Journal of Natural Gas Science and Engineering 39 (2017) 133e142

Contents lists available at ScienceDirect

Journal of Natural Gas Science and Engineering

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

Evaluation of a novel gas-liquid contactor/reactor system for natural

gas applications
Muftah H. El-Naas a, *, Ameera F. Mohammad b, Mabruk I. Suleiman c,
Mohamed Al Musharfy c, Ali H. Al-Marzouqi b
a
Gas Processing Center, College of Engineering, Qatar University, Doha, Qatar
b
Chemical and Petroleum Engineering Department, UAE University, Al Ain, United Arab Emirates
c
Takreer Research Center, Abu Dhabi, United Arab Emirates

a r t i c l e i n f o a b s t r a c t

Article history: In this study, a novel system for contacting gases and liquids, which may be suitable for many appli-
Received 1 October 2016 cations involving gas-liquid contact or gas-liquid reactions, has been developed and characterized. The
Received in revised form system consists of a vertical vessel with gas and liquid ports and inert particles. The gas is injected
16 January 2017
through a single orifice at the bottom of the vessel and leaves at the top, while the liquid feed is
Accepted 31 January 2017
Available online 1 February 2017
introduced at the top, and the effluent leaves at the bottom of the vessel through a vertical tube inserted
from the top all the way to the bottom. The system involves the use of inert particles that create circular
motion within the vessel, enhancing mixing and providing high gaseliquid interfacial area for effective
Keywords:
Gas-liquid reactions
mass transfer. The contactor system was evaluated for the capture of CO2 through reactions with
Inert particles ammonium hydroxide and saline wastewater, namely desalination reject brine. The effect of inert par-
Mixing ticles surface area on both CO2 capture and ions removal was evaluated. The liquid residence time dis-
Residence time distribution tribution (RTD) was also evaluated for experiments with and without the mixing particles, using sodium
CO2 capture hydroxide as a tracer. In addition, the system dynamic behavior was assessed through step changes in the
gas flow rate, liquid flow rate and temperature. The reactor system showed high reaction efficiency,
excellent mixing and very stable steady state.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction the bulk liquid is negligible and the rate of gas absorption is
“controlled by gaseliquid interfacial area” (Charpentier, 1981).
Gas-liquid contactors are widely used in many industries that Higher mass transfer rates and correspondingly faster reaction
involve such processes as chemical, petrochemical, biochemical, rates can be achieved through enhancing the contact area between
and metallurgical. The selection as well as the design, effectiveness the liquid and the gas.
and the overall performance of these contactors or reactors often One major drawback of some high performance gaseliquid
depend on the transport mechanisms as well as contactor hydro- contactors that require good mixing is the need for high mechanical
dynamics and reaction kinetics. These units are commonly energy. However, such mechanical energy may be utilized more
encountered as aerators or gas-liquid reactors, where the gas first efficiently in some types of gaseliquid contactors than others.
dissolves in the liquid and then reacts with the liquid or any ma- Gaseliquid contactors with diverse designs may exhibit signifi-
terials dissolved in the liquid. Reactions in these types of reactors cantly different mass transfer performance under the same oper-
are often divided into two categories: slow or fast reactions ating conditions (Gaddis, 1999). Bubble column reactors, spouted
(Charpentier, 1981). Good mass transfer and high liquid hold-up are bed reactors, and agitated reactors are known to be more suitable
often needed to keep the gas bulk concentration near saturation for for processes involving slow reactions, such as liquid-phase
slow reactions, while high gas hold-up and very small gas bubbles oxidation, hydrogenation, chlorination and fermentations
are required in fast reactions, since the concentration of the gas in (Deckwer and Schumpe, 1993). Other types of gas-liquid reactors
such as packed columns and venturi reactors are known to be more
effective for processes with fast reactions, because of their high
* Corresponding author. gaseliquid interfacial contact area (Deckwer and Schumpe, 1993);
E-mail address: [email protected] (M.H. El-Naas).

http://dx.doi.org/10.1016/j.jngse.2017.01.031
1875-5100/© 2017 Elsevier B.V. All rights reserved.
134 M.H. El-Naas et al. / Journal of Natural Gas Science and Engineering 39 (2017) 133e142

