Sie sind auf Seite 1von 9

International Journal of Greenhouse Gas Control 83 (2019) 256–264

Contents lists available at ScienceDirect

International Journal of Greenhouse Gas Control


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

Injection of in-situ generated CO2 microbubbles into deep saline aquifers for T
enhanced carbon sequestration
Seokju Seoa, Mohammad Mastiania, Mazen Hafeza, Genevieve Kunkela, Christian Ghattas Asfoura,
Kevin Ivan Garcia-Ocampoa, Natalia Linaresa, Cesar Saldanaa, Kwangsoo Yangb,
Myeongsub Kima,

a
Department of Ocean and Mechanical Engineering, Florida Atlantic University, 777 Glades Road, Boca Raton, FL33431, USA
b
Department of Computer & Electrical Engineering and Computer Science, Florida Atlantic University, 777 Glades Road, Boca Raton, FL33431, USA

ARTICLE INFO ABSTRACT

Keywords: Carbon sequestration into deep saline aquifers has been considered a promising technology for mitigating heavy
Carbon sequestration atmospheric carbon dioxide (CO2) concentration. When gaseous CO2 is continuously injected into these aquifers,
CO2 dissolution resident brine near a wellbore area is rapidly evaporated while precipitating significant amounts of salt at pores,
CO2 injectivity thereby damaging the aquifer media unfavorable for subsequent CO2 injection. In addition, the continuous
Deep saline aquifers
injection of CO2 at a large volume significantly hinders dissolution of CO2 into brine. In this study, we propose a
Microfluidics
new method of sequential water injection with gaseous CO2 for in-situ generation of micro-sized CO2 bubbles
that minimizes the brine drying-out and simultaneously accelerates CO2 dissolution. We observed that, with this
method, a partial volume of CO2 dissolves effectively into the co-injected water during pumping, thereby de-
creasing the rate of brine drying-out at pores. Another benefit of sequential injection is the significantly in-
creased rate of CO2 hydration induced by the large surface-to-volume ratio of tiny bubbles at micro to nanoscale.
To further accelerate CO2 hydration, we investigated reactive dynamics of bubble-driven CO2 hydration at
different frequencies of sequential injection and pH levels of the solution. Operation at a higher frequency with
higher basicity proved to be the most effective in decreasing the bubble size and therefore accelerating CO2
hydration into brine, which is a more feasible CO2 storage plan.

1. Introduction of saline formations in the near-wellbore (Fuller et al., 2006; Kim et al.,
2013) and the slow rate of CO2 hydration (Bhaduri and Šiller, 2013; Seo
Carbon capture and sequestration (CCS) technology has been pro- et al., 2017, 2018b). The large volume of continuous CO2 injection
posed as a promising method of global carbon dioxide (CO2) mitigation creates a dried zone near the injection well (Miri and Hellevang, 2016;
(Gunter et al., 1998; Herzog, 2001). This technology aims to capture Peysson et al., 2014). In the dried zone, the majority of the pores
CO2 from industrial and power plant combustion processes, and then contain trapped brine that will be eventually evaporated while salinity
store the captured CO2 away in safe areas (e.g., depleted oil and gas of brine increases due to continuous CO2 injection (Miri and Hellevang,
reservoirs, unmineable coal seams, and deep saline aquifers) to prevent 2016). The brine drying-out results in salt precipitation and the pre-
it from reaching the atmosphere. Among various possibilities, CO2 in- cipitated salt blocks pores leading to a significant reduction in CO2
jection into geological formations has proven to be the most attractive injectivity. Salt precipitation near the wellbore in the field-scale pro-
method at a reasonable cost. Specifically, deep saline aquifers, defined jects has been reported at the Snøhvit field (Grude et al., 2014; Hansen
as porous and permeable reservoir rocks containing brine, located at et al., 2013) and the Ketzin pilot reservoir (Baumann et al., 2014), re-
800 to 2000 m from the surface level are widely available for CO2 spectively. Especially, a rapid pressure increase by salt precipitation
storage with the injection rate at more than 1 metric megaton per year was observed after 5000 h of CO2 injection, resulting in a significant
(Chow et al., 2003). decrease in injectivity at the Snøhvit field (Grude et al., 2014). Some
Although the storage capacity of these aquifers is known to be tre- laboratory studies have investigated the rate of brine drying-out under
mendous, the potential of this option is currently limited by drying-out different injection/fluid conditions and found important controlling


Corresponding author.
E-mail address: kimm@fau.edu (M. Kim).

https://doi.org/10.1016/j.ijggc.2019.02.017
Received 17 October 2018; Received in revised form 26 February 2019; Accepted 27 February 2019
1750-5836/ © 2019 Elsevier Ltd. All rights reserved.
S. Seo, et al. International Journal of Greenhouse Gas Control 83 (2019) 256–264

