Impact of SO2 and NO on carbonated rocks submitted

Impact of SO2 and NO on carbonated rocks submitted to a geological storage of CO2: an experimental study. a

a

a

a

b

Stéphane Renard , Jérôme Sterpenich , Jacques Pironon , Aurélien Randi , Pierre Chiquet and Marc b Lescanne a Nancy-University, CNRS, CREGU, UMR G2R, B.P. 239, F-54506 Vandoeuvre-lès-Nancy, France b TOTAL, CSTJF, Avenue Larribau, F-64018 Pau, France

Abstract Geological storage of acid gases in carbonated rocks (deep saline aquifers or oil depleted reservoirs) is one of the solutions studied to limit the emissions of greenhouse gases in the atmosphere This paper is devoted to the study of the reactivity of rocks that could be submitted to CO2 and annex gases (SO2 and NO) during the injection of a CO2 rich gas in a geological storage. This experimental study focuses on the interactions that take place between carbonate rocks (dolomite and calcite rich) and CO2 co-injected annex gases. The results, interpreted in terms of petrophysical and chemical impacts of the injected gases can be used to improve thermodynamical and geochemical modelling.

Keywords: geological storage, SO2, NO, CO2, carbonate rocks, experiments Table of Contents 1. Introduction 2. Apparatus and methods 2.1. Solids and aqueous solution 2.2. Gases 3. Results and discussion 3.1. Reactivity of the blank experiments 3.2. Reactivity with SO2 3.3. Reactivity with NO 4. Conclusion Date of submission 08/31/2010

167

1. Introduction The CO2 capture and geological storage from high emitting sources (coal and gas power plants) is one of a panel of solutions proposed to reduce the global greenhouse gas emissions. Different pre- , post- or oxy-combustion capture processes are now available to separate associated gases (SOx, NOx, etc.) and the CO2. However, complete purification of CO2 is unachievable for cost reasons as well as for CO2 surplus of emissions due to the separation processes. By consequence, a non-negligible part of these gases could be coinjected with the CO2. Their impact on the chemical stability of reservoir rocks, caprocks and well has to be evaluated before any large scale injection procedure. Physico-chemical transformations could modify mechanical and injectivity properties of the site and possibly alter storage safety. The study presented here is focused on experiments of geochemical interactions between rocks and gases (SO2 and NO) which could be co-injected with CO2. The rocks we studied are carbonate rocks (dolomite and calcite rich) which are some possible analogues of reservoir rocks and cap-rocks. Samples are placed in 1cm3 gold capsules together with saline water (25 NaCl g/l). Gases are hermetically transferred by cold trap into the gold reactors that are sealed by electrical welding and placed in an autoclave during one month at 150°C and 100 bar, which represent geological conditions of a depleted deep reservoir. After experiments, solid samples are observed and analysed with different techniques (SEM, TEM, Raman and XRD). Gases are also collected and analysed by Raman spectrometry whereas the aqueous solution is analysed with ICP-MS, ICP-AES and ionic chromatography. As sampling during experiments wasn’t possible, we developed the synthetic fluid inclusions technique to trap and analyse the fluids under experimental conditions. This allows to characterise the different phases and the nature of dissolved species. Mass budgets are established in order to quantify the ratio of mineral transformation. This study shows the first results concerning the mineralogical transformation of rocks and well materials submitted to the chemical action of possible annex gases, NO and SO2. The results, interpreted in terms of petrophysical and chemical impacts of the injected gases can be used to improve thermodynamical and geochemical modelling

2. Apparatus and methods Experiments are performed on natural rock samples in batch conditions during one month at 150°C and 100 bar, which represent realist ic conditions in the context of geological storage of CO2 into depleted reservoir. The batch reactors are made of gold capsules hermetically welded. Gold is used because of its chemical inertia, and 168

its ability to conduct pressure and temperature (Seyfried et al., 1987). The volume of the reactors is around 2 cm3 (inner diameter of 0.5 cm for a length of 10 cm). After welding capsules are placed in a pressure vessel of 100 cm3 heated by a coating device (figure 1). The pressure is controlled by a hydraulic pump. The device is presented in more details in Jacquemet et al. (2005). It has been routinely employed for several experimental studies under similar pressure and temperature conditions (Landais et al., 1989; Teinturier and Pironon, 2003; – Jacquemet et al., 2005) mimicking geological environments. Mass balances are established after experiment using analytical characterization of each phase.

