Here is a special trick which helps defining exact phase fractions in initial multi-phase microstructures. Let's assume we want to simulate Ostwald ripening of gamma-prime phase in a Ni-base superalloy. We could plan to define the initial microstructure based on a Thermo-Calc equilibrium calculation for a given temperature. The spatial distribution of the gamma-prime particles can be created by random setting of initial grains or by reading experimental microstructures from a file The initial concentrations of fcc and gamma-prime then is specified according to the results of the equilibrium calculation. But how can we achive the correct phase fraction of gamma-prime in the initial microstructure which is in accordance with the equilibrium calculation? If the phase fractionm of gamma-prime is not correct - and this will be probably the case due the pixeled nature of the sharp initial structure - then the total composition is not correct!
The trick presented here uses a normally unwanted side effect of the initialisation procedure for the interfaces, i.e. the transformation of the initially sharp interfaces to smooth interfaces. MICRESS allows specifiying the number of initialisation (i.e. phase-field) steps for this smoothening. The side effect consists in a shrinkage of curved structures during this initialisation!
Using this side effect, the procedure is the following: Choose parameters for the initial microstructure such that the phase fraction of gamma-prime is higher than the value to be specified. Then, stepwise increase the number of initialisation time steps until the correct phase fraction is achieved
![Wink ;)](./images/smilies/icon_e_wink.gif)
Bernd