Hydrometallurgy 144–145 (2014) 54–62

Contents lists available at ScienceDirect

Hydrometallurgy
journal homepage: www.elsevier.com/locate/hydromet

Liquid–liquid extraction in microchannels with Zinc–D2EHPA system

Mayur Darekar, Nirvik Sen, K.K. Singh ⁎, S. Mukhopadhyay, K.T. Shenoy, S.K. Ghosh
Chemical Engineering Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400085, India

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

Article history: Liquid–liquid extraction experiments have been carried out in two different types of microchannels using water
Received 9 July 2013 (Zn) D2EHPA + dodecane system, a standard reactive extraction system recommended for evaluation of liquid–
Received in revised form 18 October 2013 liquid extraction equipment. The microchannels used in the experiments are a t-junction serpentine
Accepted 9 January 2014
microchannel and a split and recombine microchannel. Effects of O/A ratio and residence time on overall volu-
Available online 24 January 2014
metric mass transfer coefficient have been studied and an attempt has been made to explain the experimental
Keywords:
trends with the help of liquid–liquid flow patterns observed by using a high speed image acquisition system. Cor-
Microchannel relations relating overall volumetric mass transfer coefficients to O/A ratio and velocity are reported for both the
Mass transfer coefficient microchannels.
Serpentine © 2014 Elsevier B.V. All rights reserved.
Split and recombine
D2EHPA

1. Introduction attractive for carrying out liquid–liquid extraction. Interest in liquid–

liquid extraction in microchannels is evident by several studies report-
Due to several inherent advantages such as high surface to vol- ed in recent literature.
ume ratio ensuring intensified heat transfer rates, low inventories Hotokezaka et al. (2005) studied extraction of uranium from an
ensuring possibility of carrying out reactions involving toxic and haz- aqueous phase into tributyl phosphate (TBP) containing organic
ardous chemicals, compact setups allowing for distributed and in-situ phase using a microchannel etched in a glass chip. Aota et al.
production, modular and hence confident scale up, microchannels find (2007) carried out experiments to study extraction of cobalt com-
numerous potential applications in the field of chemical processing. plex from toluene to water in a microchannel fabricated in glass.
Interest in chemical processing in microchannels can be gauged by Zhao et al. (2007) performed liquid–liquid extraction experiments
a large number of studies reporting use of microchannels for reac- with water–succinic acid–butanol system in opposed T-junction
tions (Domínguez et al., 2011; Liu et al., 2012; Regatte and Kaisare, and cross T-junction microchannels drilled in stainless steel. Okubo
2013; Ren et al., 2012; Tan et al., 2011), gas–liquid and solid–liquid et al. (2008) studied extraction of phenol from dodecane into
mass transfer operations (de Loos et al., 2011; Tan et al., 2012) and water in a Y-junction microchannel etched in glass. Xu et al. (2008)
multiphase flows (Rajesh and Buwa, 2012). focused on extraction of succinic acid from n-butanol to aqueous phase
Liquid–liquid extraction is one of the potential applications of in co-flowing and flow-focusing microchannels made of polymethyl
microchannels. Some of the advantages specific to liquid–liquid ex- metacrylate (PMMA). Su et al. (2009) carried out studies on enhancement
traction in microchannels are higher overall volumetric mass trans- of mass transfer rates for 30%TBP in kerosene–acetic acid–water system
fer coefficients compared to conventional extractors (Kashid et al., due to agitation by inert gas in PMMA microchannels. Ghaini et al.
2011b; Su et al., 2010; Zhao et al., 2007), narrow drop size distribu- (2010) studied extraction of n-butyl-formate into NaOH solution in
tions (or if the flow is parallel or annular, no dispersion at all) leading microchannels made of polytetrafluoroethylene (PTFE) Y-junctions
to reduced settling problems (Mehrotra et al., 2008; Xu et al., 2006a,b) having downstream microbore capillaries. Aoki et al. (2011) studied
and confident scale-up due to numbering-up approach. The limitations extraction of phenol from dodecane to water in PTFE microbore
of liquid–liquid extraction in microchannels are relatively higher pres- tubes. Assmann and von Rohr (2011) performed experiments to
sure drop and inability to handle a feed containing solids. However, study extraction of vanillin from water to toluene intensified by
the advantages outnumber the limitations making microchannels very inert gas agitation in a cross T-junction microchannel fabricated in
polydimethylsiloxane (PDMS). Kashid et al. (2011a) used water–
acetone–toluene system to compare performance of microchannels hav-
Abbreviations: AAS, atomic absorption spectrometry; D2EHPA, Di-2-ethyl-hexyl- ing different types of microfluidic junctions (opposed T-junction with
phosphoric acid; EFCE, European Federation of Chemical Engineering; LMCD, log mean square cross section, opposed T-junction with trapezoidal cross section,
concentration difference; LMTD, log mean temperature difference; PDMS, polydimethylsi-
loxane; PMMA, polymethylmetacrylate; PTFE, polytetrafluoroethylene.
Y-junction with rectangular cross section, concentric junction and cater-
⁎ Corresponding author. Tel.: +91 22 25592601; fax: +91 22 25505151. pillar micromixer). Tamagawa and Muto (2011) studied extraction of ce-
E-mail address: [email protected] (K.K. Singh). sium from aqueous phase into organic phase containing D2EHPA