however, under certain conditions, bubble column reactors and mass transfer area between the two phases (Gaddis and Vogelpohl,
packed column reactors can be suitable for processes with highly 1992). Another similar reactor is the reciprocating jet reactor,
exothermic fast reactions, “which are widely used in the chemical, which consists of several perforated discs joined together through a
biochemical, petrochemical and metallurgical applications” central shaft. The shaft and discs are housed in a cylindrical vessel
(Deckwer and Schumpe, 1993). Bubble column reactors have that is subjected to a counter motion with high amplitude and
several advantages, including low maintenance and operating cost, frequency, pushing the mixture through the perforated discs in the
low capital cost, and excellent heat transfer and temperature con- form of strong jets (Gaddis, 1999).
trol. They also tend to have high gaseliquid interfacial area as well Although the above-mentioned gas-liquid contactors have been
as high volumetric mass transfer coefficient. These reactors, how- widely used in many industries, none of them can be applied to a
ever, suffer from some drawbacks such as back-mixing and bub- variety of unit operations with the same efficiency, and they all
bleebubble interactions in the turbulent flow regimes. In addition, seem to suffer from different drawbacks such as complexity, high
the lack of specific data on the reactor hydrodynamics and the demand for mechanical energy and difficulty to scale-up. The cur-
characteristics of mass transfer under real industrial conditions rent study describes a novel system that can provide excellent gas-
makes it rather difficult to scale-up (Lemoine et al., 2008). liquid contact, high performance efficiency and can be easily
Gaseliquid contactors are often classified into surface and vol- scaled-up. The system is suitable for numerous gas-liquid contact
ume contactors. They may also be sorted in terms of mass transfer applications, such as the sweetening of natural gas, solvent
rate. In general, contactors with low mechanical energy con- regeneration and CO2 capture. The system performance and dy-
sumption have low mass transfer rates and, consequently, low namic behavior were evaluated for the capture of CO2 through re-
performance. More mechanical energy consumption, which is actions with ammonium hydroxide at different conditions.
usually associated with more mixing inside the contactor, improves
the mass transfer rate. Surface contactors are typically used for the 2. Description of the contactor/reactor system
bio-treatment of several types of wastewater and usually have the
form of pools with moderately low depth. They often involve the The system consists of a jacketed, stainless steel cylindrical
use of impellers or liquid jets to enhance mixing and provide the vessel with 78 mm internal diameter, 700 mm height, and a total
needed interfacial area (Gaddis, 1999). On the other hand, in vol- working volume of 3000 ml. The liquid enters at the top of the
ume gas-liquid contactors, the contact area between the gas and contactor, while the gas is injected through an orifice at the bottom
liquid phases is often formed within the bulk of the liquid. Gas of the conical base of the vessel as shown in Fig. 1. The gas leaves at
dispersion in the liquid, in the form of bubbles with spherical or the top of the vessel and the effluent liquid leaves at the same level
irregular shapes, is usually achieved through the use of mixing tools as the liquid inlet through a tube inserted all the way to the bottom
such as “spargers, liquid jets, two-mixture nozzles or hollow of the vessel. The liquid effluent is pushed through the effluent tube
rotating mixers” (Gaddis, 1999). Examples of common gas-liquid by the liquid hydrostatic pressure in the vessel. This mechanism
contactors/reactors include bubble column reactors, stirred vessel simulates a counter-current flow, improves the gas-liquid contact
reactors, jet loop reactor, reciprocating jet reactor, and impinging- and eliminates the possibility of gas entrainment with the effluent
stream reactor. liquid. The pressure needed to force the gas through the liquid
A bubble column reactor is a vessel with many shapes, where outlet tube at the bottom of the reactor is very high compared to
gas bubbles are created at the bottom of the vessel through a the rest of the reactor vessel and hence no gas will be entrained
sparger. Due to their mechanical simplicity, relatively low capital with the effluent liquid. The reactor is also equipped with a fine
cost, and good heat and mass transfer performance, bubble col- screen at the top, which acts as a demister to prevent entrainment
umns are widely used in many applications in bio-processing in- of liquid droplets or particles with the effluent gas. The contactor
dustry such as aerobic fermentation (McClure et al., 2015). The vessel is equipped with inert particles ranging in size from 5 to
mass transfer coefficient in bubble columns is known to depend on 15 mm and represent 5e10% of the total vessel volume. The parti-
such factors as the physical properties of the gases and liquids used, cles must be inert in relation to the gas-liquid system and should
the gas flow rate (Gaddis, 1999), sparger design (Camarasa et al., not be reacted or affected by contact with the gas or the liquid in
1999), reactor height to diameter ratio (H/D) (Anderson and the vessel. The density of the particles must be very similar to that
Quinn, 1970), system pressure (Letzel et al., 1997), and tempera- of the liquid so that they can move easily within the vessel. The
ture (Lin et al., 2001). The bubble size within the column is believed particles create vigorous mixing within the vessel through sys-
to approach a stable size shortly after dispersion, making mass tematic circular motion and provide high gaseliquid interfacial
transfer within the vessel less sensitive to sparger design (Deckwer area for effective mass transfer between the two phases. The sys-
and Schumpe, 1993). A stirred tank reactor is a cylindrical vessel tem is suitable for gas-liquid reactions, especially those involving
that is equipped with an impeller at its center. Mixing and mass gas solubility in the liquid followed by reaction. It can also be used
transfer are often enhanced by placing a sparger under the impeller for mass transfer operations involving stripping or absorption such
to generate very small gas bubbles that have large surface area per as amine acid gas removal and amine regeneration. The size of the
unit volume and increase turbulence in the liquid (Shi et al., 2014). orifice, the gas velocity as well as the gas to liquid ratio depend on
A jet loop reactor, on the other hand, is a vessel that houses a two- the type of the gas contact or the reaction system. The good mixing
mixture nozzle, which may be fixed at the top or bottom of the and high interfacial area are essential for operations involving ab-
reactor. The vessel is also equipped with a draft tube that may sorption, stripping or gas-liquid reactions.
either be concentric with the main tube or next to it (Gaddis, 1999).
The momentum created by the liquid jet leads to good circulation of 3. Experimental evaluation
the gaseliquid mixture, which in turn leads to good mixing in these
type of reactors with no dead zones (Zhang et al., 2014). The 3.1. Gas-liquid reaction efficiency
impinging stream reactor is a special case of the jet loop reactor,
where both gas and liquid are fed to the reactor through nozzles The contactor system was evaluated for the capture of CO2
placed at the inlet of guide tubes, forming a homogeneous two- through reactions with ammonium hydroxide. The reaction was
phase stream. The kinetic energy associated with the two-phase carried out through contacting a gas mixture containing 10% CO2
streams generates a high turbulence and, consequently, a large and 90% air with ammonium hydroxide (25% NH3) mixture with
 M.H. El-Naas et al. / Journal of Natural Gas Science and Engineering 39 (2017) 133e142 135

