zhang 2011

7/25/2019 Zhang 2011

1/13

Modeling of fate and transport of coinjection of H2S with CO2

in deep saline formations

Wei Zhang,1 Tianfu Xu,2 and Yilian Li1

Received 19 April 2010; revised 13 September 2010; accepted 2 December 2010; published 8 February 2011.

[1] The geological storage of CO2 in deep saline formations is increasingly seen as aviable strategy to reduce the release of greenhouse gases into the atmosphere. However,costs of capture and compression of CO2 from industrial waste streams containingsmall quantities of sulfur and nitrogen compounds such as SO2, H2S, and N2 are veryexpensive. Therefore, studies on the coinjection of CO2 containing other acid gases fromindustrial emissions are very important. In this paper, numerical simulations were

performed to study the coinjection of H2S with CO2 in sandstone and carbonateformations. Results indicate that the preferential dissolution of H2S gas (compared withCO2gas) into formation water results in the delayed breakthrough of H2S gas. Coinjectionof H2S results in the precipitation of pyrite through interactions between the dissolvedH2S and Fe

2+ from the dissolution of Febearing minerals. Additional injection of H2Sreduces the capabilities for solubility and mineral trappings of CO2 compared to theCO2only case. In comparison to the sandstone (siliciclastic) formation, the carbonateformation is less favorable to the mineral sequestration of CO2. In sandstone and carbonateformations, the presence of Febearing siliciclastic and/or carbonate is more favorableto the H2S mineral trapping.

Citation: Zhang, W., T. Xu, and Y. Li (2011), Modeling of fate and transport of coinjection of H2S with CO2in deep saline

formations, J. Geophys. Res., 116, B02202, doi:10.1029/2010JB007652.

1. Introduction

[2] Globalemissions of greenhouse gases (GHG) especiallycarbon dioxide (CO2) have increased rapidly and led to global

climate change and ocean acidification with severe con-sequences for ecosystems and for human society [Holloway,2001; West et al., 2005]. Thus, reducing the concentrationofCO2 in the atmosphere is very important to the mitigation ofglobal climate change. Currently, research on CO2geologicalstorage as a possible method for reducing the emission of CO2from industrial point sources is being extensively carried out[Gentzis, 2000; Holloway, 2005; Gough, 2008]. The maingeologic formations include depleted or depleting oil and gasreservoirs, unmineable coal seams, and deep saline formations[Bachu et al., 1994;Hitchon et al., 1999]. Injecting CO2intosaline formations in sedimentary basins is one of the mostpromising methods of CO2 geological storage for the longterm sequestration of the gas. This is because saline forma-

tions are ubiquitous to sedimentary basins [Hitchon et al.,1999; Gunter et al., 2000; Izgec et al., 2008], they haveenough capacity to store large amounts of CO2from anthro-pogenic emissions, and there are short distances between mostlarge CO2 point sources and saline formations, which can

minimize CO2transportation costs [Soong et al., 2004;Allenet al., 2005;Zerai et al., 2006].

[3] The fate of minor quantities of sulfur and nitrogencompounds during combustion or gasification of coal is of

considerable interest,as their release into the atmosphere leadsto the formation of urban ozone and acid rain, the destructionof stratospheric ozone, and global warming. Coal also con-tains many trace elements that are potentially hazardous tohuman health and the environment, such as mercury andarsenic, and their release into the atmosphere is restrictedunder the Clean Air Act Amendments (CAAA) of 1990.Studies indicated that costs of capture and compression ofCO2 from flue gas or a coal gasification process are veryhigh, accounting for 75% of the total cost of a geologicalstorage process [Knauss et al., 2005]. Therefore it may beeconomically advantageous to sequester/store CO2 withthese constituents in deep geological formations [Knausset al., 2005; Bachu et al., 2009a; Ellis et al., 2010; Xiao

et al., 2009; Crandell et al., 2010]. In this study, we evalu-ate the coinjection of hydrogen sulphide (H2S) with CO2.

[4] In order to reduce emissions of produced H2S andCO2 as byproducts of sour gas processing, a total of morethan 40 active acidgas injection projects were performed inthe Alberta Basin of Canada to the end of 2003 [Bachu andGunter, 2005]. The acid gas can be injected as solution ordense fluid (liquid or supercritical). The injected gas com-position varies from 83% H2S and 14% CO2to 2% H2S and95% CO2 for different storage sites [Bachu et al., 2005].Since the first acidgas injection operation in a depleted

1School of Environmental Studies, China University of Geosciences,

Wuhan, China.2Earth Sciences Division, Lawrence Berkeley National Laboratory,

University of California, Berkeley, California, USA.

Copyright 2011 by the American Geophysical Union.01480227/11/2010JB007652

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 116, B02202,doi:10.1029/2010JB007652, 2011

B02202 1 of13
http://dx.doi.org/10.1029/2010JB007652http://dx.doi.org/10.1029/2010JB007652http://dx.doi.org/10.1029/2010JB007652

7/25/2019 Zhang 2011

2/13

sandstone oil reservoir on the outskirts of Edmonton beganin 1989, no leakage or other safety problems have beenreported [Bachu and Gunter, 2004; Machel, 2005]. At theend of 2003, a total of about 2 Mt H2S and 2.5 Mt CO2wereinjected in deep saline formations, and depleted oil or gasreservoirs, for all storage sites in western Canada, at averagerates that vary between 1 103 and 500 103 m3 d1

[Bachu and Gunter, 2005; Bachu et al., 2005]. The success

of these acid

gas injection projects in Canada and the UnitedStates [Bachu et al., 2009b] indicates that the coinjection ofH2S and CO2 in geological media is a mature and safetechnology.

[5] The acid gases from natural gas production contain 284 vol % H2S, but most coals contain no more than 5 wt. %sulfur. Therefore, the concentration of H2S in CO2 from theintegrated gas combined cycle (IGCC) plants is unlikely toexceed 1.5 vol % [Gunter et al., 2000; Xu et al., 2007].

[6] In order to better understand the multiphase fluidtransport and geochemical reactions during the coinjectionprocess of acid gases. Some laboratory experiments havebeen conducted for a short time. Bachu et al.[2009a, 2009b]and Bachu and Bennion [2009] used a series of static and

dynamic experiments to investigate the effects of gas con-centration and in situ conditions of pressure, temperatureand water salinity on the partitioning of CO2 and H2S.Experimental results showed that the higher solubility ofH2S than that of CO2 results in suppressed H2S concentra-tions at the leading edge of the breakthrough gas phase.They also performed 1D numerical simulations to study theeffect of flow conditions such as gas solubility, mediumabsolute and relative permeabilities, dispersion, mobilityand displacement direction on the time of CO2 and H2Sbreakthrough. In these studies, however, watergasrockchemical interactions are not addressed. The chemical pro-cesses are very important for the longterm fate and trans-port of injected acid gases, because these interactions may

cause major changes in the rock structure and chemicalcomposition of the storage formation. Gunter et al. [2005]performed autoclave batch experiments to study the reac-tion of siderite with acid gas (CO2 and H2S) at the condi-tions of 5.15 bars and 54C for 14 days. Analysis results ofthe reaction products showed that the precipitated sulfidemineral during the experiment is pyrite.