parameters for this phenomenon (Peysson et al., 2014; Pruess and the liquid−CO2 interface and the amount of gaseous CO2 at pore-scale.
Müller, 2009; Rathnaweera et al., 2016; Zhao and Cheng, 2017). CO2 The results show that the amount of mobile free-phase CO2 injected,
flow rates at 0.0083–0.83 mL/s (Miri et al., 2015) and 0.667 mL/s which is one of the main sources for the brine drying-out, can be re-
(Wang et al., 2017) have been used to study the mechanisms of salt duced by changing the frequency of sequential injection and pH levels
crystallization and the changes in formation properties, respectively. In of the injected solution.
addition, a range of salinity at 0.25–25% has been used (Tang et al.,
2015); note that the salt concentration of real brine has been found at 2. Materials and methods
0.71–35% (Michael et al., 2010).
Another technical challenge to the feasibility of CO2 sequestration 2.1. Microfluidic platforms
in saline aquifers is the slow reaction rate of carbonic acid (H2CO3)
formation from CO2 hydration (Bhaduri and Šiller, 2013; Seo et al., A microfluidic approach was used to visualize and quantify phase
2017, 2018b). This slow reaction rate can be negatively impacted fur- equilibria and fluid transport associated with brine evaporation and
ther by brine containing high salinity contents in natural formations. CO2 dissolution at the pore scale. The potential of this technique to
When CO2 dissolves into brine, the CO2-saturated brine sinks toward examine phase change analyses and fluid transport at micro- to na-
the bottom of the aquifers and thereby reduces the risk of CO2 leakage noscale has been proven in many studies due to the fast mass transfer at
since the density of CO2-saturated brine is greater than that of the these scales and its excellent capability for visualization of multiphase
original brine (Pau et al., 2010; Zhang and Song, 2014). The dissolved flow (Kim et al., 2013; Mastiani et al., 2018, 2017a, 2017b; Seo et al.,
CO2 will combine with metal cations leading to precipitation of mi- 2018a). In this approach, an experimental platform containing fluidic
nerals. Resultantly, properties of brine in porous media are changed to channels, called a microfluidic chip, is essential. Microfluidic chips
be the favorable condition for increasing permanent storage of CO2. were fabricated by laser ablation using a laser cutter (VersaLaser
Therefore, to accelerate CO2 dissolution into deep saline aquifers, a new VLS2.30, Universal Laser Systems, Inc.). Poly(methyl methacrylate)
methodology needs to be developed. (PMMA) was used for a chip material due to its high chemical resistance
The brine drying-out, salt precipitation, and slow CO2 hydration are and mechanical stability (Sher et al., 2017). As illustrated in Fig. 1, the
attributed mainly to a continuous injection of CO2 to the formation microfluidic chip consists of 1000 μm wide straight channels for sup-
fluid. Over the years, many methods other than the continuous injec- plying gaseous CO2 and isolated pore channels for trapping brine. The
tion of CO2 have been proposed to resolve these challenges. For mini- height of all channels is 1500 μm.
mizing the brine drying-out, one study suggested a preflush of the saline
formation with fresh water and it showed that this method delayed the 2.2. Experimental procedure
onset of salt precipitation and reduced its severity significantly. For
enhancing CO2 hydration, an ex-situ dissolution of CO2 via an external All tests were performed by pumping fluids under two different
reactor (Leonenko and Keith, 2008; Zendehboudi et al., 2011; Zirrahi conditions: continuous and sequential injection (Fig. 1a). For the con-
et al., 2013a, 2013b), a reverse gas-lift technology (Shafaei et al., tinuous injection of CO2 (Fig. 1b), once the microfluidic chip was filled
2012), a static mixer at the bottom of well (Zirrahi et al., 2013a, with brine at different salinities, a clamping device sealed the inlet port
2013b), the injection of nanosized CO2 bubbles (Uemura et al., 2016), for the solution to prevent backflow of trapped brine induced by in-
and catalyst addition to the injection fluid (Seo et al., 2018b, 2017) jection of CO2. Then, the CO2 gas with 99.9% purity (Airgas®, Miami,
have been proposed. Among these strategies, the injection of CO2 FL) was introduced to the microfluidic chip at different flow rates
bubbles, instead of the CO2 stream, has shown promising results to through the connected inlet port to a regulator mounted on a CO2 tank.
accelerate CO2 hydration remarkably (Leonenko and Keith, 2008; For the sequential injection of water with CO2, a fast switching me-
Uemura et al., 2016; Zendehboudi et al., 2011). One limitation of this chanism of a three-way solenoid valve operated by microcontroller-
method is that the pumping process to generate CO2 bubbles in external based (Arduino Uno) electronic control circuits (Fig. 1c) was employed.
reactors requires huge energy consumption. Although numerous studies Two inlets of this solenoid valve were connected to a syringe pump
have shown the important mechanisms associated with the brine (PHD ULTRA™ 4400, Harvard Apparatus, Natick, MA) and the CO2
drying-out and CO2 hydration, they still lack a practical strategy to regulator. This valve sequentially injected water at 0.03 mL/s from the
resolve these concerns. syringe pump and CO2 gas at 118 mL/s from the tank into the micro-
In this study, we present a practical methodology that could address fluidic platform. This ensures precisely controlled frequencies (0.1 to
these challenges: significant reduction of drying-out and acceleration of 0.5 Hz) of sequential injection of water with gaseous CO2. When gas-
CO2 hydration by sequential injection of water with CO2 into the for- eous CO2 was sequentially introduced with water, CO2 bubbles with a
mation. It is broadly known that the sequential injection of CO2-water diameter range of 40–150 μm were generated in the trapped brine. The
into a pipeline will generate various CO2 bubbles in their size due to the microbubble is defined as bubbles having a diameter less than few
gas-liquid two-phase flow (Abdulmouti, 2014; Das et al., 2009). In the hundred μm in this study similar to that in the previous literature in the
literature, it is well known that the fragmentation of large air cavities field of fluid physics (Parmar and Majumder, 2013). Once the tests for
by the turbulent flow can produce submillimeter bubbles (Rodríguez- CO2 bubble formation at different frequencies were completed, the
Rodríguez et al., 2015). Especially, after the liquid flow has established behavior of brine evaporation and the amount of mobile free-phase CO2
in a large reactor or tubing, microbubbles can be generated by injecting in brine were investigated at various pH levels of pumping water (pH 3,
a short burst of gas into the liquid flow using a solenoid valve (Millard pH 7, and pH 11). A high-speed camera (Fastec IL5S, Fastec Imaging,
et al., 2015; Ohl, 2001). We hypothesize that the sequential injection of CA) attached to an inverted microscope (IX73, Olympus Corporation,
water with CO2 into the aquifers increases the CO2 dissolution rate by Japan) was used to capture time-resolved images of the liquid-gas in-
producing microsized bubbles in the reservoirs and also effectively terface and the amount of trapped CO2 in brine.
decreases brine evaporation by supplying fresh water to salt-saturated
pores. This study aims to test this hypothesis and provide a new venue 2.3. Materials and experimental conditions
for geologic CO2 injection while minimizing the brine drying-out and
maximizing CO2 storage capacity. The time-dependent brine evapora- For the fabrication of microfluidic devices, PMMA sheets were
tion, resultant salt precipitation, and formation of CO2 bubbles were purchased from Good Fellow (Berwyn, PA). Sodium chloride (NaCl,
visualized at high-speed through a microfluidic technique. Micro- ≥99.5% purity), sodium hydroxide (NaOH, ≥98% purity), hydro-
fluidics allowed us to quantify the amounts of brine evaporation and chloric acid (HCl, 37%), and phosphate buffer powder were purchased
mobile free-phase CO2 successfully by measuring both the velocity of from Sigma − Aldrich (St. Louis, MO). One millimolar (mM) phosphate

