Conceptual Models > Database selection and modification

Selecting a database for ionic strength of 0.7 m

(1/1)

wasylenki:
Hello, I am trying to use PHREEQC for the first time in 10 years. I want to calculate the speciation of Ni in a solution that is similar to seawater. I thought I would use the pitzer.dat database, but it does not contain Ni. When I looked at Ni data in other databases, it looked like the format of data was different, and I wasn't sure how to add that to pitzer.dat. I see there is also sit.dat. Is this appropriate for high ionic strength? I tried just putting in my Ni concentration to the uranium+seawater example, and it does run and give species abundances that seem reasonable. But am I fooling myself?

Thank you for your comments!

dlparkhurst:
You can't really add data from an ion-association model database to a Pitzer database. There are too many interaction parameters between Ni and other ligands that you are ignoring.

At an ionic strength of 0.7, I would probably use wateq4f.dat. I am a little more comfortable with this database than minteq or llnl, but because none of the databases (except pitzer.dat) have really been evaluated holistically, there is not a solid reason. The SIT database should be applicable to higher ionic strengths, but I am not sure that it has been evaluated systematically either. You can always use multiple databases and compare differences. At least that gives you an idea of the uncertainty related to different choices for thermodyanamic data.

wasylenki:
Thank you very much, David. I will try my speciation with wateq4f.dat and see how it compares with the one I did with sit.dat. If the answers are comparable, that will be satisfactory.

I tried this calculation with Geochemist's Workbench first, but it struck me as very strange that <1% of Ni was present at chloro complexes. This result is in contrast to the Ni speciation for seawater described by several references, including Turner et al. (1981, GCA 45, 855) and Byrne (1988, Mar Chem. 25, 163). So I wanted to try PHREEQC to see how much chloride complexation it predicts.

I'm trying to explain why the magnitude of Ni stable isotope fractionation during adsorption to Mn oxides is really sensitive to ionic strength.

wasylenki:
Wow! I tried running with the same input file, once with the sit database and once with wateq4f. The results are frighteningly different. With wateq4f, 59% of Ni occurs as NiCO3 ion, while this species is only 0.4% with the sit database. It isn't even listed in the Geochemist's Workbench output. Ni2+ is 20% of all Ni with wateq4f, 59% with sit, and 89% with GWB.

I have no idea which one to work with as I contemplate the isotope behavior we're seeing. Any idea why these results are so different?

Thanks!

dlparkhurst:
All the databases except pitzer.dat are accumulations of thermodynamic data for individual species. Often data are "selected" from different sources to produce the final aqueous model. This final model is often not checked for consistency with the original sources, which may have used different activity-coefficient models, possibly different standard free energies, or other inconsistent approaches. What is needed is a comparison of the final model with the original experimental data, whether that is mineral solubilities, mean-activity coefficients, pH and emf measurements, or other measureable quantities. But that is a lot of work.

In your case, the different databases probably illustrate different or incomplete sources for nickel complexes.  I don't have an easy solution. To the extent you can find solubility (or other experimental) measurements, you could test each database to see which gives the best results.  But, more generally, you are stuck trying to derive a more consistent aqueous model, either by selecting one of the available databases, additional literature work, or actual selection or refitting of available data.

Navigation

[0] Message Index