[7] On the other hand, many modeling studies have con-centrated on chemical processes induced by the coinjectionof H2S with CO2. Knauss et al. [2005] used a 1D reactivetransport simulation and found that the additional injectionof dissolved H2S with CO2 as an aqueous phase does nothave a significant impact on the waterrock interactioncompared to the CO2only case.Gunter et al.[2000] used a

batch geochemical model to study the interaction of indus-trial waste streams comprising CO2 and H2S with theminerals in typical carbonate and sandstone (siliciclastic)formations from the Alberta Basin, Canada. Their simula-tion results indicated that decrease in pH caused by disso-lution of injected H2S results in the dissolution of sideritewhich provides Fe2+ to the precipitation of pyrrhotite. Xuet al. [2007] developed a 1D radial flow model for mod-eling of coinjection of dissolved H2S with gaseous CO2in asandstone formation. Their simulation results showed thatthe coinjection, compared to the CO2 injection only, does

not significantly affect pH distribution, mineral alteration,and CO2 mineral trapping. The main difference is precipi-tation of pyrite in the coinjection case.

[8] Some of previous reactive transport modeling [e.g.,Xuet al., 2007; Xiao et al., 2009] are limited in that only CO2can be injected in the gaseous state. Therefore, coinjection ofbrines carrying dissolved H2S was used for expediency. Theconsequences of this artifact on the model results, especially

in relation to CO2sequestration, are relatively minor becauseinjection occurs only in the early time, much shorter than thesimulation time (on the order of 1,000 years).

[9] In this paper, fluid flow modeling capabilities forinjection of anhydrous supercritical CO2containing H2S gasare employed owing to the improvement of a version ofTOUGHREACT (see more details in section 2.1). We per-formed twodimensional (2D) radial flow modeling tostudy coinjection of H2S with CO2 in sandstone and car-bonate formations. The 2D radial flow model used hereallows us to study buoyancy forces that would tend to drivethe gas mixture toward the top of the formation, whichcannot be demonstrated by the previous 1D well models[e.g.,Xu et al., 2007]. Compared to that ofXu et al. [2007],

this paper is a significant step forward on modeling ofinjection of H2S with CO2 and subsequent fate and trans-port. The simulation results can be used to evaluate thebehavior and effect of impurityH2S in the CO2 stream andtheir performance during the injection and storage period.

2. Numerical Approaches

2.1. Numerical Tool

[10] The present simulations employed the non-isothermal multiphase reactive geochemical transport codeTOUGHREACT [Xu and Pruess, 2001; Xu et al., 2006],which is developed by introducing reactive chemistry into themultiphase fluid and heat flow code TOUGH2 [Pruess et al.,

1999, 2004]. An improved version [Battistelli, 2008] of theTMVOC simulator [Pruess and Battistelli, 2002] was linkedto TOUGHREACT, resulting in an acid gas injection simu-lator (Xu, unpublished). TMVOC models the migration ofthreephase multicomponent inorganic gases and hydro-carbons mixtures for environmental applications. Theimproved TMVOC can simulate the twophase behavior ofsodium chloride dominated brines in equilibrium with anonaqueous phase made up of inorganic gases, such as CO2,H2S and N2 and hydrocarbons (alkanes up to decane, ben-zene, toluene, ethylbenzene and xylenes (BTEX), and pseu-docomponents). It can then be used to model injection ofacid gas mixtures in saline formations, and account for thepresence of additional gaseous species in the injected CO2,

such as contaminants like H2S or N2.[11] The numerical method for solving fluid flow and

chemical transport is based on an integral finite difference(IFD) method for space discretization [Narasimhan andWitherspoon, 1976]. The IFD method provides flexiblediscretization of geologic media by allowing the use ofirregular grids, which is well suited for simulation of flow,transport, and fluidrock interaction in heterogeneous andfracture rock systems with varying petrology. For regulargrids, the IFD method is equivalent to the ConventionalFinite Difference method. An implicit timeweighting

ZHANG ET AL.: COINJECTION OF H2S WITH CO2 B02202B02202

2 of 13

7/25/2019 Zhang 2011

3/13

scheme is used for the flow, transport and kinetic geo-chemical equations. TOUGHREACT uses a sequential

iteration approach similar to that described by Yeh andTripathi [1991]. After solution of the flow equations, thevelocities and saturations of the aqueous phase are used foraqueous chemical transport simulation. Chemical transport isthen solved on a component basis. Resulting concentrationobtained from the transport and CO2and H2S gas pressuresin the multiphase flow calculation are substituted into thechemical reaction model. The system of chemical reactionequations is solved on a gridblock basis by NewtonRaphson iteration. The program can be applied to one, two,or threedimensional porous and fractured media withphysical and chemical heterogeneity. It can accommodateany number of chemical species present in the liquid, gas andsolid phases. A wide range of subsurface thermophysico-

chemical processes are considered under various thermo-hydrological and geochemical conditions such as pressure,temperature, ionic strength, pH and Eh. Changes in porosityand permeability due to mineral dissolution and precipitationcan modify fluid flow. Feedback between flow and chem-istry can be considered in this model. Porosity changes arecalculated from volume changes due to mineral dissolutionand precipitation. Permeability changes can then be evalu-ated by consideration of several alternative models describ-ing the porositypermeability relationship, including a simplegrain model of KozenyCarman, as used in the present study.

2.2. Model Setup

[12] Much specific and detailed information is required to

assess the feasibility of the acidgas injection and to developengineering designs for the injection systems. Before con-ducting sitespecific investigations, general features and is-sues relating to the injected formation should be explored.This can be done by extracting the sitespecific features andrepresenting characteristics. Increases in geological andgeometric complexities can increase the difficulty of iden-tifying the dominant geochemical processes. Therefore, asimple twodimensional (2D) radial well model was usedin this study. The 2D model was a homogeneous formationof 50 m thickness with a cylindrical geometrical configu-

ration (Figure 1). In the vertical direction, 25 model layerswere used with a constant spacing of 2 m. In the horizontal

direction, a radial distance of 100 km was modeled with aradial grid spacing that increases gradually away from theinjection well. A total of 50 radial grid elements were used.The volume of the outer grid element is specified a largevalue of 1030 m3, representing an infinitive lateral boundary.CO2 only or CO2 containing H2S injection was applied atthe bottom portion of the well (the thickness of the injectionportion is 20 m). Injection of acid gases was applied for aperiod of 10 years, using a CO2 injection rate of 2 kg s

1

(0.063 Mt yr1) for the CO2only case, or using a CO2

injection rate of 1.9 kg s1 (0.06 Mt yr1) a n d a H2S

injection rate of 0.1 kg s1 (0.003 Mt yr1) for the coin-

jection case. The fluid flow and geochemical transportsimulation was run for a period of 500 years, which may be

a relevant timescale of interest for geological sequestrationof acid gases.