257
S. Seo, et al. International Journal of Greenhouse Gas Control 83 (2019) 256–264

Fig. 1. (a) Schematic of the experimental setup (not to scale). (b) Continuous CO2 injection to the microfluidic chip with time-dependent brine evaporation at the
trapped channel. (c) Sequential water−CO2 injection to the same channel generating CO2 microbubbles.

buffer solution used in all tests was prepared using deionized water determined through the projected area of CO2 bubble. Then, the CO2
produced from a Millipore Milli − Q integral 5 purification system bubble volume (V = (4/3)π(D/2)3) was determined using the average
(Millipore, Bedford, MA). Brine evaporation experiments were per- diameter (D) of CO2 bubbles.
formed at concentrations of 0%, 5%, 10%, 20%, and 30% NaCl added to
1 mM phosphate buffer solution at different injection rates of dry gas- 3. Results and discussion
eous CO2 (68, 97 and 118 mL/s). Then, the highest salinity at 30% with
the highest injection rate of dry CO2 at 118 mL/s was used to in- 3.1. Continuous injection of CO2
vestigate how effectively sequential water injection with CO2 can re-
duce brine evaporation and the amount of gaseous CO2 in brine. The pH The experiments on the water evaporation at various salinity levels
of the sequential water was titrated to desired levels that were created and CO2 flow rates were performed to characterize how different fluid
by adding either one M of HCl or NaOH while monitoring with a pH and flow rate conditions in deep saline aquifers influence the brine
meter (Accumet AE150, Fisher Scientific, Pittsburgh, PA). All experi- drying-out. Fig. 2 shows time-lapse optical microscopy images of
ments were performed at room temperature of 25 °C and atmospheric changes in the gas-liquid interface at an isolated pore channel. The
pressure. An atmospheric pressure condition was used at the outlet. interface was formed when dry gaseous CO2 was continuously injected
at 68, 97 and 118 mL/s at different salinities (0, 5, 10, 20, and 30%
2.4. Image processing for evaporated brine and gaseous CO2 NaCl). In these conditions, CO2 velocities are 23.51, 33.25, and
40.72 m/s, which represent assumed conditions around injection wells
The volumes of evaporated brine and gaseous CO2 in the isolated at the CO2 storage site. In general, as the volume flow rate increases, the
pore channel were analyzed on the ImageJ software, an open-source speed of interface retraction also increases; for example, compare time-
image processing program. To measure brine evaporation in various dependent changes in the location of the interface at 0% salinity in
conditions, the changes in the interface between CO2 gas and trapped Fig. 2a and c. This behavior was quantitatively analyzed by tracking
brine in the isolated pore channel corresponding to water evaporation changes in volume of evaporated brine at a T-junction (Fig. 3). Ap-
were tracked by the ImageJ. In addition, the volume fraction of the proximately 2.2 μL of volume of pure water in the channel was eva-
mobile free-phase CO2 in trapped brine was determined by changes in porated within 120, 46, and 4.6 s at 68, 97 and 118 mL/s of CO2 flow
its volume at 0.04 s intervals. The CO2 bubble diameter (D) was rates, respectively. This analysis shows that the amount of brine

258
S. Seo, et al. International Journal of Greenhouse Gas Control 83 (2019) 256–264

Fig. 2. Sequential images for changes in the CO2-water interface location at a T-junction at different salinities (0, 5, 10, 20, and 30% NaCl) with different CO2
injection rates at (a) 68, (b) 97, and (c) 118 mL/s.

evaporated exponentially increased as CO2 flow rate increased from 68


to 118 mL/s. In general, an increase in air velocity over water leads to
an increase in rate of escaping water molecules (Arnal et al., 2005). In
line with this phenomenon, we observed forced convective mass
transfer between horizontal CO2 gas flow and water increases with
increasing gaseous CO2 flow rates. Reynolds numbers (Re) at different
flow rates were determined to be 4464.37, 6313.92, and 7732.42 using
8.36 × 10−6 m2/s of kinematic viscosity of CO2 at 25 °C, 1.58 × 10-3 m
of tubing diameter, and CO2 velocities at 23.51, 33.25, and 40.72 m/s.
For the impact of salinity on brine evaporation, the kinetics of brine
evaporation at different salinity levels are plotted in Fig. 4. This figure
is determined by different slopes for the relationship between evapo-
rated brine and the elapsed time (Fig. 3). As the concentration of sali-
nity increases up to 10%, the evaporation rate decreases in all flow rate
conditions. However, by increasing the salinity from 10% to 30%, an
increase in the evaporation rate is observed in three flow rate condi-
tions. The highest rate of evaporation was determined to be 0.57 μL/s
Fig. 4. A diagram of changes in the evaporation rate of brine as a function of
under 30% salinity at the highest CO2 flow rate (118 mL/s). This de-
the salinity (0, 5, 10, 20, and 30% NaCl) with different CO2 injection rates (68,
crease-to-increase behavior can be explained by an activity coefficient
97, and 118 mL/s) and the graph (◊) of the activity coefficients of NaCl solution
of NaCl solution controlling brine evaporation dynamics. The activity at 25 °C.
coefficient is used in thermodynamics to explain the behavior of che-
mical substances. As the activity coefficient decreases, the energy re-
quired to separate constituent molecules increases. Conversely, when 0.5067 IM
log = + 0.0316 2IM log(1 + 0.036M )
the activity coefficient increases, molecules have a strong repelling 1 + 0.837 2IM
force, and thereby less energy is required to separate them (Pitzer et al.,
where IM, M, and γ are the ionic strength (mol/L), the concentration
1984). Evaporation occurs when molecules in a liquid gain enough
(mol/L), and the activity coefficient of NaCl solution at 25 °C, respec-
energy to overcome an attraction force from other molecules. The ac-
tively. We estimated each value of the activity coefficient at different
tivity coefficients of brine at different salinities can be calculated from
salinities ranging from 0% to 30% to characterize the salinity effect on
the electromotive force data by Hückel's equation (Harned and Nims,
the evaporation rate. A decrease in the activity coefficient while in-
1932).
creasing the salinity concentration up to 7% led to a decrease in brine
evaporation into the CO2 gas phase (Fig. 4). In contrast, increasing

Fig. 3. Variations of evaporated brine vs. elapsed time at different salinities and CO2 flow rates. (Error bars represent the standard error of the mean from triplicated
experiments in each condition.)

