Experimental study of CO2-brine-rock interaction

International Journal of Coal Geology 154–155 (2016) 265–274

Contents lists available at ScienceDirect

International Journal of Coal Geology journal homepage: www.elsevier.com/locate/ijcoalgeo

Experimental study of CO2–brine–rock interaction during CO2 sequestration in deep coal seams Kairan Wang, Tianfu Xu, Fugang Wang ⁎, Hailong Tian Key Laboratory of Groundwater Resources and Environment, Ministry of Education, Jilin University, Changchun 130021, China

a r t i c l e

i n f o

Article history: Received 13 August 2015 Received in revised form 19 January 2016 Accepted 19 January 2016 Available online 20 January 2016 Keywords: CO2 storage ECBM Water–rock interaction Mineralogical changes Caprock

a b s t r a c t CO2 sequestration in deep coal seams is a potential option for reducing greenhouse gas emissions. Once CO2 is injected into coal seams, sealing capability of the cap rock is critical. To investigate and quantify reactions over time between CO2, cap rocks and brine, associated with selected cap rocks of the No. 3 coalbed of the Qinshui Basin in China, batch experiments were conducted for reacting powdered rock samples (180–220 μm) with CO2 and brine, as well as CO2-free brine, at 160 °C and 15 MPa. The analysis of leachate chemistry indicated significant mobilization of major elements from dissolution of carbonate and silicate minerals in the coal measure strata. Analysis of reacted solids by XRD and SEM also revealed appreciable changes in mineralogical compositions. For lithic sandstone after reaction with CO2–brine, the contents of quartz, plagioclase, illite and chlorite increased considerably, whereas the contents of illite/smectite, biotite and kaolinite decreased more or less. The calcareous mudstone reacting with CO2–brine and CO2-free brine all showed major mineralogical alteration after 12 days of treatment. The modeling results identified key chemical processes, but they also showed that the models are not capable of covering all possible contingencies. The precipitation of carbonate minerals could also enhance the security of CO2 sequestration in deep coal seams. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Climate change has become a hot point involved in global environmental issues. Controlling greenhouse gas emissions and protecting global climate is a major issue at present. Of all types of greenhouse gases blamed for climate change, emissions of CO2 from human activities has one of the largest impacts on climate change gases, accounting for 63% of the temperature effect from all total greenhouse gases. Additionally, CO2 is retained in the atmosphere for more than 200 years (Metz et al., 2005). The question of how to reduce the content of CO2 in the atmosphere is the current problem to be solved. It is well known that geological CO2 storage (GCS) is undoubtedly the most realistic and effective disposal method. At present, storing CO2 by injection into subsurface oil and gas reservoirs (Winter and Bergman, 1993; Bachu and Shaw, 2003), deep saline aquifers (Xu et al., 2004; W. Zhang et al., 2009; X. Zhang et al., 2009; Lei et al., 2015; Tian et al., 2015) and deep unmineable coal beds (Massarotto et al., 2010; Dawson et al., 2011) is the most valid and economic choice for reducing CO2 emissions into the atmosphere. Of these methods, coal bed geological disposal is one of the most favorable, as enhanced coal bed methane recovery can be accomplished simultaneously with sequestration of

⁎ Corresponding author. E-mail address: [email protected] (F. Wang).

http://dx.doi.org/10.1016/j.coal.2016.01.010 0166-5162/© 2015 Elsevier B.V. All rights reserved.

CO2. This technique is known as CO2-ECBM (Gale and Freund, 2001; White et al., 2005; Czerw, 2011; Baran et al., 2015). Throughout the world, coal measure strata are important source layers of oil and gas, and they are important reservoirs, especially for natural gas resources. By the statistics, many large gas fields in the world are associated with coal measure strata. Coal measure strata include a large number of abandoned coal seams due to technical and economic reasons, and these unmineable coal seams are potential geological structures that can be used for geological CO2 sequestration. Only a handful of CO2-ECBM tests in the world have been implemented. These tests showed great potential for both CO2 sequestration and ECBM production. For instance, the United States built the first CO2ECBM pilot project in the San Juan basin, and more than 100,000 t of CO2 has been injected into the Fruitland coal seam since 1996 (Weber et al., 2012). Canada is in the process of a controlled trial of a CO2ECBM test in Alberta (Gentzis, 2000), and the evaluation of regional projects is widely being carried out throughout the world. In addition, Australia, Japan, The Netherlands and other countries have implemented CO2 injection to enhance CBM exploitation, and they have carried out related researches (Li et al., 2004; Golding et al., 2013). There are many coal-bearing rock series widely distributed in continental sedimentary basins (Qinshui Basin and Ordos Basin, etc.) in China, which are more suitable for CO2 sequestration. Since 2002, under the support of the governments of Canada and China, the China United Coal Bed Methane Corporation carried out the theory and technological research of CO2 injection/burial, enhanced coal bed methane

266

K. Wang et al. / International Journal of Coal Geology 154–155 (2016) 265–274

recovery, and successful implementation of a single well CO2 injection test in the southern Qinshui Basin of Shanxi Province in 2004, obtaining a series of key parameters (Wong et al., 2007, 2010; Wei et al., 2007). Under the support of the China Ministry of Science and Technology, the China United Coal Bed Methane Corporation carried out a technological research on deep coal seam CO2 injection and coal bed methane exploitation in 2007, to achieve the commercialization of this technology. This is the first research project of deep coal bed CO2-ECBM in China, and 234 t of CO2 was injected into the No. 3 coal seam. In recent years, in allusion to the basic theory of CO2-ECBM, a series of studies on multicomponent gas competitive adsorption on coal surfaces (Tang et al., 2004; Zhang et al., 2004; Zhou et al., 2013), coal matrix expansion and contraction induced by adsorption (Karacan, 2003; Qin et al., 2005; Pan and Luke, 2007), and numerical simulation (Sun, 2005; Wu and Zhang, 2007; Siriwardane et al., 2012; Vishal et al., 2013) were carried out. However, long-term studies have ignored the fact that, once CO2 is injected into coal seams, it will dissolve in formation water and form carbonic acid. Thus, the minerals in coal seams and roof-floor rocks were dissolving and the components of coal were changing. Secondary aluminosilicate and carbonate minerals could be formed at the same time, so that CO2 can be safely sealed underground for a long time (Li et al., 2013). To reveal the physicochemical processes and evaluate the security of CO2 sequestration in deep unmineable coal seams, it is necessary to carry out the related experimental investigations on CO2–brine–rock reactions, to discuss the geochemical behavior and possible mechanism of mineral traps in the long-term interaction between CO2 and coal measure strata, and to provide a natural analogy as a research object and as a geological basis for CO2-ECBM projects.

