Advances in Water Resources 77 (2015) 44–56

Contents lists available at ScienceDirect

Advances in Water Resources

journal homepage: www.elsevier .com/ locate/advwatres

Co-sequestration of SO2 with supercritical CO2 in carbonates: Anexperimental study of capillary trapping, relative permeability, andcapillary pressure

http://dx.doi.org/10.1016/j.advwatres.2014.08.0110309-1708/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.

Morteza Akbarabadi ⇑, Mohammad PiriDepartment of Chemical and Petroleum Engineering, University of Wyoming, Laramie, WY 82071-2000, USA

a r t i c l e i n f o

Article history:Received 22 October 2013Received in revised form 14 August 2014Accepted 18 August 2014Available online 27 August 2014

Keywords:CO2 + SO2 co-sequestrationMadison limestoneCapillary trappingRelative permeabilityCapillary pressure hysteresisDynamic effect

a b s t r a c t

In this study we performed three categories of steady- and unsteady-state core-flooding experiments toinvestigate capillary trapping, relative permeability, and capillary pressure, in a scCO2 + SO2/brine/lime-stone system at elevated temperature and pressure conditions, i.e., 60 �C and 19.16 MPa. We used a Mad-ison limestone core sample acquired from the Rock Springs Uplift in southwest Wyoming. We carried outtwo sets of steady-state drainage-imbibition relative permeability experiments with different initialbrine saturations to study hysteresis. We found that the final scCO2 + SO2 drainage relative permeabilitywas very low, i.e., 0.04. We also observed a rapid reduction in the scCO2-rich phase imbibition relativepermeability curve, which resulted in a high residual trapping. The results showed that between 62.8%and more than 76% of the initial scCO2 + SO2 at the end of drainage was trapped by capillary trappingmechanism (trapping efficiency). We found that at higher initial brine saturations, the trapping efficiencywas higher. The maximum initial and residual scCO2-rich phase saturations at the end of primary drain-age and imbibition were 0.525 and 0.329, respectively. Each drainage-imbibition cycle was followed by adissolution process to re-establish Sw = 1. The dissolution brine relative permeabilities for both cycleswere also obtained. We characterized the scCO2 + SO2/brine capillary pressure hysteresis behaviorthrough unsteady-state primary drainage, imbibition, and secondary drainage experiments. We observednegative imbibition capillary pressure curve indicative of possible wettability alteration throughout theexperiments due to contact with scCO2 + SO2/brine fluid system. The trapping results were compared tothose reported in literature for other carbonate core samples. We noticed slightly more residual trappingin our sample, which might be attributed to heterogeneity, different viscosity ratio, and pore-space topol-ogies. The impact of dynamic effects, i.e., high brine flow rate imbibition tests, on trapping of the scCO2-rich phase was also explored. We performed two imbibition experiments with relatively high brine flowrates. The residual scCO2 saturation dropped to 0.291 and 0.262 at the end of the first and second imbi-bition tests, i.e., 11.5% and 20.4%, respectively, compared to 0.329 under capillary-dominated regime.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

In 2012, worldwide energy-related CO2 emissions increased by1.4% in comparison with the previous year and reached the highestlevel of 31.6 Gt [1]. Power plants are the major producer of CO2, i.e.,more than 78% of the total CO2 emissions [2,3]. It is also reportedthat the consumption of coal, which is responsible for 40% of thetotal CO2 emissions globally, has increased by 5% over the past dec-ade while for natural gas and oil, these numbers were 2% and 1%,respectively [3]. In order to mitigate the CO2 emissions, the mostviable option is carbon capture and storage (CCS) [2].

The goal of CO2 capture is to produce a high-concentrationstream of CO2 ready to be transported or to be stored [2,4].Captured CO2 from various sources, e.g., refineries and steelindustry, has been used for various purposes including EnhancedOil Recovery (EOR). It can also be stored in different geologicformations [2,5,6].

Geologic mitigation occurs through CO2 injection into three dif-ferent types of formations: (1) unmineable coal seams, (2) maturehydrocarbon reservoirs, and (3) deep saline aquifers, among whichthe latter is known to have the highest total storage capacity [2,5]and comprises about 90% of the global CO2 storage space [7].

In the mitigation process, CO2 captured from commercial orindustrial operations is injected into a saline aquifer that providesa medium in which CO2 would be at supercritical state [8,9], which

M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56 45

means more mass of scCO2 can be stored into a given volume of theformation.

It is reported that the risk of CO2 leakage from deep saline for-mations through percolation of CO2 toward the surface is fairly lowas this upward movement might take thousands or even millionsof years to occur [10]. This is attributed to the storage of CO2 inthe pores of the hosting rocks with various trapping mechanisms.There are four different mechanisms to sequester CO2 in theabove-mentioned geologic formations: (1) structural and strati-graphic trapping, (2) capillary or residual trapping, (3) solubilitytrapping, and (4) mineral trapping [2]. Studies have shown thatcapillary trapping of CO2 is a very important short-term storagemechanism with a considerably low risk of leakage [11–15]. Thismechanism also contributes to the effectiveness of the solubilityand mineral trapping during later stages of the sequestration pro-cess [13]. Hence, it is critical to investigate trapping and relativepermeability and capillary pressure hystereses that control thismechanism in CO2/brine systems.

In most cases, the CO2 sources also emit impurities [16,17].Most North America coal-fired power plants emit large quantitiesof sulfur dioxide (SO2) and nitrogen oxides (NOx) along with CO2

[17–19]. Eliminating these by-products from the flue gas streamis expensive at commercial scale. Currently some power plantsare employing coal gasification methods to remove sulfur contentsfrom their flue gas emissions; and as a result, they are creatinglarge amounts of elemental sulfur [4]. Co-sequestration of SO2 withCO2 might be a feasible substitute for removal procedures [2,20].The Weyburn EOR project in Canada, which stores CO2 along withH2S, another acid gas, captured from the Great Plains SynfuelsPlant in North Dakota, is the only plant that uses this technique[2,4].

Even though sequestering SO2 and NOx along with CO2 mayhave economical advantages [21], the impact of these co-contam-inants on the trapping of CO2 in deep saline aquifers, and on thegeologic formations is widely undiscovered [22]. There have been,however, some studies on the CO2–SO2–rock interactions to under-stand the impact of highly-acidic brine, resulting from the dissolu-tion of CO2 and SO2 in aquifers, on different caprocks and on thereservoir integrity [16,22–34]. The studies show that scCO2 andSO2 react significantly with the minerals of the rock and alter thepore and matrix structures and also the wettability of the caprocks;whereupon they affect the maximum CO2 storage pressure andcapacity. For instance, SO2 dissolution in brine drastically lowersthe pH of the solution, which in turn makes the surface of the rockless water wet [27,30,35]. This leads to less trapping of thenon-wetting phase, e.g., scCO2, as it can not be trapped easily inthe larger pores [36] due to better hydraulic conductivity of thenon-wetting phase.

Some studies have shown that acidification of formation brinein the presence of SO2 might take place due to hydrolysis, oxida-tion, and SO2 disproportionation into the brine, among which thelatter is suggested to be the predominant reaction [34,28]. How-ever, it is predicted that diffusion of SO2 through the scCO2-richphase plume leads to the retention of about 73% to 90% of SO2,which in turn limits the sever acidification of the formation brinedue to SO2 disproportionation [20,21]. Therefore, its impacts onthe caprock and reservoir integrity will be minimized. Further-more, some studies indicate that the acidified brine might dissolvethe walls of the pore space and widen the flow paths, therebyaltering the pore-space topology and porosity and increasing thepermeability of the rock sample [37,31]. This causes less trappingof the non-wetting phase as its hydraulic conductivity is improved.However, it is also shown that in a fractured carbonate rocksample, the dissolution of minerals might lead to a decrease inthe permeability due to different mechanisms such as blockageof the flow path by the mobilized grains and closure of the fracture

aperture [34]. As it is evident, injection of acidic brine may havedifferent impacts on the formation properties and may generatecomplexity during core-flooding experiments. Hence, we eliminatethe effect of mineral dissolution on the trapping of the scCO2-richphase.

In this study, we investigate capillary trapping of scCO2 in thepresence of SO2 as an impurity. To simplify the experiments, weexclude the effect of mineral trapping by equilibrating the aqueousand scCO2-rich phases with the minerals of the rock sample. Weexamine relative permeability and capillary pressure of the scCO2/brine system in the presence of SO2. In addition to these properties,we study the impact of dynamic effects on trapping of scCO2 + SO2

mixture using high flow rate imbibition tests. Even though therehave been a number of studies presenting findings on scCO2/brinerelative permeability and capillary pressure in sandstones[35,36,38–53] and in carbonates [42,43,54–56], to our knowledge,no experiments have been performed with SO2 + scCO2 mixture ina carbonate core sample at elevated pressure and temperature con-ditions to investigate capillary trapping and drainage-imbibitionrelative permeability and capillary pressure hystereses. Okabeand Tsuchiya [54] carried out one set of unsteady-state (uss) drain-age-imbibition scCO2/brine experiment on a carbonate core sampleobtained from a reservoir in the Middle East to measure the initialand residual scCO2 saturations using an in-situ saturation mea-surement technique. Bennion and Bachu [55] performed series ofunsteady-state drainage-imbibition scCO2/brine core-floodingexperiments at various pressure and temperature conditions ondifferent carbonate core samples. They measured the initial andresidual scCO2 saturations using mass balance and calculated therelative permeabilities through history matching. El-Maghrabyand Blunt [56] reported the results of series of unsteady-statescCO2/brine drainage-imbibition experiments performed on anIndiana limestone core. The fluids in this work, unlike previousstudies, were equilibrated with the minerals of the Indiana lime-stone core sample. The trapping results were obtained using anisothermal expansion technique. The authors also reported thedrainage capillary pressure versus saturation curves for the Indianalimestone under gaseous and supercritical conditions. Note that inall of these studies, the experiments were carried out on horizon-tally-placed core samples. Table 1 lists these experimental studieswith their experimental conditions and procedures.

