PHREEQC simulation for the upflow percolatoin by column test

[Khyeam EANG]

Hello,

I have some questions regarding PHREEQC simulation for the column test. My column was installed with 30cm layer thickness of filled materials and 5cm of column diameter. The infilling solution (dissolved CaCl2 1mmol/L into distilled water with adjusted pH to 4) percolated down to upward direction with a flow rate of 12mL/hour as shown Fig. 1 (see the attached file).
I plan to model solute transport in a kinetic column test of massive fluorine-containing slag and a slight amount of dolomite hydroxide with potassium dihydrogen phosphate (K2HPO4). The leachate was collected every week for monitoring the releasing fluoride, pH, and temperature.

In this case, my infilling solution (Solution 0) is known for the details; however, I do not have the details of leachate (Solution 1-10, i.e., the column was assigned into 10 cells). As the leachate in the column was basically monitoring the releasing fluoride, pH, and temperature, so what should be the input for Solution 1-10 to model using TRANSPORT? This column test is now reached 532 days and we are measuring the concentrations of all existing ions and other properties before we are going to stop the test. So, what to make use of these results for the input of Solution 1-10?

*   filled materials:
   1237.86 g of fluorine containing slag (with the content of fluorine is 2500mg F/kg slag) and its particle size more than 50% between 2mm and 4mm, where the particle size of the rest is smaller than 2mm.
   37.14 g (~3wt.%) of dolomite hydroxide and K2HPO4 (31.569g of dolomite hydroxide = 19.95g Ca(OH)2 + 7.1g Mg(OH)2 + 3.16g MgO + 0.82g CaCO3; 5.571g of K2HPO4).

(Note: The fluorine containing slag consists of substantial amount of CaMgSiO4 (Monticellite) according to XRF and XRD results in Tables 1 and 2 in the attachment. The presence of 3wt.% of dolomite hydroxide and K2HPO4 was to form fluorapatite Ca5(PO4)3F for reducing the fluoride releasing from the slag.)

Thank you and I would appreciate your response and any suggestion.

Regards,
Khy Eam

[dlparkhurst]

Sounds like you have data to evaluate the rate of release of the single constituent fluoride with time. You could use TRANSPORT and KINETICS to determine a rate expression, but I don't think the overhead of full reactive transport is warranted just to develop fluoride kinetics.

When considering reactions of minerals and transport of major solutes, I think it will be difficult to incorporate fluoride. For reactive transport modeling, it helps to have well-defined minerals of known stoichiometry and thermodynamics. I suspect you do not have that for any minerals, much less the fluoride bearing phases. My first estimate is that I would not try reaction transport modeling, only a simplified kinetic expression for fluoride release with time.

The next step to consider is inverse modeling with the complete chemical analyses you have at the end of the experiment. This would be a simpler way to try to get at the major reactions in your column. It will still not be easy because of the unknown stoichiometries of reactants. But, I think you need to evaluate the reactions from inverse modeling, before you have any chance of reactive transport modeling.

[Khyeam EANG]

Dear Prof. David,

Thank you very much for your responses and suggestions. I would say yes to all your mentioned points. That sounds reasonable and makes sense. So, first of all I will model only a simplified kinetic expression for fluoride release with time. If the result is not as expected, I will next involve with reactive transport modeling when all the minerals are well-defined with their details of stoichiometry and thermodynamics.
I hope to ask you again later if I face troubles regarding the input of the kinetic expression and the reactive transport modeling part.

Best regards,

Khyeam