Hello everyone,
I’m working on modelling 35 water samples collected from an acid mine drainage site (located in the same general area but from different spots).
My main objectives are:
Calculate the distribution of fluoride species in each sample (F⁻, HF, Al–F complexes, etc.).
Identify key factors affecting fluoride speciation—such as pH, ionic strength, major cations/anions, and possible equilibrium with fluoride-bearing minerals.
I have a few specific questions:
Multiple samples: Can we run all 35 samples together in a single PHREEQC input file, or do they need to be run separately?
Output analysis: If I export results to a CSV file after running the model, will it directly identify the key factors controlling fluoride speciation, or will we need to process the output data further (e.g., statistical analysis or correlation) to find those relationships?
Equilibrium phases: Which equilibrium phases are most relevant for fluoride in acidic mine drainage (e.g., fluorite, cryolite, and aluminum fluoride minerals), and how should they be set up in the input?
Exporting log K: From the saturation indices section, how can we export the log K values (equilibrium constants) used for each mineral phase to a CSV or output file?
I have full chemical analyses for all samples (major cations/anions, pH, temperature, etc.) and can format them for PHREEQC’s SOLUTION blocks.
If anyone has example input files, guidance on structuring multiple-sample runs, or advice on interpreting the results to pinpoint fluoride controls, I’d greatly appreciate it.
Thank you!