC_A2) !
Having classified as normal TiC and TiN should avoid problems with the miscibility gap in FCC_A1.