0304-386X/$ – see front matter © 2014 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.hydromet.2014.01.010
 M. Darekar et al. / Hydrometallurgy 144–145 (2014) 54–62 55

dissolved in cyclohexane in a Y-junction microchannel fabricated 2. Experimental

in acrylic. Priest et al. (2012) studied extraction of copper from aqueous
solution into a solvent containing LIX84-IC in a glass microchannel 2.1. Experimental setup
having Y-junction. Tsaoulidis et al. (2013) studied extraction of urani-
um from an aqueous stream into an organic phase containing TBP dis- A schematic diagram of the experimental setup is shown in Fig. 1.
solved in an ionic liquid using a PTFE microchannel having cross T- The setup consists of two pumps, two feed vessels, a microchannel
junction. chip, a membrane separator, an air compressor and two collection
Many of the reported studies on liquid–liquid extraction in vials. The aqueous and organic phases are pumped by respective pos-
microchannels either aim to compare the performance of different itive displacement pumps to the microchannel at the desired flow
kinds of microchannels or compare performance of a microchannel rates. Dispersion is generated and mass transfer takes place inside
with that of conventional liquid–liquid extraction contactors. The the microchannel. A microbore tube connects the outlet of the
performance of a microchannel or any other contactor used for liq- microchannel to the membrane separator. The membrane separator
uid–liquid extraction depends, among other things, also on the phys- has a hydrophobic membrane across which a differential air pressure
ical properties of the two phases involved as the densities, viscosities is applied. Driven by this differential pressure organic phase moves
and interfacial tension affect the quality of dispersion and film coef- from the higher pressure side to the lower pressure side and thus
ficients for mass transfer. In view of the important role played by the phase separation is achieved. Experiments are carried out to study
physical properties, comparison of liquid–liquid extraction equip- the effects of residence time and O/A ratio on the overall volumetric
ment should ideally be done using the same phase system. The mass transfer coefficient. Zinc concentration in the extract and the
phase systems recommended by the European Federation of Chemi- raffinate is determined by atomic absorption spectrometry (AAS).
cal Engineering (EFCE), for which physical properties and data rele-
vant to liquid–liquid extraction are well documented, can be used 2.2. Microchannels
for this purpose. While there are several studies on liquid–liquid ex-
traction in microchannels with standard physical extraction systems Two different microchannels etched in glass chips have been used in
recommended by EFCE (Benz et al., 2001; Xu et al., 2008; Zhao et al., the experiments. These are a 250 μL t-junction serpentine microchannel
2007), to the best of our knowledge, there is no reported study on and a split and recombine microchannel having 26 μL volume. These
liquid–liquid extraction in microchannels with a standard reactive microchannels are parts of FRX-200 system procured from M/s. Syrris
extraction system such as Zinc–D2EHPA recommended by EFCE. In Ltd., UK. Fig. 2 shows the schematic layout of the t-junction serpentine
industry reactive extractions are equally important (Torkaman et al., microchannel and the t-junction at which the liquid–liquid dispersion
2013; Traeger et al., 2012) and performance of microchannels should is generated. This microchannel has three inlets. The organic and the
also be evaluated for reactive extraction. This study aims to fill this aqueous phases are applied at Inlet-1 (I-1) and Inlet-3 (I-3), respective-
gap area by reporting data of solvent extraction in microchannels with ly. Inlet-2 (I-2) is kept closed. The phase applied at the Inlet-1 is split
a standard reactive extraction system recommended by EFCE. Zinc– and comes at the junction from two sides, as shown in Fig. 2. The first
D2EHPA system has got significant industrial relevance also as evident few serpentine channels after the junction are called mixing channels
by several studies using this system reported in literature (Bart, 2005; and the serpentine channels following the mixing channels are called
Innocenzi and Vegliò, 2012; Ji et al., 2006; Principe and Demopoulos, reaction channels. The mixing channels have smaller cross section and
2004). In this study we have studied performance of two types of smaller distance between the bends compared to the reaction channels.
microchannels – a serpentine microchannel and a split and recombine The equivalent diameters of the mixing and reaction channels are
microchannel – for liquid–liquid extraction using Zinc–D2EHPA system. 278 μm and 319 μm respectively. While the role of the mixing channels

Fig. 1. Schematic diagram of the experimental setup.