2.1. Solids and aqueous solution The rock samples come from cores drilled in the Aquitania basin (France) in a fractured Portlandian dolomite, namely the Mano Dolostone, and in Early Cretaceous limestones, namely the Campanian Flysch. They were sampled respectively at 4580 m and 4500 m deep and were previously analyzed by Renard (2010.) using Scanning Electron Microscopy (SEM), Electron Probe Micro Analysis (EPMA) and Transmission Electron Microscopy (TEM). The sample of Mano Dolostone is made of a dolomitic matrix crossed by a fracture filled with Fe-dolomite and with a thin layer of calcite. For the experiments, we selected samples containing both facieses separated according to a ~ 20 um-thick layer of calcite. The Campanian Flysch is mainly calcitic. The fracture of the Mano Dolostone is made of 93% Fe-dolomite (CaMg)xFe2-x(CO3)2, 5% calcite CaCO3 and 2% dolomite CaMg(CO3)2. The matrix of the Mano Dolostone contains 92.2% dolomite,

4.2%

illite

Si3.43Al2.26Fe0.06Mg0.24K0.71Na0.07Ca0.02

and

interstratified

illite/smectite, 3% quartz SiO2, 0.5% pyrite FeS2 and 0.1% calcite. The Campanian Flysch is made of 63.2% calcite, 10.5% quartz SiO2, 8.3% illite Si3.42Al2.18Fe0.2Mg0.2K0.7Ca0.05, 6.5% interstratified chlorite/smectite, 4.5% chlorite Si2.56 Al2.7Fe3.56 Mg1.27, 4.6% ankerite Fex(Ca,M,Mn)1-xCO3, 2.1% dolomite, 0.3% pyrite.

The rock samples were cut into stick fragments of around 10 mm x 2 mm x 2 mm. They were then polished on one face in order to better detect the mineralogical changes (dissolution or precipitation) on the surface. 169

We partially filled the gold capsules with a 25g/l NaCl brine. The water/rock and water/gas mass ratios were respectively about 3 and 5 as specified in Table 2-1.

Experiment

Mass

Rock (mg)

Solution (mg)

Gas (mg)

N2

Reservoir rock

127

434

38

N2

Caprock

145

550

40

SO2

Reservoir rock

103

510

423

SO2

Caprock

130

630

220

NO

Reservoir rock

130

520

160

NO

Caprock

135

510

155

Table 2-1 : Quantities expressed in mg of rock, aqueous solution and gas used for the experiments on the reservoir rock and caprock.

For each experiment, a decrepited quartz was added to the system in order to trap the fluids during the experiment in synthetic fluid inclusions.

At the end of experiment, gold capsules were opened to collect the gas phase, the aqueous solution and the minerals for analyses.

2.2. Gases The two different types of gases selected for experiments are SO2 and NO. A blank capsule containing the same phases (aqueous solution and solid) was filled with N2 as an inert gas phase. The injected quantities for each experiment are displayed in Table 2-1. The gases are loaded in the capsules using the gas loading device adapted from Jacquemet et al., 2005 (Figure 2-1). During the loading procedure, the gold capsules are hermetically fixed on the capsule connector which is plugged to the loading device through the valve E. Knowing the volume of the loading line and controlling the pressure in the system, it is possible to fill the reactor with a known mass of gas thanks to a nitrogen cold trap. After experiment, cold capsules are pierced in an appropriate device plugged to valve C. After trapping, the gas can be driven to a Raman cell for analysis.

