Mineral Equilibrium calculation

[Sol-Chan Han]

Hello, Phreeqc users and experts!
I am a new user who has been started using Phreeqc recently.
Many comments, which are posted on this website, were very useful to me.

However, there are some remained problems annoying me. So, I ask you a help to solve some troubles.

I want to model equilibrium phases among the minerals, which are in contact with the water.
I am trying to use EQUILIBRIUM_PHASES BLOCK and REACTION BLOCK. However, they are me give different answers.

Following text is part of my model.

1. When I using Equilibrium phases block

SOLUTION 1
EQUILIBRIUM_PHASES 1
C3S   0   6
C2S   0   3
C3A   0   0.9
C4AF   0   0.2
Brucite 0   0
Hematite 0  0
Portlandite 0  0
END

2. When I using Raction block

SOLUTION 1
EQUILIBRIUM_PHASES 1
Brucite 0   0
Hematite 0  0
Portlandite 0  0
REACTION 1
C3S      6
C2S      3
C3A      0.9
C4AF      0.2
10.1 moles
END

I found that some users used reaction block when they calculated equilibrium between solid and solution, while some users used equilibrium phases block.

Which could be better modeling? Could you please give me any comments?

Thank you.
Best regards,

[dlparkhurst]

Your file has two different conceptual models. In the first input set, you are assuming the moles of each phase and that the phases will react to equilibrium (dissolving or precipitating as needed) to achieve equilibrium (or dissolve completely).

In the second input set, you are adding the specified moles of the four reactants to solution. These reactants dissolve totally increasing the dissolved concentrations of the solution. Then only the minerals in EQUILIBRIUM_PHASES will react to equilibrium. C2S, C3S, and the other two will not be in equilibrium, whereas they probably will be in the first calculation.

[Sol-Chan Han]

Hello sir, I appreciate your comments.

I am sorry that there were some mistakes in my previous post.
I actually also listed C3S, C3S, C3A, and C4AF in the equilibrium_phases block in the second input set as can be seen below.

2. When I using Reaction block

SOLUTION 1
EQUILIBRIUM_PHASES 1
Brucite 0   0
Hematite 0  0
Portlandite 0  0
C3S     0   0
C2S     0   0
C3A     0   0
C4AF   0   0
REACTION 1
C3S      6
C2S      3
C3A      0.9
C4AF      0.2
10.1 moles
END

I am wondering that what kinds of difference can occur between the first and the second input?
In my opinion, I think that the first input is more realistic because the reaction block force to dissolve all phases regardless of the reaction constant.
(I mean that for example even if C3S has very low Ksp, all amount of C3S will be dissolved into the solution by REACTION block)
Could you please give me any comment whether my opinion is right?

Thank you.

[John Mahoney]

The REACTION keyword is best used for irreversible reactions - all of it goes into solution.  EQUILIBRIUM_PHASES Keyword means that an equilibrium assemblage will be made up of the phases in that block.  It will draw components out of the solution (precipitate stable phases) to make up that assemblage.  The other issue with EQUILIBRIUM_PHASES is the activities of many of the components will be defined by the solubility reactions and the log Ksp values.  This is a Phase Rule requirement. 

I suggest you run your model with the  EQP and REACTION blocks and compare it to the model below. You have to have those phases defined for this to work.  I do not have those in my database so I cannot run it.

EQUILIBRIUM_PHASES 1
Brucite 0   0
Hematite 0  0
Portlandite 0  0
C3S   0   60.6
C2S   0   30.3
C3A    0  9.09
C4AF  0    2.02

 


I am also a little concerned about your naming conventions (this is cement terminology I assume)  I assume you have phase names such as C3S,  for a Tri Calcium Silicate solids,  defined  in the PHASES block.  But, and David Parkhurst might want to chime in here,  there might be some confusion using a single capital letter C, which is carbon and followed by a capital letter S for silica, when S represents Sulfur in most databases.  F I assume is iron, but  to the database it is Fluoride.   So you may want to spell out the names to avoid confusion.  If they are defined as phases I think you are OK but be careful.  Check what the model is actually reading.  Capital letters tend to be read as Elements.   

[Sol-Chan Han]

Thank you for Dr. David L Parkhurst and John Mahoney.

However, still, I have a remained question.

I wonder that How the Reaction data block works.
The attached file below is the database what I used. There is no SiO2, Al2O, SO3 etc in the data block as phases.
However following code works well (part of my code). Can you give me any comment that how the Phreeqc recognize the SO3 etc in the Reaction data block? How the Phreeqc dissolves the SiO2 and SO3 to the solution without any thermodynamic data?

USE SOLUTION 1
REACTION 1
SiO2      0.4
Al2O3   0.01
Fe2O3   0.02
CaO      0.7
MgO      0.05
SO3      0.04
Na2O   0.02
36.28292124 moles

[John Mahoney]

These formulas are all defined by the solution_master_species block in the database so in that case it "knows" to link SiO2 to Silicon and Oxygen.    it also knows the difference between C and Ca.  But I still think until you have checked that is why you want to be careful with formula names that include capital C and S,  but I think that will still be OK. 