56 M. Darekar et al. / Hydrometallurgy 144–145 (2014) 54–62

Fig. 2. Schematic layout of the t-junction serpentine microchannel and its microfluidic junction.

is to ensure rapid mixing, the role of the reaction channels is to provide and is found to be 11.4. Flow patterns are observed using a high speed
sufficient time for reaction or mass transfer. image acquisition system. In both the microchannels the microfluidic
The split and recombine microchannel, its mixing unit and mixing junctions are not optically accessible and hence flow patterns have
principle are shown in Fig. 3. The chip has got two microchannels, been observed at locations away from the junctions. Experiments are
only one of which is used in the experiments. This microchannel carried out to study the effect of O/A ratio and residence time on overall
has very small volume i.e. 26 μL and is basically a micromixer work- volumetric mass transfer coefficient (KLa).
ing on the principle of multi-lamination, as shown in Fig. 3. The
equivalent diameters of the main channel and the side channels are 3. Evaluation of overall volumetric mass transfer coefficient
184 μm and 71 μm respectively. The volume of the tube connecting
the outlet of the microchannel with the membrane separator is Overall volumetric mass transfer coefficients are calculated based
3 μL. The liquid holdup in the membrane separator is 75 μL. on the concentration of zinc in the organic phase. The driving force
for mass transfer, in general, will vary from the entrance to the exit
2.3. Phase system and procedure of the microchannel and hence, in an analogy with log mean temper-
ature difference (LMTD) in a heat exchanger, a log mean concentra-
The phase system used in the experiments is a reactive extraction tion difference (LMCD) can be defined as
system. Several studies on mechanism of extraction in this reactive
extraction system are reported in literature (Mörters and Bart, ðC oi −C oi Þ−ðC oo −C oo Þ
2000; Wachter et al., 1993) and hence these details are omitted ΔLMC ¼ 
ð1Þ
ln C oi −C oi =ðC oo −C oo Þ
here for brevity. The feed is an aqueous solution of zinc sulfate in
water. The organic phase is a 5% (v/v) solution of di-(2-ethylhexyl)-
phosphoric-acid (D2EHPA) in dodecane. Initially the organic phase is Using LMCD, the mean value of mass transfer flux can be written as
equilibrated with demineralized (DM) water in a stirred tank. After
equilibration, the phases are separated and zinc sulfate is added to the N ¼ K L ΔLMC ð2Þ
equilibrated water to prepare 500 ppm zinc feed solution. The equili-
bration step ensures that in subsequent experiments the transfer of
the solute from the aqueous phase to the organic phase is solely due The mean value of mass transfer flux will also be equal to
to interphase mass transfer and not due to solubility. The distribution
coefficient of the system, defined as the ratio of the concentration of Q o ðC oo −C oi Þ
N¼ ð3Þ
the solute in the organic phase to that in the aqueous phase, is measured Va

Fig. 3. The split and recombine microchannel, its mixing unit and multi-lamination principle.
 M. Darekar et al. / Hydrometallurgy 144–145 (2014) 54–62 57

Combining Eqs. (2) and (3) the overall volumetric mass transfer co- becomes finer on reducing residence time from 98 s to 33 s. When
efficient can be written as the residence time is further reduced to 5 s, small slugs separated
by very fine droplets are observed. The droplets are now so fast mov-
C oo −C oi ing that higher shutter speeds are required to capture them causing
KLa ¼ ð4Þ
V illumination to be insufficient. Even at higher shutter speeds edges
Δ
Q o LMC of the drops are blurred. These images show that dispersion does be-
come finer and velocities are significantly enhanced when residence
With the inlet and the outlet concentrations, equilibrium data, time is reduced. Both these factors contribute to enhancement in
volume of the microchannels and flow rates known, Eqs. (1) and overall volumetric mass transfer coefficients with reduction in resi-
(4) can be used to calculate the overall volumetric mass transfer co- dence time as shown in Fig. 4.
efficient (KLa). To calculate the Reynolds number, average physical Fig. 6 shows the flow pattern in the split and recombine
properties of the phases, defined by Eqs. (5), (6) are used microchannel for O/A = 1/1. At 32 s residence time parallel flow
    with intermittent droplets in both the layers is observed. As the res-
Qo Qa idence time is reduced to 6.6 s, annular flow is observed which per-
ρmix ¼ ρo þ ρa ð5Þ
Qo þ Qa Qo þ Qa sists when residence time is further reduced to 5 s. Annular flow
    −1 will have more interfacial area than parallel flow causing overall volu-
1 Qo 1 Qa metric mass transfer coefficients to increase with reduction in residence
μ mix ¼ þ ð6Þ
μo Q o þ Q a μa Q o þ Q a time.