In this work, a closed-loop, reservoir-conditions multi-phaseflow core-flooding apparatus was used to perform flow tests withaqueous and scCO2-rich phases. The fluids were equilibrated withthe minerals of the carbonate rock sample to minimize the dissolu-tion of the rock matrix into the highly-acidic brine and also theimpact of mineral dissolution on the trapping of scCO2-rich phase.A medical CT scanner was utilized to determine the in-situ satura-tions in the rock sample during the flow tests. All the experimentswere carried out at 60 �C temperature and 19.16 MPa backpressure.

In this paper, first, we provide detailed information on the rocksample, fluids, and experimental conditions, apparatus, and proce-dures used in this study. Next, we present the results of the trap-ping, relative permeability, and capillary pressure experiments.This is then followed by the study of the impact of brine flow rateon trapping of the scCO2-rich phase. We conclude the paper with aset of final remarks.

2. Experiments

In this work, we performed three categories of experiments. Incategory A, we investigated relative permeability hysteresis ofscCO2-rich/brine fluid system in a Madison limestone sample inthe presence of SO2, as a co-contaminant, using the steady-statemethod. In category B, we used an unsteady-state technique to

Tabl

e1

Expe

rim

enta

lst

udie

sw

ith

CO2/b

rine

flow

inca

rbon

ate

rock

sam

ples

;ss

:st

eady

-sta

te,u

ss:

unst

eady

stat

e,Ca

t.:Ca

tego

ry.

Item

Ref

eren

ceFl

uid

sEx

peri

men

tal

con

diti

ons

Met

hod

Satu

rati

onm

easu

rem

ent

tech

niq

ue

Roc

ksa

mpl

esM

easu

red

para

met

ers

Proc

ess

Pre

rock

-m

iner

aleq

uil

ibri

um

1O

kabe

and

Tsu

chiy

a[5

4]sc

CO

2,B

rin

e(1

4w

t%N

aI)

40�C

,9.8

MPa

uss

In-s

itu

mea

sure

men

tC

arbo

nat

efr

omM

iddl

eEa

stIn

itia

lan

dre

sidu

alsa

t.D

rain

age

imbi

biti

onN

o

2B

enn

ion

and

Bac

hu[4

3,55

]sc

CO

2,B

rin

e(1

03,7

01–

320,

847)

(ppm

)

36–5

6�C

8.73

–18.

8M

Pau

ssM

ass

bala

nce

Red

wat

erLe

duc

and

Wab

amu

nli

mes

ton

eC

ooki

ng

lake

,Nis

ku,

Mor

invi

lle

Ledu

c,an

dG

rosm

ont

dolo

mit

e

Init

ial

and

resi

dual

sat.

Rel

ativ

epe

rm.

Dra

inag

eim

bibi

tion

No

3El

-Mag

hra

byan

dB

lun

t[5

6]sc

CO

2,g

CO

2

Bri

ne

50�C

,4.

2,9.

0M

Pau

ssIs

oth

erm

alex

pan

sion

Indi

ana

lim

esto

ne

Init

ial

and

resi

dual

sat.

Cap

illa

rypr

es.

Dra

inag

eim

bibi

tion

Yes

4Th

isst

udy

scC

O2

+SO

2

Bri

ne(

NaI

,NaC

l)(8

6,50

0pp

m)

60�C

,19

.16

MPa

(Cat

.A)

ssu

ssIn

-sit

um

easu

rem

ent

Mad

ison

lim

esto

ne

Res

idu

alsa

t.R

elat

ive

perm

.D

rain

age

imbi

biti

ondi

ssol

uti

onY

es

5Th

isst

udy

scC

O2

+SO

2

Bri

ne(

NaI

,NaC

l)(8

6,50

0pp

m)

60�C

,19

.16

MPa

(Cat

.B&

C)

uss

In-s

itu

mea

sure

men

tM

adis

onli

mes

ton

eR

esid

ual

sat.

Cap

illa

rypr

es.

Dra

inag

eim

bibi

tion

Yes

Table 2Experiments performed in this study; RP: relative permeability.

Item Experiment No. –Category No.

Process Experiment Technique

1 Exp. 1-A Drainage 1st RP cycle ss2 Exp. 2-A Imbibition 1st RP cycle ss3 Exp. 3-A Dissolution 1st RP cycle uss4 Exp. 4-A Drainage 2nd RP cycle ss5 Exp. 5-A Imbibition 2nd RP cycle ss6 Exp. 6-A Dissolution 2nd RP cycle uss7 Exp. 7-B Drainage Capillary pressure uss8 Exp. 8-B Imbibition Capillary pressure uss9 Exp. 9-B Drainage Capillary pressure uss

10 Exp. 10-C Imbibition Dynamic effects uss11 Exp. 11-C Drainage Dynamic effects uss12 Exp. 12-C Imbibition Dynamic effects uss

46 M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56

characterize the scCO2 + SO2/brine capillary pressure. In categoriesA and B, we also examined trapping of the scCO2 + SO2 mixturewith variations in initial brine saturation (Swi). In category C, weexplored the impact of high brine flow rate displacement regimeon trapping of the scCO2 + SO2 mixture. Table 2 lists all the exper-iments performed under this study. The table also includes thetechnique used to carry out each experiment.

In the remainder of this section, we include information aboutthe rock sample, fluids, and experimental conditions, apparatus,and procedure.

2.1. Rock sample, fluids, and experimental conditions

The experiments were performed on a Madison limestone rocksample acquired from the Rock Springs Uplift in southwest Wyo-ming. This formation is currently being targeted as a potential sitefor storage of large quantities of CO2. The layers of interest in thisreservoir are Pennsylvanian Weber sandstone (equivalent to Ten-sleep sandstone) and Mississippian Madison limestone [57]. Thecore sample used in this study was 3.77 cm in diameter and17.4 cm long. The porosity of the sample was measured 19.03%using an X-ray CT scanner. The other petrophysical properties ofthe sample are summarized in Table 3. The core was heteroge-neous, and variation of porosity along the length of the samplewas considerable (see Fig. 1). Before performing the main experi-ments listed in Table 2, a sacrificial core was used to equilibratethe fluids with the minerals of the rock. The details of this stepare presented later in this paper. The main and sacrificial coreswere cut from the same Madison limestone sample.

The main core was wrapped with a layer of Teflon shrink tubeand several layers of Teflon tape. The sleeve used in this workwas manufactured from a highly impermeable AFLAS material spe-cifically manufactured for this study. The sleeve was wrapped withseveral layers of Teflon tape and Aluminum foil to provide furtherassurance against possible penetration of the scCO2 + SO2 mixture.The experiments were performed at 60 �C and 19.16 MPa. At theseexperimental conditions, CO2 and SO2 were at the supercritical [9]and liquid states [58], respectively. We used a brine with36,500 ppm NaCl and 50,000 ppm NaI salt concentrations as wellas highly pure CO2 and SO2 gases for the flow tests.

We introduced 1200 ppm of SO2 into the scCO2-rich phase inour experiments. This value represents the amount of SO2 in theflue gas stream from the Jim Bridger coal-fired power plant inWyoming [17] and lies within the range of the SO2 concentrationin mixtures that can be found in CCS processes [21]. The solubilityof SO2 in brine is extremely high [59], and a thermodynamic flashcalculation indicated that we had formed a two-phase system [60],i.e., (1) a scCO2-rich phase, including CO2, SO2, and water vapor and(2) an aqueous phase, containing water, salts, CO2, and SO2.

Table 3Dimensions and petrophysical properties of the Madison limestone core sample used in this study.

Sample Length (cm) Diameter (cm) Avg. porosity (%) (X-ray) Kabs,brine (mD) Pore volume (cm3)

Madison limestone 17.4 3.77 19.03 13.6 36.96

0 20 40 60 80 100 120 140 160 18012

14

16

18

20

22

24

26

28

Poro

sity

(%)

Distance from outlet (mm)

Fig. 1. Porosity distribution of the Madison limestone core sample measured usingan X-ray CT scanner.

Table 4Properties of the fluid phases used in this study at 60 �C and 19.16 MPa [9,60–66].

Fluids Density q(kg/m3)

Viscosity l(m Pa s (cP))

scCO2/Brine IFT(mN/m)

scCO2-rich phase 0.70 0.0692 28.1Aqueous phase 1.05 0.635 –scCO2 0.70 0.0582 29.8Brine 1.05 0.597 –

M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56 47

Throughout the text we will refer to these phases as scCO2-rich andaqueous phases, respectively.

Table 4 shows the estimated physical properties of the fluidphases used in this work at experimental conditions withand without impurity. The values were obtained from literatureand also thermodynamic flash calculations [60].

2.2. Core-flooding apparatus

The experiments were carried out using a reservoir-conditionscore-flooding system that includes a ten cylinders Hastelloy Quizixpumping system, a medical CT scanner (Universal HD-350E-V)modified for petrophysical applications, a Hassler-type Aluminumcore holder, 3500 cm3 three-phase separator, two 2000 cm3

Hastelloy accumulators, and a pressure array system with Rose-mount differential pressure transducers. To maintain a uniformtemperature throughout the setup, three mechanical convectionovens were used to contain the pumps, the separator, and the accu-mulators. Several thermocouples were inserted at different loca-tions of the apparatus, such as the inlet and the outlet of thecore holder, to monitor the temperature. The wetted parts of theset up are made of Hastelloy to minimize corrosion. We also usedspecial AFLAS seal kits with the equipment to leakproof the systemat high pressure and temperature with the fluids mentionedearlier. Fig. 2 presents a schematic diagram of the core-floodingsystem [67].