170

Figure 2-1: Gas loading and sampling line used during the experimental phase, adapted from Jacquemet et al. (2005). (A-E) valves. Different devices can be connected to the line: a capsule piercing device used to collect gases after experiment, a capsule loading device used to trap gases in the capsule and a cell for the Raman analysis of the gases.

3. Results and discussion This section is devoted to the description of the mineral changes observed from the solid samples of reservoir and caprock aged with N2 (blank experiments) SO2 and NO during one month at 150°C and 100 bar.

3.1. Reactivity of the blank experiments After experiment, the samples of the reservoir rock do not present any visible transformation except a slight frosted aspect of the initially polished face. SEM observations (Figure 3-1) show that the frosted aspect is due to a slight dissolution of the carbonate phases, the dolomite of the matrix and the calcite of the fracture. However the dolomite of the fracture does not seem to have undergone any

171

significant dissolution. Pyrite and quartz keep unaltered whereas the analyses of the clay fraction (Renard 2010) show a partial leaching of Na and Ca cations.

Figure 3-1 : SEM backscattered images of the reservoir rock sample after experiment with N2 and saline water (25 g/l). (I) Global view of the matrix (Ma) and the fracture (Fr), (II) zoom on the matrix, (III) zoom on close to the wall of the fracture. (cal) calcite, (dol) dolomite, (Py) pyrite.

Concerning the caprock, the optical observations show a slight frosted aspect as well as the presence of a brown-orange colour on the surface. The limited reactivity is observable with SEM (Figure 3-2Figure 3-2). The surface of the calcite is slightly dissolved. The grains of quartz and the framboidal pyrites seem to be unaltered. However EDS (Energy Dispersive Spectrometry) analyses show that the surface of the pyrites is oxidised explaining the brownish aspect of the sample. Clay minerals analysed by TEM before and after experiment do not react significantly during experiment.

172

Figure 3-2 : SEM backscattered images of the caprock sample after experiment with N2 and saline water (25 g/l). (I) Global view of the sample, (II-III) zoom on a zoned siderite, (IV) zoom on a pyrite rich zone. (ag) clay minerals, (cal) calcite, (dol) dolomite, (qtz) quartz, (sd) siderite, (py) pyrite.

The blank experiments both with the caprock and the reservoir show a very limited reactivity of the minerals corresponding to the equilibration between the initial aqueous solution and the different minerals. The pH of the solution is rapidly buffered by carbonate minerals (dolomite and calcite). The main chemical reactions considered during the experiment that can affect the pH as well as the elemental concentrations of the solution are:

(1)

(2)

(3)

(4)

CaMg(CO3)2 + 2H+ = Ca2+ + Mg2+ + 2 HCO3CaCO3 + H+ = Ca2+ + HCO3(CaMg)0,13Fe0,74CO3 + H+ = 0,13 Ca2+ + 0,13 Mg2+ + 0,74 Fe2+ + HCO3Quartz = SiO2,aq

173

(5)

Amorphous silica = SiO2,aq + n H2O

(6)

HCO3- = CO2g + H+

(7)

H2O = H+ + OH-

During the loading of the reactors, gaseous oxygen can be trapped leading to a partial oxidation of the reduced mineral such as pyrite. This phenomenon, enhanced by the framboidal shape of the mineral increasing its reactive surface area, can be resumed by the following chemical reaction leading to the formation of hematite (Fe2O3) and sulfates (mainly anhydrite CaSO4). (8)

2 FeS2 + 4 H2O+ 7.5 O2 = Fe2O3 + 4 SO42- + 8 H+

The mass balance calculated from these blank experiments confirm that the mineral dissolution is very limited with less than 5% of the initial quantity of the minerals affected by the mineral transformations. Calcite and pyrite seem to be the most sensitive minerals in our experimental conditions.

3.2. Reactivity with pure SO2 The initial reservoir and caprock samples were completely crumbled after the experiment with SO2. After drying, a powder made of fibrous crystals of anhydrite and amorphous native sulphur was observed in association with an amorphous silica rich phase (Figure 3-3, Figure 3-4) containing iron sulphur, aluminium and potassium. Quartz and pyrite couldn’t be detected. Large amounts of CO2 were released in the gas phase as a proof of the high reactivity of the carbonates towards SO2.

