Incorrect calculation of Gibbs free energy in IdealSolidSolnPhase and BinarySolutionTabulatedThermo

[speth]

Problem description

Calculations involving the calculation of Gibbs free energy when using the IdealSolidSolnPhase class (YAML name ideal-condensed) or BinarySolutionTabulatedThermo class (YAML name binary-solution-tabulated) give thermodynamically inconsistent results when the solution is not at the reference temperature and/or pressure.

>>> import cantera as ct
>>> p = ct.Solution('lithium_ion_battery.yaml', 'electrolyte')
>>> p.TP = 430, 5*101325
>>> p.gibbs_mole
-24597376.815229505
>>> p.enthalpy_mole - p.T * p.entropy_mole
-24567921.30990614
>>> sum(p.chemical_potentials * p.X)
-24567921.309906136

Since the latter two values are consistent, this suggests that the problem is in the gibbs_mole function itself.

Similar results are obtained for the binary-solution-tabulated model by replacing the phase setup steps with:

p = ct.Solution('lithium_ion_battery.yaml', 'anode')
p.TPX = 300, 10*101325, [0.2, 0.8]

System information

• Cantera version: 2.6.0 or main at 8fbb4bc

Additional context

This was discovered as part of #1299, which implements Cantera/enhancements#114.