Unable to converge at low water-rock mass ratio for anorthite dissolution

[ats5482]

Hi,

I'm trying to model anorthite dissolution in Archean-like seawater under a wide range of w/r (kg:kg) conditions. I would like to run models for w/r = 0.001 - 10,000 (expecting to see a dilution curve at high w/r). However, I can't get the model to converge below w/r ~ 0.33 (i.e., An = 0.033 kg, H2O = 0.1 kg).

I attached a copy of my input script. Can I get some advice on what I'm doing incorrectly?

 

Thanks in advance!

-- Adam

input_script_An_archean_var_wr_allsupp.txt

[Jia Wang]

Hello,

 

I took a quick look at your script and I noticed that you have suppressed all minerals. When you do this, the program won't allow any minerals to precipitate even when it's supersaturated. As you decrease the amount of water even further, the program can't numerically converge on a stable state for your system which has a lot of highly supersaturated minerals. Did you intend to suppress all minerals? If I unsuppress Anorthite in your system, I am able to run your script at a much lower water to anorthite ratio than 0.33 and have no issues with convergence.

 

I think this would also answer your questions in the other post too. If you want to allow Anorthite to accumulate in your system when it's thermodynamically favorable, then you would need to unsuppress Anorthite. The mass reacted you plot under the Variable Type is the total amount of stuff (minerals, aqueous species, etc) you have added to your system in the reaction path. In your case, you only added 100 grams of Anorthite. To see the minerals that precipitate in your system, you would want to plot the Variable Type "Minerals". Again, if you suppress all minerals, the program won't consider any minerals for precipitation in your system.

 

If you would like to see more examples of titration paths and how they work, please see the example in section 3.1 in the Reaction Modeling User Guide. For more information on the suppress feature, please see section 6.1 the GWB Command Reference.

 

Hope this helps,

Jia Wang
Aqueous Solutions LLC

 

[ats5482]

Hi Jia,

I did intend to suppress all secondary minerals just because I'm doing sensitivity tests and adding complexity as I learn how each parameter affects the system. It makes sense now why this would create an unstable system at low w-r ratios, so thank you for your explanation.

Re my other post: My confusion is more so whether React requires that all of the primary mineral be consumed to work. So if 100 g of Anorthite is "reacted", does that mean it is necessarily gone, or simply that all of it has been titrated into the system? In reality, if you have a tub of salty water and place some relatively large mass of An into it, only a finite amount of An can dissolve before it is saturated. Or would it be more correct to say that, given enough time, it would indeed all react and turn into whatever secondary minerals are stable and this is the equilibrium that React is reporting? Apologies if this isn't a well-posed question. I just want to make sure I'm interpreting the results correctly.

 

Thanks again!

-- Adam

[Jia Wang]

Hello Adam,

You're welcome. React titrates incrementally the reactants you have set in your Reactants tab and reports the equilibrium state at each step. By default, React divides titration paths into 100 steps. For example, in your script, 100 grams of Anorthite is set to titrate into the initial fluid over the course of the simulation. The program is going to add in 1 gram of Anorthite at each time step into the initial system until all 100 grams are added. The equilibrium state of your system is reported at each step.

In its calculation of the equilibrium state, React will allow supersaturated minerals that are most thermodynamically stable in your system to precipitate, if precipitation is enabled and the mineral is not suppressed from forming. If suppressed, the mineral is not considered. The mineral assemblage does not differentiate between primary (i.e. the mineral(s) you are titrating in) minerals or secondary (other minerals that were not present initially) minerals. In your case however, the program won't consider any minerals for precipitation in its calculation because they are all suppressed. If you disable precipitation (go to Config -> Options), it would also have the same effect.

I think it's also important to note here that the only thing that matters for a simple reactant is its composition. Perhaps an example will be best to demonstrate what I mean. In a new React instance, set an initial fluid with 1 mg/kg of SiO2(aq) and in the Reactants pane, add 10 mmol of Amrph' silica as a simple mineral. Include a suffix like _amrph_silica.

1 mg/kg SiO2(aq)
react 10 mmol Amrph^silica
suffix _amrph_silica
go

In another React instance, I will set the same initial system but now titrate in 10 mmol of quartz instead.

1 mg/kg SiO2(aq)
react 10 mmol Quartz
suffix _qtz
go

Comparing the results, you will see that Quartz precipitates in both cases, because it is the most thermodynamically stable silica mineral in this system. In both cases, the simple reactants are changing the composition of your system by adding SiO2. Whether it is Amrph^silica or Quartz (or another silica polymorph), does not affect the results. If you return to either of the simulations and suppress all silica minerals (Amrph^silica, Chalcedony, Cristobalite, Quartz, and Tridymite), then no minerals will precipitate and all the silica titrated in will remain in fluid. Disabling precipitation in the Options dialog will also have the same effect.

If you would like more details regarding the conceptual model behind React and other GWB apps, I would recommend checking out the Geochemical and Biogeochemical Reaction Modeling Textbook by Craig Bethke in addition to the user guides installed with the software. I think you will find chapter 2 Modeling Overview very helpful.

Hope this helps,
Jia

[ats5482]

Hi Jia,

Thank you for such a detailed explanation. I ran those two simple dissolution experiments, which were helpful. I read Ch. 2 of Craig's book a while ago, but perhaps I should re-read it now that I have more experience with the software.

One last question regarding the first comment in this post: How low of a water-to-An ratio were you able to get to work? I ran the attached script (which allows precipitation) and I can't get it to go below w/r ~ 0.14 (i.e., 1 kg H2O, 7 kg An). Ideally I'd like to be able to go down to 0.1 because there are hydrothermal alteration papers that report and model w/r that low.

Edit: I should note that the convergence error I get is:

-- Can't converge, abandoning path.
-- Xi step is too small

I tried increasing delXi to no avail.

Best,
Adam

An_diss.txt

[Jia Wang]

Hello Adam,

You're welcome. I am glad you found the examples helpful. I took a quick look at the second script and compared it to your first script. Your first script had the CO2 gas swapped in for H+, but your second script has CO2(g) swapped in for HCO3-. Is this correct?

With your original script, I unsuppressed Anorthite as a test and was able to converge with going as far as 100 kg of anorthite. Once the system is saturated with respect to the anorthite, it simply starts to accumulate. Not really sure what minerals you're expecting your system to form when titrating the anorthite. Are there any specific secondary minerals you are expecting to form in this system? It initially sounds like you were doing a test with anorthite titration until it's saturated but perhaps I am not really understanding what you are hoping to achieve ultimately. You could try unsuppressing one mineral at a time when testing. Again, the program will consider everything that's not suppressed as part of the simulation.

Best regards,
Jia