[13] We used the same hydrogeological parameters as thoseused in the research in the Songliao Basin of China [Zhanget al., 2009], which is listed in Table A1 in Appendix A.Two types of (initial) mineralogical compositions (Table A2in Appendix A) were used, sandstone from Zhang et al.[2009] and carbonate from Zerai et al. [2006].

[14] Prior to simulating reactive transport, batch geo-chemical modeling of waterrock interaction was performedto generate an aqueousphase chemical composition closelyapproaching the composition of a typical formation brine byequilibrating a 0.171 M (mol kg1 H2O) solution of sodiumchloride in the presence of the primary minerals listed in

Table A2 with CO2gas pressure of 0.01 bar at a temperatureof 50C. We are primarily interested in the region affectedby CO2 injection. The uncertainty of background CO2pressure on simulation results should be very small for ourobjectives. A reasonably short simulation time (10 years inthe present study) is needed to obtain a quasistable (ornearly steady state) aqueous solution composition (Table A3in Appendix A).

[15] Dissolution and precipitation of minerals are con-sidered under kinetic conditions. The rate laws used arepresented in Appendix B. The kinetic parameters were taken

Figure 1. Schematic representation for the 2D well flow model for the coinjection of CO2with H2S in aformation (modified from Zhang et al. [2009]).


3 of 13

7/25/2019 Zhang 2011

4/13

from Xu et al. [2007] and Zhang et al. [2009], which arelisted in Table C1 in Appendix C. A temperature of 50Cwas used, which may represent the temperature at a depth ofabout 1200 m, given a land surface temperature of 15C anda geothermal gradient of 30C km1.

2.3. Simulations

[16] As shown in Table 1, a total of four groups ofnumerical simulations were performed with different com-binations of injection scenario and rock type. The purpose isto investigate the effect of additional H2S injected with CO2in different rock types (e.g., sandstone and carbonate) on thegeochemical changes (aqueous composition, and mineral

dissolution and precipitation), fate and transport of injectedCO2 and H2S gases.

3. Results and Discussion

3.1. Sandstone Formation

[17] Figure 2 shows that CO2and H2S gases (supercritical)injected at the bottom of the deep saline formation migrateupward rapidly by buoyancy forces. Mass fraction of CO2at the advancing gas front is higher (Figures 2a and 2b).The front of H2S gas is behind that of CO2 gas (Figures 2cand 2d). This is due to the preferential solubility of H2S information water compared with that of CO2, which inducesthe delayed breakthrough of H2S gas, the separation between

CO2 and H2S gases, and suppressed H2S concentrations information water at the advancing gas front [Bachu et al.,2009a, 2009b; Bachu and Bennion, 2009]. Figure 3 showsthe spatial distribution of total dissolved sulfur (TDSu) andcarbon (TDC) in the sandstone formation. There is a greaterconcentration of TDSu in the interior of the gas plume,corresponding to the higher H2S mass fraction in the gasphase than that in the other region.

[18] In order to better understand the relationship amongthe concentration of TDC and TDSu, the mass fraction ofCO2 and H2S in the gas phase, and the partial pressure ofCO2 and H2S, we plot their changes with time on the basisof a point A in the model, which lies in the twophase zonewhere CO2 gas and aqueous phases coexist (see Figure 2b).

Figure 4 reveals that when the injected CO2gas reaches thepoint A, the mass fraction of CO2 in the gas phase is higherthan that in the initial gases injected (0.95, namely, the dash dotdotted line in Figure 4a). The temporal evolution ofmass fraction of CO2 and H2S indicates CO2 reaches thepoint A faster than H2S. This is due to the preferentialdissolution of H2S compared with CO2.

[19] Trends in concentrations of TDC in the CO2 con-taining H2S and CO2only cases, and TDSu in the coinjec-tion case are generally similar to variations in partialpressures of CO2 and H2S (Figure 4b). Trend in concentra-tions of TDC is opposite to that of TDSu after the gases

reach this point. As expected the concentration of TDC inthe coinjection case is lower than that in the CO2 alonecase. These indicate again the preferential dissolution ofH2S gas reduces the potential and capacity for the dissolution

Table 1. Four Groups of Simulations in Our Studies

Simulation Groups Injection Scenarios Rock Types

1 CO2 and H2S sandstone2 CO2 only sandstone3 CO2 and H2S carbonate4 CO2 and H2S carbonate without siderite

Figure 2. Spatial distribution of mass fraction of (a and b)CO2 and (c and d) H2S in gas phase obtained from thecase of H2S with CO2 in sandstone formation after 100 and500 years.


4 of 13

7/25/2019 Zhang 2011

5/13

of the injected CO2 gas in formation water. However, withthe decrease in TDSu concentration caused by the decreasein H2S partial pressure, the solubility of CO2 in formationwater will gradually increase.

[20] Dissolution of CO2and H2S lowered pH of formationwater, then induced dissolution and precipitation of minerals.Chlorite (Mg2.5Fe2.5Al2Si3O10(OH)8) dissolves (Figure 5a),which supplies Fe2+ and Mg2+ for the precipitation ofankerite (CaMg0.3Fe0.7(CO3)2) (Figure 5b) and pyrite (FeS2)(Figure 5c) for the case of CO2 containing H2S. Figure 6ashows cumulative CO2 sequestrated by carbonate precipita-tion (mineral trapping) in the sandstone formation. Precipi-

tation of ankerite requires Ca2+

provided by dissolution ofcalcite (CaCO3) (Figure 5d). In the case of coinjection ofH2S, pyrite precipitation occurs in the twophase zone, andankerite precipitation occurs in the aqueous phase zone. Thisis because the pH in the twophase zone is lower than that inthe aqueous phase zone (Figure 6b), and pyrite is stableunder relatively lowpH conditions.

Figure 3. Spatial distribution of concentration (unit is molkg1 H2O) of (a and b) total dissolved sulfur (TDSu) and(c and d) total dissolved carbon (TDC) obtained from thecase of H2S with CO2in sandstone formation after 100 and500 years.

Figure 4. Changes in total dissolved carbon and sulfur,mass fraction of CO2and H2S in gas phase, and partial pres-sure of CO2 and H2S with time in sandstone and carbonateformations for point A (see Figure 2b): (a) CO2and (b) H2S.