174

Figure 3-3 : SEM backscattered images of the reservoir rock sample after experiment with SO2 and saline water (25 g/l). (I, III) global view; (II) zoom on native sulfur; (IV) zoom on a zone containing native sulfur, anhydrite and amorphous silica. (S0) native sulfur, (Anh) anhydrite, (Siam) amorphous silica rich phase.

175

Figure 3-4 : SEM backscattered images of the caprock sample after experiment with SO2 and saline water (25 g/l). (S0) native sulfur, (Anh) anhydrite, (Siam) amorphous silica rich phase.

Under experimental pressure, temperature and water molar ratio, respectively 100 bar, 150°C and 0.6 to 0.9, the SO 2-H2O system is monophasic with a complete dissolution of the SO2 in the liquid water (Van Berkum et al., 1979). The effect of NaCl is not documented under the experimental conditions but the synthetic fluid inclusions analyses show that no gaseous SO2 was trapped during the experiments implying its quasi- total dissolution in the saline water. When SO2 is in solution, a reaction of disproportionation occurs leading to sulphuric acid and native sulphur according to the reaction: 2 H2O + 3 SO2,aq = 2 H+ + SO42-+ 1/8S8

(9)

This reaction accounts for the presence of native sulphur after experiment as well as the strong alteration of minerals due to high acidic conditions. The main mineral transformations can be sum up by the following reactions involving carbonate minerals:

(10)

Dolomite + SO42- + Ca2+ + 4 H+ = 2 Anhydrite + 2 CO2,aq + Mg2+

The clay minerals are also strongly affected by the high acidity of the solution. They dissolved to give mainly Si and Al in solution that can combine to form an amorphous phase called amorphous silica rich phase. This gel can incorporate sulfur and a part of the alkalis and alkaline-earth elements coming from the dissolution of carbonates and silicates. If we consider muscovite as a proxy for clay minerals, the reaction could be expressed as follows:

(11)

Si3Al3KO10(OH)2 + 10 H+ = 3 SiO2,am + 3 Al3+ + K+ + 6 H2O

Pyrite is also concerned both by the acidic attack and the oxidizing power of SO2. Pyrite is thus transformed by the following reaction enhanced in acidic conditions:

(12)

4H+ + SO2,aq + 2 FeS2 = 2 Fe2+ + 5/8 S8 + 2 H2O

176

The dissolution of clay minerals, especially illites, can release Fe3+ but the presence of SO2, as a reducing compound in this case, leads to its reduction in Fe2+ in agreement with Palandri et al. (2005) according to:

(13)

2 Fe3+ + SO2,aq + 2 H2O = 2 Fe2+ + HSO4- + 3 H+

Thus, the presence of high amounts of SO2 leads to a total dissolution of carbonates, silicates and pyrite and to the precipitation of anhydrite, native sulfur and an amorphous silica rich phase. The mass budget of the experiment was calculated thanks to the chemistry of the solution, the stoichiometry of the mineral phases and the composition of the gaseous phase (consumed SO2 and produced CO2). For both the reservoir and the caprock, the total amount of carbonates disappeared whereas it was the case only for 15 to 20% of the clayey fraction. After reaction, about 15% of the initial SO2 gave anhydrite, 25% gave native sulfur and less than 1% gave barite (BaSO4). 3.3. Reactivity with pure NO After experiment with NO the caprock and reservoir samples kept their initial shape but showed strong visible transformations on their surface. The matrix dolomite of the reservoir rock sample (Figure 3-5), disappeared from the surface and was only detectable deeper below the surface. Clay minerals and quartz are still present. The fracture wall calcite is altered, and the dolomite is partially dissolved according to its cleavages. The pyrites of the rock was completely oxidized into hematite. A part of the sulfur coming from the oxidation of the sulfides reprecipitated in anhydrite and in a lesser extent in barite (BaSO4) from the calcium of carbonates and the barium as a trace element in the calcites.