259
S. Seo, et al. International Journal of Greenhouse Gas Control 83 (2019) 256–264

Fig. 5. Representative time-lapse optical microscopy images showing the generation and coalescence of CO2 microbubbles during sequential water−CO2 injection.

salinity concentration from 7% to 30% displays an increase in the ac- microbubbles effectively with a large surface-to-volume ratio leading to
tivity coefficient to 0.869, resulting in a proportional increase in the significant improvement of a CO2 solubility trapping process because
evaporation rate of brine. The observations of our test results are in a the mass transfer of gaseous CO2 into water will further increase with
good agreement with the theory of the activity coefficient. These mi- larger interface areas between these two phases. As in Fig. 5, due to the
crofluidic tests confirm that continuous injection of gaseous CO2 into physical agitation, number and size of CO2 bubbles change over time.
deep saline aquifers may lead to an increase in risks of the brine drying- The figure shows that the onset of CO2 bubble formation is observed at
out. The extreme drying-out creates salt precipitation behind and a few locations at t = 1 s. The CO2 bubbles are accumulated im-
causes formation damage. We observed that the degree of brine drying- mediately near the T-junction and start to coalesce due to the collision
out increases as CO2 flow rate increases at salinity greater than 10%. between the CO2 bubbles (see Video 2). It should be noted that mi-
crobubbles in an aqueous solution are negatively charged under a wide
3.2. Sequential water injection with CO2 range of pH (Takahashi, 2005). However, as increasing the ionic
strength in the aqueous phase, the reduction in the zeta (ζ) potential of
We hypothesized that sequential injection of water with CO2 ad- microbubbles was reported (Takahashi, 2005). As a result, CO2 mi-
dresses the challenge with the brine drying-out and provides another crobubbles in 30% NaCl near the T-junction are easily coalesced to
benefit to accelerate the rate of CO2 hydration. To test the feasibility of minimize their surface areas when they are in touch. The size of CO2
the sequential injection strategy, microfluidic experiments were per- bubbles apart from the T-junction is significantly decreased due to ef-
formed in a simple T-junction microchannel. Fig. 5 shows re- fective CO2 hydration.
presentative images of CO2 microbubble generation in the micro- As typical natural saline aquifers are characterized as random
channel taken by an optical microscope from selected time points porous structures, the directional water−CO2 injection into pore
(t = 1, 20, and 50 s) at a controlled frequency of 0.1 Hz. As pure water structures is certainly of interest. The pore-throat and pore sizes of ty-
and CO2 were sequentially injected, CO2 microbubbles were generated pical sandstones range from 1 to 500 μm in conventional reservoirs
due to the nature of two-phase flow (Abdulmouti, 2014; Das et al., (Nelson, 2009; Radlinski et al., 2004; Wardlaw and Cassan, 1979). To
2009). Locations of the gas-liquid interface were maintained constant examine the behaviors of CO2 bubble generation in different geome-
near the T-junction during sequential injection, which consequently tries, transient CO2 bubble formation during sequential water and CO2
delayed the evaporation rate of water. Not only does decreasing the injection was recorded at different pore angles (30°, 90°, and 120°,
brine drying-out and therefore reducing a risk of salt precipitation, but Fig. 6 and see Video 3). The results show that an insignificant variation
the sequential injection of pure water−CO2 also increases CO2 hydra- in the amount of CO2 bubbles was observed in the pore channel at these
tion into brine. The physical absorption of gaseous CO2 by the liquid angles.
phase is determined by Henry’s law.
Pco2 = kH [CO2 ]aq 3.3. Frequency effect

where kH is Henry’s constant for CO2 in pure water (3.45 × 103 kPa L/ The high solubility coefficient of CO2 makes it immediately diffuse
mol), [CO2]aq is the concentration of dissolved CO2 in the aqueous into an aqueous phase when it is in contact with water. In addition, as
phase, and PCO2 is the partial pressure of the CO2. The rate for this aforementioned the high surface-to-volume ratio will improve CO2
physical absorption can be increased further by agitation in the liquid hydration into water. To further enhance CO2 gas dissolution and safe
phase (Noyes et al., 1996). The agitation can generate CO2 storage potential, generation of a much smaller volume of CO2 bubbles

Fig. 6. Representative microscopy images showing the generation and coalescence of CO2 microbubbles at different pore angles (30, 90, and 120° between an inlet
channel with isolated pore channel) during sequential water−CO2 injection.

260
S. Seo, et al. International Journal of Greenhouse Gas Control 83 (2019) 256–264

Fig. 7. Time-dependent images of CO2 microbubble generation and coalescence near the T-junction at different water−CO2 injection frequencies (0.1, 0.2, and
0.5 Hz).

is important. We found that the frequency of water-gas injection plays a trapped brine (VCO2/VChannel) in the microchannel at different fre-
role in varying the gas bubble sizes. Fig. 7 shows the total volume of quencies. In agreement with the observation in Fig. 7, as the frequency
CO2 bubbles generated at t = 0˜60 s in a T-junction channel at fre- increases, the volume of trapped water is slowly replaced by that of CO2
quencies of 0.1, 0.2, and 0.5 Hz. At a frequency of 0.5 Hz, number of bubbles. Particularly at the 0.5 Hz, the volume of CO2 bubbles at 60 s
CO2 micro-bubbles was increased as time increases while their sizes are was estimated to be approximately 2-fold smaller than that in the
less than 100 μm at t = 60 s. On the other hand, at 0.1 Hz number of 0.1 Hz condition. These significant differences in the volume of CO2
CO2 bubbles was decreased while their sizes are greater than 500 μm at bubbles trapped in brine at different frequencies can be characterized
t = 60 s. This implies that an increase in frequency leads to a decrease by bubbly two-phase flow. The volume of CO2 bubbles in bubbly two-
in the volume of CO2 bubbles occupying the brine at the channel (VCO2/ phase flow is likely to be dominated by the ratio of liquid-gas flow rates.
VChannel) owing to improved CO2 hydration. Fraga and Stoesser (Fraga and Stoesser, 2016) revealed that smaller
Fig. 8 shows the change in ratio of the volume of CO2 bubbles to bubbles are generated when a high flow rate of an aqueous phase and a

261
S. Seo, et al. International Journal of Greenhouse Gas Control 83 (2019) 256–264

results can be explained by the behavior of the effective Henry’s con-


stant (Seinfeld and Pandis, 2016). The total quantity of dissolved CO2 in
brine includes the amount of all species of aqueous CO2, the inter-
mediate bicarbonate (HCO3−) form, and the equilibrium carbonate
(CO32-) form in water. First, physical absorption of CO2 into brine is
governed by Henry’s law.