Fig. 1. A schematic diagram of the contactor/reactor system.

saline wastewater, namely desalination reject brine. The water had

7% salinity, which consisted of different ions including sodium, CO2 capture efficiency
magnesium, and calcium. However, sodium has the highest con- Moles of CO2 captured (2)
centration in the feed brine compared to calcium and magnesium
¼  100%
Moles of CO2 loaded to the reactor
(23,712 ppm for sodium, 2,794 ppm for magnesium and 1,375 ppm
for calcium); therefore, sodium removal has been used to represent
Moles of CO2 captured
the reaction efficiency as described in the reaction below. Z t
Volume of CO2 captured ðL=minÞ dt (3)
NaCl þ NH3 þ CO2 þ H2O / NaHCO3 þ NH4Cl ¼ 0
Molar volume of CO2 ðL=molÞ
Increasing ammonia concentration increases the pH of the brine
Where Xi is the initial sodium concentration in the feed brine
mixture and, consequently, increases the CO2 capture efficiency and
(mg/L), Xf is the final sodium concentration in the treated brine
sodium removal. The CO2 capture also increases with decreasing
(mg/L) and t is the time in minutes.
temperature and increasing the gauge pressure due to increasing
It is worth noting here that the effect of the orifice diameter on
the CO2 solubility in the brine and decreasing the solubility of so-
CO2 capture and ions removal was evaluated for three different
dium bicarbonate (NaHCO3). When increasing the gas flow rate, the
sizes (2, 3, 4 mm) and found to be negligible with the 3 mm giving
CO2 loading increases, and hence the sodium removal increases.
slightly better performance than the other two; therefore, all ex-
However, high gas flow rates have a negative effect on the reaction,
periments were carried out using an orifice diameter of 3 mm. The
due to the reduction of the CO2 residence time for CO2 inside the
inert particles were made from plastic with an average particle size
reactor.
of 13 mm and a density of 1,020 kg/m3. Table 1 summarize the
The following equations have been used to calculate the sodium
specifications for the inert particles used in this study. It is impor-
removal (1) and CO2 capture efficiency (2e3).
tant to note that different particle sizes were tested using a Plexi-
glas reactor with the same dimensions as the stainless steel reactor
to observe the mixing patterns inside the reactor. Using sizes less
Xi  Xf than 13 mm makes some particles accumulate at the top of the
Sodium removal % ¼  100% (1) reactor, while using bigger sizes makes some of the particles settle
Xi
at the bottom of reactor. Particles with an average diameter of
13 mm gave a relatively uniform movement and very good mixing
136 M.H. El-Naas et al. / Journal of Natural Gas Science and Engineering 39 (2017) 133e142

Table 1
Inert particles specifications.

Material Poly methyl methacrylate (PMMA)

Average size 13 mm
Average weight 0.511 g/particle
Wet density 1.022 g/cm3
Volume 0.5 cm3/particle
Surface area 4.92 cm2/particle
Sphericity 0.62
Particle to reactor volume fraction 6.6 ± 1 vol%
Particle area to reactor volume ratio 0.655 cm2/cm3.

