Processes > Dissolution and precipitation

Help with Thermodynamic Modeling of Fluoride Removal via Cryolite Precipitation

(1/1)

Norman_Jimenez:
Hi everyone,

I?m currently working on a thermodynamic model using MacPHREEQC as part of my graduate research. My goal is to simulate and predict the percentage removal of fluoride (F⁻) in a fluidized-bed reactor through the precipitation of cryolite (Na₃AlF₆).

Context:
The system involves the reaction of fluoride and aluminum to form cryolite. Sodium is assumed to be non-limiting and in excess. I want to simulate how varying initial concentrations of fluoride and aluminum, as well as initial pH, affect the extent of cryolite precipitation.

Known Inputs:
Initial [F⁻] (mg/L or mol/L)

Initial [Al?⁺] (mg/L or mol/L)

Initial pH
(Na⁺ assumed to be abundant)

My Goals:
Model cryolite (Na₃AlF₆) precipitation using the Saturation Index (SI) or equilibrium phase reaction.

Estimate % fluoride removal based on what precipitates as cryolite.

Eventually incorporate this into a dynamic model, but for now, I?m focused on static equilibrium simulations.

My Questions:
How should I write the PHREEQC input prompt for this scenario?

Is it correct to treat cryolite as a PHASES entry and set its saturation index or amount to zero to see how much would precipitate?

Should I manually define cryolite in the database, or is it available in phreeqc.dat or minteq.v4.dat?

Any tips on selecting a database for accurate fluoride-aluminum speciation?

Any help on structuring the input file or general guidance would be greatly appreciated! Thank you in advance.

Best,
Norman
Graduate Student ? Environmental Engineering
Philippines

dlparkhurst:
The databases have extremely strong complexes between Al and F, which can produce weird results, like saturation indices decreasing as concentrations increasing. So, I am not convinced that calculations with Al and F are accurate, especially at high concentrations and high ionic strengths. I will be how your calculations compare to experiment.

With that caution, here is a script that calculates the precipitation of cryolite as NaCl is added to a solution of Al and F. I took the cryolite definition from minteq.v4.dat and used phreeqc.dat for the calculation.


--- Code: ---PHASES
Cryolite
Na3AlF6 = 3 Na+ + Al+3 + 6 F-
log_k -33.84
delta_h 38 kJ
END
SOLUTION
-units mmol/kgw
pH 7
Al 1
F  3
Na 1 charge
END
USE solution 1
EQUILIBRIUM_PHASES
Cryolite 0 0
REACTION
NaCl 1
10*0.01 10*0.1 5*1
INCREMENTAL_REACTIONS true
USER_GRAPH 1
    -headings               rxn Al F
    -axis_titles            "NaCl added, moles" "Molality" ""
    -axis_scale x_axis      auto auto auto auto log
    -axis_scale y_axis      auto auto auto auto
    -initial_solutions      false
    -connect_simulations    true
    -plot_concentration_vs  x
  -start
10 GRAPH_X TOT("Cl")
20 GRAPH_Y TOT("Al"), TOT("F")
  -end
    -active                 true
END

--- End code ---

Norman_Jimenez:
Thank you so much!

I just wanted to also ask for some advice on how to approach my thermodynamic modeling.

I?ve seen several publications use PHREEQC for this kind of work, but some of them also derive or manipulate speciation equations outside of PHREEQC. I do have the speciation equations, but since I?m planning to rely mainly on PHREEQC, I?m unsure how much of that I still need to manually incorporate.

My primary goal is to predict fluoride recovery under equilibrium conditions (via cryolite precipitation). Would the free F⁻ concentration output from PHREEQC be enough for this type of modeling? Or would I still need to manually use the equations outside the software to calculate the extent of recovery?

Any guidance would be appreciated ? especially on whether using PHREEQC?s built-in equilibrium results would suffice for estimating recovery percentages.

Thanks in advance!

dlparkhurst:
I'm not sure what manipulations you would do outside of PHREEQC. My opinion is that PHREEQC would be a more complete approach than simplified equations, especially considering the strong ion pairing between Al and F- and Al and OH-.

You should probably consider the total dissolved F, not just the free F-. The Basic function TOT("F") gives the amount of F remaining in solution. The Basic Function EQUI("Cryolite") gives the moles of cryolite precipitated.

You may need to consider an Al-OH phase like gibbsite or Al(OH)3(a) (gibbsite being less soluble than Al(OH)3(a)), which could precipitate in your system. You also need to consider pH control to maximize the precipitation of cryolite. Again, I am not completely confident in the thermodynamic data in the databases, but you will have to do the calculations and compare to any experimental data that are available.


Here is a variation of the calculation that adjusts pH.


--- Code: ---PHASES
Cryolite
Na3AlF6 = 3 Na+ + Al+3 + 6 F-
log_k -33.84
delta_h 38 kJ
Fix_H+
H+ = H+
log_k 0
END
SOLUTION
-units mmol/kgw
pH 7
Al 1
F  3
Na 1 charge
EQUILIBRIUM_PHASES
Cryolite 0 0
Gibbsite 0 0
Fix_H+ -6 NaOH 10
REACTION
NaCl 1
5 in 20 steps
USER_GRAPH 1
    -headings               rxn Al F
    -axis_titles            "NaCl added, moles" "Molality" ""
    -axis_scale x_axis      auto auto auto auto #log
    -axis_scale y_axis      auto auto auto auto
    -initial_solutions      true
    -connect_simulations    true
    -plot_concentration_vs  x
  -start
10 GRAPH_X RXN
20 GRAPH_Y TOT("Al"), TOT("F")
  -end
    -active                 true
END

--- End code ---

Navigation

[0] Message Index