Hydrogen dissolution in brine

[ylg]

Hi,
I am a reservoir enginneer from France working on Hydrogen underground storage.
The lab ran an experiment withi Hydrogen (gas) in contact with brine at atmospheric pressure (actually 0.2 MPa). Experiment was done using 125 mL which contained a total of 60 mL of formation water and with addition of 30 mM HCO3- supplied by NaHCO3. Without bacteria, they obesrve CO2 and no CH4 in the gas and pH of about 8.
When runing PHREEQC with phreeqc.dat, the model reach equilibrium and forecast CH4 and no CO2. The only way I could model CO2 was by changing significantly the log of partial pressure equilibrium for CO2 (-35) as shown in the attached data deck. The lab told me CH4 generation is not possible under this conditions.
I found this numerical trick not quite satisfactory and am wandering if this might be due to a wrong definition of the gas liquid equilibrium and hydrogen dissolution in brine which appear to be quite significant
Any clue how to match model and experiment?

[nickliam]

Hi YLG,
I observe similar results in that the hydrogen is not reacting with bicarbonate to form methane (despite phreeqc telling me so). Did you do more H2-bicarbonate experiments to follow up on potential methane vs. carbon dioxide formation, and did you publish your results somewhere?

[dlparkhurst]

By using H2(g) and H2(aq), PHREEQC will calculate thermodynamic equilibrium. So, thermodynamically, the reaction CO2 + 4H2 = CH4 + 2H2O is favored; just like the atmosphere should be nitric acid. Kinetically, the reactions may not occur. It sounds to me like you should use Hdg and Hdg(g), which are defined in phreeqc.dat, Amm.dat, and pitzer.dat. Hdg simulates unreactive H2, so that total thermodynamic equilibrium will not occur. The hydrogen (Hdg) will be inert unless you use an explicit KINETICS reaction to react the hydrogen.

[nickliam]

Thanks for the advise, will do. BTW, I tried to dodge the bullet by adding 2 H2O + 2 e- = H2 + 2 OH- (log_k -17.15) to the phreeqc database. While it should thermodynamically not matter, (provided I got the log_k for that reaction right which I calculated as -3.15 + -14 from water hydrolysis), I wanted to tone down the hideous 2H+ + 2 e- = H2 (log_k -3.15) whilst taking the fact into account that carbonate dissolution will raise the pH. And indeed, when I run my hydrogen-carbonate simulation with these two extra lines in phreeqc.dat, no methane forms.....why, if it is all still thermodynamically consistent? Why have I suddenly tamed hydrogen reactivity? I also tried S(6) reduction in the presence of hydrogen with this modified database and again, no reduction and HS- formation (and H2 loss). What am I missing?

[dlparkhurst]

Now, now. Let's not be harsh. The H2 reaction is not hideous. It is our old friend the Standard Hydrogen Electrode to which all electromotive measurements are referenced.

And yes, you have made a mistake in the log K. You should add -28, not -14, because you are adding 2 H2O to the reaction.

If you want H2 not to react, then I think using Hdg is the way to go. It has the same solubility as H2, but is inert until you impose reactions. Do a search in the forum for Hdg(g). There are several recent posts about Hdg with examples that allow kinetic sulfate reduction, methanogenesis, and acetogenesis.