To prepare for the experiments, the apparatus was thoroughlycleaned and vacuumed for several hours to remove any trappedair in the lines. The separator and accumulators were then filledwith brine using pump 3 (P3), see Fig. 2. We added 1200 ppmSO2 to the separator and accumulators using the brine pump(P1). After saturating the setup with the appropriate fluids, the sys-tem was pressurized by injecting CO2 using pump 2 (P2). At thesame time, the ovens were turned on to heat the equipment to60 �C. This also helped with pressurization of the system.

After establishing the pre-specified temperature and pressure,the fluids were re-circulated (by-passing the core sample) toachieve thermodynamic equilibrium. P1 and P2 were retractingbrine and scCO2 + SO2 mixture from the separator, respectively,and injecting them at constant flow rate to point P (see Fig. 2) atwhich two phases were mixed and retracted by the back pressureregulation (BPR) pump (P3). We used the Constant PressureReceive mode in P3 to retract effluent mixture from the core holderat a constant back pressure. This mode allowed us to establish avery stable boundary condition, which was essential for the exper-iments as the fluids were highly miscible. Variations in the backpressure could introduce inaccuracies into the experimentalresults (e.g., in-situ saturations).

It is important to note that all the experiments listed in Table 2were carried out in a closed-loop system in which we maintainedthe equilibrium between the phases and minimized the amount ofmass transfer between them in the core. A dual cylinder 5000Quizix pump was utilized to adjust the overburden pressure onthe core at 21.37 MPa. Additional details about the core-floodingsetup can be found elsewhere [51].

2.3. Experimental procedure

As it is listed in Table 2, steady- and unsteady-state techniqueswere used to carry out different series of experiments. In this sec-tion, first, we explain the preparation steps taken prior to the maintests. We then describe the procedures used to perform three dif-ferent categories of flow experiments.

Before initiating the main experiments, a sacrificial core wasprepared to equilibrate the fluids with the minerals of the lime-stone core. The fluids (i.e., scCO2 + SO2 mixture and brine) wereinjected into the sacrificial core and re-circulated throughout thesystem to achieve the thermodynamic equilibrium. This stagewas necessary as the interactions between the minerals of the car-bonate rock and the acidic fluids prevented matrix dissolution dur-ing the main core-flooding experiments. Therefore, the core used inthe main experiments remained intact.

The main core was placed in the core holder and flushed withdry CO2 to displace trapped air in the pore space and lines. Thesample was then vacuumed for several hours and pressurizedusing P2 by injecting pure CO2. At this stage, when temperatureand pressure were 60 �C and 19.16 MPa, respectively, the outletof the core was connected to the BPR pump (P3). We then started,using P2, to inject water-saturated scCO2 into the core while P3was retracting the effluent mixture from the core holder. Afterinjecting about 15 pore volumes of scCO2-rich phase, the corewas imaged to obtain the reference scan (CTgc) for saturation calcu-lations. Thereafter, the core was flushed with fresh brine.

A 1000 cm3 of fresh brine was prepared and equilibrated withthe minerals of the crushed Madison limestone rock in a separate

Fig. 2. Schematic flow diagram of the experimental setup used in this work. The apparatus includes a medical CT scanner and a Vertical Positioning System, ten Quizix Hastelloy cylinders, a three-phase separator, an AluminumHassler core holder, differential pressure transducers, two compensation accumulators, and several temperature measurement devices [67].

48M

.Akbarabadi,M

.Piri/Advances

inW

aterR

esources77

(2015)44–

56

Table 5Aqueous and scCO2-rich phases flow rates (cm3/min) and the fractional flows usedduring steady-state experiments (1, 2, 4, and 5).

Item Drainage Imbibition

Aqueous phase flow rate 0.03–0.2 0.03–0.16scCO2-rich phase flow rate 0–0.1 0–0.08Fractional flow 0.05–3.33 0.06–2.67

0 20 40 60 80 100 120 140 160 1800.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

S CO

2 (C

O2 s

atur

atio

n)

Distance from outlet (mm)

SCO2i-Exp.1 SCO2r-Exp.2

SCO2i-Exp.4 SCO2r-Exp.5

Fig. 3. scCO2 saturation profiles along the length of the Madison limestone coresample at the end of experiments 1, 2, 4, and 5.

M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56 49

cell placed in an oven maintained at 60 �C and 19.16 MPa. Thefresh brine was re-circulated through the crushed minerals at thepressure and temperature of the experiments. It was then injectedinto the core using a 6000 Quizix cylinder. More than 20 pore vol-umes of the brine was injected to displace and dissolve all thescCO2 + SO2 mixture to establish Sw = 1; the core was then scanned.The main goal of this step was to obtain a reference scan (CTfb) forcomparison in later stages of the study. The core was then sub-jected to injection of scCO2-saturated brine using P1. After injec-tion of about 20 pore volumes of the brine, the core was imagedto find another reference scan (CTwc). CTwc was identical to CTfb.At this point, the first drainage-imbibition cycle was started.

2.3.1. Porosity, permeability, and in-situ saturation measurementsThe resolution per slice of the scanner was 250 lm. Throughout

the experiments, the core was scanned with 4 mm slice thicknessand 2 mm spacing, producing 44 slices. The porosity was calcu-lated using the CT numbers obtained from the scanner:

/ ¼ CT�wc � CT�gc

� �.ðCTw � CTgÞ ð1Þ

where CT�wc and CT�gc are the CT numbers of the core fully saturatedwith fresh brine and CO2 at ambient temperature and 0.14 MPapressure, respectively, whereas CTw and CTg were obtained by scan-ning the core holder while it was filled with brine and CO2,respectively.

The permeability of the sample was calculated using the rela-tionship between flow rate and pressure drop in Darcy’s law. Tomeasure the absolute permeability, when the core was saturatedwith scCO2-saturated brine, we recorded steady-state pressuredrops at several brine flow rates. The permeability was obtainedby calculating the slope of the line generated by plotting the pres-sure drop data versus the flow rate values.

To calculate two in-situ fluid phase saturations, the core wasscanned at one energy level (130 kV, 100 mA) during the experi-ments, and also when it was fully saturated with each of the fluidsto obtain the reference scans. The saturations were determinedusing:

Sw ¼ ðCTc � CTgcÞ=ðCTwc � CTgcÞ ð2Þ

Sg ¼ 1� Sw ð3Þ

where CTc is the CT number of the core containing two fluid phasesduring the experiments, and CTwc and CTgc are the CT numbers ofthe core fully saturated with aqueous and scCO2-rich phases atthe temperature and pressure of the experiments, respectively.

2.3.2. Relative permeability, capillary pressure, and dynamic-effectsexperiments

To perform the category A of the experiments, fully brine-satu-rated core was flooded with both fluid phases simultaneously to per-form a test using the steady-state technique. In this approach, theratio of scCO2-rich and aqueous phase flow rates (QCO2/Qw) wasmonotonically increased and decreased during drainage and imbibi-tion processes, respectively. At each fractional flow, the saturationdistribution was carefully monitored and pressure difference alongthe length of the core was recorded. When the difference between

two consecutive measurements of the saturation distribution andpressure difference were less than 1%, the step was considered atsteady state. The core was scanned, and the flow rates were changed.Meanwhile, we used capillary number at each step to select the flowrates in such a way that the process was always under the capillary-dominated displacement regime (i.e., Nc 6 10�5). Nc was calculatedusing:

Nc ¼krw � Kabs � DP

/ � rgw � Lð4Þ

where krw, Kabs, /; rgw, DP, and L are brine relative permeability,absolute permeability, porosity, interfacial tension between scCO2

and brine, pressure drop, and length of the core, respectively. Theflow rates of each phase and the fractional flows used in this studyduring steady-state experiments are listed in Table 5.

After establishing the maximum scCO2 saturation(Smax

CO2 = 1 � Swi) at the end of drainage test, the imbibition experi-ment was started. It was performed analogously to the drainagetest with this difference that the QCO2/Qw was monotonicallydecreased during the experiment until QCO2 became zero. At thispoint, small increases in Qw did not affect the saturations. Aftercompleting the imbibition and determining the residual scCO2 sat-uration (SCO2r), the Qw was increased gradually to dissolve all thetrapped scCO2 + SO2, and re-establish Sw = 1. We then started anew drainage-imbibition cycle with shorter saturation span. Afterthe second cycle, the dissolution process was carried out again.As a result of the thermodynamic equilibrium established betweenthe phases, the dissolution was a time-consuming process. Eachdissolution run took about 7 days to be completed. About 250 porevolumes of brine were injected to re-create a fully brine-saturatedcore. After the second steady-state drainage-imbibition cycle, thecapillary pressure measurement experiments (category B) wereperformed. This category included primary drainage, imbibition,and secondary drainage tests using the unsteady-state approach.

In category B, to accomplish primary drainage, scCO2-rich phasewas injected at a low flow rate, i.e., 0.025 cm3/min into a fullybrine-saturated core. After reaching the steady state, the brine sat-uration was measured. The pressure difference was also recordedthroughout the whole process. This information was then pro-cessed to obtain capillary pressure as we will explain later in thispaper. For the next point, the scCO2 flow rate was increased, andsimilar procedure was repeated until upon increase in the scCO2

flow rate, no significant decrease in brine saturation was observed.At this point, we had constructed a full drainage capillary pressurecurve.