5 of 13

7/25/2019 Zhang 2011

6/13

[21] Dissolution of chlorite in the CO2only case (Figure 7a)is similar to that in the CO2containing H2S case (Figure 5a).Because the amount of precipitated ankerite in the coinjectioncase (Figure 5b) is lower than that in the CO2only case

(Figure 7b), the abundance of dissolved calcite in the formercase (Figure 5d) is also lower than that in the later case(Figure 7c). In the CO2 alone case, the CO2 mineral trap-ping (Figure 7d) is greater than those in the CO2 containingH2S case. This is because the precipitation of pyrite sup-presses the precipitation of Febearing carbonate mineralssuch as ankerite.

[22] Because the mineral trapping of CO2 is a slow

process, there is no significant increase in the amount ofinjected CO2 in solid phase compared with those in gas(supercritical) and aqueous phases under the present min-eralogical composition. Therefore, we only investigatechanges in gas, and aqueous+solid phases for differentcases, respectively. Temporal changes in fractions ofinjected CO2 in gas, and aqueous and solid phases obtainedfrom cases of CO2 containing H2S and CO2 only in sand-stone formations (Figure 8) confirm that preferential disso-lution of H2S can suppress CO2 dissolution.

[23] The 1D simulation results from Xu et al. [2007]indicated that additional injection of H2S causes morepyrite precipitation than the CO2only case. And comparedto the CO2 alone case, precipitation of pyrite in the lowpH

region for the case of H2S coinjection reduces ankeriteprecipitation in this zone. These modeling results are inagreement with our 2D study. However, owing to con-straints of the previous version of TOUGHREACT, theinjected H2S is assumed as aqueous phase. The acid gasinjection simulator used in current study can model bothCO2and H2S in gaseous state. The higher solubility of H2S

Figure 5. Spatial distribution of change of (a) chlorite,(b) ankerite, (c) pyrite, and (d) calcite obtained fromthe case of H2S with CO2 in sandstone formation after500 years (unit for minerals is change in volume fraction;positive values indicate precipitation, and negative valuesindicate dissolution).

Figure 6. Spatial distribution of (a) CO2 mineral trapping(in kg of CO2 per m

3 medium) and (b) pH value obtainedfrom case of H2S with CO2 in sandstone formation after500 years.


6 of 13

7/25/2019 Zhang 2011

7/13

gas in formation water than CO2, affects the fate andtransport of injected CO2 and H2S gases, including thedelayed breakthrough of H2S, the separation between thetwo gases, and suppressed H2S concentrations in formationwater at the advancing gas front. These were not obtained

from the study ofXu et al. [2007] because injected H2S isin the aqueous phase. Compared to the previous 1D model[e.g.,Knauss et al., 2005;Xu et al., 2007;Bachu and Bennion,2009; Bachu et al., 2009b], the current 2D radial flowmodel allows us to study buoyancy forces that would tend todrive the gas mixture toward the top of the formation, totrack transport and dissolution of CO2 and H2S at the frontof the gas plume, and to investigate variations of physical

and chemical processes over a vertical cross section.

3.2. Carbonate Formation

[24] Figure 9a shows that decrease of H2S amount in thegas phase for the carbonate (with siderite) formation ismore significant than that for the sandstone formation. Thisis because dissolution of siderite available in carbonate(Figure 10a) is faster than that of chlorite available insandstone, accelerating precipitation of pyrite (Figure 10b)and driving more H2S dissolution (Figures 11a and 11b)into formation water. The predicted precipitation of pyriteis consistent with the experiment results reported byGunter et al. [2005]. As expected, the decrease in TDSuconcentration (Figures 12a and 12b) induced by the

decrease in H2S partial pressure should enhance the moreCO2 dissolution due to the increasing contact between CO2gas and formation water (Figures 11c and 11d). However,Figure 9b indicates that decrease in CO2 gas in the car-bonate formation is less than that in the sandstone for-mation. This is because dissolution of carbonates such asdolomite (Figure 10c) and siderite increases the concen-tration of total dissolved carbon in formation water, whichcan suppress the CO2 dissolution. Some amounts of calciteprecipitate (Figure 10d) and ankerite precipitation arelimited.

[25] Some typical carbonate formations do not have sid-erite. Therefore, we performed a sensitivity simulation toinvestigate the effect of coinjection of CO2 containing H2S

in a no Febearing carbonate formation. Figure 9a showsthat dissolution of H2S in the carbonate (without siderite)formation is much less than that in sandstone and carbonateformations. Temporal changes of H2S gas in this case

Figure 7. Spatial distribution of change of (a) chlorite,(b) ankerite, (c) calcite, and (d) CO2 mineral trappingobtained from the case of CO2 only in sandstone formationafter 500 years.

Figure 8. Comparison of time evolution of injected CO2indifferent phases obtained from cases of H2S with CO2 andCO2 only in sandstone formations.


7 of 13

7/25/2019 Zhang 2011

8/13

confirm again that dissolution of Febearing minerals suchas chlorite and siderite has a strong impact on reduction ofH2S in the gas phase because of formation of pyrite.

[26] As expected, the significant decrease of H2S amountin the gas phase for the carbonate (with siderite) formationshould enhance CO2dissolution (compared to the carbonatecase without siderite). However, temporal changes of CO2gas for two cases in carbonate formations do not have a

significant difference (Figure 9b). This should be becausethe siderite dissolution in the carbonate case provides Fe2+

to precipitation of pyrite, and increases the concentration ofTDC in formation water, which can inhibit the dissolution ofCO2 gas. Changes in dolomite and calcite volume fractions(Figure 13) in this case are similar to the previous case withsiderite (Figure 10).

[27] The increased concentration of H+ in aqueous phaseinduced by acid gas injection interacts with aluminosilicateand/or silicate minerals such as feldspars and clay minerals

releasing cations such as Ca2+, Mg2+, and Fe2+. The dis-solved bicarbonate species reacts with these divalent cations

to precipitate carbonate minerals, sequestrating CO2 per-manently. Therefore, siliciclastic (sandstone) formations aremore favorable for CO2 mineral trapping than carbonateformations [Hitchon et al., 1999; Gentzis, 2000]. This doesnot mean that carbonate formations are not suitable for the

Figure 9. Comparison of time evolution of injected (a) H2Sand (b) CO2in different phases obtained from cases of H2Swith CO2 in sandstone, and carbonate with and without sid-erite formations.

Figure 10. Spatial distribution of change of (a) siderite,(b) pyrite, (c) dolomite, and (d) calcite obtained fromthe case of H2S with CO2 in carbonate formation after500 years.


