[OLD] Charge Balance

[webmaster]

From: Greg Anderson

Subject: charge balance

How about some discussion on the question of forcing a charge balance while modeling solutions. Real solutions are of course charge balanced. However due to analytical error or unanalyzed constituents, analyses are generally not perfectly balanced. Craig in his text ("Geochemical Reaction Modeling", p.46) says "it is customary ... " to force a balance, usually on Cl-. Clearly for a good, complete analysis with random analytical error, adjusting some element a little will not make much difference. My question is whether NOT balancing ("balance off") is ever justified. Under what circumstances? If the analyzed elements are without error, but some element is missing, do we not make things worse by charge balancing? What experience do you have on this question?

 

From: Mark J. Logsdon

Subject: Re: charge balance

As Greg suggests, the question arises in two different ways in dealing with high I solutions, such as ARD waste streams, from "real" data sets. Consider a stiff ARD solution (e.g., a copper heap-leach effluent), containing very high SO4 (maybe x0 g/L) and g/L concentrations of Fe, Al, Cu, etc. In most such solutions, Cl- would be quite minor. One problem lies in sampling and analytical chemistry, including the intrinsic problems of precision in messy matrices (e.g., due to metastable solution chemistries, interferences, multiple dilutions, ...). The other problem, as Greg notes, is the possibility that the analysis is missing some component(s). For routine monitoring data at mine sites, this almost always is the case - maybe Sr on the cation side, or F on the anion side, various trace metals and metalloids, or whatever. It seems probable to me that the strategy for dealing with charge balance needs to consider both types of probelms, as well as the principal focus of the simulation. Perhaps it's just a prejudice on my part, but I'm usually more concerned about affecting the apparent solution chemistry (i.e., in the model) in ways I don't understand with respect to anions than cations. For the ARD solutions, this means that I would be inclined to balance on SO4, not Cl or Sr, because (a) the absolute deviation in SO4 is probably greater; ( I actually care about SO4 as part of the problem on the ground as well as inthe model; and © I want to represent the complexation at least qualitatively as well as I can - not force "Cl complexing" where I have no reason to suppose that it would really be much of an issue. If I had a modest I system and was worried about U mobility, I might well take a different approach - maybe model the system three ways (unbalanced; balance on SO4; balance on F; maybe even a fourth and balance on U, or add V if it were missing) and see what my modeled differences were, which ones seemed plausible, which ones croaked the model run, and which assumptions led to qualitatively different views of the solutions. I might even use such a simple test to go back to my client with a rationale for additional testing and analysis.

 

From: Vlassopoulos, Dimitri

Subject: RE: charge balance

That's a question that has occasionally bothered me as well. In my experience with REACT, I've noticed that it is not generally a problem except under what I would call 'extreme' modeling conditions (with a large initial imbalance or at later stages of reaction progress when adding a large mass of reactants) when problems with convergence arise.

One way to possibly get around this (though I haven't actually tried it yet) may be to define a 'dummy' ionic basis species that does not participate in any reactions, and can therefore be used for setting charge balance without affecting the composition. One could argue that this may lead to errors in the ionic strength which would affect activity coefficients and therefore speciation. If you start with an incomplete (not charged balance) water analysis, however, the error in the calculated ionic strength may be even greater. I'd be interested in your comments/suggestions.