50 M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56

To establish an imbibition capillary pressure curve, one can usetwo different methods: (1) spontaneous imbibition test in whichthe capillary pressure is decreased gradually to zero allowing thewetting phase to imbibe and (2) forced imbibition test that canbe performed by decreasing the capillary pressure to a potentiallylarge negative number [68]. In this study, we used the latter wherethe imbibition process was started just after the primary drainagestep with a very low brine flow rate (0.005 cm3/min). This processwas continued by increasing the brine flow rate in small incre-ments. When slight increase in brine flow rate did not changethe saturation profile, the imbibition process was considered com-plete, and the residual scCO2 saturation was reported. This allowedus to build an imbibition capillary pressure curve. After completingthe imbibition test, the secondary drainage was started by inject-ing scCO2 at a much lower flow rate, i.e., 0.005 cm3/min, comparedto that of the primary drainage. The procedure was identical to theprimary drainage process to construct the capillary pressure curveand ultimately establish the same Swi as primary drainage (i.e.,�0.49).

Following the secondary drainage, in order to perform the lastgroup of experiments (category C), an imbibition test was startedwith a relatively high brine flow rate of 1.01 cm3/min, which corre-sponded to Nc = 1.3 � 10�5. After reaching the steady-state condi-tion, we recorded the pressure drop and measured the SCO2r. Thisimbibition test was then followed by another set of drainage andimbibition experiments, with this difference that we used an evenhigher brine flow rate, i.e., 2.02 cm3/min. Before each imbibitiontest, we made sure that Swi was similar to previous imbibition tests.All the saturation values used in this study were averaged over theentire length of the core, unless stated otherwise.

3. Results and discussion

In this section, first, we present the capillary trapping resultsobtained from experiments 1–8, see Table 2, followed by theresults of the relative permeability tests using steady- andunsteady-state methods; then we explain the characteristics ofthe capillary pressure curves (exp. 7–9), and finally we demon-strate the results of the high brine flow rate imbibition experi-ments (exp. 9–12).

3.1. Capillary trapping

Two steady-state (experiments 1, 2, 4, and 5) and one unsteady-state (experiments 7 and 8) drainage-imbibition cycles wereperformed on the Madison limestone core sample under capillary-dominated displacement regime.

Fig. 3 demonstrates the variation of the CO2 saturation acrossthe core sample during steady-state experiments. This figure alsoshows the effect of heterogeneity on the saturation profiles alongthe length of the core sample. Fig. 4 exhibits the variation of theSCO2r with Smax

CO2 for these series of experiments. In this figure, theresults are compared with the Spiteri et al. [69] empirical correla-tion and Land model [70], which provide a relationship betweeninitial and residual non-wetting phase saturations. The resultscompare well using the model parameters listed in Table 6. Thecoefficients a and b were calculated through fitting the experimen-tal data into the quadratic equation provided by Spiteri et al. [69].Note that each point in this figure represents the slice-averagedsaturations obtained along the length of the core sample. For thefirst steady-state drainage-imbibition cycle, the Smax

CO2 was about0.32. The subsequent waterflood led to SCO2r of about 0.245, mean-ing that approximately 76.6% of the initial scCO2 was trapped per-manently through capillary trapping. The Smax

CO2 , SCO2r, and trappingefficiency (SCO2r/S

maxCO2 ) for the second cycle were 0.164, 0.126, and

76.8%, respectively. For the last drainage and imbibition cycle,i.e., experiments 7 and 8, these values were 0.525, 0.329, and62.7%, respectively. Fig. 5 illustrates the trapping efficiencyobtained from each slice along the length of the Madison limestonecore sample. It indicates that significant fractions of scCO2-richphase can be trapped, i.e., from 62.7% to 76.8%, due to chase brineflooding. It also shows that as the initial brine saturation increases,the trapping efficiency improves. The scatter in the data is attrib-uted to the heterogeneity of the sample. The trapping resultsobtained from the second cycle of steady-state experiments withlow Smax

CO2 are particularly significant due to the fact that at reservoirscale, initial CO2 saturation in grid blocks away from the injectionwells may not be high. This reduces the saturation span that thoseblocks experience during a drainage-imbibition cycle.

These results show that the maximum initial scCO2-rich phasesaturation obtained at the end of unsteady-state experiments(i.e., 0.525) is less than its counterparts in Berea and Nugget SScores, i.e., 0.59 and 0.68, respectively [51]. The lower initial satura-tion may be attributed to higher drainage capillary pressureneeded. Furthermore, compared to sandstone rock samples, lessamount of scCO2-rich phase is trapped in the Madison limestone,which might be related to the latter being less water wet leadingto better hydraulic conductivity of the scCO2-rich phase. In Fig. 4,we also plot our data against the trapping results of the variouscarbonate rock samples reported in the literature. In some cases,higher residual trapping observed in our study may be attributedto heterogeneity [35], lower viscosity ratio between two phasesunder our experimental conditions as well as different pore-spacetopologies impacting pore-scale displacement mechanismsresponsible for trapping. In Table 7, the initial and residual scCO2

saturations measured on different carbonate rock samples pre-sented in the literature are compared to those of this study. Itcan be seen that our results lie within the range of data that havebeen reported by other researchers (see also Fig. 4). This could bein part owing to the fact that small amount of SO2 did not affectparameters such as IFT and contact angle significantly [65], andhence SO2 + scCO2 trapping results compare favorably with thosewithout SO2. This is consistent with the results of another studysuggesting that small amount of SO2 does not decrease the volumeof the scCO2-rich phase trapped in the rock sample [16]. Note thatto simplify this work, the fluid phases were equilibrated with theminerals of the rock sample to minimize the impact of acidic brineon the trapping of scCO2-rich phase, for instance, by altering thepore and matrix structures.

3.2. Relative permeability measurements

In this section, we demonstrate the results of two sets of drainage-imbibition relative permeability experiments with different satura-tion spans in the Madison limestone core sample (category A). Wealso present the brine relative permeability results generatedthrough dissolution processes.

3.2.1. Steady-state experimentsFigs. 6 and 7 illustrate the results of the drainage-imbibition rel-

ative permeability tests for both cycles. As it is shown in Fig. 6, theinitial and residual scCO2 saturations for the first and second cycleswere 0.321, 0.245 and 0.164, 0.136, respectively. We observerather rapid reduction of scCO2-rich phase relative permeabilityduring imbibition. Furthermore, scCO2 + SO2 relative permeabilityat the end of drainage reaches relatively small values, e.g., 0.04.This is consistent with the results reported by various researchgroups in the literature [39,46,48,49,51,55] (see also Fig. 7). Thelow scCO2-rich phase relative permeability in scCO2/brine systemat the end of the steady-state drainage process might be attributed

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0Steady- & unsteady-state Exp.Experiments 1,2,4,5,7, and 8

This study; scCO 2+SO2, Slice-averaged saturation Spiteri et al., 2008 Land model, 1968 Bennion & Bachu, 2010; Carbonate samples

S CO

2r (R

esid

ual C

O2 s

atur

atio

n)

SmaxCO2 (Initial CO2 saturation)

Fig. 4. Variation of SCO2r with SmaxCO2 for capillary-dominated experiments on the

Madison limestone core sample.

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

SmaxCO2 (Initial CO2 saturation)

Steady- and unsteady-state Exp.Experiments 1,2,4,5,7, and 8

Trapping efficiency, Slice-averaged saturation Curvefit

S CO

2r/S

max

CO

2 (Tr

appi

ng e

ffici

ency

)

Fig. 5. Variation of trapping efficiency with SmaxCO2 for capillary-dominated experi-

ments on the Madison limestone core sample.

M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56 51

to several factors, such as heterogeneity of the sample, viscosityratio, interfacial tension, and wettability of the rock [71].

It is clear that unlike the results of the drainage-imbibition rel-ative permeability experiments on, for instance, Berea sandstonecore [51], we do not observe a sharp decrease in brine relative per-meability during drainage. This might be attributed to poor pore-level connectivity of large brine-filled elements that get invadedby scCO2 + SO2 compared to those of the scCO2/brine system inthe Berea SS core. This figure suggests that the brine maintaineda good hydraulic conductivity while the non-wetting phase gradu-ally invaded the pores. The trends observed in this category agreewell with those of the carbonate core samples published in the lit-erature [43,55].

For the imbibition process we noticed that for a given brine sat-uration, the drainage brine relative permeability is slightly smallerthan that of the imbibition. This phenomenon was also observed inthe results presented by Akbarabadi and Piri [51], Bennion andBachu [43,55], and Oak et al. [72] and might have been causedby slight re-configuration of pore fluid occupancy. Numerical val-ues of the measured steady-state relative permeabilities for exper-iments 1, 2, 4, and 5 are listed in Table 8.

3.2.2. Dissolution experimentsAn unsteady-state dissolution step was performed as an

approach to re-establish Sw = 1, and initiate a new set of experi-ments. At the end of each dissolution test, the core was fullyscanned and CT numbers were compared with both CTfb and CTwc.Comparing the scans showed that we had successfully re-estab-lished Sw = 1. The end-point dissolution brine relative permeabilityfor both cycles are demonstrated in Fig. 8. The slight differencebetween the results of the two dissolution experiments might havebeen caused by the vastly different brine flow rates used for disso-lution in experiment 3 compared to those in test 6. The differencein brine flow rate may have resulted in a slightly different order bywhich scCO2 + SO2-filled pores were invaded by brine through dis-solution of the mixture. This may have then led to a different fluid

Table 6Parameters used with Spiteri [69] empirical correlation and Land model [70].

