Differennce in simple Al-Mg Phasediagram

[ChRen95]

Hello,
So I was thinking about trying pycalphad for binary systems.
I went through the examples (https://pycalphad.org/docs/latest/examples/BinaryExamples.html) and wanted to reproduce the shown phase diagram for Al-Mg. Because the original used could not be downloaded from TDBDB I used the newest one from MatCalc mc_al_v_2036.

I implemented the code

import matplotlib.pyplot as plt
from pycalphad import Database, binplot
import pycalphad.variables as v

database_name = 'mc_al_v2036.tdb'
database = Database(database_name)
components = ['AL', 'MG', 'VA']
possible_phases = database.phases.keys()

number_of_moles = 1
atmospheric_pressure_pascal = 101325

start_temperature_kelvin = 20 + 273.15
end_temperature_kelvin = 700 + 273.15
temperature_step = 10

binplot(database, components, possible_phases, {v.N: number_of_moles, v.P: atmospheric_pressure_pascal, v.T: (start_temperature_kelvin, end_temperature_kelvin, temperature_step), v.X('MG'): (0, 1, 0.02)})

plt.show()

The result is this

So I was wondering what makes the phase diagram from the example so different when using the mc Database?
I always thought that the .tdb files provided by mc where like top notch.

[bocklund]

The database defines the CL_FCC phase as:

 PHASE CL_FCC % 2 1 1 >
Early metastable Mg-Si Co-Cluster. [REF:C31,C32]
Solely use in thermokinetic precipitation simulations (comment pov23).
>> 4 !

so when you do

possible_phases = database.phases.keys()

it includes this phase in the set of phases PyCalphad considers, even though I don't think the database author intended that to be used when computing normal phase equilibria. If I change your definition of possible_phases, I get the following phase diagram.

# possible_phases = database.phases.keys()
possible_phases = list(set(database.phases.keys()) - {"CL_FCC"})