within the reactor vessel. Choosing the suitable particle size,

however, depends on the liquid density as well as the gas flow rate
and hence different feed mixtures and gas flow rates require
different particles sizes to ensure uniform mixing within the
reactor. The CO2 capture and ions removal (Na, Mg, Ca) % were
optimized through Response Surface Methodology (RSM) using
Minitab 17.0 (Mohammad et al., 2016). Based on the response
Fig. 2. CO2 capture at optimum conditions for different gas flow rates.
surface methodology modeling, the optimum operating conditions
were found to be at a gas flow rate of 1,543 ± 2 ml/min and inert
particles volume of 6.6 ± 0.1 vol%. Experiments were carried out at particles on CO2 capture and ions removal were investigated since
optimum conditions for two gas flow rates 1,542 ml/min and good mixing and high interfacial area are essential for an efficient
4,000 ml/min, all at a fixed liquid flow rate of 12.5 ml/min. Both absorption and stripping process. The volume of inert particles
experimental runs were carried out at atmospheric pressure. The used inside the reactor was optimized using RSM. Fig. 4 (a) and (b)
experimental results are summarized in Table 2 and plotted in show the effect of inert particles total surface area as well as vol-
Figs. 2 and 3 for CO2 capture and ions removal, respectively. For the ume at optimum conditions on both CO2 capture efficiency and
low gas flow rate (1,542 ml/min), the experimental run lasted for Sodium removal, respectively. Fig. 4 (a) shows that there is an in-
11 h; whereas for the other flow rate (4,000 ml/min), the reactor crease in the CO2 capture efficiency with increasing the total sur-
was operated for 6 h. The experimental results clearly indicated face area of inert particles up to 0.6 cm2/cm3. This increase can be
that the reactor system was very stable and reached stable steady explained by improving the mixing inside the reactor and hence
state. The CO2 removal reached steady state after 3 h for the low gas increasing the reaction rate due to the increase in CO2 loading.
flow rate and about 4 h for the high gas flow rate. Similarly, ions However, increasing the surface area of inert particles more than
removal reached steady after 3 h and 6 h for low and high gas flow 0.7 cm2/cm3 decreases the CO2 capture efficiency, which is believed
rates, respectively. As expected, the gas residence time inside the to be due to the decrease in the gas residence time in the reactor.
reactor seemed to have significant effect on the CO2 capture, but not Increasing the amount of inert particles creates vigorous mixing
as important effect on the ions removal as shown in Table 2. As within the vessel that can have a negative impact on the residence
more gas passes through the reactor system, less CO2 gets captured time through extraordinary circular motion and reducing the
(about 90%). Nevertheless, the difference between the two cases in accessible gas volume, which decreases the residence time of gas
CO2 capture efficiency is not substantial given the large difference bubbles inside the reactor (Yeh et al., 2005). However, the increase
in gas-to-liquid ratio (G/L). For the low gas flow rate, the gas to in surface area of the inert particles seems to have insignificant
liquid ratio (G/L) was 123, and the CO2 capture efficiently was effect on ions removal despite its effect on the CO2 capture effi-
97.7%; whereas, for the higher gas flow rate, the G/L ratio was 320 ciency. This may be attributed to the mechanism of the involved
and the CO2 capture efficiency was 90.1%. These experiments have chemical reactions, where carbon dioxide first dissolves in the
proven that the contactor/reactor system described in this study is
very effective in gas-liquid contact/reaction and can achieve a very
stable steady state operation.

3.2. Effect of inert particle surface area on CO2 capture and ions
removal

In this part of the study, the effect of total surface area of inert

Table 2
Summary of experimental results at optimum conditions for different gas flow rates.

Conditions Gas flow rate Gas flow rate

1542 ml/min 4000 ml/min

Liquid flow rate (ml/min) 12.5 12.5

Inert particles vol% 6.6 6.6
Gas/liquid flow ratio 123 320
CO2 capture efficiency% 97.7 90.1
Na removal% 32.5 30.0
Mg removal% 97.2 98.4
Ca removal% 87.3 85.1
Run time (min) 660 360
Time to reach steady state (min) 240 180
Fig. 3. Ions removal at optimum conditions for different gas flow rates.
 M.H. El-Naas et al. / Journal of Natural Gas Science and Engineering 39 (2017) 133e142 137

Fig. 4. (a) CO2 capture efficiency and (b) ions removal versus inert particles volume and surface area to reactor volume at temperature of 19  C and gas flow rate of 1,542 ml.

brine to form carbonic acid (H2CO3), which dissociates into bicar- vessel; the circulation of the inert mixing particles forces more gas
bonate ions at high pH level (Greenwood and Earnshaw, 1996). within the liquid flow patterns inside the vessel, which in turn
Consequently, the CO2 capture efficiency and sodium removal can enhances mixing, facilitates gas-liquid contact, and improves mass
only increase to certain limit; therefore, the increase in CO2 solu- transfer.
bility is not necessarily associated with more ions removal as
shown in Fig. 4 (b).
3.4. Effect of inert particles on residence time distribution (RTD)

3.3. Effect of inert particles on gas hold-up Residence time is a measure of the time fluids spend inside a
reactor, which can be found from the inlet and outlet concentration
Gas hold-up is a key hydrodynamic factor and an elementary profiles (Fogler, 2016). The exit-age distribution function E(t) is
measure of gas-liquid contacting in any gas/liquid contactor sys- defied in Equation (5).
tem, and it is defined as the ratio of gas phase volume to the total
volume (Hebrard et al., 2001). The maximum gas hold-up is CðtÞ
EðtÞ ¼ Z ∞ (5)
correlated to high gas population, interfacial area and mass transfer
CðtÞ* dt
rate; accordingly, measuring the gas hold-up is essential for scaling 0
up and design any gas/liquid contactor system. The following
equation was used to measure the gas hold-up (εG ) in the contactor The mean residence time takes into account any dead volumes or
system: short-circulating in the reactor and is defined in Equation (6).