Sample Experiment Fluid system

Madison limestone Exp. 1, 2, 4, 5, 7, and 8 scCO2 + SO2/BrMadison limestone Exp. 9 and 10 scCO2 + SO2/BrMadison limestone Exp. 11 and 12 scCO2 + SO2/Br

occupancy at the pore scale for a given saturation producing differ-ent relative permeability.

3.3. Capillary pressure

In this section, we present the results of the capillary pressuremeasurement experiments performed on the Madison limestonecore using the unsteady-state method. We used inlet saturationsto characterize the capillary pressure behavior of the rock sample.The method of using the inlet saturations during the 100%non-wetting phase injection to calculate relative permeabilityand capillary pressure was first introduced by Leas et al. [73] andRamakrishnan and Capiello [74], and then adopted by Pini et al.[35,52] to study the capillary pressure versus saturation of thescCO2/brine core-flooding experiments.

During unsteady-state core-flooding experiments, scCO2-richphase is injected at a constant flow rate (QCO2) into the fullybrine-saturated core. The boundary conditions are: (1) Pjx=L = Pback

pressure = constant (back pressure is kept constant using back pres-sure regulation pump) and (2) QCO2jinlet = constant (inlet flow rateis constant). Therefore, a pressure gradient exists along the lengthof the core. The inlet and outlet pressures, as well as pressuredrops, are carefully monitored and recorded. At the steady state,there is no brine production and the pressure drop is constant. Thismeans that brine pressure in the core reaches to a fixed value, i.e.,dPw/dx = 0, and Pwjx=0 = Pwjx=L. As long as brine remains a continu-ous phase across the outlet face of the core, which is a validassumption for the wetting phase, the CO2 flow creates the pres-sure drop across the core. At this point, if CO2 stays as a continuousphase across the upstream of the core, it can be assumed thatPinlet = PCO2 and Poutlet = Pbrine. This means that before reaching tozero capillary pressure outside outlet face of the core, Pc takes afinite value at the end faces, i.e., Pcjx=0 and Pcjx=L – 0.

For a CO2/brine system, the capillary pressure is calculatedusing:

Pc ¼ PCO2 � Pbrine ð5Þ

a b SmaxCO2r C

ine 0.85 0.35 0.62 0.613ine 0.80 0.42 – –ine 0.70 0.40 – –

Table 7Initial and residual scCO2 saturations measured in various carbonate rock samples obtained from the literature.

Rock sample Porosity(%)

Kabs

(mD)Temperature(�C)

Pressure(MPa)

Viscosityratio(lw/lnw)

Initial scCO2

saturationResidual scCO2

saturationReferences

Wabamun#3 15.4 54.3 41 11.9 15.41 0.148 0.045 [42,43]Nisku#2 10.4 21.02 56 17.4 11.80 0.508 0.218 [42,43]Nisku#3 10.9 74.4 56 17.4 11.80 0.603 0.207 [42,43]Grosmont 11.8 153.9 41 11.9 – 0.480 0.365 [42,43]Morinville Leduc 11.6 371.9 40 11.4 – 0.470 0.131 [42,43]Redwater Leduc 16.8 353.6 36 9.2 – 0.335 0.208 [42,43]Cooking Lake#2 16.7 4.87 55 15.5 15.43 0.431 0.268 [42,43]Slave Point 9.9 0.217 43 18.8 – 0.454 0.256 [42,43]Winnipegosis 14.8 3.09 36 8.73 – 0.789 0.415 [42,43]Carbonate from Middle

East14.0 6.7 40 9.8 – 0.400 0.230 [54]

Indiana 19.7 244 50 9.0 3.0 0.730 0.230 [56]Madison 19.03 13.6 60 19.16 9.18 0.475 0.329 This study

0.4 0.5 0.6 0.7 0.8 0.9 1.00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0Steady-state Exp.

krw-Exp.1 krg-Exp.1 krw-Exp.2 krg-Exp.2 Curvefit

k r (Br

ine/

scC

O2 re

lativ

e pe

rmea

bility

)

Sw (Brine saturation)

0.4 0.5 0.6 0.7 0.8 0.9 1.00.00

0.02

0.04

0.06

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Sw (Brine saturation)

Steady-state Exp.

krw-Exp.4 krg-Exp.4 krw-Exp.5 krg-Exp.5 Curvefit

k r (Br

ine/

scC

O2 re

lativ

e pe

rmea

bility

)

(a)

(b)

Fig. 6. scCO2 + SO2/brine drainage-imbibition relative permeability results gener-ated using steady-state method on the Madison limestone core sample for (a) first(b) second cycles.

0.4 0.5 0.6 0.7 0.8 0.9 1.00.000.020.040.060.080.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0 Drainage; This study; scCO

2+SO2

Imbibition; This study; scCO2

+SO2

Drainage; Bennion & Bachu, 2010; Carbonate samples Imbibition; Bennion & Bachu, 2010; Carbonate samples

k rg (C

O2 r

elat

ive

perm

eabi

lity)

Sw (Brine saturation)

0.4 0.5 0.6 0.7 0.8 0.9 1.00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0 Drainage; This study; scCO

2+SO2

Imbibition; This study; scCO2

+SO2

Drainage; Bennion & Bachu, 2010; Carbonate samples Imbibition; Bennion & Bachu, 2010; Carbonate samples

k rw (B

rine

rela

tive

perm

eabi

lity)

Sw (Brine saturation)

(a)

(b)

Fig. 7. Comparison between the relative permeability data of this study and thosefrom the literature.

52 M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56

Using the above-mentioned boundary conditions with Eq. (5),the final Pdra

c can be found through pressure drop measurement:

Pdrac jx¼0 ¼ Pinlet � Poutlet ¼ DP ð6Þ

To calculate the capillary pressure during the imbibition process,we only injected brine and implemented the following procedure:at the steady-state condition, we use Eq. (5) where the pressures

are measured at the inlet face of the core. Considering the pore-scale physics of the displacement mechanism during the imbibitiontests: Pinlet

CO2 = Pback pressure + DPdramax and Pinlet

brine = Pback pressure + DPimb,where DPdra

max is the maximum pressure drop at the end of the pri-mary drainage process preceding imbibition and DPimb is the pres-sure drop at the end of each imbibition step. Hence, the finalequation would be:

Table 8Steady-state relative permeability data of experiments 1, 2, 4, and 5.

Experiments 1,2 Experiments 4,5

Sw krw krg Sw krw krg

Drainage 1.000 1.000 0.000 1.000 1.000 0.0000.8274 0.32712 0.0106 0.9039 0.5361 0.00350.7892 0.2340 0.0152 0.8873 0.4669 0.00610.7509 0.1980 0.0205 0.885 0.4629 0.00670.7254 0.1266 0.0263 0.8727 0.4213 0.00960.6792 0.0825 0.0357 0.8364 0.3779 0.0131

Imbibition 0.6792 0.0825 0.0357 0.8407 0.3830 0.00990.6848 0.0958 0.0331 0.8511 0.4063 0.00700.6887 0.1035 0.0307 0.8539 0.4093 0.00660.7004 0.1161 0.0215 0.8653 0.4181 0.00340.7145 0.1383 0.0179 0.8744 0.4381 0.00000.7278 0.1400 0.01090.7429 0.1556 0.00610.7554 0.1797 0.0000

0.4 0.5 0.6 0.7 0.8 0.9 1.00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Dissolution Exp.

krw-Exp.3 krw-Exp.6 Curvefit

k r (Br

ine

rela

tive

perm

eabi

lity)

Sw (Brine saturation)

Dissolution

Fig. 8. End-point dissolution brine relative permeability for experiments 3 and 6performed on the Madison limestone core sample.

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0-30-25-20-15-10-505

101520253035404550

Unsteady-state Exp.Madison limestone

Exp.7-Primary drainage Exp.8-Primary imbibition Exp.9-Secondary drainage Curvefit

P c (C

apilla

ry p

ress

ure

, kPa

)

Sw-inlet (Inlet brine saturation)

Fig. 9. scCO2 + SO2/brine capillary pressure versus saturation curves for theMadison limestone core sample.

M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56 53

Pimbc jx¼0 ¼ DPdra

max � DPimb ð7Þ

Therefore, through measuring the inlet saturation using the CTscanner and repeating this procedure for various CO2 and brineflow rates, the capillary pressure versus saturation curves can beconstructed [52].

Two sets of experiments were performed on the Madison lime-stone core sample to characterize the scCO2 + SO2/brine primarydrainage, imbibition, and secondary drainage capillary pressuresversus saturation. In category C, to obtain more accurate valuesfor the saturations, the inlet face of the core was scanned 20 timesafter reaching steady state, and the final value was calculatedthrough averaging.

Fig. 9 depicts the results of the scCO2 + SO2/brine capillary pres-sure data obtained using the core-flooding apparatus. As it is illus-trated, the maximum initial and residual scCO2 saturations are0.692 and 0.532, respectively. These values agree well with thetrend shown in Fig. 4.

During the imbibition process, if the sample was stronglywater-wet, imbibition would establish the final residual CO2 satu-ration at capillary pressures close to zero; otherwise, additionalCO2 could be displaced by more brine injection [68]. This phenom-enon was observed in the Madison limestone core sample. Ourresults show that the wettability of the core sample might havebeen altered due to being in contact with scCO2 + SO2 mixture. Thisis consistent with the observations reported in Chiquet et al. [26],Chalbaud et al. [27], and Be [30]. The numerical values of the pri-mary drainage, imbibition, and secondary drainage capillary pres-sure versus saturation curves are presented in Table 9.

