Modeling a batch experiment of CO2-rich water with CaCO3

[lrossi]

Hello all,
I try to reproduce by modeling an experiment that I made in the laboratory. In this experiment, I placed a volume of CaCO3 powder (98% +/- 1%, according to the analyzes of the samples) in a beaker which I filled with a volume of water enriched with CO2. This beaker is closed and there is no headspace. I see over time a gradual increase in pH until the equilibrium, and an increase and decrease in major and traces ions concentrations.

First, I modeled this taking into account only the major ions. I obtain a correct stabilization of the pH  but the HCO3- and Ca2 + ions concentrations increase until reaching a high threshold and they do not decrease.

Here is the script:
   

Code: [Select]

SOLUTION 1
    temp      17.89
    pH        7.89
    pe        4
    redox     pe
    units     mg/l
    density   1
    Alkalinity 247 as HCO3
    Ca        102
    Cl        23
    F         3
    Mg        39
    N(5)      48
    Na        14
    S(6)      53
    water    1 # kg
    K         16

EQUILIBRIUM_PHASES 1
    CO2(g)    -0.1
   
Save solution 1
end


use solution 1
RATE 1
Calcite
KINETICS 1
Calcite
    -formula  Calcite  8
    -m        0.003
    -m0       0.003
    -parms    15380 0.6
    -tol      1e-08
-steps       3024000 in 105 steps # seconds
-step_divide 1
-runge_kutta 3
-bad_step_max 500
save solution 1
end

I feel that the way I am modeling the batch is wrong.
Do you have a solution to my problem?

Thank you in advance,
Regards,
L. Rossi

[dlparkhurst]

I have a few comments on your input file.

Code: [Select]

RATE 1
Calcite

This section has no effect. The keyword RATES (with an S) is used to define the rate expression for kinetic reactions. The word RATE is not interpreted correctly, an is ignored. So, the rate expression used in the calculation is the RATES definition for Calcite in the database that you are using.

The coefficient of 8 in the -formula definition is odd. You define -m 0.003 in the KINETICS definition. Each mole of reactant has 8 moles of calcite, so effectively you have defined 0.024 moles of calcite available to react kinetically. With the 8 coefficient, there is enough calcite to react to equilibrium; about 8.7e- x 8 ~ 0.007 moles of calcite react to reach equilibrium. If you use a coefficient of 1, only 0.003 moles are available, and all the calcite dissolves to reach a saturation index of -0.7. Perhaps this is why you have higher concentrations than you expected. Note your initial solution has a P(CO2) of 10^-0.1, or a partial pressure of about 0.8 atm; hopefully that is what you wanted.

If you add the keyword INCREMENTAL_REACTIONS true, the calculation will proceed faster. The default (INCREMENTAL_REACTIONS false) is a little brain dead in that it runs the kinetic reaction for 1/105 of the total time, then goes back to the start and runs for 2/105 of the total time, etc. With INCREMENTAL_REACTIONS true, it runs 1/105 of the time, stores the result, and runs for the next 1/105 of the time; so, it does not keep repeating the early parts of the integration.

[lrossi]

Hello,
Thank you for your response and your helpful advice. I can correctly model the pH and the increase in Ca concentration is similar to what I get experimentally.
However, I experimentally observe a reprecipitation of Ca over time as the pH increases, but not with the model.
Is it possible to also model this reprecipitation at the same time?

Thank you in advance,
Regards,
L. Rossi

[dlparkhurst]

Theoretically, a closed system at constant temperature would not reprecipitate calcite.

If CO2 escapes, or temperature rises, then you could model those effects.

I suppose there could be an effect from fine grain size, where small grains are more soluble than larger grains because of surface energies. You could look at the literature about recrystallization. You could force the effect by changing the log K with time in the RATES definition, but without some mechanistic basis, you may be simply data fitting.

[lrossi]

Hello

I am coming back to you for advice concerning a new carbonate dissolution model on which I am working on the same experiment (putting a carbonate rock (98% CaCO3 + 2% SiO2 and amorphous iron oxides) in contact with water gasified with CO2).
I observed an increase in the concentrations of most of the major and trace elements correlated with the decrease of the pH. I also observed a decrease in the concentration of SO4, NO3 ions correlated with the decrease in pH.