2. Geological setting The Qinshui Basin is a near longitudinal-trending tectonic basin between Taihang Mountain and Lvliang Mountain, located in the middle of the North China Platform. The basin has experienced the tectonic movements of the Indosinian Period (Late Permian to Triassic), the Yanshan Period (Jurassic to Early Cretaceous) and the Himalayan Period (Late Tertiary). The basin's fold and fault structures are well developed, and most of the tectonic lines are in the NE-NNE direction. The study area, in the north block of Shizhuang, is located in the southeast of the Qinshui Basin. The main geological structure is folding. Faulting is not well developed, and existing faults are no longer than 100 m. Collapse columns occur occasionally. The overall structure of the study area is a western-leaning monocline (Fig. 1). The block is divided by the longitudinal-trending syncline structure into three tectonic units. The east unit is a gentle monoclinal structure, with only a few small faults and with dip angles generally less than 5°, and it slowly uplifts towards the SE. The central unit is a severely squeezed anticline-syncline pair with a range of 210 m, and the axis orientation is NNE. The west unit is a complex groove structure, with several large faults, and the coal seam has a discernible undulation (Huang et al., 2010; Ye et al., 2012). The Upper Carboniferous Taiyuan Formation and the Lower Permian Shanxi Formation are the main coal-bearing strata in the study area, containing a total of 6–11 coal seams. Among them, the thicknesses of the No. 3 coal seam of the Shanxi Group and the No. 15 coal seam of the Taiyuan Group are larger and have a stable spatial distribution. They are the main coal seams for coal bed methane exploration. The burial depth of the No. 3 coal seam is from 830 m to more than 1600 m due to dipping strata, and the thickness is 4–6 m with an

Fig. 1. Burial depth contour map of the No. 3 coal seam of the Qinshui Basin and the stratigraphic section of coal-bearing strata. Modified from Wei et al. (2007) and Cai et al. (2015).


average of 5.82 m. The No. 3 coal is mainly bright coal with vitrain bands, metallic luster, vertical joints, and endogenous fissure development. Plant fossil fragments can be found in part of the boreholes, and there is a sandy mudstone or shale interlayer with a thickness of 0.24– 0.7 m (average 0.44 m) in the coal seam. Its roof is mainly composed of mudstone, siltstone and silty mudstone, which locally is fine and medium grained sandstone. Its floor is mainly mudstone and siltstone, which locally is silty mudstone (Lu et al., 2012). The No. 3 coal seam is considered to be suitable for geological CO2 storage.

267

of the two samples, as determined by X-ray fluorescence (XRF), are given in Table 2. The XRD analysis was conducted using a PANalytical X'Pert PRO powder diffractometer with a Cu X-ray source at 45 kV and 40 mA. Powder X-ray diffractograms were recorded between 5° and 52° 2-Theta. XRF analysis was conducted using a Rigaku ZSX Primus II automatic order scanning X-ray fluorescence spectrometer. EMP analyses were conducted on a JEOL JXA-8230 at 15 kV and 15 nA. SEM-EDS observations were made using a Hitachi JSM-6700F field-emission scanning electron microscope equipped with an INCA ENERGY 350 EDS microanalytical system.

3. Materials and methods 3.2. Experimental setup and methods 3.1. Materials The roof rocks of the No. 3 coal seam used in the experiments were taken from the north block of Shizhuang of the Qinshui Basin in Shanxi Province. Microscopic observations in transmitted light were made on polished thin sections with a standard thickness of 0.02 mm. These examinations were carried out using a Leica DM2500 polarizing microscope. Two types of lithology samples were selected: lithic sandstone and calcareous mudstone (Fig. 2). X-ray diffraction (XRD) semi-quantitative analysis, Electron Microprobe (EMP) analysis and Scanning Electron Microscopy (SEM) incorporating the use of a quantitative energy dispersive spectrometer (EDS) confirm that the main mineral composition in the lithic sandstone includes quartz, clay minerals, plagioclase (albite dominated), and a small amount of chlorite, and the main mineral composition in the calcareous mudstone includes clay minerals, quartz, and small amounts of K-feldspar and calcite (Table 1). The chemical compositions

The experiments were conducted in several uniform hightemperature and high-pressure airtight reaction kettles, with liners made of Hastelloy C-276 (Fig. 3). The advantages of the kettles are their resistance to acid and alkali corrosion. The volume of the reaction kettles was 300 mL. Other instruments, such as the incubator chamber, electronic scales, and pH meters, were also used. The crushed rock, with a grain size between 180 and 220 μm, was rinsed with distilled water and dried in an oven at 80 °C for 48 h prior to the experiments. Before the experiments, a mixture of rock and brine solution (1 M NaCl, Merck analytical grade) with a water/rock ratio of 20:1 was poured into the reactor, and then the mixture was bubbled with (N2) to remove O2 from the solution to imitate the conditions found in geological reservoirs. For easier observation of mineral morphology, a polished thin section was also added in the reactor. The system was then heated to the desired temperatures before pressurizing with CO2 from a commercially supplied bottle (≥ 99.9 vol.%). In the experiments with CO2–brine, a

Fig. 2. Thin-section imaging of the roof rock samples with a polarizing microscope. (a and b) Lithic sandstone, 10× magnification: (a) with plane-polarized light and (b) with crosspolarized light. These two images show quartz, sandstone fragments and feldspar phenocrysts. (c and d) Calcareous mudstone, 10× magnification: (c) with plane-polarized light and (d) with cross-polarized light. These two images show mudstone fragments and clay minerals. Carbonate cement is present in the fractures.

268


Table 1 Mineralogical composition by XRD (wt.%) (semi-quantitative). Sample

Quartz

Plagioclase

K-feldspar

Calcite

I/S

Kaolinite

Chlorite

Biotite

Lithic sandstone Calcareous mudstone

48 19

13 2

N/D 2

N/D 2

25 54

10 20

2 N/D

2 1

I/S: illite/smectite mixed-layer minerals. N/D: not detected.

