coupled substitution in solid solution minerals

[Brendan]

Hi,
In using PHREEQC for inverse and mixing models, I used coupled substitution (knowing the molar Na/Ca ratio) to specify plagioclase compositions (An#) for different waters. I made up my own "fictive phases" as I have seen them called, thinking that maybe I could run the models with smaller uncertainties. It seems to work well, so I also specified mafic minerals by molar Fe/Mg, and volcanic glasses by Al/Si. The reason for this post is that I would like to know if this is a common approach, and if anyone knows of any paper (or author) that I could cite where a similar thing was done.
Thanks,
Brendan

[dlparkhurst]

Sounds like you are specifying a mineral composition, which is perfectly legitimate. However, specifying a particular ratio form element stoichiometry usually increases the adjustments to solution composition to match that ratio, which might require the opposite of decreasing the uncertainty specification.

Frequently, I have tried putting in the end member compositions, say anorthite and albite, expecting the program to determine an optimal plagioclase composition for reaction. Usually, the models that are found use one or the other phase, and less commonly both phases.  So, maybe it is best to pick the composition that is most reasonable as you are doing.

Note too, that if you have two models, one with albite and one with anorthite, linear combinations of the models are also feasible models.