3.4. Dynamic effects

To investigate the impact of dynamic effects, corresponding toviscous-dominated displacement regime, on the trapping of thescCO2-rich phase, two separate imbibition tests with relatively highbrine flow rates (i.e., 1.01 and 2.02 cm3/min) were conducted. Theinitial scCO2 saturation for both tests were identical (i.e., �0.49).Fig. 10 displays the scCO2 + SO2 trapping results for individual slicesfor both experiments in the Madison limestone core. We also com-pared these results with those obtained from the capillary-domi-nated experiments as well as the Spiteri et al. [69] empiricalcorrelation. The results compare well utilizing the model parame-ters listed in Table 6. We observed that the scCO2-rich phase satu-ration was significantly reduced because of the dynamic effects. Theaverage residual scCO2 saturation decreased from 0.329 at the endof experiment 8 to 0.291 and 0.262 at the end of experiments 10and 12, respectively. It means that the residual scCO2-rich phasesaturation dropped about 11.5% for the first and 20.4% for the sec-ond imbibition test. This decline is attributed to high brine flowrates producing a considerable pressure drop across the core. Underthese conditions, brine invades accessible scCO2-filled pores notonly based on the pore threshold capillary pressures but also theviscous pressure drops associated with each displacement. This

Table 9Capillary pressure data of experiments 7 to 9.

Primary drainage Imbibition Secondary drainage

Qg (cm3/min) DP (kPa) Sw�inlet Qw(cm3/min) DP (kPa) Sw�inlet Qg (cm3/min) DP (kPa) Sw�inlet

0.000 0.00000 1.0000 0.0100 8.29710 0.35740 0.005 12.1876 0.46510.025 20.5374 0.5486 0.0150 �2.86230 0.40040 0.009 15.0733 0.41400.040 21.8607 0.4294 0.0200 �12.1661 0.40980 0.013 14.9520 0.39320.050 22.2994 0.3959 0.0250 �20.7562 0.43430 0.017 16.2138 0.37560.065 23.6311 0.3741 0.0275 �22.2562 0.45880 0.022 17.6727 0.35920.075 24.7552 0.3620 0.0300 �26.7706 0.46770 0.030 18.4670 0.33800.085 25.3560 0.3467 0.035 18.9899 0.33000.100 26.9562 0.3476 0.040 19.6030 0.32180.120 27.7999 0.3315 0.045 20.0791 0.31970.140 29.4642 0.3103 0.055 21.5183 0.30930.175 30.8489 0.3098 0.065 22.6772 0.30540.200 32.6594 0.3086

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0Steady- and unsteady-state Exp.Slice-averaged saturation

Experiments 9,10 Experiments 11,12

Exp.1,2,4,5,7, and 8- Spiteri et al.,2008 (capillary dominated) Exp.9,10- Spiteri et al.,2008 Exp.11,12- Spiteri et al.,2008

S CO

2r (R

esid

ual C

O2 s

atur

atio

n)

SmaxCO2 (Initial CO2 saturation)

Fig. 10. Variation of SCO2r with SmaxCO2 for high flow rate brine floods and comparison

with the capillary-dominated counterparts.

54 M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56

creates displacements similar to frontal piston-like advance, whichsignificantly reduces trapping. These reductions were more signifi-cant specifically through the first half of the core. One could imag-ine the possibility of high pressure brine, created by high flow rate,at the inlet side of the core dissolving some of the scCO2-rich phaseand releasing it at the second half of the core, which was at a muchlower pressure due to considerable heterogeneity. This phenome-non, however, is expected to have negligible impact on the satura-tion difference observed in the first half of the core, because thedifference in pressure values could not create a meaningful differ-ence in additional dissolution and hence saturations. Flash calcula-tions [60] showed that only 0.3% and 0.5% (saturation percentage)reductions in residual saturation during the above-mentionedexperiments were due to additional dissolution of scCO2-rich phasein brine as a result of pressure increase. Note that this amountrepresents the maximum dissolution when the pressure is at itshighest value, i.e., at the inlet of the core sample. This pressure gra-dient decreases significantly toward the outlet of the sample, lead-ing to even less additional dissolution of the scCO2 + SO2 in brine.Furthermore, the reservoir core sample used in this study had alower porosity and permeability in the middle (see Fig. 1); there-fore, this created a significant pressure gradient in the first half ofthe core sample while the rest of the sample did not experiencelarge pressure drop. This led to more decrease in saturation throughthe first half of the core.

3.5. Conclusions

A heterogeneous Madison limestone core sample acquired fromRock Springs Uplift in southwest Wyoming was used to performreservoir-conditions core-flooding experiments with scCO2 + SO2/brine fluid system at 60 �C temperature and 19.16 MPa pressure.A state-of-the-art experimental apparatus was utilized to carryout three categories of flow tests to characterize relative perme-ability, capillary pressure, and capillary trapping of scCO2 + SO2

mixture using steady- and unsteady-state techniques. We alsoinvestigated the impact of dynamic effects on capillary trappingof the above-mentioned mixture due to chase brine injection.

The results of three different categories of experiments werepresented and discussed. In category A, we employed a steady-statemethod to perform two drainage-imbibition relative permeabilitycycles. At the end of each cycle, a dissolution process was imple-mented to re-establish Sw = 1. The comparison between the dissolu-tion scans and the fresh brine scans confirmed the Sw = 1 state.Different ranges of fractional flow during drainage experimentswere applied to establish various initial scCO2-rich phase satura-tions for the subsequent imbibition tests. The results showed that62.7% to 76.8% of the initial scCO2 + SO2 volume was trappedthrough capillary trapping, meaning that significant portion of theCO2 in-placed at the end of the drainage process can be trapped per-manently due to chase brine injection with the lowest risk of leak-age. The amount of scCO2-rich phase trapping was comparable tothose of pure CO2 in other carbonates reported in the literature. Thismight be in part due to the fact that small amount of SO2 did notaffect IFT, contact angles, and storage capacity of the rock sample,which was also reported by other researchers. In this study, theimpact of mineral dissolution into the acidic brine formed by SO2

disproportionation was almost eliminated as we had equilibratedthe fluid phases with the minerals of the rock sample.

The results show that the higher initial brine saturations led tohigher trapping efficiencies. We found that the final scCO2-richphase drainage relative permeability was very low, i.e., 0.04, whichis consistent with the results that have been reported by otherresearch studies. We observed a rapid reduction in imbibition rel-ative permeability of the scCO2-rich phase, resulting in high resid-ual trapping. For the imbibition process, we noticed that for a givenbrine saturation, the imbibition brine relative permeability ishigher than that of the drainage. This might be due to the re-con-figuration of the fluid occupancy. The end-point dissolution brinerelative permeabilities were also obtained for two dissolutionexperiments. A slight difference between the dissolution resultswas attributed to difference in brine flow rates used, resulting ina different order by which scCO2 + SO2-filled pores were invadedby brine through dissolution of the mixture, which in turn could

M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56 55

lead to a different fluid occupancy at the pore scale for a given sat-uration. This result indicates that dissolution plays an effective roleon the permanent mitigation of the scCO2-rich phase at the laterstages of the geological storage process.

In category B of the experiments, unsteady-state technique wasused to perform primary drainage, imbibition, and secondarydrainage tests to obtain scCO2 + SO2/brine capillary pressures.The results showed that the lowest initial brine saturation wecould establish in the Madison limestone core sample was 0.475.The subsequent imbibition process resulted in 0.329 residual scCO2

saturation, meaning that more than 62% of the scCO2-rich phasewas trapped through capillary trapping mechanism. By comparingthe trapping results (i.e., initial and residual saturations, and trap-ping efficiency) obtained from both categories (A and B) of exper-iments in this study to those available in the literature forsandstone samples (e.g., Berea and Nugget), it can be concludedthat the capacity of the rock sample to receive scCO2-rich phaseduring the drainage was slightly lower in the Madison limestonecompared to those of the sandstones; furthermore, the residualsaturations of the scCO2-rich phase was also lower. This could bedue to the fact that Madison limestone is less water wet than theabove-mentioned sandstone rock samples. However, in compari-son with the results of carbonate cores, our data showed slightlyhigher residual scCO2-rich phase saturations. This might be attrib-uted to heterogeneity, different viscosity ratio, and dissimilar pore-space topologies impacting pore-scale displacement mechanismsresponsible for trapping. The imbibition capillary pressure curveshowed a negative value at the residual scCO2 saturation, whichis an indication of wettability alteration of the rock sample. Thiswas consistent with findings of other authors reported in theliterature.

In category C, two imbibition tests with relatively high brineflow rates were conducted to study the impact of dynamic effectson capillary trapping. Brine flow rates during the first and secondfast imbibition tests were 1.01 and 2.02 cm3/min. Residual scCO2

saturations at the end of each imbibition test were reduced by11.5% (0.291) and 20.4% (0.262) compared to 0.329 obtained undercapillary-dominated displacement regime. This phenomenon canbe attributed to the large viscous pressure drop, created by highbrine flow rate, along the core, causing the sweeping of the scCO2-rich phase from the pores. The extra scCO2 was produced by weak-ening the non-wetting phase trapping processes. This was done bychanging the order by which the displacement mechanisms takeplace at the pore scale, which in turn is achieved by changing theirrelative threshold pressure values. Flash calculations showed thatonly 0.3% and 0.5% (saturation percentage) of the reductions inresidual saturation during the first and second fast imbibition testswere due to high pressures created by high brine flow rates. There-fore, the contribution of the dissolution to the overall reduction intrapping was relatively insignificant.

