Quartz Solubility

[H_King]

I am using GWB 6.0.3 on Windows XPSP2

 

Using REACT, I calculate SiO2-aqueous for 1kg water and 100 g quartz from 25 to 150C. pH is set to 7

 

The values obtained are consistently below those reported by Rimstidt in his 1997 review paper on quartz solubility.

 

 

Rimstidt(1997)

T molal SiO2 H4SiO4(PPM)

 

25 0.000182 17.47

105 0.001111 106.78

150 0.002277 218.82

200 0.004305 413.77

 

 

The ultimate goal is to calculate the solubility at T=250C, pH=10 and NaCl=0.5 molar. However, must first get correct answer for fresh water

[Tom Meuzelaar]

I am using GWB 6.0.3 on Windows XPSP2

 

Using REACT, I calculate SiO2-aqueous for 1kg water and 100 g quartz from 25 to 150C. pH is set to 7

 

The values obtained are consistently below those reported by Rimstidt in his 1997 review paper on quartz solubility.

 

 

Rimstidt(1997)

T molal SiO2 H4SiO4(PPM)

 

25 0.000182 17.47

105 0.001111 106.78

150 0.002277 218.82

200 0.004305 413.77

 

 

The ultimate goal is to calculate the solubility at T=250C, pH=10 and NaCl=0.5 molar. However, must first get correct answer for fresh water

 

Hello:

 

Assuming that you are using thermo.dat (LLNL), the solubility data is likely derived from a different source. The solubility of quartz at relatively low temperatures is poorly known. If you want to reproduce Rimstidt's results, you'll need to use his solubility data in your database.

 

You can either modify the solubility constants for Quartz in the database directly, or use the Config - Alter Log Ks... option in the React (or any other) module. The latter option modifies the log K values for the course of your simulation only.

 

Regards,

 

Tom Meuzelaar

RockWare, Inc.

[H_King]

Hello:

 

Assuming that you are using thermo.dat (LLNL), the solubility data is likely derived from a different source. The solubility of quartz at relatively low temperatures is poorly known. If you want to reproduce Rimstidt's results, you'll need to use his solubility data in your database.

 

You can either modify the solubility constants for Quartz in the database directly, or use the Config - Alter Log Ks... option in the React (or any other) module. The latter option modifies the log K values for the course of your simulation only.

 

Regards,

 

Tom Meuzelaar

RockWare, Inc.

 

Thanks for the reply. I think your suggestion would work if I was only interested in fresh water at pH circum-neutral. That's where Rimstidt's results apply. If I understand correctly , extension to different pH's and ionic strengths would require additional inputs to the data base. In other words, these effects are not calculated from a model and then applied to the neutral-water solubility.

 

Perhaps I am misstating the procedure and it's possible to combine the two approaches?

 

Perhaps someone could recommend another source for a data base?

 

The folks at Penn State use a program PHREEQC to calculate solubilities of amorphous silica for pH<7 and salt contents up to 0.24 molal. Their values are in good agreement with expected concentrations, but they do not comment on the application of this program to higher pH or for crystalline quartz.

 

Hubert King

[Tom Meuzelaar]

Thanks for the reply. I think your suggestion would work if I was only interested in fresh water at pH circum-neutral. That's where Rimstidt's results apply. If I understand correctly , extension to different pH's and ionic strengths would require additional inputs to the data base. In other words, these effects are not calculated from a model and then applied to the neutral-water solubility.

 

Perhaps I am misstating the procedure and it's possible to combine the two approaches?

 

Perhaps someone could recommend another source for a data base?

 

The folks at Penn State use a program PHREEQC to calculate solubilities of amorphous silica for pH<7 and salt contents up to 0.24 molal. Their values are in good agreement with expected concentrations, but they do not comment on the application of this program to higher pH or for crystalline quartz.

 

Hubert King

 

Hi Hubert:

 

The activity model specified for the thermo dataset modifies concentrations of aqueous species to account for the effects of variable solution concentrations.

 

The dissolution reaction for Quartz is written in terms of SiO2(aq) in the LLNL database, and in terms of H4SiO4 in the PHREEQC database. You can compare results by loading the PHREEQC database (thermo_phreeqc.dat) into React (via the File - Thermo Data... menu option). Note that the temperature range in thermo_phreeqc.dat is from 0-100C.

 

Regards,

 

Tom Meuzelaar

RockWare, Inc.

[H_King]

Hi Hubert:

 

The activity model specified for the thermo dataset modifies concentrations of aqueous species to account for the effects of variable solution concentrations.

 

The dissolution reaction for Quartz is written in terms of SiO2(aq) in the LLNL database, and in terms of H4SiO4 in the PHREEQC database. You can compare results by loading the PHREEQC database (thermo_phreeqc.dat) into React (via the File - Thermo Data... menu option). Note that the temperature range in thermo_phreeqc.dat is from 0-100C.

 

Regards,

 

Tom Meuzelaar

RockWare, Inc.

[H_King]

Tom,

 

Thanks for your reply. As I understand it, I can input the Log K values from Rimstidt and then let GWB use solution models to calculate the effects of NaCl and pH. Using thermo.dat, I followed your suggestion to modify the Log K values and verified that the fresh water results for quartz solubility are reproduced--that works fine.

Here are the values in case anyone wants to duplicate this

 

T, C Log K

0 -4.0808

25 -3.7406

60 -3.3501

100 -2.9936

150 -2.6427

200 -2.3660

250 -2.1423

300 -1.9575

 

 

I then set the conditions to NaCl=0.5 molar and pH=10. GWB does not complain about these conditions, and I believe that these conditions are appropriate for Debye-Huckel. The elevation in aqueous silica is significant, but the values are within the expected range.

 

Thanks for the advice

 

Hubert King

[Tom Meuzelaar]

Tom,

 

Thanks for your reply. As I understand it, I can input the Log K values from Rimstidt and then let GWB use solution models to calculate the effects of NaCl and pH. Using thermo.dat, I followed your suggestion to modify the Log K values and verified that the fresh water results for quartz solubility are reproduced--that works fine.

Here are the values in case anyone wants to duplicate this

 

T, C Log K

0 -4.0808

25 -3.7406

60 -3.3501

100 -2.9936

150 -2.6427

200 -2.3660

250 -2.1423

300 -1.9575

 

 

I then set the conditions to NaCl=0.5 molar and pH=10. GWB does not complain about these conditions, and I believe that these conditions are appropriate for Debye-Huckel. The elevation in aqueous silica is significant, but the values are within the expected range.

 

Thanks for the advice

 

Hubert King

 

Hi Hubert:

 

Thanks for posting your data and results so that others can benefit down the line.

 

Regards,

 

Tom

[Helge]

NaCl = 0.5 molal is well beyond Debye-Hückel and Davies should be used instead.

Thermodynamic modeling of the solubility of an anhydrous oxide phase might lead to results significantly (even orders or magnitude) lower concentrations than are experimentally obtained. This is because the phase that is actually controlling solubility is a hydrous metastable phase on the surface of the oxide. In the case reported this might be something like "Si(OH)4(am)". logK for this "phase" would be quite different.

 

A (Pitzer-) model for calculations of Si-containing solutions in brines can be found in Reardon, E. J. (1990): An ion interaction model for the determination of chemical equilibria in cement/water systems, Cement and Concrete Research (20), 175-192.

 

Helge