Hydrogen Storage

[kadie89]

Hello!

I am currently using PHREEQC as part of my Master’s thesis and was hoping for some insight.

I am trying to build a system representing the injection of hydrogen into a sandstone reservoir, in order to assess possible fluid-rock interactions. I have constructed an input file containing the brine fluid composition and a mineral assemblage in an equilibrium batch reaction, and then into this introduced a gas phase containing pure hydrogen.

Code: [Select]

DATABASE c:\phreeqc\database\PHREEQC.dat

TITLE Mineral Assemblage Douglas Field

Solution 1
temp 55
pH 6.66
-units mg/l
Na 104140 charge
Ca 2365
K 398
Mg 836
Sr 58
Li 6.6
Mn 0.3
Fe 2.5
B 2.6
Al 5.8
Cl 155610
Br 145

   
EQUILIBRIUM_PHASES 1
Quartz 0 51.06
K-Feldspar 0 2.895         
Hematite 0 3.537         
Pyrite 0 3.826             
Dolomite 0 1.457           
Anhydrite 0 1.063         
Calcite 0 1.375           
Halite 0 0.961             
Illite 0 0.358             
Chlorite(14A) 0 0.0495     
     
Save Solution 1
Save Equilibrium_Phases 1
END     

Use Solution 1
Use Equilibrium_Phases 1

GAS_PHASE 2
-fixed_pressure
-pressure 81
-volume 7.5
-temp 55
H2(g) 81

END

I am interested in understanding the relationship between the carbonate minerals and the pH and pe as I increase the partial pressure of hydrogen in the gas phase. It appears as though at hydrogen partial pressures greater than 20atm, the calcite begins to dissolve, and methane generation begins, also coinciding with the consumption of CO2. The dissolution of calcite interestingly coincides with a large increase in pH and sharp decrease in pe value. I initially thought of the Sabatier reaction (CO2 + 4H2 = CH4 + 2H2O) as an explanation for the methane generation but after looking in the database (phreeqc.dat) saw that perhaps the following reaction is facilitating the methane generation, and perhaps is also influencing the pH and pe:

CO32- + 10H+ + 8e- = CH4 + 3H2O

And thus, the removal of carbonate from solution, according to Le Chatelier’s principle would push the following reaction to the left, explaining the consumption of CO2:

CO32- + 2H+ = CO2 + H2O   

However, I am unsure of these results. Is it likely that large quantities of methane generation would be favoured in reality? I am also curious about the dissolution and precipitation of other minerals as the partial pressure of H2 is increased e.g. hematite precipitates at partial pressures greater than 20atm despite a very low pe?

PHREEQC also calculates a positive saturation index for several serpentine minerals (chrysotile, talc, sepiolite), but am I correct in thinking that if kinetics do not permit the precipitation of these minerals (which from literature, it seems unlikely at the temperature in my system) they would not precipitate in reality? 

Any feedback is greatly appreciated!

Kind Regards,

Kate

[dlparkhurst]

Here is a simulation that may help. It adds H2 incrementally. Initially, there is no gas phase, but as more H2 is added, a methane-dominated gas phase evolves to a hydrogen-dominated gas phase.

The addition of hydrogen reduces sulfate and carbonate until anhydrite, dolomite, and calcite are completely dissolved. That much is clear; reactions with silicates, the formation of portlandite, and other details are subject to assumptions.

This calculation simulates pure thermodynamics. The rate that the reduction reactions occur is another question, which I have no experience. The opposite extreme is that H2(g) is essentially inert, and none of the reduction reactions occur. 

Code: [Select]

PHASES
Brucite             19
        Mg(OH)2 + 2H+ = Mg+2 + 2H2O
        log_k           16.84
        delta_h -27.1 kcal
Portlandite         539
        Ca(OH)2 + 2H+ = Ca+2 + 2H2O
        log_k           22.8
        delta_h -31.0 kcal
TITLE Mineral Assemblage Douglas Field

Solution 1
pressure 81
temp 55
pH 6.66
-density 1 calc
-units mg/l
Na 104140 charge
Ca 2365
K 398
Mg 836
Sr 58
#Li 6.6
#Mn 0.3
Fe 2.5
#B 2.6
Al 5.8
Cl 155610
#Br 145

   
EQUILIBRIUM_PHASES 1
Quartz 0 51.06
K-Feldspar 0 2.895         
Hematite 0 3.537         
Pyrite 0 3.826             
Dolomite 0 1.457           
Anhydrite 0 1.063         
Calcite 0 1.375           
Halite 0 0.961             
Illite 0 0.358             
Chlorite(14A) 0 0.0495     
Brucite 0 0
Portlandite 0 0
   