177

Figure 3-5 : SEM backscattered images of the reservoir rock sample after experiment with NO and saline water (25 g/l). (I) global view; (II) zoom on the limit between matrix and fracture; (III) zoom on the matrix. (Dol) dolomite, (Anh) anhydrite, (Ba) barite, (Ag) clay minerals.

Concerning the caprock (Figure 3-6), the calcite was strongly dissolved. Fecontaining minerals (siderite and pyrite) were oxidized leading to the precipitation of hematite. The sulfur from pyrites partially precipitated into anhydrite and barite. The ferriferous chlorites were also oxidized. The observations of both the reservoir and the caprock show two main chemical mechanisms responsible for the mineral transformations: reactions under acidic conditions and oxydo-reduction reactions.

178

Figure 3-6 : SEM backscattered images of the caprock sample after experiment with NO and saline water (25 g/l). (I) global view; (II and III) successive zooms on the matrix. (Ex-Si) exsiderite transformed in hematite, (Ag) clay minerals.

There are very few thermodynamical data for NO under the experimental pressure and temperature range. The analyses performed onto the gaseous and aqueous phase indicate that NO is not stable under these conditions. The chemistry of nitrogen oxides is complex and numerous phases appear during the experiment such as N2O, NO2, N2, O2, NH4+, NO3-. In the gaseous phase some reactions of oxydoreduction can run such as:

(14)

3 NO = NO2 + N2O

(15)

2 NO = N2 + O2

(16)

2 NO = ½ N 2 + NO2

(17)

2 NO = ½ O 2 + N2O

In the queous phase, the following reaction can explain the presence of N2O and the nitrates and leading to a very acidic solution:

179

4 NO + ½ H 2O = 3/2 N2O + H+ + NO3-

(18)

The dissociation of N2O in N2 and O2 was also descibed in the littretaure but under different conditions (Li et al., 1992, Rivallan et al., 2009).

Whatever the occuring reactions, the presence of NO in an aqueous system leads to a dual reactivity due to the presence of protons H+ and oxidising agents such as O2. In this case, several chemical mechanisms can be written to explain the complete oxidation of iron bearing phases (pyrite and siderite) as well as the presence of ammonium or nitrates in the aqueous solution. The following reactions can be proposed athough they are not exhaustive. The presence of CO2 in the fluid phase after experiment proves that the carbonates phases are altered by the acidic solution due to the initial presence of NO. The same reactions as for the dissolution of carbonates in an acidic solution (reactions 1 to 3) can be also envisaged here:

(19)

CaCO3 + 2H+ = Ca2+ + H2O + CO2

(20)

CaMg(CO3)2 + 4 H+ = Ca2+ + Mg2+ + 2 H2O + 2 CO2

The reactions lead to emission of gaseous CO2 and the release of Ca2+ and Ba2+. The oxidation of pyrites cans be described differently if considering either the presence of ammonium or di-nitrogen, or considering the oxidising agent to be nitrate or di-oxygen:

(21)

8 FeS2 + 31 H2O + 15 NO3- = 4 Fe2O3 + 15 NH4+ + 16 SO42- + 2 H+

(22)

2 FeS2 + H2O + 6 NO3- = Fe2O3 + 3 N2 + 4 SO42- + 2 H+

(23)

2 FeS2 + 4 H2O+ 7,5 O2 = Fe2O3 + 4 SO42- + 8 H+

For each reaction used, sulphates are formed that combine with Ca and Ba to form anhydrite and barite:

(24)

Ca2+ + SO42- = CaSO4

(25)

