You are here

dataset modifications

3 posts / 0 new
Last post
sina
Joined: 2015 Oct 27
dataset modifications

What bothers me often and where I have absolutely no knowledge about is, how much can I alter the datasets without producing nonsense? That is, my main concern is, lets say in the Holland & Powell dataset, what if I e.g. don't trust the solution for Mn in biotite: am I allowed to exclude the Mn in biotite from the solution or would the result get inconsistent / wrong if I do so? My best guess is that if I alter anything concerning the solution model for a phase, it will result in utter nonsense (?) but sometimes I would like to exclude some of the solutions for (more exotic) things as e.g. Fe3+ or Mn in biotite etc.

 

DougTinkham
Joined: 2015 Oct 30

sina wrote:

What bothers me often and where I have absolutely no knowledge about is, how much can I alter the datasets without producing nonsense? That is, my main concern is, lets say in the Holland & Powell dataset, what if I e.g. don't trust the solution for Mn in biotite: am I allowed to exclude the Mn in biotite from the solution or would the result get inconsistent / wrong if I do so? My best guess is that if I alter anything concerning the solution model for a phase, it will result in utter nonsense (?) but sometimes I would like to exclude some of the solutions for (more exotic) things as e.g. Fe3+ or Mn in biotite etc.

 

 

Hi Sina, 

Whether or not you can simply exclude phase components from a model is dependent on how the a-x model is formulated and the chemical system you are working in. Where you can run in to problems are with the reciprocal style solution models because the models are formulated with a linearly independent set of phase components.

Take KFMASH biotite for example. The incorporation of octahedral Al is done using the eastonite phase component, and there is therefore no siderophyllite component in the model (it would be a linearly dependent member as it can be constructed from eastonite, annite, and phlogopite). If you wanted to do a calculation using biotite in the KFASH system, you would need to rewrite the entire model in terms of the components phl-ann-sdph because if you simply eliminate eastonite from the existing model, the model would not allow any octahedral Al because you would not have a phase component that incorporates octahedral Al.  

Extend that rationale to Mn-bearing systems, and you can easily determine if eliminating a Mn member is possible or not. For most phases, you can simply remove the Mn member and you will be fine, but always make sure first by looking at the substitution vectors and the phase components present.

Doug

sina
Joined: 2015 Oct 27

Dear Doug,

thank you for your answer, it is very helpful.

Best,

Sina

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer