Alkalinity balance

[GeeqC]

Hi everyone, dear David Parkhurst,

I am coming back with a problem of alkalinity:
In some solutions (e.g. the one below), the alkalinity of the output is changed. In other contributions to the forum it says that an alert appears if not enough weak acids are available that could act as carrier of the alkalinity.

But how is the alkalinity actually modified? Based on what assumptions? At the same time, C(4) and charge balance remained more or less unchanged.

Besides, also in solutions with sufficient weak acids, alkalinity is sometimes changed, although not as strongly. Apparently, there is a command to display all the species contributing to alkalinity (-alk to PRINT?) but it didn't work so far.

SOLUTION_SPREAD
    -temp     10
    -units    mmol/l

Alkalinity  13.612
C(4)         6.17
N(-3)        0.68
S(6)         0.00
S(-2)        4.57
Si             0.65
P              0.44
Ca            3.02
Mg           11.25
Na            91.58
K              1.95
Cl             109.14    charge

SELECTED_OUTPUT 1
    -file                 Porewater_solution.xls
    -simulation           true
    -solution             true
    -pH                   true
    -reaction             true
    -alkalinity           true
    -ionic_strength       true
    -charge_balance       true
    -totals               Cl  C(4)
    -saturation_indices   Calcite  CO2(g)

Best wishes
Patrick

[dlparkhurst]

To print the contributions to alkalinity, use the following:

Code: [Select]

PRINT
-alk

The alkalinity is different because of the conversion from mmol/L to mol/kgw (actually eq/L and eq/kgw). I assume that is the answer, but post again if you still have questions.

[GeeqC]

The PRINT -alk function works now. A list with all species contributing to alkalinity is included in the output file.

I am using mmol/l because the data are given in this unit. Adding:
    -density 1 calc
    -water 1 # kg
returns totals close to the input values.

A message appears: "pH will be adjusted to obtain desired alkalinity".
Indeed, the alkalinities in the unreacted output solutions are fine. Only after the reaction step, alkalinities are by up to 2 meq/l lower:

REACTION_PRESSURE 1-36
    2.0000
REACTION_TEMPERATURE 1-36
    10
END

RUN_CELLS
    -cells 1-36
END

Something seems to modify the alkalinity there.

[dlparkhurst]

I did not notice that you have defined both C(4) and Alkalinity in the SOLUTION definition. PHREEQC treats that as a special case. Of pH, Alkalinity, and C(4), only two of these items can be defined independently; the third can be calculated from the other two. So, in your case, the pH is adjusted to be consistent with the Alkalinity and C(4) that you have defined.

Usually, you do not want to make this calculation if you have a reliable pH. The most common situation is pH and Alkalinity, which you should probably use if you have these two measurements; TDIC will be calculated from the two values.

The change in alkalinity is due to redox reactions when a "reaction" calculation occurs. For all calculations other than the SOLUTION (initial solution) calculation, redox equilibrium will be calculated. Initial solution calculations allow for redox disequilibrium; for example, O2(aq) can coexist with S(-2) in the SOLUTION definition. If you run the following, you will find that the reaction solution contains C(-4), N(0), and S(6), which were not present in the initial solution. The result of these reactions changes the alkalinity. You can see the change in alkalinity species in the detailed print from PRINT; -alk.

Code: [Select]

SOLUTION 1
    -temp     10
    -units    mmol/l

Alkalinity  13.612
C(4)         6.17
N(-3)        0.68
S(6)         0.00
S(-2)        4.57
Si             0.65
P              0.44
Ca            3.02
Mg           11.25
Na            91.58
K              1.95
Cl             109.14    charge
END
RUN_CELLS
-cell 1
END


[GeeqC]

... I ran your example and compared the species contributing to alkalinity before and after reaction. Further sulphur species appear, as you say. I assume this is because sulphur is the only element for which two redox states are given with the input data.

The question is now, what can we do to bring back the alkalinity to the original value?
- Option 1: not running the reaction step.
- Option 2: is there a way to keep the alkalinity constant or to re-adjust the alkalinity after reaction (and change the pH instead)?

P.S. The reason why I use alkalinity is because it was measured. The measured pH may not be reliable due to escape of CO2.

[dlparkhurst]

Just to note, the pH that is calculated results in a partial pressure of CO2(g) of 10^-4.23, which is less than atmospheric. You will have to decide if that is reasonable and how best to handle the carbonate system.

You can remove most of the redox reactions if you use the Amm database and specify the N(-3) as Amm. No N2(aq) will be generated in a reaction calculation. There still will be some reduction of C(4) to C(-4) and oxidation of S(-2) to S(6) to obtain redox equilibrium in a reaction calculation.

[GeeqC]

In this simulation pCO2 should be calculated based on measured alkalinity, DIC, and ion concentrations in a brackish porewater system.

Closer examination revealed that during the reaction step, the totals are shifted from S(-2) to S(6). This oxidation is compensated by Cl as indicated for charge balance.

Oxidation of sulphide to sulphate should not occur because no other redox-active elements are defined. Under these conditions, I do not understand how the system would reach a "redox equilibrium" (equilibrium with respect to what?).

A workaround could be not to use any reaction steps. Or is there an option to "kinetically inhibit" the oxidation of sulphide? I would prefer to use the SIT database to have a more accurate activity calculation in seawater-like solutions.

[dlparkhurst]

Your choice on how you want to handle pH, alkalinity and TDIC.

Cl concentration does not change during the reaction calculation.

Carbon is the other redox couple. So, C(4)/C(-4) andS(6)/S(-2) are in redox equilibrium.

I don't know what you are trying to do. Phreeqc is calculating equilibrium reactions. If you don't like that you will have to disable some of the reaction by decreasing the log Ks for secondary master species reactions.

[GeeqC]

So far, my question concerning alkalinity balance is mostly resolved.

There still remain some questions about redox equilibria. But this can be discussed another time, as part of a new topic, so that other forum readers can follow better.

Thanks so much for your patience!

[dlparkhurst]

Just a couple other comments. If your pH from Alkalinity and TDIC is correct, then if you were measuring the pH of your sample as it exchanged with the air, CO2 would be entering the solution and the pH would be decreasing.

Your sample has ammonia, sulfide, and no sulfate. Given those characteristics, there is a good chance that there is dissolved methane in the solution as well. If there is enough methane in the the SOLUTION definition, then when the solution is used in a reaction, there will not be any significant oxidation of S(-2) in the reaction calculation. I'm not sure where your sample is from. The P(CH4) would need to be fairly high for it to be in equilibrium with an insignificant concentration of sulfate.

[GeeqC]

Further investigation of the problem may have revealed some insight:

The alkalinity I used for the input solution in Phreeqc was calculated based on a conservative ion balance according to the definition of Wolf-Gladrow et al. (2007). This definition is compatible with titration alkalinity as defined by Dickson (1981), but deviates from other definitions, such as by Peng et al. (1987).

The problem was that using the explicit conservative alkalinity as input in Phreeqc always resulted in a charge imbalance and a shift in alkalinity upon reaction.

Thanks to the PRINT -Alk function, I could now see how Phreeqc calculates alkalinity. Indeed, the calculation differs from the definition of Dickson (1981) in the stoichiometic factors for phosphates (they are as in Peng et al., 1987).

If I leave out total phosphate in the calculation for explicit conservative alkalinity, the final solution is charge balanced, and the alkalinity remains unchanged even after the reaction step.


Reference:
Wolf-Gladrow, D.A., Zeebe, R.E., Klaas, C., K?rtzinger, A. and Dickson, A.G.
(2007). Total alkalinity: the explicit conservative expression and its application to biogeochemical processes. Mar. Chem. 106, 287?300.

[dlparkhurst]

I don't know which database you are using. The alkalinity definitions are in SOLUTION_MASTER_SPECIES. For phreeqc.dat, the definition is such that H2PO4- is the species with 0 alkalinity; PO4-3 has an alkalinity of 2. A quick read looked like that is consistent with the reference that you give. However, other databases are different--llnl.dat has H3PO4 as the zero-alkalinity species.

You can use a SOLUTION_MASTER_SPECIES data block in your input file (or change the the defiinition in the database file) to redefine the alkalinity contribution of the master species for the database.

[GeeqC]

Yes, indeed. In the phreeqc.dat database, PO4-3 is defined at the level of 2 protons, i.e. H2PO4- is set as zero level of protons. However, in the sit.dat database, H2PO4- is set as 1. Replacing 1 by 0 for the Master Species solves the problem.

But there is another problem with the sit.dat database for mac. Several sulphur species are listed under total O(0) in the output file: S5-2, S4-2, S3-2, etc. For this reason, a total O(0) content of 4 mM appear, even though no oxygen is given in the input file. This causes strong shifts in sulphur species and also in alkalinity and pH.

I have noticed that this problem only occurs in the sit.dat database and only in the mac version. If I run the same solution with the sit.dat database from the PC version, everything runs fine.

[dlparkhurst]

The Mac version is in the process of being updated to version 3.8.0. Try transferring sit.dat from the PC version to the Mac. I think it is just that there have been updates to the sit.dat database.

(I never use the polysulfides anyway.)

[patersonb]

Is there a timescale for the updated 3.8.0 version for Mac and through what route will it be available? The USGS website only allows downloads of older versions for PC (3.7.3) etc. at present. And the most recent Mac version I can find is 3.5.0.

[dlparkhurst]

Brian is working on it this week. Brian's site is https://visions-of-quality.com/macphreeqc/.

Download will be available at https://sourceforge.net/projects/macphreeqc/.