Table 2 Chemical composition by XRF (%). Sample

SiO2

Al2O3

TFe2O3

CaO

MgO

K2O

Na2O

TiO2

P2O5

MnO

L.O.I

SUM

Lithic sandstone Calcareous mudstone

76.31 56.3

10.38 20.47

6.67 4.01

0.33 2.73

1.01 2.23

0.63 2.9

2.01 0.75

0.26 0.88

0.02 0.15

0.02 0.07

2.21 9.29

99.86 99.78

L.O.I: loss on ignition.

pressure of 15 MPa was achieved by adjusting the initial CO2 pressure using the air compressor and booster pump after heating. The system was closed for the duration of the experiment; therefore, no sampling was carried out until the end of the experiment. In all experiments particles in the solution were kept in suspension by turbulent stirring at a speed of 200 rpm. There was a ± 1 °C variation in temperature, and the pressure stayed within ± 3 bar. Once stabilized, the temperature stayed relatively stable throughout the experiments. A summary of the experimental parameters is given in Table 3. After the reaction, liquid and solid samples were collected after quenching and releasing the pressure in the reactor. The solution was filtered through a 0.45 μm filter to separate the solution from the solids. The liquid samples for metal analysis were acidified to pH 2 with pure HNO 3 to prevent the precipitation of metals. Dissolved SiO2 was determined using a Seal-AutoAnalyzer 3™. Basic metals were determined with an inductively coupled plasma mass spectrometry (ICP-MS) (Agilent 7500C), and dissolved anions were

determined by ion chromatography (IC) (Metrohm 861). All reported pH and alkalinity measurements were taken 3 h after the samples were equilibrated with atmospheric P CO2 at room temperature (20 °C). Solid samples were washed with distilled water on a 0.45 μm filter and then oven dried at 80 °C for 48 h before they were analyzed. The mineral compositions of powder samples were analyzed by XRD and XRF, and mineral morphology observations were made on polished thin sections with SEM-EDS. 4. Results and analyses 4.1. Reaction of solid phases 4.1.1. Lithic sandstone The content of solid minerals in lithic sandstone changed extensively at the end of the experiment after 12 days. The primary minerals alterated appreciably under the effect of CO2 acid fluid during the

Fig. 3. Schematic diagram of the experimental device.

Table 3 List of experiments and experimental parameters. Sample

Experiment

Starting fluid

Temperature (°C)

Pressure (MPa)

Solids (g)

Water/rock ratio

Duration (days)

Lithic sandstone

Ext. 1 Ext. 2 Ext. 3 Ext. 4 Ext. 5 Ext. 6

B+C B+C B+C B+C B+C B

160 160 160 160 160 160

15 15 15 15 15 0.62a

12.5 12.5 12.5 12.5 12.5 12.5

20:1 20:1 20:1 20:1 20:1 20:1

3 6 9 12 12 12

Calcareous mudstone

B + C: reacted with brine and CO2. B: reacted with CO2-free brine. a Saturated vapor pressure of water at 160 °C (~0.62 MPa).


269

plagioclase was slightly corroded (Fig. 5b). During the lithic sandstone experiment, the plagioclase composition as determined by EMPA did not significantly change. The plagioclase composition for untreated and CO2-treated samples was generally clustered about the albite endmember (Table 5). In addition, kaolinite and chlorite were also precipitated (Fig. 5c and d). However, through the SEM observation, and combined with XRD analyses of the solid phase after the experiments, there were no discernable new mineral phases (especially secondary carbonate minerals) precipitated in the whole experimental process. Therefore, after different reaction times, the content and morphological structure of the initial minerals changed appreciably. The relative contents of quartz, illite and chlorite clearly increased, whereas the contents of illite/smectite, biotite and kaolinite decreased more or less. (Table 4).

Fig. 4. Bulk XRD plots of the lithic sandstone: unreacted and reacted for 3, 6, 9, 12 days with CO2–brine. Qtz—quartz, Pl—plagioclase, Kao—kaolinite, Chl—chlorite.

reaction time, where ions leached from minerals into the solution and then formed new minerals through geochemical reactions (Fig. 4). From the SEM images of lithic sandstone after the experiment, we can see that quartz exhibited overgrowth phenomena (Fig. 5a) and

4.1.2. Calcareous mudstone The mineral composition of the calcareous mudstone alterated severely after the reaction compared with the original material. The peaks of plagioclase, siderite, and dolomite were not clearly discernible in the unreacted material, but they all increased after reaction with CO2–brine and brine only, and they exhibited an analogical trend (Fig. 6). This was due to the presence of large amounts of clay minerals in the calcareous mudstone. The dissolution of K-feldspar and clay minerals can release K, Al, Fe, Mg and other metal ions, and the dissolution of a small amount of calcite could supply HCO− 3 , which can generate secondary carbonate minerals when combined with the metal ions (Fig. 7d). As in the experiments of lithic sandstone, illite and chlorite were all generated in large quantities (Fig. 7b and c). However, plagioclase experienced slight growth under the experimental conditions (Fig. 7a). In this process, CO2 could not be a decisive factor in the formation of secondary carbonate minerals, although it may be at longer reaction times.

Fig. 5. SEM images of the lithic sandstone after 12 days of reaction with CO2–brine. (a) authigenic quartz, (b) corroded albite, (c) kaolinite in the cement, (d) authigenic chlorite.

270


Table 4 Mineral composition by XRD after reaction (semi-quantitative). Minerals

Quartz Plagioclase K-feldspar Chlorite I/S Illite Kaolinite Calcite Biotite Siderite Dolomite

Lithic sandstone

Table 6 Saturation index of minerals involved in geochemical reactions.

Calcareous mudstone

Initial

B+C (3d)

B+C (6d)

B+C (9d)

B+C (12d)

Initial

B+C (12d)

B (12d)

48 13 N/D 2 25 N/D 10 N/D 2 N/D N/D

57 17 N/D 12 N/D 8 6 N/D N/D N/D N/D

42 25 N/D 15 N/D 9 9 N/D N/D N/D N/D

52 18 N/D 17 N/D 11 8 N/D N/D N/D N/D

57 11 N/D 14 N/D 9 9 N/D N/D N/D N/D

19 2 2 N/D 54 N/D 20 2 1 N/D N/D

44 4 2 6 N/D 26 8 N/D N/D 4 6

46 5 2 6 N/D 24 8 N/D N/D 2 7

B + C: reacted with brine and CO2. B: reacted with CO2-free brine. All values are percentages of total dry weight. I/S: illite/smectite mixed-layer minerals. N/D: not detected.