Acknowledgments

This work was supported by DOE Financial Assistance Agree-ment DE-FE0004832. The authors also acknowledge the financialsupport of Encana and the School of Energy Resources and theEnhanced Oil Recovery Institute at the University of Wyoming(UW). The Carbon Management Institute (CMI) at UW is thankedfor providing the core sample.

References

[1] Birol F et al. Redrawing the energy-climate map: world energy outlook specialreport. France: International Energy Agency; 2013.

[2] Metz B, Davidson O, de Coninck HC, Loos M, Meyer LA. IPCC special report oncarbon dioxide capture and storage. Cambridge University Press; 2005.

[3] Olivier Jos GJ, Janssen-Maenhout G, Peters JAHW. Trends in global CO2

emissions. 2012 Reports, PBL Netherlands Environmental Assessment Agencythe Hague/Bilthoven; 2012.

[4] Bell DA, Towler BF, Fan M. Coal gasification and its applications. 1sted. UK: Oxford; 2011 [ISBN: 978-0-8155-2049-8].

[5] Gale J. Geological storage of CO2: what’s known, where are the gaps, and whatmore needs to be done. Energy 2004;29(9–10):1329–38. http://dx.doi.org/10.1016/j.energy.2004.03.068.

[6] Szulczewski ML, MacMinn CW, Herzog HJ, Juanes R. Lifetime of carbon captureand storage as a climate-change mitigation technology. PNAS 2012;109(14):5185–9. http://dx.doi.org/10.1073/pnas.1115347109.

[7] ZEP. The EU flagship programme for CO2 capture and storage (CCS). EuropeanTechnology Platform for Zero Emission Fossil Fuel Power Plants (ZEP);February 2008.

[8] Nordbotten JM, Celia MA, Bachu S. Injection and storage of CO2 in deep salineaquifers: analytical solution for CO2 plume evolution during injection. TranspPorous Media 2007;58:339–60. http://dx.doi.org/10.1007/s11242-004-0670-9.

[9] Span R, Wagner W. A new equation of state for carbon dioxide covering thefluid region from the triple-point temperature to 1100 K at pressure up to800 MPa. J Phys Chem Ref Data 1996;25(6):1509–96. http://dx.doi.org/10.1063/1.555991.

[10] Damen K, Faaij A, Turkenburg W. Health, safety and environmental risks ofunderground CO2 storage – overview of mechanisms and current knowledge.Clim Change 2006;74(1–3):289–318. http://dx.doi.org/10.1007/s10584-005-0425-9.

[11] Kumar A, Ozah R, Noh M, Pope GA, Bryant S, Sepehrnoori K, et al. Reservoirsimulation of CO2 storage in deep saline aquifers. SPE J 2005;10(3):33648.http://dx.doi.org/10.2118/89343-PA.

[12] Ennis-King J, Paterson L. Role of convective mixing in the long-term storage ofcarbon dioxide in deep saline formations. SPE J 2005;10(3):349–56. http://dx.doi.org/10.2118/84344-PA.

[13] Juanes R, Spiteri EJ, Orr Jr FM, Blunt MJ. Impact of relative permeabilityhysteresis on geological CO2 storage. Water Resour Res 2006;42:W12418.http://dx.doi.org/10.1029/2005WR004806.

[14] Ide TS, Jessen K, Orr Jr FM. Storage of CO2 in saline aquifers: effects of gravity,viscous, and capillary forces on amount and timing of trapping. Int J GreenhGas Control 2007;1(4):48191. http://dx.doi.org/10.1016/S1750-5836(07)00091-6.

[15] Ran Q, LaForce TC, Blunt MJ. Design of carbon dioxide storage in aquifers. Int JGreenh Gas Control 2009;3(2):195–205. http://dx.doi.org/10.1016/j.ijggc.2008.08.004.

[16] Nogueira M, Mamora DD. Effect of flue gas impurities on the process ofinjection and storage of CO2 in depleted gas reservoirs. J Energy ResourTechnol 2008;30:2339–53. http://dx.doi.org/10.1115/1.2825174.

[17] Miller PJ, Atten CV, Bradley MJ. North American power plant air emissions.Commission for Environmental Cooperation of North America; 2004.

[18] Lloyd Evan. North American power plant air emissions. Commission forEnvironmental Cooperation of North America; 2011.

[19] Reddy KJ, Weber H, Bhattacharyya P, Argyle M, Taylor D, Christensen M, et al.Instantaneous capture and mineralization of flue gas carbon dioxide: pilotscale study. Nat Preced 2010. http://dx.doi.org/10.1038/npre.2010.5404.1.

[20] Ellis BR, Crandell LE, Peters CA. Limitations for brine acidification due to SO2

co-injection in geologic carbon sequestration. Int J Greenh Gas Control2010;4:575–82. http://dx.doi.org/10.1016/j.ijggc.2009.11.006.

[21] Chialvo AA, Vlcek L, Cole DR. Acid gases in CO2-rich subsurface geologicenvironments. Rev Mineral Geochem 2013;77:361–98. http://dx.doi.org/10.2138/rmg.2013.77.10.

[22] Crandell LE, Ellis BR, Peters CA. Dissolution potential of SO2 co-injected withCO2 in geologic sequestration. Environ Sci Technol 2010;44(1):349–55. http://dx.doi.org/10.1021/ja00040a007.

[23] Kaszuba John P, Janecky DR, Snow Marjorie G. Experimental evaluation ofmixed fluid reactions between supercritical carbon dioxide and NaClbrine: relevance to the integrity of a geologic carbon repository. ChemGeol 2005;217(3–4):277–93. http://dx.doi.org/10.1016/j.chemgeo.2004.12.014.

[24] Knauss Kevin G, Johnson James W, Steefel Carl I. Evaluation of the impact ofCO2, co-contaminant gas, aqueous fluid and reservoir rock interactions on thegeologic sequestration of CO2. Chem Geol 2005;217(3–4):339–50. http://dx.doi.org/10.1016/j.chemgeo.2004.12.017.

[25] Rosenbauer RJ, Koksalan T, Palandri JL. Experimental investigation of CO2–brine–rock interactions at elevated temperature and pressure: implications forCO2 sequestration in deep-saline aquifers. Fuel Process Technol 2005;86(14–15):1581–97. http://dx.doi.org/10.1016/j.fuproc.2005.01.011.

[26] Chiquet P, Broseta D, Thibeau S. Wettability alteration of caprock minerals bycarbon dioxide. Geofluids 2007;7:112–22. http://dx.doi.org/10.1111/j.1468-8123.2007.00168.x.

[27] Chalbaud C, Robin M, Lombard J-M, Martin F, Egermann P, Bertin H. Interfacialtension measurements and wettability evaluation for geological CO2 storage.Adv Water Resour 2009;32:98–109. http://dx.doi.org/10.1016/j.advwatres.2008.10.012.

[28] Kaszuba JP, Navarre-Stichler A, Thyne G, Chopping C, Meuzelaar T.Supercritical carbon dioxide and sulfur in the Madison limestone: a naturalanalog in southwest Wyoming for geologic carbon-sulfur co-sequestration.Earth Planetary Sci Lett 2011;309:131–40. http://dx.doi.org/10.1016/j.epsl.2011.06.033.

56 M. Akbarabadi, M. Piri / Advances in Water Resources 77 (2015) 44–56

[29] Lu P, Fu Q, Syfried Jr E. Navajo sandstone–brine–CO2 interaction: implicationsfor geological carbon sequestration. Environ Earth Sci 2011;62(14–15):101–18. http://dx.doi.org/10.1007/s12665-010-0501-y.

[30] Be A. Multiphase flow in porous media with emphasis on CO2 sequestration.PhD Dissertation, University of Bergen, Norway; 2011.

[31] Ellis BR, Peters CA, Fitts J, Bromhal G, McIntyre D, Warzinski R, et al. Modelingand analysis: deterioration of a fractured carbonate caprock exposed to CO2-acidified brine flow. Greenh Gas Sci Technol 2011;1:248–60. http://dx.doi.org/10.1002/ghg.

[32] Chopping C, Kaszuba JP. Supercritical carbon dioxide–brine–rock reactions inthe Madison limestone of southwest Wyoming: an experimental investigationof a sulfur-rich natural carbon dioxide reservoir. Chem Geol 2012;322–323:223–36. http://dx.doi.org/10.1016/j.chemgeo.2012.07.004.

[33] Wang J, Ryan D, Anthony EJ, Wigstone A, Basavda-Reddi L. The effect ofimpurities in oxyfuel flue gas on CO2 storage capacity. Int J Greenh Gas Control2012;11:158–62. http://dx.doi.org/10.1016/j.ijggc.2012.08.002.

[34] Ellis BR, Fitts JP, Bromhal GS, McIntyre DL, Tappero R, Peters CA. Dissolution-driven permeability reduction of a fractured carbonate caprock. Environ EngSci 2013;30(4):187–93. http://dx.doi.org/10.1089/ees.2012.0337.

[35] Pini R, Krevor SCM, Benson SM. Capillary pressure and heterogeneity for theCO2/water system in sandstone rocks at reservoir conditions. Adv WaterResour 2012;38:48–59. http://dx.doi.org/10.1016/j.advwatres.2011.12.007.

[36] Pentland CH, El-Maghraby R, Iglauer S, Blunt MJ. Measurements of thecapillary trapping of supercritical carbon dioxide in Berea sandstone. GeophysRes Lett 2011;38:L06401. http://dx.doi.org/10.1029/2011GL046683.