Reynolds number is calculated as 4.2. Comparison of microchannels

De ρmix U Fig. 4 shows that for the same residence time, KLa values for the
Re ¼ ð7Þ
μ mix serpentine microchannel, are, in most cases, more than KLa values
in the split and recombine microchannel. However, it will be useful
To calculate U and De the cross section of the reaction channels of to compare KLa values in the microchannels for the same Reynolds
serpentine microchannel and the cross section of the main channel number and average velocity also. Figs. 7 and 8 show this compari-
of split and recombine microchannel are used. Physical properties son. As seen from Fig. 7, for the same Reynolds number KLa values
of the equilibrated phases used in calculation of Reynolds number are higher in the split and recombine microchannel than in the ser-
are taken from the literature (Bart and Slater, 2001) and given in pentine microchannel. For the same residence time, the flow rate
Table 1. for the serpentine microchannel having 250 μL volume is more
than the flow rate for the split and recombine microchannel which
4. Results and discussion has much smaller volume. This causes the average velocity, hence
KLa which is directly related to the velocity, to be more in the serpen-
4.1. Effect of residence time on KLa tine microchannel, when comparison is done for the same residence
time. However, for the same Reynolds number, the velocities will be
Fig. 4 shows the variation of the overall volumetric mass transfer higher in the split and recombine microchannel due to its equivalent
coefficient with residence time for different O/A ratios. As can be diameter being smaller. This causes the K La values, for the same
seen, the overall volumetric mass transfer coefficient increases Reynolds number, to be the higher in the split and recombine
with reduction in residence time. In these experiments, residence microchannel than in the serpentine microchannel.
time is reduced by increasing flow rate. With reduction in residence Fig. 8 shows that for the same average velocity KLa values are higher
time, the flow rate increases causing kinetic energy, the only source in the split and recombine microchannel. Comparison for the same aver-
of energy available for generating dispersion, to increase. Increased age velocity seems to be a better way of comparing the performance of
kinetic energy makes the dispersion finer. Thus specific interfacial the microchannels as the average velocity is directly related to the kinet-
area increases with reduction in residence time. Also on reducing ic energy (the energy responsible for dispersion) and the film coeffi-
residence time, velocities prevailing inside the microchannel in- cients inside the microchannel. With kinetic energy and velocity the
crease causing film mass transfer coefficients to increase. With same, difference in the KLa values can solely be attributed to the geome-
both specific interfacial area and film mass transfer coefficients in- try of the microchannel. Therefore, among the two microchannels com-
creasing with reducing residence time, overall volumetric mass pared here, the split and recombine microchannel is found to be better.
transfer coefficient also increases with reducing residence time, as Although for the same residence time overall volumetric mass transfer
seen in Fig. 4. coefficients are higher in the serpentine microchannel than in the split
To confirm that dispersion indeed becomes finer on reducing resi- and recombine microchannel, if the later is made longer to have volume
dence time, few imaging experiments, in which flow patterns inside equal to the serpentine microchannel it is likely to show higher overall
the microchannels were observed using a high speed image acquisition volumetric mass transfer coefficients than the serpentine microchannel
system, were performed. The flow patterns observed in the serpentine even when the comparison is done on the basis of the same residence
microchannel and split and recombine microchannel for O/A = 1/1 time. The reason for better performance of split and recombine
are shown in Figs. 5, 6 respectively. Fig. 5 shows that when residence microchannel can be attributed to chaotic advection and improved
time is 98 s, long slugs with a few large droplets in between are ob- mixing in both phases due to multi-lamination. Our finding is similar
served. When residence time is reduced to 33 s, length of the slugs to that of an earlier study in which a split and recombine micromixer
reduces and droplets become smaller indicating that dispersion is found to perform better than a T-junction microchannel for a mass
transfer controlled liquid–liquid two-phase reaction (Guan et al., 2008).
Table 1
Physical properties of the phases. 4.3. Effect of O/A ratio on KLa
3
Phase Density (kg/m ) Viscosity (mPa·s)
Variation of KLa with Reynolds number for different O/A ratios for
Organic phase 758.6 1.37 the serpentine microchannel is shown in Fig. 9. In these experiments
Aqueous phase 998.5 0.958
the organic phase is the dispersed phase. As O/A ratio is increased the
58 M. Darekar et al. / Hydrometallurgy 144–145 (2014) 54–62

0.1 0.1
O/A = 4/1 O/A = 2/1

0.01 0.01

KLa (1/s)
KLa (1/s)

0.001 0.001

Serpentine Serpentine
Split & recombine Split & recombine
0.0001 0.0001
1 10 100 1000 1 10 100 1000
Residence Time (sec) Residence Time (sec)

0.1 0.1
O/A = 1/1 O/A = 1/2

0.01 0.01

KLa (1/s)
KLa (1/s)

0.001 0.001

Serpentine Serpentine
Split & recombine Split & recombine
0.0001 0.0001
1 10 100 1000 1 10 100 1000
Residence Time (sec) Residence Time (sec)

0.1
O/A = 1/4

0.01
KLa (1/s)

0.001

Serpentine
Split & recombine
0.0001
1 10 100 1000
Residence Time (sec)

Fig. 4. Variation of overall volumetric mass transfer coefficient with residence time for different O/A ratios.

holdup of the dispersed phase increases causing specific interfacial causing dispersion to become coarser and specific interfacial area
area to increase with O/A. Also with increase in O/A ratio, the flow to reduce. With these two factors affecting the specific interfacial
rate and hence dispersing power of the continuous phase reduces area in opposite ways, on increasing O/A ratio the specific interfacial
area may not change monotonically. Similarly, when O/A ratio is re-
duced, the dispersed phase holdup reduces causing specific interfa-
cial area to reduce. However, with reduced O/A ratio, flow rate and
hence dispersing power of the continuous phase increases causing
dispersion to become finer and specific interfacial area to increase.
Again with the two factors affecting the specific interfacial area in
opposite ways, specific interfacial area and hence K La may not
change monotonically as O/A is reduced.
On changing the O/A ratio, flow pattern may also change as
shown in Fig. 10. While for O/A = 1/4 loosely packed large droplets
are observed, for O/A = 1/1 closely packed slug and small droplets
are observed. For O/A = 4/1 annular flow with occasional tiny slug
in the outer layer is observed. If the flow pattern changes on chang-
ing the O/A ratio, the relationship between specific interfacial area
and the O/A ratio may become even more complex. The complex re-
lation between the O/A ratio and the specific interfacial area reflects
in the complex relation between the O/A ratio and KLa as shown in
Fig. 9. As can be seen in Fig. 9, the overall volumetric mass transfer
coefficients are the highest for O/A = 4/1 and then for O/A = 2/1.
For other O/A ratios there is no clear trend. Thus at higher O/A ratios
(2/1 and 4/1) the increase in holdup due to increase in O/A ratio seems to
5 sec 33 sec 98 sec
be dictating KLa values. While for low O/A ratios (1/1, ½ and ¼) some-
Fig. 5. Flow patterns in the serpentine microchannel for O/A = 1/1. times holdup appears to dominate and sometimes fineness of dispersion
 M. Darekar et al. / Hydrometallurgy 144–145 (2014) 54–62 59