HG  H
εG ¼ (4)
HG

Where: HG: is the height of liquid with gas flow (cm).H: is the
height of liquid without gas flow (cm).εG: is the fractional gas hold-
up.
Gas hold-up is a dimensionless parameter that usually ranges in
value from 0.0 to 0.20 (Saravanan et al., 2009). It was determined at
different gas flow rates through measuring the difference in liquid
level with and without gas flow for two cases: with inert mixing
particles and without mixing particles. A plot of the gas hold-up
versus gas flow rate (ml/min) and superficial gas velocity (mm/
sec) with and without mixing particles is presented in Fig. 5. The
superficial gas velocity was calculated by dividing the gas flow rate
by the cross sectional area of the contactor vessel. As expected, the
gas hold-up increases as the superficial gas velocity and gas flow
rate increase. This is due to the considerable surge in the gas input
and hence gas volume inside the vessel, which tends to increase the
hold-up within the system. The presence of the mixing particles
seemed to enhance the gas hold-up at all gas flow rates, which is Fig. 5. Gas hold-up with and without mixing particles versus gas flow rate and su-
attributed to the circular movement of the particles within the perficial gas velocity.
138 M.H. El-Naas et al. / Journal of Natural Gas Science and Engineering 39 (2017) 133e142

calculated using Equation (7).

Z∞
tm ¼ tEðtÞdt (6) Z∞
0 s2 ¼ ðt  tm Þ2 EðtÞdt (7)
0
The variance indicates the spread of the distribution function and is

Fig. 6. The effect of mixing particles presence on (a) change in conductivity (b) residence time distribution.

Table 3
Residence time distribution analysis for reactor system with mixing particles.
1
t min C(t) ms/cm E(t) min tE(t) (t  tm) min (t  tm)2 min2 E(t)*(t  tm)2 min

10.0 3.5 6.4E-03 6.4E-02 90.0 8094.6 51.6

20.0 4.0 7.3E-03 1.5E-01 80.0 6395.2 46.4
30.0 3.7 6.7E-03 2.0E-01 70.0 4895.8 32.8
40.0 3.6 6.5E-03 2.6E-01 60.0 3596.4 23.4
50.0 3.4 6.2E-03 3.1E-01 50.0 2497.0 15.5
60.0 3.4 6.1E-03 3.6E-01 40.0 1597.6 9.7
70.0 3.3 6.0E-03 4.2E-01 30.0 898.2 5.4
80.0 3.0 5.5E-03 4.4E-01 20.0 398.8 2.2
90.0 2.8 5.2E-03 4.6E-01 10.0 99.4 0.5
100.0 2.4 4.4E-03 4.4E-01 0.0 0.0 0.0
110.0 2.3 4.2E-03 4.7E-01 10.0 100.6 0.4
120.0 2.2 3.9E-03 4.7E-01 20.0 401.2 1.6
130.0 2.0 3.7E-03 4.8E-01 30.0 901.8 3.3
140.0 1.9 3.4E-03 4.8E-01 40.0 1602.4 5.5
150.0 1.7 3.1E-03 4.7E-01 50.0 2503.0 7.9
160.0 1.7 3.1E-03 4.9E-01 60.0 3603.6 11.1
170.0 1.6 2.9E-03 5.0E-01 70.0 4904.2 14.4
180.0 1.6 2.8E-03 5.1E-01 80.0 6404.8 18.0
190.0 1.4 2.6E-03 4.9E-01 90.0 8105.4 20.9
200.0 1.1 2.0E-03 4.1E-01 100.0 10006.0 20.3
210.0 1.0 1.9E-03 3.9E-01 110.0 12106.6 22.6
220.0 0.9 1.7E-03 3.7E-01 120.0 14407.2 24.1
230.0 0.9 1.6E-03 3.7E-01 130.0 16907.8 27.0
240.0 0.7 1.3E-03 3.1E-01 140.0 19608.4 25.6
250.0 0.6 1.0E-03 2.5E-01 150.0 22509.0 22.9
260.0 0.3 4.5E-04 1.2E-01 160.0 25609.6 11.6
270.0 0.2 3.1E-04 8.3E-02 170.0 28910.2 8.9
280.0 0.1 2.0E-04 5.6E-02 180.0 32410.8 6.5
290.0 0.1 1.8E-04 5.3E-02 190.0 36111.4 6.6
300.0 0.1 1.1E-04 3.3E-02 200.0 40012.0 4.4
310.0 0.0 7.3E-05 2.3E-02 210.0 44112.6 3.2
320.0 0.1 9.1E-05 2.9E-02 220.0 48413.2 4.4
330.0 0.1 9.1E-05 3.0E-02 230.0 52913.8 4.8
340.0 0.0 7.3E-05 2.5E-02 240.0 57614.4 4.2
 M.H. El-Naas et al. / Journal of Natural Gas Science and Engineering 39 (2017) 133e142 139

Table 4
The residence time distribution analysis for reactor system without mixing particles.
1
t min C(t) ms/cm E(t) min tE(t) (t  tm) min (t  tm)2 min2 E(t)*(t  tm)2 min