8 of 13

7/25/2019 Zhang 2011

9/13

geological storage of CO2, but that the dominant trappingmechanisms in carbonate formations are solubility andhydrodynamic trappings [Zerai et al., 2006].

[28] However, if there are Febearing siliciclastic and/orcarbonate minerals occurred in primary minerals, sulfide

minerals such as pyrite can precipitate in the sandstone andcarbonate formations, increasing mineral trapping of H2S.

4. Findings and Conclusions

[29] We have developed a 2D radial reactive transportmodel for injection of H2S with CO2 in sandstone andcarbonate formations, using mineralogical composition and

water chemistry encountered in Songliao Basin of Chinaand Ohio. Major findings and conclusions from simulationsare summarized as follows.

[30] 1. The preferential dissolution of H2S gas intoformation water (compared with CO2 gas) results in thedelayed breakthrough of H2S gas, and the separationbetween CO2 and H2S gases at the moving front. InjectedCO2gas moves faster than H2S gas. More H2S is containedin the interior of the gas plume.

[31] 2. Coinjection of H2S reduces CO2 solubility incomparison with the CO2only case. However, the prefer-ential dissolution of H2S can enhance CO2dissolution at thegas moving front.

[32] 3. Coinjection of H2S with CO2 in the sandstone

formation causes the precipitation of pyrite through theinteractions between the dissolved H2S and Fe2+ from the

dissolution of Febearing minerals. Precipitation of pyritereduces ankerite precipitation in the coinjection case, thusthe mineral trapping of CO2.

[33] 4. In general, sandstone formations are more favor-able for CO2 mineral trapping than carbonate formations.The presence of Febearing siliciclastic and/or carbonateminerals in geological formations can significantly promote

Figure 11. Spatial distribution of mass fraction of (a and b)H2S and (c and d) CO2 in gas phase obtained from thecase of H2S with CO2 in carbonate formation after 100 and500 years.

Figure 12. Spatial distribution of concentration (unit is molkg1 H2O) of total dissolved sulfur obtained from the case ofH2S with CO2in carbonate formation after (a) 100 years and(b) 500 years.


9 of 13

7/25/2019 Zhang 2011

10/13

the H2S dissolution through precipitation of sulfide miner-als, increasing H2S mineral trapping.

[34] The range of problems concerning waterrockgasinteractions is very broad. The present simulation results andconclusions are specific to the conditions and parametersconsidered. The numerical experiments presented heregive a detailed view of the dynamical interplay betweencoupled hydrologic and chemical processes, albeit in anapproximate fashion. A critical evaluation of modelingresults can provide useful insight into the spatial and tem-poral evolution of injected CO2 and H2S, and associatedformation alteration in typical sandstone and carbonateformations, and better understand the behavior and effect ofimpurityH2S in the CO2 stream.

Appendix A: Hydrogeological Parameters andMineral and Water Chemical Compositions

[35] In the present paper, we used the same hydro-geological parameters (Table A1) as those used in theresearch in the Songliao Basin of China [Zha ng et al. ,

2009]. Two types of (initial) mineralogical compositions(Table A2) were used, sandstone from Zhang et al. [2009]and carbonate from Zerai et al. [2006]. The initial aque-ous solution compositions (Table A3) for different miner-alogical compositions were obtained by batch geochemicalmodeling.

Appendix B: Kinetic Rate Law for MineralDissolution and Precipitation

[36] A general kinetic rate law for mineral dissolutionand precipitation is used [Lasaga et al., 1994; Steefel and

Lasaga., 1994]:

rm k T mAm 1 Qm

Km

; B1

where m is the kinetic mineral index, rm is the dissolution/precipitation rate (positive values indicate dissolution, neg-ative values indicate precipitation), k(T)mis the rate constant

depending on the temperature (mol m2 s1),Tis the absolutetemperature,Am is the specific reactive surface area per kgwater, Km is the equilibrium constant for the mineralwaterreaction written for the destruction of one mole of mineral m,and Qm is the corresponding ion activity product. Theparameters and hare two positive numbers determined byexperiments; usually, but not always, they are taken to beequal to 1 (like in the present work).

[37] For many minerals the kinetic rate constantk(T) canbe summed from three mechanisms [Lasaga et al., 1994;Palandri and Kharaka, 2004]:

k T knu25expEnua

R

1

T

1

298:15

kH25expEHa

R

1

T

1

298:15

nHH

kOH25 expEOHa

R

1

T

1

298:15

nOHOH; B2

where superscripts and subscripts nu, H, and OH indicateneutral, acid and base mechanisms, respectively, Ea is theactivation energy,k25is the rate constant at 25C,Ris the gasconstant (8.31 J (mol K)1), and Tis the absolute tempera-ture, a is the activity of the species, and n is a power term(constant). Notice that parametersand h(see equation (B1))are assumed to be the same for each mechanism. For all

Figure 13. Spatial distribution of change of (a) dolomiteand (b) calcite obtained from the sensitivity simulation after500 years.

Table A1. Hydrogeological Parameters Used in the Simulations

Parameters Geological Formation

Porosity 0.30Horizontal permeability (m

2) 1.0 10

13

Vertical permeability (m2

) 0.5 1013

Pore compressibility (Pa1

) 4.5 1010

Diffusivity (m2

s1

) 1.0 109

Rock grain density (kg m3

) 2600Formation heat conductivity

(W (m C)1

)2.51

Rock grain specific heat (J (kg C)1) 920Temperature (C) 50Pressure (bar) 120

Relative Permeability Model: Liquid (Van Genuchten)krl=

ffiffiffiffiffiffiS*

p {1 (1 [S*]1/m)m}2 S* = (Sl Slr)/(1 Slr)

Slr, residual water saturation Slr= 0.30m, exponent m = 0.457

Relative Permeability Model: Gas (Corey)

krg= (1 S)2 (1 S2) S = (Sl Slr)/(Sl Slr Sgr)

Sgr, residual gas saturation Sgr= 0.05

Capillary Pressure Model (Van Genuchten)

Pcap = P0([S*]1/m

1)1m S* = (Sl Slr)/(1 Slr)

Slr, residual water saturation Slr= 0.00m, exponent m = 0.457

P0, strength coefficient (kPa) P0 = 19.61


10 of 13

7/25/2019 Zhang 2011

11/13

minerals it is assumed that the precipitation rate equals thedissolution rate.

Appendix C: Parameters for Calculating KineticRate of Minerals in the Modeling Studies

[38] Parameters for calculating kinetic rate of minerals aregiven in Table C1. Calcite and anhydrite were assumed toreact with aqueous species at local equilibrium because theirreaction rates are typically quite rapid. The dissolution andprecipitation of other minerals are kinetically controlled.Rate law parameters for kaolinite, illite, chlorite, albitelow,

oligoclase, K

feldspar, magnesite, and dolomite were takenfrom Palandri and Kharaka [2004], who compiled andfitted experimental data reported by many investigators. Thedetailed list of the original data sources is given by Palandriand Kharaka [2004]. Chalcedony kinetic data were referredto the work ofTester et al.[1994]. Illite kinetic data were setto those of smectite. Siderite kinetic data were from Steefel[2001]. Ankerite and dawsonite kinetic data were set tothose of siderite.