[37] Ovaysi S, Piri M. Pore-space alteration induced by brine acidification insubsurface geologic formations. Water Resour Res 2014;50(1):440–52. http://dx.doi.org/10.1002/2013WR014289.

[38] Jerauld GR, Salter SJ. The effect of pore structure on hysteresis in relativepermeability and capillary pressure: pore-level modeling. Transp PorousMedia 1990;5:103–51. http://dx.doi.org/10.1007/BF00144600.

[39] Bennion B, Bachu S. Relative permeability characteristics for supercritical CO2

displacing water in a variety of potential sequestration zones in westernCanada sedimentary basin. In: SPE 95547; Presented at the SPE technicalconference and Exhibition, Dallas, TX; 9–12 Oct. 2005. http://dx.doi.org/10.2118/95547-MS.

[40] Bennion B, Bachu S. Dependence on temperature, pressure and salinity of theIFT and relative permeability displacement characteristics of CO2 injected indeep saline aquifers. In: SPE 102138; Presented at the SPE annual technicalconference and exhibition, San Antonio, TX; 24-27 Sep. 2006. http://dx.doi.org/10.2118/102138-MS.

[41] Plug W-J, Bruining J. Capillary pressure for the sand–CO2–water system undervarious pressure conditions. Application to CO2 sequestration. Adv WaterResour 2007;30:2339–53. http://dx.doi.org/10.1016/j.advwatres.2007.05.010.

[42] Bachu S, Bennion B. Effects of in-situ conditions on relative permeabilitycharacteristics of CO2–brine systems. Environ Geol 2008;54:1707–22. http://dx.doi.org/10.1007/s00254-007-0946-9.

[43] Bennion B, Bachu S. Drainage and imbibition relative permeabilityrelationships for supercritical CO2/brine and H2S/brine systems inintergranular sandstone, carbonate, shale and anhydrite rocks. SPE ResevEval Eng 2008;11(3):487–96. http://dx.doi.org/10.2118/99326-PA.

[44] Suekane T, Nobuso T, Hirai S, Kiyota M. Geological storage of carbon dioxide byresidual gas and solubility trapping. Int J Greenh Gas Control 2008;2(1):58–64.http://dx.doi.org/10.1016/S1750-5836(07)00096-5.

[45] Shi JQ, Xue Z, Durucan S. History matching of CO2 core flooding CT scansaturation profiles with porosity dependent capillary pressure. EnergyProcedia 2009;1:3205–11. http://dx.doi.org/10.1016/j.egypro.2009.02.104.

[46] Perrin JC, Krause M, Kuo CW, Milijkovic L, Charoba E, Benson SM. Core-scaleexperimental study of relative permeability properties of CO2 and brine inreservoir rocks. Energy Procedia 2009;1:3515–22. http://dx.doi.org/10.1016/j.egypro.2009.02.144.

[47] Perrin JC, Benson SM. An experimental study on the influence of sub-core scaleheterogeneities on CO2 distribution in reservoir rocks. Transp Porous Media2010;82:93–109. http://dx.doi.org/10.1007/s11242-009-9426-x.

[48] Krevor SCM, Pini R, Li B, Benson SM. Capillary heterogeneity trapping of CO2 ina sandstone rock at reservoir condition. Geophys Res Lett 2011;38:L15401.http://dx.doi.org/10.1029/2011GL048239.

[49] Krevor SCM, Pini R, Zou L, Benson SM. Relative permeability and trapping ofCO2 and water in sandstone rocks at reservoir conditions. Water Resour Res2012;48:W02532. http://dx.doi.org/10.1029/2011WR010859.

[50] Zuo L, Krevor SCM, Falta RW, Benson SM. An experimental study of CO2

exsolution and relative permeability measurements during CO2 saturatedwater depressurization. Transp Porous Media 2012;91(2):459–78. http://dx.doi.org/10.1007/s11242-011-9854-2.

[51] Akbarabadi M, Piri M. Relative permeability hysteresis and capillary trappingcharacteristics of supercritical CO2/brine system: an experimental study at

reservoir conditions. Adv Water Resour 2013;52:190–206. http://dx.doi.org/10.1016/j.advwatres.2012.06.014.

[52] Pini R, Benson SM. Simultaneous determination of capillary pressure andrelative permeability curves from core-flooding experiments with variousfluid pairs. Water Resour Res 2013;49:3516–30. http://dx.doi.org/10.1002/wrcr.20274.

[53] Levine JS, Goldberg DS, Lackner KS, Matter JM, Supp MG, Ramakrishnan TS.Relative permeability experiments of carbon dioxide displacing brine and theirimplications for carbon sequestration. Environ Sci Technol 2014;48:811–8.http://dx.doi.org/10.1021/es401549e.

[54] Okabe H, Tsuchiya Y. Experimental investigation of residual CO2 saturationdistribution in carbonate rocks. In: International symposium of the Society ofthe Core Analysts in Abu Dhabi; 29 October–2 November, 2008.

[55] Bennion B, Bachu S. Drainage and imbibition CO2/brine relative permeabilitycurves at reservoir conditions for carbonate formations. In: Proceedings of theSPE Annual Technical Conference and Exhibition, Florence, Italy; September19–22, 2010, http://dx.doi.org/10.2118/134028-MS.

[56] El-Maghraby RM, Blunt MJ. Residual CO2 trapping in Indiana limestone.Environ Sci Technol 2013;2(1):58–64. http://dx.doi.org/10.1021/es304166u.

[57] Campbel-Stone E, Lynds Raine, Frost Carol, Becker Thomas P, Diem Bridget.The Wyoming carbon underground storage project: geologic characterizationof Moxa Arch and Rock Springs Uplift. Energy Procedia 2011;4:4656–63.http://dx.doi.org/10.1016/j.egypro.2011.02.426.

[58] Ihmels EC, Lemmon EW, Gmehling J. An equation of state and compressedliquid and supercritical densities for sulfur dioxide. Fluid Phase Equilib2003;207:111–30. http://dx.doi.org/10.1016/S0378-3812(03)00004-9.

[59] Sayegh SG, Najman J. CO2–SO2–brine phase behavior studies, PetroleumRecovery Institute, Alberta Department of Energy Library Services; 1984.

[60] Tan SP, Yao Y, Piri M. A PC-SAFT/PMSA model for the solubility of SO2 + CO2

mixtures in brine at elevated pressures and temperatures. Ind Eng Chem Res2013;52:10864–72. http://dx.doi.org/10.1021/ie4017557.

[61] Bachu S, Bennion B. Interfacial tension between CO2, freshwater, and brine inthe range of pressure from (2 to 27) MPa, temperature from (20 to 125) �C, andwater salinity from (0 to 334000) mg L�1. J Chem Eng Data 2009;54:765–75.http://dx.doi.org/10.1021/je800529x.

[62] Batzle M, Zhijing W. Seismic properties of pore fluids. Geophysics1992;57(11):1396–408. http://dx.doi.org/10.1190/1.1443207.

[63] Chiquet P, Draidon J-L, Broseta D, Thibeau S. CO2/water interfacial tensionunder pressure and temperature conditions of CO2 geological storage. EnergyConvers Manage 2007;48:736–44. http://dx.doi.org/10.1016/j.enconman.2006.09.011.

[64] Saraji S, Goual L, Piri M, Plancher H. Wettability of sc-CO2/brine/quartzsystems: simultaneous measurement of contact angle and interfacial tensionat reservoir conditions. Langmuir 2013;29(23):6856–66. http://dx.doi.org/10.1021/la3050863.

[65] Saraji S, Piri M, Goual L. The effect of brine salinity, SO2 contamination,pressure, and temperature on contact angle and interfacial tension of sc-CO2/brine/quartz systems. Int J Greenh Gas Control 2014;28:147–55. http://dx.doi.org/10.1016/j.ijggc.2014.06.024.

[66] Fenghour A, Wakeham WA, Vesovic V. The viscosity of the carbon dioxide. JPhys Chem Ref Data 1998;27(1):31–44. http://dx.doi.org/10.1021/la3050863.

[67] Piri M. Recirculating. Constant back pressure core flooding apparatus andmethod. US2012/0211089 A1; August, 2012.

[68] Anderson WG. Wettability literature survey. Part 4: Effects of wettability oncapillary pressure. J Petrol Technol 1987;39(10):1283–300. http://dx.doi.org/10.2118/15271-PA.

[69] Spiteri EJ, Juanes R, Blunt MJ, Orr Jr FM. A new model of trapping and relativepermeability hysteresis for all wettability characteristics. SPE J2008;13(3):277–88. http://dx.doi.org/10.2118/96448-PA.

[70] Land CS. Calculation of imbibition relative permeability for two- and three-phase flow from rock properties. SPE J 1968;8:149–56. http://dx.doi.org/10.2118/1942-PA.

[71] Benson S, Pini R, Reynolds C, Krevor S. Relative permeability analysis todescribe multi-phase flow in CO2 storage reservoirs. Global CCS InstituteTargeted Report No. 2; 2013.

[72] Oak MJ, Baker LE, Thomas DC. Three-phase relative permeability of Bereasandstone. J Petrol Technol 1990;42(8):1054–61. http://dx.doi.org/10.2118/17370-PA.

[73] Leas WJ, Jenks JH, Russel CD. Relative permeability to gas. J Petrol Technol1950;2(3):65–72. http://dx.doi.org/10.2118/950065-G.

[74] Ramakrishnan T, Capiello A. A new technique to measure static and dynamicproperties of a partially saturated porous medium. Chem Eng Sci1991;16(4):1157–63. http://dx.doi.org/10.1016/0009-2509(91)85109-B.