10.0 3.2 4.8E-03 4.8E-02 102.3 10465.3 50.5

20.0 4.1 6.1E-03 1.2E-01 92.3 8519.3 52.0
30.0 4.0 6.0E-03 1.8E-01 82.3 6773.3 40.7
40.0 3.9 5.8E-03 2.3E-01 72.3 5227.3 30.5
50.0 3.8 5.7E-03 2.8E-01 62.3 3881.3 22.1
60.0 3.7 5.5E-03 3.3E-01 52.3 2735.3 15.0
70.0 3.5 5.2E-03 3.7E-01 42.3 1789.3 9.4
80.0 3.4 5.0E-03 4.0E-01 32.3 1043.3 5.2
90.0 3.2 4.8E-03 4.3E-01 22.3 497.3 2.4
100.0 3.1 4.6E-03 4.6E-01 12.3 151.3 0.7
110.0 3.0 4.4E-03 4.9E-01 2.3 5.3 0.0
120.0 2.7 4.0E-03 4.9E-01 7.7 59.3 0.2
130.0 2.6 3.8E-03 5.0E-01 17.7 313.3 1.2
140.0 2.3 3.4E-03 4.8E-01 27.7 767.3 2.6
150.0 2.2 3.3E-03 4.9E-01 37.7 1421.3 4.7
160.0 2.0 3.0E-03 4.8E-01 47.7 2275.3 6.8
170.0 1.9 2.9E-03 4.9E-01 57.7 3329.3 9.5
180.0 1.8 2.7E-03 4.9E-01 67.7 4583.3 12.5
190.0 1.7 2.5E-03 4.7E-01 77.7 6037.3 15.1
200.0 1.6 2.3E-03 4.6E-01 87.7 7691.3 17.9
210.0 1.5 2.2E-03 4.5E-01 97.7 9545.3 20.6
220.0 1.3 2.0E-03 4.4E-01 107.7 11599.3 23.0
230.0 1.2 1.8E-03 4.1E-01 117.7 13853.3 24.7
240.0 1.1 1.7E-03 4.1E-01 127.7 16307.3 27.7
250.0 0.9 1.4E-03 3.5E-01 137.7 18961.3 26.5
260.0 0.8 1.2E-03 3.1E-01 147.7 21815.3 26.0
270.0 0.6 8.3E-04 2.2E-01 157.7 24869.3 20.7
280.0 0.4 6.5E-04 1.8E-01 167.7 28123.3 18.4
290.0 0.4 5.7E-04 1.6E-01 177.7 31577.3 17.9
300.0 0.3 4.9E-04 1.5E-01 187.7 35231.3 17.3
310.0 0.3 4.5E-04 1.4E-01 197.7 39085.3 17.4
320.0 0.3 4.2E-04 1.3E-01 207.7 43139.3 18.0
330.0 0.3 4.5E-04 1.5E-01 217.7 47393.3 21.2
340.0 0.3 4.3E-04 1.5E-01 227.7 51847.3 22.4

Fig. 7. Ions removal versus reaction time with step changes of (a) gas flow rate, (b) liquid flow rate.
140 M.H. El-Naas et al. / Journal of Natural Gas Science and Engineering 39 (2017) 133e142

linear relationship between its conductivity and concentration

(Trade Mark of the Dow Chemical Company). The tracer was
injected into the system in a very short period, while water and
zero air were used as liquid and gas phase respectively to eliminate
the possibility of any side reactions. The analysis was carried out
assuming a perfect injection and flow distribution. The effect of
inert particles on the change in conductivity, residence time dis-
tribution, mean residence time and variance were investigated. As
shown in Fig. 6 (a), the presence of the inert particles shifted the
change in conductivity toward shorter time values for each exact
conductivity value. The residence time inside the mixing zone
became smaller rendering a narrower E(t) distribution as shown in
Fig. 6 (b). The dimensionless time t E(t) and spread (t  tm)2 E(t)
profiles were used to calculate the mean residence time and vari-
ance according to Equations (6) and (7), respectively (Fogler, 2016).
Graph software was used to find the best fit for each profile using
10th order polynomial (Johansen, 2012). The mean residence time
was found to be 99.9 and 112.3 min and the variance was found to
be 67.5 and 76.5 min for the reactor system with and without
mixing particles, respectively. Part of RTD analysis results for the
reactor system with and without mixing particles are listed in
Tables 3 and 4, respectively. Using inert particles decreases the
mean residence time and variance by about 11%. Which clearly
indicates that using inert particles can significantly decrease the
dead zones and segregation inside the reactor as well as improving
mixing distribution (Pant et al., 2015; Common et al., 2014).

3.5. Hydrodynamics of gas-liquid contactor

To evaluate the reactor performance and response to a step

change in each of the main reaction parameters, continuous ex-
periments were carried out at the optimum conditions
(Mohammad et al., 2016). First, each experiment was carried out for
at least 60 min to reach steady state in CO2 capture and ions
removal. Then, a step change was performed for each parameter in
separate experiments (gas flow rate from 1432 ml/min to 3084 ml/
Fig. 8. CO2 capture versus reaction time with change of (a) gas flow rate, (b) liquid flow min; liquid flow rate from 12.5 to 20 ml/min; and temperature from
rate.
19 to 40  C). A second steady state condition of CO2 capture and
ions removal was reached and stabilized for at least another 60 min,
after this time, each parameter was returned to its previous value to
The RTD analysis was performed through tracer injection, using
reach the third steady state condition. Figs. 7 and 8 show the
sodium hydroxide (NaOH) as a tracer. NaOH is known to have a
decreasing or increasing in ions removal and CO2 capture through

Fig. 9. CO2 capture efficiency with (a) gas flow step change of (1542e3084 and back to 1542 ml/min) (b) Liquid flow rate step change of (12.5e20 and back to 12.5 ml/min).
 M.H. El-Naas et al. / Journal of Natural Gas Science and Engineering 39 (2017) 133e142 141