Ba2+ + SO42- = BaSO4

180

Contrary to the experiments with SO2, silicate minerals, mainly quartz and clay minerals, are slightly affected by NO with only a few percents dissolved. To sum up, experiments with NO are complex and lead to a complete oxidation of iron bearing phases (mainly pyrite and siderite), to a partial dissolution of carbonates with an enhanced reactivity of calcite by comparison with dolomite, and keep the silicates phases almost free of dissolution. For the chosen conditions of experiment, the mass budget shows that between 20 and 50% percent of the calcite is dissolved as against 15 to 20% of the dolomite. 100% of the siderite and the pyrite are oxidised in hematite. Less than a few percent of the silicates is affected by NO. 4. Conclusion The experiments performed in the context of the injection of CO2 and co-injected gases in a geological storage have demonstrated that SO2 and NO should play a role on the mineralogy of both the reservoir and the caprock. First, this study has shown that SO2 and NO have a complex behaviour with a dual action, oxidising and acidic, on the minerals. Second, many disproportionation reactions can occur when SO2 and NO are placed under geological conditions of pressure and temperature. These oxydo-reduction reactions complicate the system by multiplying the possible oxidising agents and thus the possible reactions and products of reactions. Third, the reactivity of both the reservoir rock and the caprock is strongly dependent on the nature of the mineral phases (silicates, carbonates, sulphides, etc.) but also on the nature of the reacting gas. For example it is noticeable that the presence of SO2 should lead to the formation of sulphate mineral and native sulphur, when the presence of NO should be responsible for the strong oxidation of iron bearing phases. In any case, since the molar volumes of initial minerals are different of those of secondary products (as an example the molar volume of calcite is 36.93 cm3.mol-1 against 45.16 cm3.mol-1 for anhydrite), the minerals transformations occurring with the injection of reacting gases should be interpreted in terms of petrophysical properties (porosity and permeability) of the hosting rock. This study shows also that experiments with the gases of interest under geological conditions of storage are necessary to predict the evolution of the storage submitted to the injection of CO2 and co-injected gases. References

181

Seyfried, W. E. J., Janecky, D. R., and Berndt, M. E. (1987). Rocking autoclaves for hydrothermal experiments. The flexible reaction-cell system. Hydrothermal experimental techniques (ed. Ulmer, G. C. and Barnes, H. L.), pp. 216-239. John Wiley & Sons. Jacquemet, N., J. Pironon and E. Caroli (2005). A new experimental procedure for simulation of H2S + CO2 geological storage. Application to well cement aging. Oil and Gas Science and Technology 60(1), pp. 193-206. Landais, P., Michels, R., and Poty, B. (1989) Pyrolysis of organic matter in cold-seal pressure autoclaves. Experimental approach and applications. Journal of analytical and applied pyrolysis 16, 103-115. Teinturier, S., Pironon, J. (2003). Synthetic fluid inclusions as recorders of microfracture healing and overgrowth formation rates. American Mineralogist, 88 (89), pp. 1204-1208. Renard, S (2010) Rôle des gaz annexes sur l'évolution géochimique d'un site de stockage de dioxyde de carbone. Application à des réservoirs carbonatés., PhD thesis Nancy Université INPL, p. 422. Van Berkum, J.G., Diepen, G.A.M. (1979). Phase equilibria in SO2 + H2O: the sulfur dioxide gas hydrate, two liquid phases, and the gas phase in the temperature range 273 to 400 K and at pressures up to 400 MPa. The Journal of Chemical Thermodynamics, 11 (4), pp. 317-334. Palandri, J. L., R. J. Rosenbauer and Y. K. Kharaka (2005). Ferric iron in sediments as a novel CO2 mineral trap: CO2-SO2 reaction with hematite. Applied Geochemistry 20(11), pp. 2038-2048. Li, Y., Armor, J.N. (1992). Catalytic decomposition of nitrous oxide on metal exchanged Zeolites. Applied Catalysis B, Environmental, 1 (3), pp. L21-L29. Rivallan, M., Ricchiardi, G., Bordiga, S., Zecchina, A. (2009). Adsorption and reactivity of nitrogen oxides (NO2, NO, N2O) on Fe-zeolites. Journal of Catalysis, 264 (2), pp. 104-116. Acknowledgments This work is supported by TOTAL and ADEME (France). It is included in the project "Gaz Annexes" of the French National Agency for Research (ANR).

182