PCO2
CO2(g ) + H2 O CO2(aq) , kH =
[CO2 ]aq

CO2 (aq) reacts with water to form carbonic acid (H2CO3), then
dissociates into H+ and HCO3− ions at the intermediate stage.
CO2(aq) + H2 O H2 CO3

H2 CO3 H+ + HCO3 ,
H+ HCO3 7
K1 = = 4.45 × 10
Fig. 8. Variations in the ratio of the volume of CO2 microbubbles to the total CO2(aq)
volume of brine at the isolated channel at different injection frequencies. The
inset figures represent images of the channel taken at t = 60 s. Then, HCO3− dissociates further into H+ and CO32- ions.

HCO3 H+ + CO32 ,
low flow rate of a gaseous phase are mixed. In our tests, when a valve is
H+ CO32
operating at low frequency (0.1 Hz), CO2 gas is compressed, thereby K2 = = 4.69 × 10 11

creating a high flow rate of a CO2 phase. Due to the incompressibility of HCO3
the continuous aqueous phase (i.e., water), its flow rate remains con- The total amount of CO2 in the aqueous phase can be calculated by
stant, which thereby increases the ratio of liquid-gas flow rates in the solving for the equilibrium constants (K1 and K2) with Henry’s constant.
system. The generation of much smaller CO2 bubbles at 0.5 Hz fre-
quency accelerates CO2 hydration. These results imply that sequential PCO2 K1 KK
CO2(total) = CO2(aq) + HCO3 + CO32 = 1 + + 1+ 22
water injection with CO2 at high frequency would facilitate CO2 hy- kH [H+] [H ]
dration in brine leading to rapid mineral precipitation. Another im-
An effective Henry’s constant, kH*, which accounts for the total
portant observation is that a decreasing drying-out effect by sequential
dissolved CO2, can be described by the following equation:
water injection with gaseous CO2 leads to 15-fold slower brine eva-
poration rate than continuous injection of CO2 (Fig. 2a vs. Fig. 7). PCO2 kH
kH * = =
3.4. The pH effect
CO2(total)
(1 + K1
[H +]
+
K1 K2
[H +]2 )
Each value of the effective Henry’s constant at different pH levels
The pH of an aqueous solution is considered as another parameter was determined to identify the pH effect on VCO2/VChannel (Fig. 9). In
which controls the CO2 solubility trapping process (Seinfeld and Pandis, theory, CO2 dissolution into water increases as the value of effective
2016). Fig. 9 shows a time-dependent ratio of the volume of CO2 Henry’s constant decreases; for example, kH* = 1.36 × 10−2 at pH 11,
bubbles to trapped brine (VCO2/VChannel) at different pH levels. When pH 6.33 × 102 at pH 7, and 3.44 × 103 kPa L/mol at pH 3, respectively.
changes from pH 7 to pH 3, the relative volume of CO2 bubbles at the Taking this fact into consideration, Fig. 9 shows significant increases in
isolated pore channel increases. In contrast, no significant differences in the value of effective Henry’s constant from pH 7 to pH 3 led to a de-
VCO2/VChannel were found between pH 7 and pH 11. These experimental crease in the CO2 dissolution. In contrast, as noted earlier, no sub-
stantial difference was observed at pH 7 and 11.

3.5. Discussion and outlook

The impacts of sequential CO2 and water injection into geologic


aquifers are significant with an important perspective of water eva-
poration and resultant brine drying-out when compared with the con-
ventional methods. Up to this point, several methods other than the
direct injection of dry CO2 through injection wells have been proposed
to improve formation damage during carbon sequestration due to the
brine drying-out. One method is an ex-situ approach, where brine is
extracted from the saline aquifers, sparged with pressurized gaseous
CO2 to produce CO2-saturated brine in reservoir conditions, and then
reinjected through the pipeline into the original aquifer (Leonenko and
Keith, 2008; Zendehboudi et al., 2011). The main benefit of this
methodology is that CO2 dissolution into an aqueous phase can be ac-
celerated by the generation of CO2 bubbles in a mixing chamber due to
the gas-liquid two-phase flow. When gaseous CO2 is introduced into
water, two-phase flow occurs along with generation of CO2 bubbles in a
Fig. 9. Bottom-X left-Y: Variations in the ratio of the volume of CO2 micro- continuous aqueous phase (Abdulmouti, 2014; Das et al., 2009), ulti-
bubble to the total volume of brine at the isolated pore channel at 0.5 Hz fre- mately enhancing the CO2 trapping within deep saline aquifers. Al-
quency at different pH. Top-X right-Y: A diagram of effective Henry’s constant though this approach has shown a potential for great reduction of a
as a function of pH. mobile CO2 phase causing dreadful leakage, the pumping process that