Table 5 case of decreasing gas flow rate, while it decreased when the liquid
Summary of step change results. flow rate was decreased. However, temperature step change did not
Condition Step Change t1 t2 t63.2% td tp show any significant effect on the value of dead time, which may be
Gas Flow rate (ml/min) 1542e3084 330 336 353 6 17
due to the to the insignificant effect of this temperature change
Gas Flow rate (ml/min) 3084e1542 480 488 507 8 19 (19e40  C) on the reaction kinetics (Pelkie et al., 1992). Results
Liquid Flow rate (ml/min) 12.5e20 300 309 336 9 27 show that the reactor time constant (tp) was not only influenced by
Liquid flow rate (ml/min) 20e12.5 420 424 442 4 18 reactor hydrodynamics and reaction kinetics, it was also found to
Temperature ( C) 19e40 255 271 298 16 27
be affected by the final value of the changed parameter; tp was
Temperature ( C) 40e19 375 391 503 16 112
decreased with decreasing liquid flow rate and increased with

Fig. 10. The dead time and reactor time constants for the performed step change experiments.

these step changes of gas and liquid flow rates. Results are decreasing gas flow rate and decreasing temperature. These results
consistent with the performance of the Solvay process, where the are very promising with regards to the flexibility as well as appli-
efficiency of CO2 capture increased with increasing the liquid flow cability of the reactor system for large scale processes.
rate and decreasing the gas flow rate (Mohammad et al., 2016).
A data logger was connected to the CO2 gas analyzer to record
one minute time sample as shown in Fig. 9. The dead time (td),
which is the time from when the step change is made until the 4. Conclusions
measured CO2 capture efficiency first responds, and the reactor
time constant (tp), which indicates to the reactor response to each A new gas-liquid contactor/reactor system involving a vertical
parameter step change (Liptak, 2005), were calculated according to vessel with conical base and inert particles was developed and
Equations (8) and (9), respectively. evaluated for CO2 capture and brine desalination. The reactor was
assessed at optimum conditions as well as a high gas to liquid ratio.
The experiments showed that the reactor system is very effective in
td ¼ t1  t2 (8)
CO2 capture and brine desalination and can achieve a very stable
steady state operation. Besides, the reactor performance was
tp ¼ t63:2%  t2 (9) evaluated at different conditions, where 97.7% and 90.1% in CO2
capture efficiency were reached at gas to liquid ratio (G/L) of 123
Where: t1: is the time when the step change is made. t2: is the time and 320, respectively. The effect of inert particles on CO2 capture,
when the measured CO2 capture efficiency first responds to the ions removal and residence time distribution inside the reactor was
step change. t63.2%: is the time when the measured CO2 capture examined. The inert particles showed a significant effect on
efficiency reaches 63.2% of its total final change. improving mixing inside the reactor by decreasing the mean resi-
The results presented in Table 5 and Fig. 10 show that td ˂ tp in all dence time and variance by about 11%. Finally, the dynamic
experiments, which indicates a tight and easy overall process behavior of the reactor system during step changes operation was
control (Liptak, 2005). The decrease or increase in CO2 capture with carried out. Gas flow rate, liquid flow rate and temperature step
the gas and liquid flow rate step changes can be detected faster changes were executed to estimate the dead time and reactor time
compared to the decrease in the CO2 capture with the temperature constant for each transient state condition. The dead time was
step. After the dead time, the CO2 capture decreased or increased found to be less than the reactor time constant in all cases, which
and then stabilized in all experiments. However, different time indicates, to a high probability, a tight and easy overall process
periods were required for each of the step changes. The dead time control. The reactor time constant (tp) was found to be related to
was found to be related to the step change direction in gas flow rate both, reactor hydrodynamics and reaction kinetics, where higher
and liquid flow rate. For example, the dead time increased in the time constant correlates with higher CO2 capture efficiency.
142 M.H. El-Naas et al. / Journal of Natural Gas Science and Engineering 39 (2017) 133e142

Acknowledgment fluctuation signals. Chem. Eng. Sci. 56, 6241e6247.