5 sec 6.6 sec 32 sec

Fig. 6. Flow patterns in the split and recombine microchannel for O/A = 1/1.

appears to dominate. Similar trends of variation of KLa with O/A ratio volumetric mass transfer coefficient as the dependent variable to ob-
were observed for the split and recombine microchannel. tain correlations for both the microchannels. Selection of average ve-
locity is based on the fact that it is directly related to the overall
4.4. Correlations for KLa volumetric mass transfer coefficient through its effect on the specific
interfacial area and film mass transfer coefficients. Velocity has been
There are several studies on liquid–liquid extraction in microchannels used as an independent variable in the correlations reported in a re-
in which experimental data have been used to obtain correlations for cent study which compares different types of microchannels (Kashid
overall volumetric mass transfer coefficients. The correlations reported et al., 2011a). In the present study, LABFit (Silva and Silva Cleide,
in these studies are usually not dimensionless. Here we use average 1999-2010) was used to obtain the correlations for overall volumetric
velocity and O/A ratio as the independent variable and overall mass transfer coefficient. The best fit correlations for the serpentine

0.1 0.1
O/A = 4/1 O/A = 2/1

0.01 0.01
KLa (1/s)

KLa (1/s)

0.001 0.001

Serpentine Serpentine
Split & recombine Split & recombine
0.0001 0.0001
1 10 100 1000 1 10 100 1000
Re (-) Re (-)

0.1 0.1
O/A = 1/1 O/A = 1/2

0.01 0.01
KLa (1/s)

KLa (1/s)

0.001 0.001

Serpentine Serpentine
Split & recombine Split & recombine
0.0001 0.0001
1 10 100 1000 1 10 100 1000
Re (-) Re (-)

0.1
O/A = 1/4

0.01
KLa (1/s)

0.001

Serpentine
Split & recombine
0.0001
1 10 100 1000
Re (-)

Fig. 7. Variation of overall volumetric mass transfer coefficient with Reynolds number for different O/A ratios.
60 M. Darekar et al. / Hydrometallurgy 144–145 (2014) 54–62

0.1 0.1
Serpentine O/A = 4/1 Serpentine O/A = 2/1
Split & recombine Split & recombine
0.01 0.01

KLa (1/s)
KLa (1/s)

0.001 0.001

0.0001 0.0001
0.01 0.1 1 0.01 0.1 1
U (m/s) U (m/s)

0.1 0.1
Serpentine O/A = 1/1 Serpentine O/A = 1/2
Split & recombine Split & recombine

0.01 0.01

KLa (1/s)
KLa (1/s)

0.001 0.001

0.0001 0.0001
0.01 0.1 1 0.01 0.1 1
U (m/s) U (m/s)

0.1
Serpentine O/A = 1/4
Split & recombine
0.01
KLa (1/s)

0.001

0.0001
0.01 0.1 1
U (m/s)

Fig. 8. Variation of overall volumetric mass transfer coefficients with average velocity for different O/A ratios.

 
microchannel and the split and recombine microchannels are given −4 O 4
K L a ¼ −1:2092  10 þ 0:051939U−0:075017U ð9Þ
by Eqs. (8) and (9) respectively. A

The average deviations of KLa values predicted using the above

1
KLa ¼   ð8Þ correlations from the experimental KL a values are about 8% and
O 25:795
0:84998 þ 0:35556  þ
A U

0.035 Serpentine Microchannel 250 µl

0.030

0.025
KLa (1/s)

0.020

0.015

0.010 1/1 1/2

1/4 4/1
0.005
2/1
0.000
0 50 100 150 200 250
Re (-)
O/A = 4/1 O/A = 1/1 O/A = ¼
Fig. 9. Effect of O/A ratio on the overall volumetric mass transfer coefficient for the serpen-
tine microchannel. Fig. 10. Flow patterns in 250 μL serpentine microchannel at residence time of about 33 s.
 M. Darekar et al. / Hydrometallurgy 144–145 (2014) 54–62 61

0.035
Experimental
microchannel and split and recombine microchannel are within the
0.030 range of KLa values calculated for the hollow fiber contactor. With
KLa (1/s) - correlation