[39] Mineral reactivesurface areas (the second column ofTable C1) are based on the work ofSonnenthal et al.[2005]and were calculated assuming a cubic array of truncatedspheres constituting the rock framework. The larger surfaceareas for clay minerals (kaolinite, illite and smectite) are due

to smaller grain sizes. In conformity with White andPeterson [1990] and Zerai et al. [2006], a surface rough-ness factor of 10 is incorporated and defined as the ratio ofthe true (BET) surface area to the equivalent geometricsurface area. Interaction with the minerals is generallyexpected to occur only at selective sites of the mineralsurface, and the actual reactive surface area could bebetween 1 and 3 orders of magnitude less than the surfaceroughnessbased surface area [Lasaga, 1995; Zerai et al.,2006]. The difference is attributed to the fact that onlypart of the mineral surface is involved in the reaction owingto coating or armoring, a small area exposed to the brine,

and channeling of the reactive fluid flow. To account forthese effects, the actual reactive surface areas given inTable C1 are decreased by 2 orders of magnitude from thesurface roughnessbased surface areas. The reactive surfaceareas used here for most minerals are similar to those ofZerai et al. [2006], who used a surface area of 10 cm2 g1

for all minerals.[40] If the aqueous phase supersaturates with respect to a

potential secondary mineral, a small volume faction such as1 106 is used for calculating the seed surface area for thenew phase to grow. The precipitation of secondary mineralsis represented using the same kinetic expression as that for

dissolution. However, because precipitation rate data formost minerals are unavailable, parameters for neutral pHrates only, as given in Table C1, were employed to describeprecipitation. Multiple kinetic mechanisms for precipitation

Table A3. Initial Concentrations of the Formation Water at

Reservoir Conditions of 50C and 120 Barsa

Elements

Concentration (mol kg1

H2O)

Sandstone Formation,Songliao Basin, China

Carbonate Formation,Ohio

Al 0.4235 1009 0.9955 1016

Carbon 0.8150 1002 0.2741 1001

Ca 0.2977 1002

0.2636 1003

Cl 0.1710 10+00

0.1710 10+00

Iron 0.8915 1004

0.2612 1005

K 0.1979 1003

0.1000 1015

Mg 0.1141 1004

0.1297 1001

Na 0.1718 10+00

0.1710 10+00

Sulfur 0.1000 1015 0.9983 1016

Si 0.1800 1002 0.1000 1015

O2(aq) 0.1774 1066 0.1774 1066

pH 6.891 7.446

aIron is the sum of Fe2+, Fe3+, and their related complexes. Carbon is thesum of CO2(aq), CH4(aq), and their related species such as HCO3

andacetic acid(aq). Sulfur is the sum of sulfate and sulfide species. Redoxreactions are set using O2(aq).

Table A2. Initial Mineral Volume Fractions Introduced in the Model and Possible Secondary Mineral Phases

(With a Zero Initial Volume Fraction) Used in the Simulations

Minerals Chemical Composition

Volume Fraction

Sandstone Formation,Songliao Basin, China

Carbonate Formation,Ohio

Albitelow NaAlSi3O8 0.415 0Calcite CaCO3 0.030 0.390Chalcedony SiO2 0.258 0

Chlorite Mg2.5Fe2.5Al2Si3O10(OH)8 0.027 0Illite K 0.6Mg0.25Al1.8(Al0.5Si3.5O10)(OH)2 0.028 0Kaolinite Al2Si2O5(OH) 0.009 0Kfeldspar KAlSi3O8 0.233 0Alunite KAl3 (OH)6(SO4)2 0 0Anhydrite CaSO4 0 0Ankerite CaMg0.3Fe0.7(CO3)2 0 0Casmectite Ca 0.145Mg0.26Al1.77Si3.97O10(OH)2 0 0Dawsonite NaAlCO3(OH)2 0 0Dolomite CaMg(CO3)2 0 0.600Hematite Fe2O3 0 0Magnesite MgCO3 0 0

Nasmectite Na 0.290Mg0.26Al1.77Si3.97O10(OH)2 0 0Oligoclase Ca 0.2Na0.8Al1.2Si2.8O8 0 0Pyrite FeS2 0 0Siderite FeCO3 0 0.010


11 of 13

7/25/2019 Zhang 2011

12/13

can be specified in an input file of the TOUGHREACTprogram, should such information become available.

[41] Acknowledgments. The authors would like to thank two anon-ymous reviewers and the Associate Editor for their constructive commentsand suggestions during the review process, which greatly improved thequality of the paper. The first author (Wei Zhang) was supported by theChina Scholarship Council (CSC) and by the National Natural ScienceFoundation of China (NSFC, 40872158). The second author (Tianfu Xu)was supported by the Zero Emission Research and Technology project(ZERT) of the U.S. Department of Energy under contract DEAC02

05CH11231 with Lawrence Berkeley National Laboratory. The third author(Yilian Li) was supported by the National Natural Science Foundation ofChina (NSFC, 40672168 and 40872158) and the Global Climate andEnergy Project (GCEP, subcontract 238463843106A). The support pro-vided by Lawrence Berkeley National Laboratory during Wei Zhangs visitto Berkeley is also acknowledged.

ReferencesAllen, D. E., B. R. Strazisar, Y. Soong, and S. W. Hedges (2005), Model-

ing carbon dioxide sequestration in saline aquifers: Significance of ele-vated pressures and salinities,Fuel Process. Technol., 86, 15691580,doi:10.1016/j.fuproc.2005.01.004.

Bachu, S., and D. B. Bennion (2009), Chromatographic partitioning ofimpurities contained in a CO2stream injected into a deep saline aquifer:Part 1. Effect of gas composition and in situ conditions, Int. J. Greenh.Gas Control, 3, 458467, doi:10.1016/j.ijggc.2009.01.001.

Bachu, S., and W. D. Gunter (2004), Acidgas injection in the Albertabasin, Canada: A CO2storage experience, Geol. Soc. Spec. Publ., 233,225234, doi:10.1144/GSL.SP.2004.233.01.15.

Bachu, S., and W. D. Gunter (2005), Overview of acidgas injection opera-tions in Western Canada, in Proceedings of the 7th International Confer-ence on Greenhouse Gas Control Technologies, pp. 443448, Elsevier,Oxford, U. K.