Liptak, B.G., 2005. Instrument Engineers' Handbook, Process Control and Optimi-
zation, vol. 2. CRC Press, ISBN 9781420064001 ed. 4.
The authors would like to acknowledge the financial support McClure, D.D., Aboudha, N., Kavanagh, J.M., Fletcher, D.F., Barton, G.W., 2015. Mixing
provided by Takreer Research Center as part of the TRC Project. The in bubble column reactors: experimental study and CFD modeling. Chem. Eng. J.
authors would also like to thank Dr. Hussain Mousa and Ms. Maida 264, 291e301.
Mohammad, A., El-Naas, M.H., Suleiman, M., Al Musharfy, M., 2016. Optimization of
Osman, from the Chemical and Petroleum Engineering Depart- a solvay-based approach for CO2 capture. Int. J. Chem. Eng. Appl. 7, 230e234.
ment, UAE University, and Ms. Abeer Dar Saleh, from the Archi- Pant, H.J., Goswami, S., Samantray, J.S., Sharma, V.K., Maheshwari, N.K., 2015.
tectural Engineering Department at the Al Albayt University, for Residence time distribution measurements in a pilot-scale poison tank using
radiotracer technique. Appl. Radiat. Isotopes 103, 54e60.
their help. Pelkie, J.E., Concannon, P.J., Manley, D.B., Poling, B.E., 1992. Product distributions in
the CO2eNH3eH2O system from liquid conductivity measurements. Industrial
References Eng. Chem. Res. 31, 2209e2215.
Saravanan, K., Ramamurthy, V., Chandramohan, K., 2009. Gas hold up in multiple
impeller agitated vessels. Mod. Appl. Sci. 3 (no. 2).
Anderson, J.L., Quinn, J.A., 1970. The transition to slug flow in bubble columns.
Shi, D., Liang, Y., Eaglesham, A., Gao, Z., 2014. Effect of the impeller imbalance on the
Chem. Eng. Sci. 25, 340e341.
bending moment acting on a shaft in a stirred vessel. Chem. Eng. Res. Des. 92,
Camarasa, E., Vial, C., Poncin, S., Wild, G., Midoux, N., Bouillard, J., 1999. Influence of
2191e2200.
coalescence behaviour of the liquid and of gas sparging on hydrodynamics and
Trade Mark of the Dow Chemical Company, “Conductivity of solutions,” FILMTEC
bubble characteristics in a bubble column. Chem. Eng. Process. Process Intensif.
Membr., 609-02127-804, Dow Liquid separation business.
38, 329e344.
Yeh, J.T., Resnik, K.P., Rygle, K., Pennline, H.W., 2005. Semi-batch absorption and
Charpentier, J.-C., 1981. “Mass-Transfer Rates in Gas-Liquid Absorbers and Reactors,”
regeneration studies for CO2 capture by aqueous ammonia. Fuel Process.
Advances in Chemical Engineering, vol. 11. Academic press, pp. 1e133. ISBN.
Technol. 86, 1533e1546.
0065-2377.
Zhang, W., Yong, Y., Zhang, G., Yang, C., Mao, Z., 2014. Mixing characteristics and
Common, A., Rodier, E., Sauceau, M., Fages, J., 2014. Flow and mixing efficiency
bubble behavior in an airlift internal loop reactor with low aspect ratio. Chin. J.
characterisation in a CO2-assisted single-screw extrusion process by residence
Chem. Eng. 22, 611e621.
time distribution using Raman spectroscopy. Chem. Eng. Res. Des. 92,
1210e1218.
Deckwer, W.D., Schumpe, A., 1993. Improved tools for bubble column reactor design Glossary
and scale-up. Chem. Eng. Sci. 48, 889e911.
Fogler, H.S., 2016. Elements Of Chemical Reaction Engineering: Perntice Hall In-
ternational Series In The Physical And Chemi. Pearson Education, ISBN C(t): Exit-age fluid concentration (ms/cm)
9780133887518. E(t): Exit-age distribution function (min1)
Gaddis, E.S., 1999. Mass Transf. gaseliquid contactors,” Chem. Eng. Process. Process G/L: Gas to liquid flow ratio
Intensif. 38, 503e510. H: Height of liquid without gas flow (cm)
Gaddis, E.S., Vogelpohl, A., 1992. The impinging-stream reactor: a high performance H/D: Reactor length to diameter ratio
loop reactor for mass transfer controlled chemical reactions. Chem. Eng. Sci. 47, HG: Height of liquid with gas flow (cm)
2877e2882. RSM: Response surface methodology
Greenwood, N.N., Earnshaw, A., 1996. Chemistry of the Elements. Elsevier Science & RTD: Residence time distribution
Technology Books, ISBN 9780080379418. t1: Time when the step change is made (min)
Hebrard, G., Bastoul, D., Roustan, M., Bouaifi, M., 2001. A comparative study of gas t2: Time when the measured value first responds to the step change (min)
hold-up, bubble size, interfacial area and mass transfer coefficients in stirred t63.2%: Time when the measured value reaches 63.2% of its total final change (min)
gaseliquid reactors and bubble columns”. Chem. Eng. Process. 40, 97e111. td: Time from when the step change is made until the measured value first responds
Johansen, I., 2012. “Graph,” GNU General Public Licenses, vol. 543 ed, 4.4.2. (min)
Lemoine, R., Behkish, A., Sehabiague, L., Heintz, Y.J., Oukaci, R., Morsi, B.I., 2008. An tm: Mean residence time (min)
algorithm for predicting the hydrodynamic and mass transfer parameters in
bubble column and slurry bubble column reactors. Fuel Process. Technol. 89, Greek symbols
322e343.
Letzel, H.M., Schouten, J.C., Krishna, R., van den Bleek, C.M., 1997. Characterization of εG: Fractional gas hold-up
regimes and regime transitions in bubble columns by chaos analysis of pressure s: Variance (min)
signals. Chem. Eng. Sci. 52, 4447e4459. tp: Reactor time constant (min)
Lin, T.J., Juang, R.C., Chen, C.C., 2001. Characterizations of flow regime transitions in
a high-pressure bubble column by chaotic time series analysis of pressure