Save Solution 1
Save Equilibrium_Phases 1
END     

Use Solution 1
Use Equilibrium_Phases 1
GAS_PHASE
-fixed_pressure
-pressure 81
CO2(g) 0
CH4(g) 0
H2(g) 0
H2S(g) 0
H2O(g) 0
REACTION 1
H2 1
50 in 50 steps
USER_GRAPH 1
    -headings               rxn pe pH
    -axis_titles            "H2, moles added" "pH, pe" ""
    -axis_scale x_axis      0 50 auto auto
    -initial_solutions      false
    -connect_simulations    true
    -plot_concentration_vs  x
  -start
10 GRAPH_X RXN
20 GRAPH_Y -LA("e-"), -LA("H+")
  -end
USER_GRAPH 2
    -headings               rxn P(H2) P(CH4) P(CO2) P(H2S) P(H2O)
    -axis_titles            "H2, moles added" "Partial pressure"
    -axis_scale x_axis      0 50 auto auto
    -initial_solutions      false
    -connect_simulations    true
    -plot_concentration_vs  x
  -start
5 if (GAS_VM <= 0) then goto 100
10 GRAPH_X RXN
20 GRAPH_Y PR_P("H2(g)"), PR_P("CH4(g)"), PR_P("CO2(g)"), PR_P("H2S(g)"), PR_P("H2O(g)")
100 REM
  -end
USER_GRAPH 3
    -headings               rxn Quartz/10 K-Feldspar Hematite Pyrite \         
Dolomite Anhydrite  Calcite Halite Illite  Chlorite(14A) Brucite Portlandite
    -axis_titles            "H2, moles added" "Moles"
    -axis_scale x_axis      0 50 auto auto
    -initial_solutions      false
    -connect_simulations    true
    -plot_concentration_vs  x
  -start
10 GRAPH_X RXN
20 GRAPH_Y EQUI("Quartz")/10, EQUI("K-Feldspar"), EQUI("Hematite"), \
EQUI("Pyrite"), EQUI("Dolomite"), EQUI("Anhydrite"), EQUI("Calcite"), \
EQUI("Halite"), EQUI("Illite"), EQUI("Chlorite(14A)"), EQUI("Brucite"), EQUI("Portlandite")
  -end
END


[kadie89]

Thank you so much for your help - I found it very useful and interesting!

This might be a very simple question, but I was wondering if you could help me understand how the pH is buffered prior to the carbonate reduction reaction taking effect? I assume it is related to the presence of the carbonates, as when I remove either dolomite or calcite the pH is not buffered, but I was confused about the mechanism, or reaction directly involved in buffering the pH? 

Thank you!

[dlparkhurst]

When Anhydrite, Calcite, and Dolomite are present, you have a dedolomitization reaction. See works by Plummer for examples in Florida and North Dakota where organic material is the reducing agent.

Basically, sulfate is reduced, which causes dissolution of anhydrite; increased Ca+2 concentration causes precipitation of calcite; decreased carbonate concentration causes dissolution of dolomite. The net reaction is relatively pH neutral. Once anhydrite or dolomite is gone, the pH increases substantially.

There is some buffering if calcite or dolomite are absent from the calculation, it is simply at much smaller additions of H2.

As a general rule, consider pH as the ratio of CO3-2:HCO3-. Reactions that increase the ratio increase pH; reactions that decrease the ratio lower the pH. In the absence of the dedolomitization reaction, you are adding CO3-2 from calcite or dolomite, and the pH increases.

[kadie89]

Again, thank you so much for your help!

I am also a little confused about the role of portlandite and the relationship between halite precipitation and sulphate and carbonate reduction. When portlandite is excluded from the initial assemblage I see that halite precipitation appears to be favoured by the carbonate reduction (though I'm not sure why), and when portlandite is included, halite precipitation is suppressed - I don't suppose you could help me to understand why that is the case? Similarly the precipitation of portlandite appears to 'stabilise' hematite and pyrite, whereas both these mineral phases exhibit dissolution and precipitation in its absence.

Thanks again!

Kate

[dlparkhurst]

Unfortunately, your calculations are probably on shaky ground because of both the ionic strength and the temperature.

Halite precipitates when portlandite is removed because Ca+2 accumulates in solution producing a much higher ionic strength. At the high strength, the activity coefficient of sodium gets large and halite precipitates. I don't have much confidence in this result. You are well past the range of reliability for the ion-association model incorporated in phreeqc.dat. The pitzer.dat database would be better (although it too has limits), but it does not include redox processes.  I would try running the results of the phreeqc.dat calculation with pitzer.dat using only the concentrations Ca, Mg, K, Na, S(6), Cl, C(4), at least to see the corresponding saturation indices.