262
S. Seo, et al. International Journal of Greenhouse Gas Control 83 (2019) 256–264

circulates brine from the reservoir to the ground pipeline and the amounts of brine evaporation and mobile free-phase CO2 and proven to
pressurization of CO2 in aquifer conditions requires significant energy be a potential solution to mitigate the concerns about the brine drying-
consumption, making this approach impractical. In addition to this out. We found that the brine evaporation rate in the new strategy was
method, simultaneous injection of CO2 and geologic saline water into approximately 15-fold slower than the conventional gaseous CO2 in-
reservoirs has been considered as a substitute of the ex-situ approach jection. In addition, the new strategy generated large surface-to-volume
(Hassanzadeh et al., 2009; Leonenko and Keith, 2008; Shafaei et al., ratio microbubbles and these CO2 bubbles effectively accelerated their
2012). It has been considered that this method not only increases the dissolution in brine due to the increased gas-liquid interfaces. To fur-
possibility of permanent storage of CO2 but also reduces economic ther enhance CO2 dissolution, operating condition parameters, i.e., in-
strain notably, as the energy cost associated with this process is less jection frequency and solution pH, were controlled. The results show
than 20% of the cost for compressing the CO2 to reservoir conditions that sequential of water with CO2 injection at a low frequency at higher
(Leonenko and Keith, 2008). However, many bottlenecks restricting basicity accelerated CO2 dissolution effectively and decreased the vo-
field applications of this technology still remain, including the re- lume fraction of CO2 bubbles occupying the pore.
quirement of the additional pipeline installation for brine circulation.
On the other hand, the proposed sequential water injection with gas- Acknowledgments
eous CO2 could be a solution by combining advantages of the afore-
mentioned ex-situ approach with simultaneous injection of CO2 with The authors gratefully acknowledge the funding provided by Janke
water. The sequential injection of water via a single pipeline can Research Fund at Florida Environmental Studies. We also thank
eliminate the need for an additional pipeline for economical tech- Heather Crawford for her help with laboratory experiments; Babak
nology. Mosavati and Minh Nguyen for their comments on the manuscript.
In the context of a methodology, a microfluidic-based approach
provides new perspectives regarding the brine drying-out at the pore Appendix A. Supplementary data
scale in geologic aquifers in a simple and cost-effective manner.
However, it should be noted that there still exist opportunities to im- Supplementary material related to this article can be found, in the
prove our current results in terms of the material and surface properties online version, at doi:https://doi.org/10.1016/j.ijggc.2019.02.017.
of microfluidic models and a variety of ions other than NaCl. In addi-
tion, although the ratio of the volume of CO2 bubbles to trapped brine References
(VCO2/VChannel) strongly depends on the flow rate ratio of the two
phases, the operation pressure is still not relevant to the actual oper- Abdulmouti, H., 2014. Multiphase flow, bubble plume, bubble, surface flow, turbulence,
ating condition. Temperature is an important parameter to change the buoyant flow, free surface flow, and bubbly flow. Am. J. Fluid Dyn. 47.
Arnal, J.M., Sancho, M., Iborra, I., Gozálvez, J.M., Santafé, A., Lora, J., 2005.
rates of brine evaporation and CO2 solubility. The reservoir-specific Concentration of brines from RO desalination plants by natural evaporation.
temperature conditions (e.g., 40 °C (Rathnaweera et al., 2016) and Desalination 182, 435–439. https://doi.org/10.1016/j.desal.2005.02.036.
62 °C (Holt et al., 1995)) can decrease the brine drying-out and CO2 Baumann, G., Henninges, J., De Lucia, M., 2014. Monitoring of saturation changes and
salt precipitation during CO2 injection using pulsed neutron-gamma logging at the
dissolution rates. The water content of CO2 at room temperature of Ketzin pilot site. Int. J. Greenh. Gas Control 28, 134–146. https://doi.org/10.1016/j.
25 °C and atmospheric pressure is approximately 23 times higher than ijggc.2014.06.023.
that at 10 MPa and 40 °C (Salari et al., 2011). Due to this higher CO2 Bhaduri, G.A., Šiller, L., 2013. Nickel nanoparticles catalyse reversible hydration of
carbon dioxide for mineralization carbon capture and storage. Catal. Sci. Technol. 3,
water content, the brine drying-out in our experiments is possibly 1234. https://doi.org/10.1039/c3cy20791a.
greater than that in the actual reservoir-specific condition. Never- Chow, J.C., Watson, J.G., Herzog, A., Benson, S.M., Hidy, G.M., Gunter, W.D., Penkala,
theless, the microfluidic approach, for the first time, demonstrates the S.J., White, C.M., 2003. Separation and capture of CO2 from large stationary sources
and sequestration in geological formations. J. Air Waste Manag. Assoc. 53,
efficacy of a new novel injection strategy for CO2 into geologic for-
1172–1182. https://doi.org/10.1080/10473289.2003.10466274.
mations. To test the feasibility of the proposed sequential injection Das, A.K., Das, P.K., Thome, J.R., 2009. Transition of Bubbly Flow in Vertical Tubes: New
approach to the actual saline aquifers, water evaporation and the brine Criteria Through CFD Simulation. J. Fluids Eng. 131, 091303. https://doi.org/10.
drying-out experiments in the 2D and 3D porous media that mimic the 1115/1.3203205.
Fraga, B., Stoesser, T., 2016. Influence of bubble size, diffuser width, and flow rate on the
natural aquifers in terms of surface properties and geometry will be integral behavior of bubble plumes. J. Geophys. Res. Oceans 121, 3887–3904.
performed at higher pressures. The complex nature of these random https://doi.org/10.1002/2015JC011381.
networks will offer dynamic physicochemical processes close to the Fuller, R.C., Prevost, J.H., Piri, M., 2006. Three-phase equilibrium and partitioning cal-
culations for CO2 sequestration in saline aquifers: EQUILIBRIUM AND
actual phenomena and direct visualization of these interactions would PARTITIONING FOR CO2 SEQUESTRATION. J. Geophys. Res. Solid Earth 111.
provide new understandings of the pertinent multiphase interplay in https://doi.org/10.1029/2005JB003618.
geologic formations. Grude, S., Landrø, M., Dvorkin, J., 2014. Pressure effects caused by CO2 injection in the
Tubåen Fm., the Snøhvit field. Int. J. Greenh. Gas Control 27, 178–187. https://doi.
org/10.1016/j.ijggc.2014.05.013.
4. Conclusion Gunter, W.D., Wong, S., Cheel, D.B., Sjostrom, G., 1998. Large CO2 sinks: their role in the
mitigation of greenhouse gases from an international, national (Canadian) and pro-
vincial (Alberta) perspective. Appl. Energy 19.
We performed pore-scale experiments of geofluid dynamics oc- Hansen, O., Gilding, D., Nazarian, B., Osdal, B., Ringrose, P., Kristoffersen, J.-B., Eiken,
curred in deep saline aquifers when gaseous CO2 is sequentially in- O., Hansen, H., 2013. Snøhvit: The History of Injecting and Storing 1 Mt CO2 in the
troduced with water using microfluidic platforms. A simple microfluidic Fluvial Tubåen Fm. Energy Procedia 37, 3565–3573. https://doi.org/10.1016/j.
egypro.2013.06.249.
imaging technique enabled quantification of a volume fraction of in-
Harned, H.S., Nims, L.F., 1932. The thermodynamic properties of aqueous sodium
jected CO2 and resultant evaporated brine to open a new opportunity chloride solutions from 0 to 40°. J. Am. Chem. Soc. 54, 423–432. https://doi.org/10.
for strategic carbon sequestration plans into deep saline aquifers. 1021/ja01341a002.
Specifically, the principal factors contributing to the brine drying-out Hassanzadeh, H., Pooladi-Darvish, M., Keith, D.W., 2009. Accelerating CO2 dissolution in
saline aquifers for geological storage — mechanistic and sensitivity studies. Energy
by traditional dry CO2 injection can be identified through quantifica- Fuels 23, 3328–3336. https://doi.org/10.1021/ef900125m.
tion of the amount of evaporated brine at the pore-scale. The results Herzog, H.J., 2001. Peer reviewed: what future for carbon capture and sequestration?
from this observation are well aligned with the conventional theory of Environ. Sci. Technol. 35, 148A–153A. https://doi.org/10.1021/es012307j.
Holt, T., Jensen, J.-I., Lindeberg, E., 1995. Underground storage of CO2 in aquifers and oil
the activity coefficient and offer reliable modeling of interactions be- reservoirs. Energy Convers. Manage. 36, 535–538. https://doi.org/10.1016/0196-
tween CO2-brine occurred in actual saline aquifers. To overcome the 8904(95)00061-H.
brine drying-out challenge, this study suggests a new strategy of se- Kim, M., Sell, A., Sinton, D., 2013. Aquifer-on-a-Chip: understanding pore-scale salt
precipitation dynamics during CO2 sequestration. Lab Chip 13, 2508. https://doi.
quential water injection with CO2. This scenario was experimentally org/10.1039/c3lc00031a.
tested through the microfluidic technique for quantification of the