Split & recombine

Serpentine
KL a values in these microchannels comparable to that in a hollow
0.025 fiber contactor (a contactor considered to be process intensifying
0.020 contactor due to extremely high specific interfacial area), KLa values
in microchannels for Zinc–D2EHPA system are likely to be much
0.015 higher than in conventional extractors such as mixer settlers, packed
0.010 columns, spray columns. The comparison of extraction yield in the
microchannels and the hollow fiber contactor is not done because
0.005 the mode of operation of the two contactors – the microchannels
0.000
operated in co-current mode, the hollow fiber contactor operated
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 in counter-current mode – is different.
KLa (1/s) - experimental
5. Conclusions
Fig. 11. Comparison of experimental values of KLa with the values predicted from the
correlations.
Liquid–liquid extraction experiments have been performed in a
t-junction serpentine microchannel and a split and recombine
microchannel using Zinc–D2EHPA system. The following are the
11.2%, respectively. Fig. 11 graphically shows the quality of fit of the conclusions emanating from this experimental study:
above correlations. It may again be noted that the above correlations
are not dimensionless and require value of the velocity to be put in SI a) When compared on the basis of residence time, the t-junction
units to give KLa values in SI units. serpentine microchannel shows higher KLa values than the split
and recombine microchannel. If comparison is based on equal
4.5. Comparison with other contactors Reynolds number or equal average velocity, the split and recom-
bine microchannel exhibits higher KLa values than the serpentine
Though there are several studies on liquid–liquid extraction in microchannel.
microchannels, most of them have been carried out using phase sys- b) For both microchannels, KLa values are found to increase with re-
tems involving physical extraction. None of these studies uses Zinc– ducing residence time (or with increasing flow rate/Reynolds
D2EHPA system. The quality of dispersion generated at the microfluidic number/average velocity). This can be attributed to increase in
junction of the microchannel and hence specific interfacial area for mass specific interfacial area and film mass transfer coefficients due
transfer depends significantly on the physical properties. Also whether to increase in velocities.
the system involves physical or reactive extraction makes a difference c) While dependence of KLa on residence time exhibits monotonic
as evident by the finding of an earlier study in which height of a liq- behavior, dependence of KLa on O/A ratio is not monotonic.
uid–liquid extraction column for desired separation was calculated to d) KLa values for the microchannels used in the present study are
be more when both mass transfer and reactions were considered com- comparable with KLa values reported for a hollow fiber contactor
pared to when only mass transfer was considered (Vegliò and Slater, in experiments with Zinc–D2EHPA system. This suggests that for
1996). Thus the need of comparing different designs of microchannels Zinc–D2EHPA system, KLa values for the microchannels are likely
among themselves and with conventional contactors strictly using to be much higher than the same for conventional contactors
identical phase system cannot be overemphasized. Therefore, extensive such as mixer settler, centrifugal extractor, and spray column.
literature survey was done to find out the studies in which overall vol-
umetric mass transfer coefficients for liquid–liquid extraction with This study thus highlights the possibility of intensification of liquid–
Zinc–D2EHPA system are reported. Studies reporting overall volumetric liquid extraction processes by using microchannel contactors. However,
mass transfer coefficients for this system in conventional liquid–liquid there are a couple of technological challenges that need to be addressed
extraction equipment such as mixer settlers, spray columns, and before liquid–liquid extraction in microchannels can become a potential
centrifugal extractors could not be found. However, there are several candidate for industrial deployment. These are realizing a multistage
studies carried out in hollow fiber contactors with this phase system counter-current cascade using microchannels but without or minimum
(Fouad and Bart, 2007, 2008; Juang and Huang, 2002). In one of these inter-stage pumps and uniform distribution of feed and solvent in a
studies (Fouad and Bart, 2007) the hollow contactor has been used in large number of parallel microchannels using a very compact header.
non-dispersive counter-current mode with the aqueous phase in the
lumen and the organic phase in the shell side. The range of overall Nomenclature
volumetric mass transfer coefficients can be calculated from the range a Specific interfacial area (m−1)
of the overall mass transfer coefficient (6.2 × 10−6–25.3 × 10−6 m/s) Coi Inlet concentration of the organic phase (mol m−3)
and value of specific interfacial area (29.3 cm2/cm3) reported in the Coi⁎ Concentration of the organic phase in equilibrium with the
study. Table 2 shows the comparison of the calculated range of KLa in inlet concentration of the aqueous phase (mol m−3)
the hollow fiber contactor operated in non-dispersive mode with the Coo Outlet concentration of the organic phase (mol m−3)
range of KLa observed in the microchannels used in this study. As can be Coo⁎ Concentration of the organic phase in equilibrium with the
seen, the maximum values of K La observed in the serpentine outlet concentration of aqueous phase (mol m−3)

Table 2
Comparison of KLa values in microchannels and hollow fiber contactor for Zn–D2EHPA system.

