Ideal solid solutions can have as many components as desired in PHREEQC, regular solid solutions have only two components. The properties of regular solid solutions can be defined by various means, but perhaps it cannot describe the experimental data that you have.
If a non-ideal solid solution with more than 2 components must be modeled, or a more general activity model for the components of a solid solution is required, try a kinetic approach suggested by Colin Walker for cement minerals.
The activity of a component in an ideal solid solution is χi, and Σ χi = 1.
The saturation ratio is, for example, for AFm in a solid solution of AFm, Cl_AFm and OH_AFm:
SRχAFm = IAPAFm / (KAFm * χAFm) = SRAFm / χAFm,
where χAFm = mol AFm / tot_AF, and tot_AF = mol (AFm + OH_AFm + Cl_AFm).
In kinetics, use
SAVE k * A/V * χAFm * tot_AF * (1 - SRAFm / χAFm) * TIME,
k in m·mol/s, A in m²/mol, V in m³ solution.
Or:
SAVE k * A/V * (mol_AFm - SRAFm * tot_AF) * TIME.
Figure 1 shows the increase of Cl_AFm when NaCl is added to a solution initially in equilibrium with portlandite and an ideal (SO4, OH2)-calcium-aluminate solid solution, calculated with KINETICS and SOLID_SOLUTIONS (PHREEQC input file kin_ss.phr). The results are the same, but KINETICS is more general.
Any activity model for a component in the solid solution can be added to the rate definitions. The blue line in Figure 1 results when the excess energy of mixing depends on the Cl-component only:
eG = -χCl_AFm * (1 - χCl_AFm)0.3.
The activity coefficients of the components in the solid solution follow from the Gibbs-Duhem equation:
ln(λi) = eG - Σj!=i{χj * ∂eG / ∂χj}.
| Figure 1 | |