Temperature

Albite Ankerite Calcite Chlorite Dolomite Dawsonite Illite Kaolinite K-feldspar Quartz Siderite Smectite pHm pH

Lithic sandstone

Calcareous mudstone

B+C (3d)

B+C (6d)

B+C (9d)

B+C (12d)

B+C (12d)

B (12d)

160 °C

160 °C

160 °C

160 °C

160 °C

160 °C

SI

SI

SI

SI

SI

SI

0.156 −0.580 −1.137 −0.024 1.623 −1.259 −0.008 1.261 −3.010 0.123 0.980 1.340 5.72 4.791

0.375 0.211 −0.499 −0.060 1.426 −1.030 0.009 1.310 −2.868 0.120 1.388 1.354 5.67 4.988

0.425 0.638 0.000 −0.062 1.152 −0.938 0.005 1.286 −2.842 0.101 1.411 1.349 5.86 5.124

0.504 0.726 0.002 −0.065 1.932 −0.829 0.002 1.281 −2.836 0.096 1.413 1.347 5.93 5.174

0.934 −6.515 0.000 −35.822 10.208 −1.207 −0.236 1.445 −0.439 0.360 −5.813 0.188 6.52 5.003

1.501 −4.823 0.000 −6.980 6.509 −3.166 −0.011 −0.476 −0.555 −0.136 −5.038 −0.016 6.97 7.516

B + C: reacted with CO2–brine. B: reacted with brine only. m: measured. SI: log (IAP/K), where IAP = ion activity product and K = equilibrium constant.

4.2. Fluid chemistry 4.2.1. Reaction with lithic sandstone Because the composition of the reaction liquid is 1 M NaCl solution, some ions (K, Ca, Mg, Fe and Si) did not exist in the initial solution. For the four experiments at 3, 6, 9 and 12 days, all of these ions were detected, indicating dissolution of reacting solids. A large amount of CO2 quickly dissolved into the water under high temperature and high pressure, generating a large amount of HCO− 3 , and pH also decreased rapidly. After 6 days of reaction, the concentration of HCO− 3 increased to 11.41 mmol/L, and pH was reduced to 5.67. Then, the concentration of HCO− 3 began to decline, and pH began to rise (Fig. 8). Because CO2 dissolved into the solution in this process, forming an acidic environment, mineral dissolution occurred and then precipitation. This mainly involved the dissolution of aluminosilicate minerals. The calcium concentration in solution continued to increase. After

Table 5 Representative plagioclase compositions before and after CO2 treatment, as determined by EMP analysis. Untreated

CO2 treated for 3 days

CO2 treated for 12 days

67.29 20.69 0.01 0.15 0.01 0.02 0.29 10.42 0.18 99.06

67.77 20.19 0.02 0.11 0.02 0.01 0.18 10.67 0.23 99.20

66.89 20.15 0.01 0.21 0.01 0.01 0.51 10.24 0.07 98.10

Normalized to 8 oxygen Si 3.14 Al 1.10 ∑ 4.24 K 0.01 Ca 0.02 Na 0.77 Ti 0.00 Mg 0.00 Fe 0.01 Mn 0.00 ∑ 0.81 0.01 xkfs 0.03 xan 0.96 xab

3.16 1.07 4.23 0.02 0.01 0.79 0.00 0.00 0.01 0.00 0.83 0.03 0.01 0.96

3.12 1.07 4.19 0.01 0.04 0.76 0.00 0.00 0.02 0.00 0.83 0.01 0.05 0.94

SiO2 Al2O3 MgO FeO TiO2 MnO CaO Na2O K2O ∑

Minerals

6 days, it was increasing faster, and reached 2.02 mmol/L at the end of the experiments. Magnesium and SiO2 had the same trend: increasing at the beginning, then decreasing gradually until they reached their lowest values after 6 days, and then increasing. Fe concentrations increased almost five-fold after 6 days of reaction, and then it declined sharply (Fig. 9). 4.2.2. Reaction with calcareous mudstone The release of major ions into solution varied, mainly depending on the reacting fluids. To evaluate the effects of CO2 in the reaction process, two experiments were conducted: Ext. 5 CO2–brine and Ext. 6 CO2-free brine with calcareous mudstone. As shown in Fig. 10, in solutions from reactions with CO2–brine, the concentration of Ca was almost double, whereas concentrations of K, Fe, Si were higher by 54, 94 and 27%, respectively, compared to their corresponding cases of brine only. The concentration of Mg in the CO2–brine experiment was 65 times higher than that in the brine-only experiment.

Fig. 6. Bulk XRD plots of the calcareous mudstone: unreacted, reacted with CO2–brine, and reacted with brine only.


271

Fig. 7. SEM images of the calcareous mudstone after 12 days of reaction with CO2–brine. (a) albite secondary enlargement boundary, (b) authigenic chlorite, (c) illite, and (d) carbonate minerals.

Based on a mass balance, the saturation index (SI) of the primary and secondary minerals can reflect the mineral dissolution and precipitation to a certain extent. The saturation index of minerals can be calculated by geochemical modeling. The simulation was done with the nonisothermal multiphase reactive transport code TOUGHREACT, which was developed by US Lawrence Berkeley National Laboratory (Xu et al., 2006). A broad range of subsurface thermal–physical–chemical processes are considered under various thermohydrological and geochemical

conditions of temperatures (0–300 °C), pressures (1 to hundreds bars), water saturation, ionic strength (up to 6 mol/kg H2O), and pH. The program can be applied to the related numerical simulation study for pore media and fractured rock for one, two, and three-dimensional domain. On the basis of the measured final solution composition of fluids, equilibrium speciation calculations were made for all experiments for the two rock samples. In addition, a predictive kinetic calculation was also run for the lithic sandstone that was reacted with CO2–brine for 3, 6, 9, and 12 days at 160 °C and 15 MPa. Saturation indices are given in Table 6.

Fig. 8. Changes in pH and HCO− 3 concentration during the reaction time for the lithic sandstone experiments.

Fig. 9. Changes of ion concentrations during the reaction time for the lithic sandstone experiments.

4.3. Geochemical modeling

272


Fig. 10. Leachate chemistry of the calcareous mudstone reacted with CO2–brine and brineonly for 12 days at 160 °C.