Reference Contactor KLa (1/s)

Present study 250 μL serpentine microchannel 5.41 × 10−4–2.96 × 10−2

Present study Split and recombine microchannel 7.36 × 10−4–2.18 × 10−2
Fouad and Bart, 2007 Hollow fiber contactor 1.80 × 10−2–7.40 × 10−2
62 M. Darekar et al. / Hydrometallurgy 144–145 (2014) 54–62

De Equivalent diameter (m) Ji, J., Mensforth, K.H., Perera, J.M., Stevens, G.W., 2006. The role of kinetics in the extraction
of zinc with D2EHPA in a packed column. Hydrometallurgy 84, 139–148.
KL Overall mass transfer coefficient (m s−1) Juang, R.S., Huang, H.L., 2002. Modeling of nondispersive extraction of binary Zn(II) and
KLa Overall volumetric mass transfer coefficient (s−1) Cu(II) with D2EHPA in hollow fiber devices. J. Membr. Sci. 208, 31–38.
N Molar flux (mol m−2 s−1) Kashid, M.N., Renken, A., Kiwi-Minsker, L., 2011a. Influence of flow regime on mass
transfer in different types of microchannels. Ind. Eng. Chem. Res. 50,
Qa Volumetric flow rate of the aqueous phase (m3s−1) 6906–6914.
Qo Volumetric flow rate of the organic phase (m3 s−1) Kashid, M.N., Renken, A., Kiwi-Minsker, L., 2011b. Gas–liquid and liquid–liquid mass
Re Reynolds number (−) transfer in microstructured reactors. Chem. Eng. Sci. 66, 3876–3897.
Liu, H., Huang, J., Sun, D., Lin, L., Lin, W., Li, J., Jiang, X., Wu, W., Li, Q., 2012. Microfluidic
U Average velocity in microchannel (m s−1) biosynthesis of silver nanoparticles: effect of process parameters on size distribution.
V Volume of the microchannel contactor (m3) Chem. Eng. J. 209, 568–576.
Mehrotra, R., Jing, N., Kameoka, J., 2008. Monodispersed polygonal water droplets in
microchannel. Appl. Phys. Lett. 92, 213109.
Mörters, M., Bart, H.J., 2000. Extraction equilibria of Zinc with Bis (2-ethylhexyl)phospho-
Greek letters ric acid. J. Chem. Eng. Data 45, 82–85.
ΔLMC Log mean concentration difference (mol m−3) Okubo, Y., Maki, T., Aoki, N., Khoo, T.H., Ohmukai, Y., Mae, K., 2008. Liquid–liquid extrac-
tion for efficient synthesis and separation by utilizing micro spaces. Chem. Eng. Sci.
μa Viscosity of the aqueous phase (kg m−1 s−1) 63, 4070–4077.
μmix Viscosity of the mixed phase (kg m−1 s−1) Priest, C., Zhou, J., Kink, S., Sedev, R., Ralston, J., 2012. Microfluidic solvent extraction of
μo Viscosity of the organic phase (kg m−1 s−1) metal ions and complexes from leach solutions containing nanoparticles. Chem.
Eng. Technol. 35 (7), 1312–1319.
ρa Density of the aqueous phase (kg m−3) Principe, F., Demopoulos, G.P., 2004. Comparative study of iron(III) separation from zinc
ρmix Density of the mixed phase (kg m−3) sulphate–sulphuric acid solutions using the organophosphorous extractants, OPAP
ρo Density of the organic phase (kg m−3) and D2EHPA. Part I. Extraction. Hydrometallurgy 74, 93–102.
Rajesh, V.M., Buwa, V.V., 2012. Experimental characterization of gas–liquid–liquid flows
in T-junction microchannels. Chem. Eng. J. 207–208, 832–844.
Regatte, V.R., Kaisare, N.S., 2013. Hydrogen generation in spatially coupled cross-flow
Acknowledgment microreactors. Chem. Eng. J. 215–216, 876–885.
Ren, J., He, S., Ye, C., Chen, G., Sun, C., 2012. The ozone mass transfer characteristics and
ozonation of pentachlorophenol in a novel microchannel reactor. Chem. Eng. J. 210,
The authors acknowledge the analytical support received from LSS, 374–384.
ChED, BARC. Silva, W.P., Silva Cleide, M.D.P.S., 1999–2010. LAB fit curve fitting software (nonline-
ar regression and treatment of data program) V 7.2.47 (1999–2010). online,
available at the world wide web at www.labfit.net (date of access: 2013-03-08).
References Su, Y., Chen, G., Zhao, Y., Yuan, Q., 2009. Intensification of liquid–liquid two-phase mass
transfer by gas agitation in a microchannel. AIChE J 55 (8), 1948–1958.
Aoki, N., Tanigawa, S., Mae, K., 2011. A new index for precise design and advanced oper- Su, Y., Zhao, Y., Chen, G., Yuan, Q., 2010. Liquid–liquid two-phase flow and mass transfer
ation of mass transfer in slug flow. Chem. Eng. J. 167, 651–656. characteristics in packed microchannel. Chem. Eng. Sci. 65, 3947–3956.
Aota, A., Nonaka, M., Hibara, A., Kitamori, T., 2007. Countercurrent laminar microflow for Tamagawa, O., Muto, A., 2011. Development of cesium ion extraction process using a slug