263
S. Seo, et al. International Journal of Greenhouse Gas Control 83 (2019) 256–264

Leonenko, Y., Keith, D.W., 2008. Reservoir engineering to accelerate the dissolution of CO srep19362.
2 stored in aquifers. Environ. Sci. Technol. 42, 2742–2747. https://doi.org/10.1021/ Rodríguez-Rodríguez, J., Sevilla, A., Martínez-Bazán, C., Gordillo, J.M., 2015. Generation
es071578c. of Microbubbles with Applications to Industry and Medicine. Annu. Rev. Fluid Mech.
Mastiani, M., Mosavati, B., Kim, M.(Mike), 2017a. Numerical simulation of high inertial 47, 405–429. https://doi.org/10.1146/annurev-fluid-010814-014658.
liquid-in-gas droplet in a T-junction microchannel. RSC Adv. 7, 48512–48525. Salari, H., Hassanzadeh, H., Gerami, S., Abedi, J., 2011. On Estimating the Water Content
https://doi.org/10.1039/C7RA09710G. of CO2 in Equilibrium With Formation Brine. Pet. Sci. Technol. 29, 2037–2051.
Mastiani, M., Seo, S., Jimenez, S.M., Petrozzi, N., Kim, M.M., 2017b. Flow regime map- https://doi.org/10.1080/10916461003681778.
ping of aqueous two-phase system droplets in flow-focusing geometries. Colloids Surf. Seinfeld, J.H., Pandis, S.N., 2016. Atmospheric Chemistry and Physics: From Air Pollution
Physicochem. Eng. Asp. 531, 111–120. https://doi.org/10.1016/j.colsurfa.2017.07. to Climate Change. John Wiley & Sons.
083. Seo, S., Nguyen, M., Mastiani, M., Navarrete, G., Kim, M., 2017. Microbubbles loaded
Mastiani, M., Seo, S., Mosavati, B., Kim, M., 2018. High-throughput aqueous two-phase with nickel nanoparticles: a perspective for carbon sequestration. Anal. Chem. 89,
system droplet generation by oil-free passive microfluidics. ACS Omega 3, 10827–10833. https://doi.org/10.1021/acs.analchem.7b02205.
9296–9302. https://doi.org/10.1021/acsomega.8b01768. Seo, S., Mastiani, M., Mosavati, B., Peters, D.M., Mandin, P., Kim, M., 2018a. Performance
Michael, K., Golab, A., Shulakova, V., Ennis-King, J., Allinson, G., Sharma, S., Aiken, T., evaluation of environmentally benign nonionic biosurfactant for enhanced oil re-
2010. Geological storage of CO2 in saline aquifers—A review of the experience from covery. Fuel 234, 48–55. https://doi.org/10.1016/j.fuel.2018.06.111.
existing storage operations. Int. J. Greenh. Gas Control 4, 659–667. https://doi.org/ Seo, S., Perez, G.A., Tewari, K., Comas, X., Kim, M., 2018b. Catalytic activity of nickel
10.1016/j.ijggc.2009.12.011. nanoparticles stabilized by adsorbing polymers for enhanced carbon sequestration.
Millard, T.P., Endrizzi, M., Everdell, N., Rigon, L., Arfelli, F., Menk, R.H., Stride, E., Olivo, Sci. Rep. 8. https://doi.org/10.1038/s41598-018-29605-1.
A., 2015. Evaluation of microbubble contrast agents for dynamic imaging with x-ray Shafaei, M.J., Abedi, J., Hassanzadeh, H., Chen, Z., 2012. Reverse gas-lift technology for
phase contrast. Sci. Rep. 5. https://doi.org/10.1038/srep12509. CO2 storage into deep saline aquifers. Energy 45, 840–849. https://doi.org/10.1016/
Miri, R., Hellevang, H., 2016. Salt precipitation during CO2 storage—a review. Int. J. j.energy.2012.07.007.
Greenh. Gas Control 51, 136–147. https://doi.org/10.1016/j.ijggc.2016.05.015. Sher, M., Zhuang, R., Demirci, U., Asghar, W., 2017. Paper-based analytical devices for
Miri, R., van Noort, R., Aagaard, P., Hellevang, H., 2015. New insights on the physics of clinical diagnosis: recent advances in the fabrication techniques and sensing me-
salt precipitation during injection of CO2 into saline aquifers. Int. J. Greenh. Gas chanisms. Expert Rev. Mol. Diagn. 17, 351–366. https://doi.org/10.1080/14737159.
Control 43, 10–21. https://doi.org/10.1016/j.ijggc.2015.10.004. 2017.1285228.
Nelson, P.H., 2009. Pore-throat sizes in sandstones, tight sandstones, and shales. AAPG Takahashi, M., 2005. ζ Potential of microbubbles in aqueous solutions: electrical prop-
Bull. 93, 329–340. https://doi.org/10.1306/10240808059. erties of the gas−water Interface. J. Phys. Chem. B 109, 21858–21864. https://doi.
Noyes, R.M., Rubin, M.B., Bowers, P.G., 1996. Transport of carbon dioxide between the org/10.1021/jp0445270.