From the simulation results, it can be seen that, for the lithic sandstone reacted with CO2–brine at 160 °C, the solution gradually became saturated with respect to ankerite and illite over time. Dawsonite and K-feldspar in all four experiments at different reaction times were not saturated. Calcite was unsaturated in the early stage of the reaction, then it reached a saturated state after 12 days of reaction. The solution was saturated with respect to kaolinite, quartz, siderite and smectite in all four experiments. For the calcareous mudstone, the main change was the emergence of plagioclase and dolomite, which were detected by XRD and SEM. This was consistent with the positive saturation indices. The solution was saturated with respect to kaolinite, quartz and smectite under the CO2–brine condition, whereas it was unsaturated with respect to these minerals under the CO2-free condition. For both lithic sandstone and calcareous mudstone, the solution was unsaturated with respect to chlorite in all six experiments, indicating that its dissolution would continue if the experiment was been allowed to run for longer periods. 5. Discussion 5.1. Dissolution and precipitation of minerals induced by injected CO2 To speed up the rate of reactions, and to overcome the kinetic barrier of mineral dissolution and precipitation, the experiments were run at higher temperature (160 °C) than the current reservoir temperature (40 °C). This is because the rate of silicate mineral reactions increases with increased temperature (Gunter et al., 1997; Gislason and Oelkers, 2003; Credoz et al., 2009). In addition, the reactive surface area of the caprock material was increased by crushing it to 180–220 μm grain size to increase the reactivity. A brine solution of 1 M NaCl was used as a substitute for formation water to encourage the dissolution process, as the presence of more ions in the initial brine solution, especially divalent cations, has a pH-buffering effect and decreases the acid-induced reactions. The water rock ratio was also increased for reactivity compared to practical geological storage conditions. Further enhancement of dissolution was guaranteed by continuous stirring. Results of the six experiments provided data that contribute to the qualitative understanding of fluid rock interactions under geochemical settings of CO2 sequestration in deep coal seams. From X-ray diffraction analysis and SEM images, dissolution and precipitation of minerals due to reaction of acidic fluid with formation rock after CO2 injection could be clearly observed. However, secondary carbonate minerals were found in the intergranular pore space only in the experiment of calcareous mudstone

reacted with CO2 and brine for 12 days. The generation of secondary carbonate minerals was not found in any of the other experiments. The main reason may be because the reaction time was shorter. The experiments of other studies lasted for longer times, e.g., several weeks, months or more than 1 year (Alemu et al., 2011; Tarkowski and Wdowin, 2011; Fischer et al., 2013; Wdowin et al., 2014). With the progress of the experiments, various elements released into the reaction liquid from primary materials, resulting in dramatic changes in ion concentrations in the aqueous phase, which can be used as an effective monitoring tool of CO2 leakages from underground reservoirs to shallow aquifers, This makes it valuable and necessary to obtain the chemical background characteristics of aquifers. Geochemical simulation can reflect the results of the experiments to some extent, but there were also some differences between the simulation and experimental results (Gaus, 2010). This is because a number of parameters were used in the simulation, and the actual formation condition is very complex. There are many factors involved in geochemical reactions, and the differences between simulation and experimental results are larger when an actual field situation is considered. Therefore, the fitting of experimental and simulation results is very challenging. However, if considering the thermal dynamic parameters of actual minerals, geochemical modeling is also very necessary (Fischer et al., 2013). In view of these reasons, the geochemical modeling needs a more comprehensive database that includes various minerals, particularly for clay minerals with unique qualities. Under the condition of geologic sequestration, CO2 injected into the target reservoir will tend to migrate upwards towards the cap rock, because the density of the CO2 is lower than that of the formation water. Because of the overlying cap rock with a lower permeability, much of the supercritical CO2 will accumulate at the bottom of the cap rock and migrate slowly with the flow of the formation water (Bachu and Shaw, 2003; Xu et al., 2010). The accumulated CO2 under the cap rock will gradually dissolve into brine, lowers the pH and change the aqueous complexation, thereby inducing mineral alteration. In turn, the mineralogical composition could impose significant effects on the evolution of the solution, and on the CO2 storage. The sealing ability of the cap rock is always the key in the study of the safety of GCS, and it is also an important security index in the evaluation of GCS (W. Zhang et al., 2009; X. Zhang et al., 2009). A large number of experiments (Kaszuba et al., 2005; Wigand et al., 2008; Busch et al., 2009; Gaus, 2010; Yalcinkaya et al., 2011; Liu et al., 2012) and numerical simulations (Gaus et al., 2005; Gherardi et al., 2007; Fischer et al., 2013, 2014; Tian et al., 2014) were carried out to study the effects of CO2–brine–rock geochemical reactions on the self-sealing of cap rock. The results show that the influence mainly depends on mineral composition, pH conditions and reaction time, etc. Secondary precipitation of minerals could block pore channels, further reducing the rate of CO2 diffusion and restricting the process of interaction between fluid and rock. 5.2. Potential mobilization of organic compounds due to the presence of CO2 Supercritical CO2 is a good solvent for organic compounds such as benzene, toluene, ethyl-benzene, and xylene (BTEX), phenols, and polycyclic aromatic hydrocarbons (PAHs) (Anitescu and Tavlarides, 2006; Kolak and Burruss, 2006). Monitoring results from GCS field tests have shown that organic compounds are mobilized following CO2 injection (Monin et al., 1988; Olukcu et al., 1999; Okamoto et al., 2005; Kharaka et al., 2009, 2011; Scherf et al., 2011). Such results have raised concerns regarding the potential for groundwater contamination by toxic organic compounds mobilized during GCS. Knowledge of the mobilization mechanism of organic compounds and their transport and fate in the subsurface is essential for assessing risks associated with GCS. Especially for coal measure strata, which contain large amounts of organic matter, once CO2 is injected into the coal seams for long term sequestration, it will inevitably extract the organic matter from the coal and typical cap rock such as shale or mudstone.