The reaction keyword block is even sensitive to valence.   For example, I once added Ni to a model and the pe dropped to a negative value so i realized I had to add the oxides.  if I added it as an oxide like NiO. there would have been less change to the redox condition ( maybe no change I cannot recall). 

You should also watch out for charge imbalances using simple oxides avoids most of that problem.   But adding anions can get tricky.  Do you add them as part of a salt such as NaCl so pH changes are minimal or as an Acid, such as HCl,  and in that case the pH will drop. 

In all of these cases I suggest you prepare some very simple models using different reactants and see what happens, and figure out what rules apply to your situation.  It looks like if you name the Phase using the cement nomenclature CS or C3S you will be OK.  But I got a couple of responses with the CS reactant (not named in phases, both as capital letters) in some cases the program selected Cs cesium that was unexpected;  and now it is showing up as C and S, carbon and sulfur,  which make sense for an undefined phase, so use caution and check your outputs, and make sure all your reactants are defined as phases before you add the reactants.  I used a wollastonite reaction for my CS phase.  if you want to properly define those phases you will need log K values for their solubilities.

 Remember to check the added concentrations as well as changes in pH and pe.    Good Luck

[Sol-Chan Han]

Hello, John Mahoney.

I am really thank you for the reply.
I missed explaining that I already defined the phases such as C2S etc.

However, I apologize that my short understanding of how Phreeqc works.
I want to ask you an additional question. Would you please give me advice again?

By following your comment I understand a little bit about how the Reaction data block works.
For example,
1. If I added 1 mole of SiO2 by using Reaction data block, Phreeqc recognizes that 1 mole of Si is added to the solution. And Phreeqc adds 1 mole of Si(OH)4 to the solution (because Solution master species of Si is defined as Si(OH)4 in my database.). Is it right?
2. However, I am confusing following circumstance. If I add 1 mole of Fe2O3 by using Reaction data block, Phreeqc add just 1 mole of Fe to the solution? If Phreeqc adds just 1 mole of Fe, the Fe+2 will be added (because the solution master species of Fe is Fe+2) even though the actual oxidation state of Fe is +3. Is it right? Or Phreeqc recognize that 1 mole Fe(3) is added?

I really thank you for your advice.
I really sorry for interrupting you, however, I hope you give me some comments.

[dlparkhurst]

REACTION is not dependent on the master species. It depends only on the amounts of elements added (BTW if you add Fe+2 in REACTION, the charge is ignored and it is equivalent to adding Fe).

A solution is defined by the total moles of H, O, and all other dissolved elements. So the calculation has a mole-balance equation for every one of these elements. It also uses a charge balance equation, where the charge imbalance is kept internally (based primarily on the charge imbalance obtained by the SOLUTION calculation).

Thus, when you add Fe2O3 in REACTION, the total amounts of Fe and O are incremented, not just Fe. So, adding FeO will be different than adding Fe2O3.

When you add SiO2 in REACTION, the total amounts of Si and O are incremented. The set of equations includes also mass-action equations for each species other than "master" species. In solving these equations (simultaneously with the mole-balance and charge-balance equations), the SiO2 ends up as Si(OH)4(aq), which means that there has been a reaction

SiO2 + 2H2O = Si(OH)4

So, in fact the mass of water will have decreased slightly. If you added Si(OH)4 (or both SiO2 and 2H2O) in REACTION, then the mass of water would probably be very, very small.

[Sol-Chan Han]

Hello prof. David L Parkhurst.

I always thank you for your prompt comments.
This might be a silly question, however, I cannot understand how Phreeqc forces the below reaction.

SiO2 + 2H2O = Si(OH)4

Because there are no related reaction or data either in the database nor my model.

If it is okay for you, would you please explain more details?

In addition, in my thought, if SiO2 is rarely dissoluble solid, I should not use REACTION block (because REACTION dissolves all amount of reactant). But I should use EQUILIBRIUM PHASES block. Will it be right?

[dlparkhurst]

When you add SiO2 as a REACTION it is distributed among the aqueous species that are defined in the database in SOLUTION_SPECIES. Let's suppose as a simplification that the only Si species is Si(OH)4 and that SiO2 is added to pure water. So then all of the Si added must be Si(OH)4(aq); it is the only possible aqueous species with our assumptions. That means that 4 H and 2 O must be taken from somewhere to form Si(OH)4(aq), the only place to get the H and O is water. Basically, the mole balance on H and O require the a slight decrease in solvent water.

The choice of REACTION or EQUILIBRIUM_PHASES depends on your conception of the reactions. I cannot tell you what is right.

[Sol-Chan Han]

Thank you, Prof. David L Parkhurst.

This question would be my last question for this modeling.
I want to make clear the usage of REACTION block.
Can I use rarely soluble solid phase as a reactant in REACTION block?

I really thank you for your comments on my questions.

[dlparkhurst]

My last reply. Do you think the mineral should be assumed to be at equilibrium, or do you want to see reactions as it dissolves or precipitates to approach equilibrium?