Bachu, S., W. D. Gunter, and E. P. Perkins (1994), Aquifer disposal ofCO2: Hydrodynamic and mineral trapping, Energy Convers. Manage.,35, 269279, doi:10.1016/0196-8904(94)90060-4.

Bachu, S., J. M. Nordbotten, and M. A. Celia (2005), Evaluation ofthe spread of acidgas plumes injected in deep saline aquifers in westernCanada as an analogue for CO2 injection into continental sedimentary

basins, in Proceedings of the 7th International Conference on Green-house Gas Control Technologies, pp. 479487, Elsevier, Oxford, U. K.

Bachu, S., D. B. Bennion, M. PooladiDarvish, and H. Hong (2009a), Fac-tors affecting the chromatographic partitioning of CO2 and H2S injectedinto a watersaturated porous medium, Energy Procedia,1, 31653172,doi:10.1016/j.egypro.2009.02.099.

Bachu, S., M. PooladiDarvish, and H. Hong (2009b), Chromatographicpartitioning of impurities (H2S) contained in a CO2 stream injected intoa deep saline aquifer: Part 2. Effects of flow conditions, Int. J. Greenh.Gas Control, 3, 468473, doi:10.1016/j.ijggc.2009.01.002.

Battistelli, A. (2008), Modeling multiphase organic spills in coastal siteswithTMVOC V.2.0,Vadose Zone J.,7, 316324, doi:10.2136/vzj2006.0119.

Crandell, L. E., B. R. Ellis, and C. A. Peters (2010), Dissolution potential

of SO2 c o

injected with CO2 in geologic sequestration, Environ. Sci.Technol., 44, 349355, doi:10.1021/es902612m.

Ellis, B. R., L. E. Crandell, and C. A. Peters (2010), Limitations for brineacidification due to SO2 coinjection in geologic carbon sequestration,

Int. J. Greenh. Gas Control, 4, 575582, doi:10.1016/j.ijggc.2009.11.006.Gentzis, T. (2000), Subsurface sequestration of carbon dioxide: An over-

view from an Alberta (Canada) perspective, Int. J. Coal Geol., 43,287305, doi:10.1016/S0166-5162(99)00064-6.

Gough, C. (2008), State of the art in carbon dioxide capture and storage inthe UK: An experts review, Int. J. Greenh. Gas Control, 2, 155168,doi:10.1016/S1750-5836(07)00073-4.

Gunter, W. D., E. H. Perkins, and I. Hutcheon (2000), Aquifer disposal ofacid gases: Modeling of waterrock reactions for trapping of acid wastes,

Appl. Geochem., 15, 10851095, doi:10.1016/S0883-2927(99)00111-0.Gunter, W. D., A. Pratt, B. E. Buschkuehle, and E. H. Perkins (2005), Acid

gas injection in the Brazeau Nisku Q carbonate reservoir: Geochemicalreactions as a result of the injection of an H 2SCO2waste stream, inPro-ceedings of the 7th International Conference on Greenhouse Gas Control

Technologies, pp. 469

477, Elsevier, Oxford, U. K.Hitchon, B., W. D. Gunter, T. Gentzis, and R. T. Bailey (1999), Sedimen-tary basins and greenhouse gases: A serendipitous association, EnergyConvers. Manage., 40, 825843, doi:10.1016/S0196-8904(98)00146-0.

Holloway, S. (2001), Storage of fossil fuelderived carbon dioxide beneaththe surface of the Earth, Annu. Rev. Energy Environ. , 26, 145166,doi:10.1146/annurev.energy.26.1.145.

Holloway, S. (2005), Underground sequestration of carbon dioxide: A viablegreenhouse gas mitigation option,Energy,30, 23182333, doi:10.1016/j.energy.2003.10.023.

Izgec, O., B. Demiral, H. Bertin, and S. Akin (2008), CO2 injection intosaline carbonate aquifer formation I: Laboratory investigation, Transp.

Porous Media, 72, 124, doi:10.1007/s11242-007-9132-5.Knauss, K. G., J. W. Johnson, and C. I. Steefel (2005), Evaluation of

the impact of CO2, cocontaminant gas, aqueous fluid and reservoir rock

Table C1. List of Parameters for Calculating Kinetic Rate of Mineralsa

MineralsSurface

Area (cm2 g1)

Neutral Mechanism Parameters Acid Mechanism Parameters Base Mechanism Parameters

k25(mol m2 s1)

Ea(kJ mol1)

k25(mol m2 s1)

Ea(kJ mol1) n (H+)

k25(mol m2 s1)

Ea(kJ mol1) n (H+)

Calcite equilibrium Anhydrite equilibrium Chalcedony 9.8 1.2500e14 87.5Kaolinite 151.6 6.9183e14 22.2 4.8978e12 65.9 0.777 8.9125e18 17.9 0.472

Illite 151.6 1.6596e13 35.0 1.0471e11 23.6 0.34 3.0200e17 58.9 0.40Chlorite 9.8 3.0200e13 88.0 7.7624e12 88.0 0.5Albitelow 9.8 2.7542e13 69.8 6.9183e11 65.0 0.457 2.5119e16 71.0 0.572Oligoclase 9.8 1.4451e12 69.8 2.1380e10 65.0 0.457Kfeldspar 9.8 3.8905e13 38.0 8.7096e11 51.7 0.5 6.3096e22 94.1 0.823Magnesite 9.8 4.5709e10 23.5 4.1687e07 14.4 1.0Dolomite 9.8 2.9512e08 52.2 6.4565e04 36.1 0.5Siderite 9.8 1.2598e09 62.76 6.4565e04 36.1 0.5Dawsonite 9.8 1.2598e09 62.76 6.4565e04 36.1 0.5Ankerite 9.8 1.2598e09 62.76 6.4565e04 36.1 0.5

Nasmectite 151.6 1.6596e13 35.0 1.0471e11 23.6 0.34 3.0200e17 58.9 0.40Casmectite 151.6 1.6596e13 35.0 1.0471e11 23.6 0.34 3.0200e17 58.9 0.40Hematite 12.87 2.5119e15 66.2 4.0738e10 66.2 1.0Alunite 9.8 1.0000e12 57.78 1.0000e12 7.5 1.00Pyrite 12.87 k 25 = 2.8184e5,

Ea= 56.9,n(O2(aq)) = 0.6

k25 = 3.024e8,Ea= 56.9,

n(H+

) = 0.5,

n(Fe3+

) = 0.5aNote that (1) all rate constan ts are liste d for dissol ution; (2) k25 is kinetic constant at 25C, Ea is activation energy, and n is the power term

(equation (B2)); (3) the power terms n for both acid and base mechanisms are with respect to H+; (4) for pyrite, the neutral mechanism has n with

respect to O2(aq), the acid mechanism has two species involved: one n with respect to H+ and anothern with respect to Fe

3+.