highly efficient solvent extraction. Angew. Chem. Int. Ed. 46, 878–880. flow microreactor. Chem. Eng. J. 167, 700–704.
Assmann, N., von Rohr, P.R., 2011. Extraction in microreactors: intensification by adding Tan, J., Du, L., Lu, Y.C., Xu, J.H., Luo, G.S., 2011. Development of a gas liquid microstructured
an inert gas phase. Chem. Eng. Process. 50, 822–827. system for oxidation of hydrogenated 2-ethyltetrahydroanthraquinone. Chem. Eng. J.
Bart, H.J., 2005. Extraction columns in hydrometallurgy. Hydrometallurgy 78, 21–29. 171 (3), 1404–1414.
Bart, H.J., Slater, M.J., 2001. Standard test system for reactive extraction — Zinc/D2EHPA. Re- Tan, J., Lu, Y.C., Xu, J.H., Luo, G.S., 2012. Mass transfer performance of gas–liquid segment-
port of European Federation of Chemical Engineering, Working Party on Distillation, Ab- ed flow in microchannels. Chem. Eng. J. 181–182, 229–235.
sorption and Extraction (http://www.processnet.org (accessed on 18th March 2013)). Torkaman, R., Moosavian, M.A., Torab-Mostaedi, M., Safdari, J., 2013. Solvent extraction of
Benz, K., Jäckel, K.P., Regenauer, K.J., Schiewe, J., Drese, K., Ehrfeld, W., Hessel, V., Löwe, H., samarium from aqueous nitrate solution by Cyanex301 and D2EHPA. Hydrometallur-
2001. Utilization of micromixers for extraction processes. Chem. Eng. Technol. 24 (1), gy 137, 101–107.
11–17. Traeger, J., König, J., Städtke, A., Holdt, H.J., 2012. Development of a solvent extraction
de Loos, S.R.A., van der Schaaf, J., de Croon, M.H.J.M., Nijhuis, T.A., Schouten, J.C., 2011. En- system with 1,2-bis(2-methoxyethylthio) benzene for the selective separation
hanced liquid–solid mass transfer in microchannels by a layer of carbon nanofibers. of palladium(II) from secondary raw materials. Hydrometallurgy 127–128,
Chem. Eng. J. 167 (2–3), 671–680. 30–38.
Domínguez, M., Cristiano, G., López, E., Llorca, J., 2011. Ethanol steam reforming over co- Tsaoulidis, D., Dore, V., Angeli, P., Plechkova, N.V., Seddon, K.R., 2013. Dioxouranium(VI)
balt talc in a plate microreactor. Chem. Eng. J. 176–177, 280–285. extraction in microchannels using ionic liquids. Chem. Eng. J. http://dx.doi.org/
Fouad, E.A., Bart, H.J., 2007. Separation of Zn by a non-dispersion solvent extraction pro- 10.1016/j.cej.2012.08.064 (Article in Press).
cess in a hollow fiber contactor. Solvent Extr. Ion Exch. 25, 857–877. Vegliò, F., Slater, M.J., 1996. Design of liquid–liquid extraction columns for the possible
Fouad, E.A., Bart, H.J., 2008. Emulsion liquid membrane extraction of zinc by a hollow- test system Zn/D2EHPA in n-dodecane. Hydrometallurgy 42, 177–195.
fiber contactor. J. Membr. Sci. 307, 156–168. Wachter, B., Bart, H.J., Moosbrugger, T., Marr, R., 1993. Reactive liquid–liquid test system
Ghaini, A., Kashid, M.N., Agar, D.W., 2010. Effective interfacial area for mass transfer in the Zn/Di (2-ethyl hexyl) phosphoric acid/n-dodecane. Equilibrium and kinetics. Chem.
liquid–liquid slug flow capillary microreactors. Chem. Eng. Process. 49, 358–366. Eng. Technol. 16, 413–421.
Guan, G., Kusakabe, K., Moriyama, K., Sakurai, N., 2008. Continuous production of biodie- Xu, J.H., Li, S.W., Tan, J., Wang, Y.J., Luo, G.S., 2006a. Preparation of highly monodisperse
sel using a microtube reactor. Chem. Eng. Trans. 14, 237–244. droplet in a T-junction microfluidic device. AIChE J 52 (9), 3005–3010.
Hotokezaka, H., Tokeshi, M., Harada, M., Kitamori, T., Ikeda, Y., 2005. Development of in- Xu, J.H., Luo, G.S., Li, S.W., Chen, G.G., 2006b. Shear force induced monodisperse droplet
novative nuclide separation system for high level radioactive waste using formation in a microfluidic device by controlling wetting properties. Lab Chip 6,
microchannel chip-extraction behavior of metal ions from aqueous phase to organic 131–136.
phase in microchannel. Prog. Nucl. Energy 47, 439–447. Xu, J.H., Tan, J., Li, S.W., Luo, G.S., 2008. Enhancement of mass transfer performance of liq-
Innocenzi, V., Vegliò, F., 2012. Separation of manganese, zinc and nickel from leaching so- uid–liquid system by droplet flow in microchannels. Chem. Eng. J. 141, 242–249.
lution of nickel-metal hydride spent batteries by solvent extraction. Hydrometallurgy Zhao, Y., Chen, G., Yuan, Q., 2007. Liquid–liquid two-phase mass transfer in the T-junction
129–130, 50–58. microchannels. AIChE J 53 (12), 3042–3053.