Gas Phase and water under well-stirred conditions: rate constants and mass accom- Tang, Y., Yang, R., Du, Z., Zeng, F., 2015. Experimental study of formation damage caused
modation coefficients. J. Phys. Chem. 100, 4167–4172. https://doi.org/10.1021/ by complete water vaporization and salt precipitation in sandstone reservoirs.
jp952382e. Transp. Porous Media 107, 205–218. https://doi.org/10.1007/s11242-014-0433-1.
Ohl, C.D., 2001. Generator for single bubbles of controllable size. Rev. Sci. Instrum. 72, Uemura, S., Matsui, Y., Kondo, F., Tsushima, S., Hirai, S., 2016. Injection of nanosized
252–254. https://doi.org/10.1063/1.1329900. CO2 droplets as a technique for stable geological sequestration. Int. J. Greenh. Gas
Parmar, R., Majumder, S.K., 2013. Microbubble generation and microbubble-aided Control 45, 62–69. https://doi.org/10.1016/j.ijggc.2015.12.011.
transport process intensification—A state-of-the-art report. Chem. Eng. Process. Wang, L., Yao, B., Xie, H., Kneafsey, T.J., Winterfeld, P.H., Yin, X., Wu, Y.-S., 2017.
Process Intensif. 64, 79–97. https://doi.org/10.1016/j.cep.2012.12.002. Experimental investigation of injection-induced fracturing during supercritical CO2
Pau, G.S.H., Bell, J.B., Pruess, K., Almgren, A.S., Lijewski, M.J., Zhang, K., 2010. High- sequestration. Int. J. Greenh. Gas Control 63, 107–117. https://doi.org/10.1016/j.
resolution simulation and characterization of density-driven flow in CO2 storage in ijggc.2017.05.006.
saline aquifers. Adv. Water Resour. 33, 443–455. https://doi.org/10.1016/j. Wardlaw, N.C., Cassan, J.P., 1979. Oil recovery efficiency and the rock-pore properties of
advwatres.2010.01.009. some sandstone reservoirs. Bull. Can. Pet. Geol. 27, 117–138.
Peysson, Y., André, L., Azaroual, M., 2014. Well injectivity during CO2 storage operations Zendehboudi, S., Khan, A., Carlisle, S., Leonenko, Y., 2011. Ex Situ dissolution of CO2: a
in deep saline aquifers—part 1: Experimental investigation of drying effects, salt new engineering methodology based on mass-transfer perspective for enhancement
precipitation and capillary forces. Int. J. Greenh. Gas Control 22, 291–300. https:// of CO2 sequestration. Energy Fuels 25, 3323–3333. https://doi.org/10.1021/
doi.org/10.1016/j.ijggc.2013.10.031. ef200199r.
Pitzer, K.S., Peiper, J.C., Busey, R.H., 1984. Thermodynamic properties of aqueous so- Zhang, D., Song, J., 2014. Mechanisms for geological carbon sequestration. Procedia
dium chloride solutions. J. Phys. Chem. Ref. Data 13, 1–102. https://doi.org/10. IUTAM 10, 319–327. https://doi.org/10.1016/j.piutam.2014.01.027.
1063/1.555709. Zhao, R., Cheng, J., 2017. Effects of temperature on salt precipitation due to formation
Pruess, K., Müller, N., 2009. Formation dry-out from CO2 injection into saline aquifers: 1. dry-out during CO2 injection in saline aquifers: modeling and Analysis: effects of
Effects of solids precipitation and their mitigation: formation dry-out from CO2 in- temperature on salt precipitation due to formation dry-out during CO2 injection.
jection, 1. Water Resour. Res. 45. https://doi.org/10.1029/2008WR007101. .Greenh. Gases Sci. Technol. 7, 624–636. https://doi.org/10.1002/ghg.1672.
Radlinski, A.P., Ioannidis, M.A., Hinde, A.L., Hainbuchner, M., Baron, M., Rauch, H., Zirrahi, M., Hassanzadeh, H., Abedi, J., 2013a. Modeling of CO2 dissolution by static
Kline, S.R., 2004. Angstrom-to-millimeter characterization of sedimentary rock mi- mixers using back flow mixing approach with application to geological storage.
crostructure. J. Colloid Interface Sci. 274, 607–612. https://doi.org/10.1016/j.jcis. Chem. Eng. Sci. 104, 10–16. https://doi.org/10.1016/j.ces.2013.08.056.
2004.02.035. Zirrahi, M., Hassanzadeh, H., Abedi, J., 2013b. The laboratory testing and scale-up of a
Rathnaweera, T.D., Ranjith, P.G., Perera, M.S.A., 2016. Experimental investigation of downhole device for CO2 dissolution acceleration. Int. J. Greenh. Gas Control 16,
geochemical and mineralogical effects of CO2 sequestration on flow characteristics of 41–49. https://doi.org/10.1016/j.ijggc.2013.02.020.
reservoir rock in deep saline aquifers. Sci. Rep. 6. https://doi.org/10.1038/

264

Das könnte Ihnen auch gefallen