12 of 13

7/25/2019 Zhang 2011

13/13

interactions on the geologic sequestration of CO2, Chem. Geol., 217,339350, doi:10.1016/j.chemgeo.2004.12.017.

Lasaga, A. C. (1995), Fundamental approaches in describing mineral disso-lution and precipitation rates, in Chemical Weathering Rates of Silicate

Minerals. Rev. Mineral, vol. 31, edited by A. F. White and S. L. Brantley,pp. 2386, Mineral. Soc. of Am., Washington, D. C.

Lasaga, A. C., J. M. Soler, J. Ganor, T. E. Burch, and K. L. Nagy (1994),Chemical weathering rate laws and global geochemical cycles, Geochim.Cosmochim. Acta, 58, 23612386, doi:10.1016/0016-7037(94)90016-7.

Machel, H. G. (2005), Geological and hydrogeological evaluation of the

Nisku Q

Pool in Alberta, Canada, for H2S and/or CO2 storage, OilGas Sci. Technol., 60, 5165, doi:10.2516/ogst:2005005.Narasimhan, T. N., and P. A. Witherspoon (1976), An integrated finite dif-

ference method for analyzing fluid flow in porous media,Water Resour.Res., 12, 5764, doi:10.1029/WR012i001p00057.

Palandri, J. L., and Y. K. Kharaka (2004), A compilation of rate parametersof watermineral interaction kinetics for application to geochemical mod-eling,U. S. Geol. Surv. Open File Rep., 20041068, 64 pp.

Pruess, K., and A. Battistelli (2002), TMVOC: A numerical simulator forthreephase nonisothermal flows of multicomponent hydrocarbon mix-tures in saturatedunsaturated heterogeneous media, Rep. LBNL49375,Lawrence Berkeley Natl. Lab, Berkeley, Calif.

Pruess, K., C. Oldenburg, and G. Moridis (1999), TOUGH2 Users Guide,V2,Rep. LBNL43134, Lawrence Berkeley Natl. Lab, Berkeley, Calif.

Pruess, K., J. Garca, T. Kovscek, C. Oldenburg, J. Rutqvist, C. Steefel, andT. Xu (2004), Code intercomparison builds confidence in numerical sim-ulation models for geologic disposal of CO2, Energy, 29 , 14311444,doi:10.1016/j.energy.2004.03.077.

Sonnenthal, E., A. Ito, N. Spycher, M. Yui, J. Apps, Y. Sugita, M. Conrad,and S. Kawakami (2005), Approaches to modeling coupled thermal, hydro-logical, and chemical processes in the drift scale heater test at Yucca Moun-tain, Int. J. Rock Mech. Min. Sci., 42, 698719, doi:10.1016/j.ijrmms.2005.03.009.

Soong, Y., A. L. Goodman, J. R. McCarthyJones, and J. P. Baltrus (2004),Experimental and simulation studieson mineraltrapping of CO2with brine,

Energy Conver s. Manage ., 45 , 18451859, doi:10.1016/j.enconman.2003.09.029.

Steefel, C. I. (2001), CRUNCH: Software for modeling multicomponent,multidimensional reactive transport, Users Guide, Rep. UCRLMA143182, Lawrence Livermore Natl. Lab., Livermore, Calif.

Steefel, C. I., and A. C. Lasaga (1994), A coupled model for transport ofmultiple chemical species and kinetic precipitation/dissolution reactionswith application to reactive flow in single phase hydrothermal systems,

Am. J. Sci., 294, 529592, doi:10.2475/ajs.294.5.529.

Tester, J. W., G. W. Worley, B. A. Robinson, C. O. Grigsby, and J. L.Feerer (1994), Correlating quartz dissolution kinetics in pure water from25 to 625C, Geochim. Cosmochim. Acta, 58, 24072420, doi:10.1016/0016-7037(94)90020-5.

West, J. M., J. Pearce, M. Bentham, and P. Maul (2005), Issue profile:Environmental issues and the geological storage of CO2, Eur. Environ.,15, 250259, doi:10.1002/eet.388.

White, A. F., and M. L. Peterson (1990), Role of reactive surface area char-acterization in geochemical models, in Chemical Modeling of AqueousSystems II, edited by D. C. Melchior and R. L. Bassett, ACS Symp.

Ser., 416, 461

475, doi:10.1021/bk-1990-0416.ch035.Xiao, Y., T. Xu, and K. Pruess (2009), The effects of gas fluidrock inter-actions on CO2 injection and storage: Insights from reactive transportmodeling, Energy Procedia, 1 , 17831790, doi:10.1016/j.egypro.2009.01.233.

Xu, T., and K. Pruess (2001), Modeling multiphase nonisothermal fluidflow and reactive geochemical transport in variably saturated fracturedrocks: 1. Methodology,Am. J. Sci.,301, 1633, doi:10.2475/ajs.301.1.16.

Xu, T., E. Sonnenthal, N. Spycher, and K. Pruess (2006), TOUGHREACT:A simulation program for nonisothermal multiphase reactive geochem-ical transport in variably saturated geologic media: Applications to geo-thermal injectivity and CO2 geological sequestration, Comput. Geosci.,32, 145165, doi:10.1016/j.cageo.2005.06.014.

Xu, T., J. A. Apps, K. Pruess, and H. Yamamoto (2007), Numerical mod-elingof injectionand mineral trapping of CO2 with H2S and SO2 ina sand-stone formation,Chem. Geol., 242 , 319346, doi:10.1016/j.chemgeo.2007.03.022.

Yeh, G. T., and V. S. Tripathi (1991), A model for simulating transport of

reactive multispecies components: Model development and demonstra-tion,Water Resour. Res., 27, 30753094, doi:10.1029/91WR02028.Zerai, B., B. Z. Saylor, and G. Matisoff (2006), Computer simulation of CO2

trapped through mineral precipitation in the Rose Run Sandstone, Ohio,Appl. Geochem., 21, 223240, doi:10.1016/j.apgeochem.2005.11.002.

Zhang, W., Y. Li, T. Xu, H. Cheng, Y. Zheng, and P. Xiong (2009), Longterm variations of CO2 trapped in different mechanisms in deep salineformations: A case study of the Songliao Basin, China, Int. J. Greenh.Gas Control, 3, 161180, doi:10.1016/j.ijggc.2008.07.007.

Y. Li and W. Zhang, School of Environmental Studies, China Universityof Geosciences, Wuhan, Hubei 430074, China. ([email protected];[email protected])

T. Xu, Earth Sciences Division, Lawrence Berkeley National Laboratory,University of California, Mail Stop 901116, 1 Cyclotron Rd., Berkeley,CA 94720, USA. ([email protected])


13 of 13

zhang 2011

Documents