[OLD] Precipitation of primary minerals at 25 C

[webmaster]

From: Megan Elwood Madden

Subject: Precipitation of primary minerals at 25 C

I have been running some simple low temperature (25 C) basalt-water reactions and I am a bit puzzled by the mineral assemblages which are predicted-- Diopside, Epidote, Grossular, Albite, Annite etc. I would assume that these high temperature minerals would not be precipitating in a chemical weathering environment at 25 C. Why does this happen? Any ideas about how I can correct for this? I've suppressed many of the minerals I find particularly troubling (Diopside, grossular...) but in some cases that would include the majority of the minerals predicted and many times other questionable minerals are predicted instead.

 

From: Mark J. Logsdon

Subject: Re: Precipitation of primary minerals at 25 C

Your question is a good one. We had an attempt at a discussion of this a number of months ago in the GWB user group, but it didn't get as far as I would have hoped. Later today I'll look through my archive and see if I can find that - I will forward it to you if I do. The short answer is that the program is converging on the lowest G assemblage that is permitted by the database. (See Chapters 3 and 5 of the Green Book.) So long as the database you are working with contains these phases, they must be part of the computation. The situation is equivalent to the one outlined in several places in the Green Book and the user's manuals where the proposed approach for ferric oxides at low-T is to use external information (your geological knowledge) to **assume** that ferric iron will not precipitate initially as hematite or goethite, but rather as an amorphous ferric oxyhydroxide Fe(OH)3 ("ferrihydrite"). Again, one can usefully **assume** that, at 20C or less, SiO2 will not precipitate as quartz, tridymite, or cristobalite, and perhaps not even as chalcedony, so you would suppress those phases, and the model will "precipitate" only amorphous silica. There are lots of other examples - dolomite as a plausible phase with respect to Mg controls, etc. The difficulty, as you are seeing, in working with a complex solution is that the database may contain some hundreds of minerals, and it can take a long while to clean up the database for your purposes. The way React is coded currently there is not any easy way around this. [You have to do the same thing (i.e., subjectively determine or at least constrain the acceptable mineral phases) in PHREEQC, but you build from the bottom up, so to speak, by adding phases for the model to consider rather than by subtracting them from the superset. This seems easier, but in some ways places more of a burden on the modeler because you may not know in advance how to build the set and so could artificially limit the solution. In GWB (assuming you are satisfied with the mineral database) you can screen downward by looking at the saturation indices and comparing them to your prior-knowledge base. It just takes a while.

Here's a possible strategy that I sometimes find helpful:

1. Run React with "precip off". This will simply give you the apparent saturation state of the solution. You may or may not want to set "species" and "saturations" to "long" - you'll probably have to experiment with that. Then critically evaluate the output and set up a list of phases to be suppressed.

2. Re-enter the model, including the suppressed minerals (remember to reset "precip").

3. Check the output again (and again) until you have purged the active database of all the phases you wish the calculations **not** to consider.

4. Save this input model as a "script" so that you don't have to go through all this pain on every simulation. (Then save your subsequent simulations with other names, of course.) You can easily make changes to the script at the beginning of the simulation runs if, for example, you decide to look at the differences that would arise from assuming chalcedony saturation instead of amorphous silica (by "unsuppress"=ing the phases you wish to add back in).

This may seem pedantic, but I have found that the systematic approach actually (a) is more efficient in the long run and ( serves to focus my attention on the subjective nature of the modeling process. [You could also create a custom database with only the mineral phases that you wish to consider -- but be **sure** to follow the guidance on how to do this and also how to re-name your custom database. I get nervous messing about with databases and personally prefer to deal with the problem on a more ad-hoc basis. Also, the code is so fast on any plausible First-World computer that there is no great advantage in terms of modeling time to working with the sparse database. However, if you are going to be using this for thesis research, for example, you may want to consider this in the interests of elegance. Obviously, if you do this, it becomes harder to add phases back in if you change your mind or problem, because you will have to re-enter the database again before you begin modeling. This all strikes me as a matter of taste.]

 

From: Craig Bethke

Subject: Re: Precipitation of primary minerals at 25 C

Mark gave you some good advice. It looks to me, however, like you may be having a different kind of problem. You haven't included an example of your input, so it's difficult to say for sure, but it looks to me like your reaction path is producing an alkaline fluid. If you are trying to model the weathering of basalt by reaction with water, this is the inevitable result. The hydrogen ions in solution are consumed by the dissolution reaction, which proceeds by proton attack. Real weathering, on the other hand, is driven by CO2 from the soil zone and atmosphere. Have you included CO2 in your model? If you do, your reaction path will remain acidic and you will not get the unexpected phases you mention.