There is still very little study devoted specifically to organic material migration caused by CO2 sequestration in deep coal seams or abandoned oil and gas field formations. Kolak and Burruss (2006) conducted laboratory experiments to evaluate the potential for mobilizing nonmethane hydrocarbons during CO2 storage or enhanced coal bed methane recovery from deep coal beds, where coal samples of different rank were extracted with supercritical CO2 (40 °C; 10 MPa). The experimental results showed that supercritical CO2 is capable of mobilizing hydrocarbons from the coal matrix. The amount and type of hydrocarbons mobilized varied in response to coal rank. The hydrocarbon distributions in samples extracted indicate that this influence of coal rank is likely derived from the chemical changes associated with bitumen generation and cracking during increasing degrees of coalification and thermal maturation. The coal samples continued to yield hydrocarbons during consecutive supercritical CO2 extractions, indicating that the extent of hydrocarbon mobilization is not controlled solely by the solubility of hydrocarbons in supercritical CO2, but it may also be affected by other factors, including hydrocarbon partitioning between bitumensupercritical CO2 phases. Zhong et al. (2014) conducted column experiments using a water-wetted sandstone core installed in a tri-axial core holder to study the potential impact on groundwater of toxic organic compounds mobilized from coal by supercritical CO2 under simulated GCS conditions. Their results indicated that the mobility though the core sample was much higher for BTEX compounds than for naphthalene. Retention of organic compounds from the vapor phase to the core appeared to be primarily controlled by partitioning from the vapor phase to the aqueous phase. Adsorption to the surfaces of the wetted sandstone was also significant for naphthalene. Reduced temperature and elevated pressure resulted in greater partitioning of the mobilized organic contaminants into the pore water. In this study, we focused on inorganic minerals, rather than organic matter, in coal seams, despite the lack of theoretical or experimental studies regarding the transport of organic compounds mobilized by CO2 from a coal seam to the shallow aquifer or ground surface. Therefore, the mobilization of organic compounds and its effects on the physic-chemical properties of coal and rock should be considered in future studies. 6. Conclusions Experimental study indicated that, in the process of CO2 storage in deep coal seams, the cap rock actively participated in chemical reactions and played an important role in geological CO2 sequestration. The chemical composition of the fluid and minerals alteration reflected the process of CO2–brine–rock interaction. The change of mineral components was clearly observed, mainly including dissolution of silicate minerals in lithic sandstone experiments, and precipitation of carbonate minerals in calcareous mudstone experiments. Overall, the calcareous mudstone was more reactive than the lithic sandstone, as documented by the dissolution of calcite and illite/smectite and the formation of dolomite, siderite, illite and chlorite. CO2 was also permanently trapped as dolomite and siderite. The lithic sandstone showed significant dissolution of silicate minerals (illite/smectite, kaolinite and biotite) and the formation of chlorite and illite. Formation of clay minerals could cause intergranular and intragranular pore blockage, reducing the porosity of cap rock and thereby preventing permeation of CO2 through the cap rock into the shallow aquifer, which would otherwise lead to groundwater pollution. Thus, the formation of clay minerals would increase the security of geological CO2 sequestration. Geochemical simulation can reflect the dissolution and precipitation state of minerals to some extent, but it does not entirely coincide with the experimental results. Therefore, to guarantee the reliability of the results, simulation requires many of the physical properties of the actual formation and the thermal dynamic parameters of the rocks. Finally, these types of water-rock interaction studies should be applied to the real project sites with 3-D models. As much as possible,

273

consideration should be given to the effects of relevant factors, including fluid flow, geochemical reactions and geomechanics. The conjunction of these methods can yield more comprehensive prediction of the physico-chemical changes due to mineral dissolution and precipitation of caprock in geological CO2 storage. Acknowledgments This work was supported by the China Geological Survey working project (Grant No. 12120113006300). We would like to thank the editor Dr. Karacan, and three anonymous reviewers for their constructive comments and suggestions for improving the quality of this paper. References Alemu, B.L., Aagaard, P., Munz, I.A., Skurtveit, E., 2011. Caprock interaction with CO2: a laboratory study of reactivity of shale with supercritical CO2 and brine. Appl. Geochem. 26 (12), 1975–1989. Anitescu, G., Tavlarides, L.L., 2006. Supercritical extraction of contaminants from soils and sediments. J. Supercrit. Fluids 38, 167–180. Bachu, S., Shaw, J., 2003. Evaluation of the CO2 sequestration capacity in Alberta's oil and gas reservoirs at depletion and the effect of underlying aquifers. Pet. Technol. 42 (9), 51–61. Baran, P., Zarębska, K., Bukowska, M., 2015. Expansion of hard coal accompanying the sorption of methane and carbon dioxide in isothermal and non-isothermal processes. Energy Fuel 29 (3), 1899–1904. Busch, A., Alles, S., Krooss, B.M., Stanjek, H., Dewhurst, D., 2009. Effects of physical sorption and chemical reactions of CO2 in shaly caprocks. Energy Procedia 1, 3229–3235. Cai, Y., Liu, D., Pan, Z., Yao, Y., Li, C., 2015. Mineral occurrence and its impact on fracture generation in selected Qinshui Basin coals: an experimental perspective. Int. J. Coal Geol. 150, 35–50. Credoz, A., Bildstein, O., Jullien, M., Raynal, J., Pétronin, J., Lillo, M., Pozo, C., Geniaut, G., 2009. Experimental and modeling study of geochemical reactivity between clayey caprocks and CO2 in geological storage conditions. Energy Procedia 1, 3445–3452. Czerw, K., 2011. Methane and carbon dioxide sorption/desorption on bituminous coal—experiments on cubicoid sample cut from the primal coal lump. Int. J. Coal Geol. 85 (1), 72–77. Dawson, G.K.W., Golding, S.D., Massarotto, P., Esterle, J.S., 2011. Experimental supercritical CO2 and water interactions with coal under simulated in situ conditions. Energy Procedia 4, 3139–3146. Fischer, S., De Lucia, M., Liebscher, A., 2014. Kinetic modeling of laboratory CO2-exposure experiments performed on whole rock reservoir samples. Greenh. Gases Sci. Technol. 4 (2), 244–256. Fischer, S., Liebscher, A., De Lucia, M., Hecht, L., 2013. Reactivity of sandstone and siltstone samples from the Ketzin pilot CO2 storage site—laboratory experiments and reactive geochemical modeling. Environ. Earth Sci. 70 (8), 3687–3708. Gale, J., Freund, P., 2001. Coal-bed methane enhancement with CO2 sequestration worldwide potential. Environ. Geosci. 8 (3), 210–217. Gaus, I., 2010. Role and impact of CO2–rock interactions during CO2 storage in sedimentary rocks. Int. J. Greenh. Gas Control 4 (1), 73–89. Gaus, I., Azaroual, M., Czernichowski-Lauriol, I., 2005. Reactive transport modeling of the impact of CO2 injection on the clayey cap rock at Sleipner (North Sea). Chem. Geol. 217 (3), 319–337. Gentzis, T., 2000. Subsurface sequestration of carbon dioxide—an overview from an Alberta (Canada) perspective. Int. J. Coal Geol. 43 (1), 287–305. Gherardi, F., Xu, T.F., Pruess, K., 2007. Numerical modeling of selflimiting and selfenhancing caprock alteration induced by CO2 storage in a depleted gas reservoir. Chem. Geol. 244 (1), 103–129. Gislason, S.R., Oelkers, E.H., 2003. Mechanism, rates, and consequences of basaltic glass dissolution: II. An experimental study of the dissolution rates of basaltic glass as a function of pH and temperature. Geochim. Cosmochim. Acta 67 (20), 3817–3832. Golding, S.D., Uysal, I.T., Bolhar, R., Boreham, C.J., Dawson, G.K.W., Baublys, K.A., Esterle, J.S., 2013. Carbon dioxide-rich coals of the Oaky Creek area, central Bowen Basin: a natural analogue for carbon sequestration in coal systems. Aust. J. Earth Sci. 60 (1), 125–140. Gunter, W.D., Wiwchar, B., Perkins, E.H., 1997. Aquifer disposal of CO2-rich greenhouse gases: extension of the time scale of experiment for CO2 sequestering reactions by geochemical modelling. Mineral. Petrol. 59 (1–2), 121–140. Huang, X., Sun, Q., Yan, B., Xiong, D., Wang, W., 2010. Coalbed methane potential of Shizhuang North Block in Qinshui Basin, Shanxi. China Coalbed Methane 7 (5), 3–9. Karacan, C.Ö., 2003. Heterogeneous sorption and swelling in a confined and stressed coal during CO2 injection. Energy Fuel 17 (6), 1595–1608. Kaszuba, J.P., Janecky, D.R., Snow, M.G., 2005. Experimental evaluation of mixed fluid reactions between supercritical carbon dioxide and NaCl brine: relevance to the integrity of a geologic carbon repository. Chem. Geol. 217 (3), 277–293. Kharaka, Y.K., Campbell, P., Thomas, R.B., 2011. Environmental impacts of geologic sequestration of CO2: mobilized toxic organic compounds. International Symposium on Isotopes in Hydrology, Marine Ecosystems, and Climate Change Studies. IAEA, Monaco (March 29, 2011). Kharaka, Y.K., Thordsen, J.J., Hovorka, S.D., Nance, H.S., Cole, D.R., Phelps, T.J., Knauss, K.G., 2009. Potential environmental issues of CO2 storage in deep saline aquifers:

274


geochemical results from the Frio-I Brine Pilot test, Texas, USA. Appl. Geochem. 24, 1106–1112. Kolak, J.J., Burruss, R.C., 2006. Geochemical investigation of the potential for mobilizing non-methane hydrocarbons during carbon dioxide storage in deep coal beds. Energy Fuel 20, 566–574. Lei, H., Xu, T., Jin, G., 2015. TOUGH2Biot—a simulator for coupled thermal–hydrodynamic–mechanical processes in subsurface flow systems: application to CO2 geological storage and geothermal development. Comput. Geosci. 77, 8–19. Li, H., Shimada, S., Zhang, M., 2004. Anisotropy of gas permeability associated with cleat pattern in a coal seam of the Kushiro coalfield in Japan. Environ. Geol. 47 (1), 45–50. Li, W., Cheng, Y.P., Wang, L., Zhou, H.X., Wang, H.F., Wang, L.G., 2013. Evaluating the security of geological coalbed sequestration of supercritical CO2 reservoirs: The Haishiwan coalfield, China as a natural analogue. Int. J. Greenh. Gas Control 13, 102–111. Liu, F., Lu, P., Griffith, C., Hedges, S.W., Soong, Y., Hellevang, H., Zhu, C., 2012. CO2–brine– caprock interaction: reactivity experiments on Eau Claire shale and a review of relevant literature. Int. J. Greenh. Gas Control 7, 153–167. Lu, X., Huang, W., Tang, X., Ao, W., Zhang, S., Liu, S., Jia, Y., Liu, H., 2012. Impact of limestone of no. 15 coal seam roof on the coal bed methane exploitation, Southern Qinshui Basin. Geoscience 26 (3), 518–526. Massarotto, P., Golding, S.D., Bae, J.S., Iyer, R., Rudolph, V., 2010. Changes in reservoir properties from injection of supercritical CO2 into coal seams—a laboratory study. Int. J. Coal Geol. 82 (3), 269–279. Metz, B., Davidson, O., De Coninck, H.C., Loos, M., Meyer, L.A., 2005. IPCC special report on carbon dioxide capture and storage. Prepared by Working Group III of the Intergovernmental Panel on Climate Change. Cambridge University Press, New York. Monin, J.C., Barth, D., Perrut, M., Espitalie, M., Durand, B., 1988. Extraction of hydrocarbons from sedimentary rocks by supercritical carbon dioxide. Org. Geochem. 13, 1079–1086. Okamoto, I., Li, X., Ohsumi, T., 2005. Effect of supercritical CO2 as the organic solvent on cap rock sealing performance for underground storage. Energy 30, 2344–2351. Olukcu, N., Yanik, J., Saglam, M., Yuksel, M., Karaduman, M., 1999. Solvent effect on the extraction of Beypazari oil shale. Energy Fuel 13, 895–902. Pan, Z., Luke, C., 2007. Effects of coal matrix shrinkage/swelling on enhanced CBM recovery through CO2 sequestration. China Coalbed Methane 4 (1), 7–10. Qin, Y., Fu, X.H., Wu, C.F., 2005. Self-adjusted elastic action and its CBM pool-forming effect of the high rank coal reservoir. Chin. Sci. Bull. 50 (B10), 99–103. Scherf, A.K., Zetzl, C., Smirnova, I., Zettlitzer, M., Vieth-Hillebrand, A., 2011. CO2SINKgroup mobilisation of organic compounds from reservoir rocks through the injection of CO2—comparison of baseline characterization and laboratory experiments. Energy Procedia 4, 4524–4531. Siriwardane, H.J., Bowes, B.D., Bromhal, G.S., Gondle, R.K., Wells, A.W., Strazisar, B.R., 2012. Modeling of CBM production, CO2 injection, and tracer movement at a field CO2 sequestration site. Int. J. Coal Geol. 96, 120–136. Sun, K., 2005. Model and numerical simulation about multispecies fluid diffusion and seepage in exploitation coalbed methane by gas injection. J. Liaoning Tech. Univ. 24 (3), 305–308. Tang, S., Tang, D., Yang, Q., 2004. Variation regularity of gas component concentration in binary-component gas adsorption–desorption isotherm experiments. J. China Univ. Min. Technol. 33 (4), 448–452. Tarkowski, R., Wdowin, M., 2011. Petrophysical and mineralogical research on the influence of CO2 injection on Mesozoic reservoir and caprocks from the Polish lowlands. Oil Gas Sci. Technol. Rev. IFP Energies Nouvelles 66 (1), 137–150. Tian, H., Pan, F., Xu, T., McPherson, B.J., Yue, G., Mandalaparty, P., 2014. Impacts of hydrological heterogeneities on caprock mineral alteration and containment of CO2 in geological storage sites. Int. J. Greenh. Gas Control 24, 30–42. Tian, H., Xu, T., Li, Y., Yang, Z., Wang, F., 2015. Evolution of sealing efficiency for CO2 geological storage due to mineral alteration within a hydrogeologically heterogeneous caprock. Appl. Geochem. 63, 380–397.