I was able to model the increase in concentrations of all trace elements using this script with SOLID_SOLUTIONS:

Code: [Select]

REACTION 1
   
    CO2(g)     1
    4 moles in 2000 steps
INCREMENTAL_REACTIONS True


SOLUTION 1
    temp      15
    pH        7.8
    pe        4
    redox     pe
    units     mg/l
    density   1
    Alkalinity 300 as HCO3
    As        0.0009
    Ba        0.023
    Ca        139.6
    Cl        54.1
    Co        0.0003
    Cr        0.0036
    Cu        0.0028
    F         4.55
    Fe        0.01
    K         3
    Mg        4
    Mn        0.005
    N(5)      20
    Na        10
    Ni        0.0008
    Pb        0.0002
    S(6)      59.65
    Sr        0.1
    Zn        0.021
    Sb        0.0003
    Sn        0.0001
    Cd        0.0001
    V         0.0013
    Se        0.0002
    Li        0.0008
    Si        0.0046
    Al        0.0048
    Mo        0.0018
    -water    1 # kg




SOLID_SOLUTIONS 1
    CaSrBaCO3
        -comp Calcite 1
        -comp Witherite 0.0005
        -comp Magnesite 0.005
        -comp Celestite 0.01
        -comp K-Jarosite 0.005
        -comp CuCO3 0.0009
        -comp CoCO3 0.0003
        -comp Rhodochrosite 0.0001
        -comp Cerussite 0.0001
        -comp Smithsonite 0.0002
        -comp Natron 0.0003
        -comp CaCrO4 3e-05
        -comp NiCO3 0.0003
        -comp Otavite 0.0003
        -comp Ca(VO3)2 2e-07
        -comp Quartz 0.002
        -comp NiCO3 0.0002
        -comp Li2MoO4 2e-07
        -comp Fe2(SeO3)3:2H2O 0.0001
        -comp Ca3(AsO4)2:4H2O 1e-08
        -comp Na-Jarosite 0.0002

end





However I am unable to model the increase in Na, K , Cl concentrations and the decrease in SO4 and NO3 concentrations with this method.


I have a doubt about this model, because for me I don't take into account the sorption on iron oxides or on the surface of calcite.
Do you think I am doing wrong? Also, do you know how I could model the decrease of NO3 and SO4?


Thank you in advance
Sincerely,

[dlparkhurst]

What exactly are you trying to model? Are you interested in the major ions, trace elements, or both? If you are interested in the majors, I would ignore the elements with small concentrations.

For major elements, after equilibrating with calcite, barite and fluorite are supersaturated, and you could precipitate a small amount of SO4 and Ca (not much). Goethite and other iron oxides are supersaturated. All of these could be included in equilibrium phases. There could be some sorption of SO4-2 or NO3- on iron oxides, but my guess is that it would not be enough to make much difference in the concentration. Similarly, there is no obvious way to increase the concentrations of Na, K, and Cl. I would want to know that there is a clear, reproducible, significant change in these concentrations  before I would go too far with modeling.

For trace elements, you are talking about very small concentrations, and analytical error becomes important. In theory, solid solutions do account for the fact that there are no pure compounds; all reagents have trace impurities because solid solutions prevent formation of an absolutely pure compound. However, I seriously doubt that you would be able to model these trace elements accurately. There are too many uncertainties in the equilibrium constants for the phases, and more importantly, the solid solutions are probably not ideal (as you have modeled the SOLID_SOLUTION). Further, sorption on the iron oxides and SiO2 may be important, and accurately modeling sorption of trace elements is probably not possible.

Even with ideal solid solutions, I am skeptical of some of the minerals that you have included in the SOLID_SOLUTION. The most likely phases would be those that replace calcium with a cation or carbonate with an anion. I don't think it makes sense to include quartz, jarosite, and a couple of the other phases.

[lrossi]

Hello,
Thank you for your answer.

I would like to model major ions and trace elements.
With the script I sent previously I got precisely the same concentrations that I find in batch or field experiments for Fe, Co, Cu, Mn, Sr, Ca, Mg and HCO3. To achieve this I simply modified the number of moles of each mineral phase in SOLID_SOLUTIONS.
I understand that some phases are not suitable or representative of the carbonate rock I am studying. I had mainly chosen in a first time phases which would allow to represent the impurities in the carbonates.


But with this technique, I can't find the same concentrations for Na, K and Cl which are increasing:
Na: 10 mg/L (pH 7) to 25 mg/L (pH 6)
K: 0.5 mg/L (pH 7) to 7 mg/L (pH 6)
Cl: 15 mg/L (pH 7) to 30 mg/L (pH 6)

and NO3 and SO4 decreasing:
NO3: 30 mg/L (pH 7) to 20 mg/L (pH 6)
SO4 : 65 mg/L (pH 7) to 35 mg/L (pH 6)


I don't see how to model NO3, SO4, Na, K and Cl. I thought that they could be adsorbed on the surface of the calcite or come from small amounts of clays that would have gone unnoticed by XRD.