Vishal, V., Singh, L., Pradhan, S.P., Singh, T.N., Ranjith, P.G., 2013. Numerical modeling of Gondwana coal seams in India as coalbed methane reservoirs substituted for carbon dioxide sequestration. Energy 49, 384–394. Wdowin, M., Tarkowski, R., Franus, W., 2014. Supplementary studies of textural and mineralogical changes in reservoir and caprocks from selected potential sites suitable for underground CO2 storage. Arab. J. Sci. Eng. 39 (1), 295–309. Weber, M., Wilson, T.H., Akwari, B., Wells, A.W., Koperna, G., 2012. Impact of geological complexity of the Fruitland formation on combined CO2 enhanced recovery/sequestration at San Juan Basin pilot site. Int. J. Coal Geol. 104, 46–58. Wei, C., Qin, Y., Wang, G.G., Fu, X., Jiang, B., Zhang, Z., 2007. Simulation study on evolution of coalbed methane reservoir in Qinshui basin, China. Int. J. Coal Geol. 72 (1), 53–69. White, C.M., Smith, D.H., Jones, K.L., Goodman, A.L., Jikich, S.A., LaCount, R.B., DuBose, S.B., Ozdemir, E., Morsi, B.I., Schroeder, K.T., 2005. Sequestration of carbon dioxide in coal with enhanced coalbed methane recovery — a review. Energy Fuel 19 (3), 659–724. Wigand, M., Carey, J.W., Schütt, H., Spangenberg, E., Erzinger, J., 2008. Geochemical effects of CO2 sequestration in sandstones under simulated in situ conditions of deep saline aquifers. Appl. Geochem. 23 (9), 2735–2745. Winter, E.M., Bergman, P.D., 1993. Availability of depleted oil and gas reservoirs for disposal of carbon dioxide in the United States. Energy Convers. Manag. 34 (9), 1177–1187. Wong, S., Law, D., Deng, X., Robinson, J., Kadatz, B., Gunter, W.D., Zhiqiang, F., 2007. Enhanced coalbed methane and CO2 storage in anthracitic coals—micro-pilot test at South Qinshui, Shanxi, China. Int. J. Greenh. Gas Control 1 (2), 215–222. Wong, S., Macdonald, D., Andrei, S., Gunter, W.D., Deng, X., Law, D., Ho, P., 2010. Conceptual economics of full scale enhanced coalbed methane production and CO2 storage in anthracitic coals at South Qinshui basin, Shanxi, China. Int. J. Coal Geol. 82 (3), 280–286. Wu, S., Zhang, A., 2007. A three-dimension multicomponent flow model of enhanced coalbed methane recovery by gas injection. Nat. Gas Geosci. 18 (4), 580–583. Xu, T., Apps, J.A., Pruess, K., 2004. Numerical simulation of CO2 disposal by mineral trapping in deep aquifers. Appl. Geochem. 19 (6), 917–936. Xu, T., Kharaka, Y.K., Doughty, C., Freifeld, B.M., Daley, T.M., 2010. Reactive transport modeling to study changes in water chemistry induced by CO2 injection at the Frio-I Brine Pilot. Chem. Geol. 271 (3), 153–164. Xu, T., Sonnenthal, E., Spycher, N., Pruess, K., 2006. TOUGHREACT—a simulation program for non-isothermal multiphase reactive geochemical transport in variably saturated geologic media: applications to geothermal injectivity and CO2 geological sequestration. Comput. Geosci. 32, 145–165. Yalcinkaya, T., Radonjic, M., Willson, C.S., Bachu, S., 2011. Experimental study on a single cement-fracture using CO2 rich brine. Energy Procedia 4, 5335–5342. Ye, J., Zhang, B., Wong, S., 2012. Test of and evaluation on elevation of coalbed methane recovery ratio by injecting and burying CO2 for 3# coal seam of north section of Shizhuang, Qingshui Basin, Shanxi. Chin. Eng. Sci. 14 (2), 38–44. Zhang, W., Li, Y., Xu, T., Cheng, H., Zheng, Y., Xiong, P., 2009a. Long-term variations of CO2 trapped in different mechanisms in deep saline formations: a case study of the Songliao Basin, China. Int. J. Greenh. Gas Control 3 (2), 161–180. Zhang, X., Lu, X., Liu, Q., 2009b. The effect of the characteristics of cap on the escaping velocity of CO2. Soil Eng. Found. 23 (3), 67–70. Zhang, X., Qian, C., Peng, J., 2004. Adsorption and desorption characteristics of coal. Well Test. Prod. Technol. 25 (3), 23–24. Zhong, L., Cantrell, K.J., Bacon, D.H., Shewell, J., 2014. Transport of organic contaminants mobilized from coal through sandstone overlying a geological carbon sequestration reservoir. Int. J. Greenh. Gas Control 21, 158–164. Zhou, F., Hussain, F., Cinar, Y., 2013. Injecting pure N2 and CO2 to coal for enhanced coalbed methane: experimental observations and numerical simulation. Int. J. Coal Geol. 116, 53–62.