Chemical Engineering Science 155 (2016) 38–44

Contents lists available at ScienceDirect

Chemical Engineering Science

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

Gas–liquid mass transfer in a falling film microreactor: Effect of reactor

orientation on liquid-side mass transfer coefficient
David Lokhat n, Ashveer Krishen Domah, Kuveshan Padayachee, Aman Baboolal,
Deresh Ramjugernath
Reactor Technology Research Group, School of Engineering, University of KwaZulu-Natal, Durban 4041, South Africa

H I G H L I G H T S G R A P H I C A L A B S T R A C T

 Lower mass transfer rates in inclined

microchannels.
 Curved interface approximation
gives more realistic mass transfer
coefficients.
 Cellular convection in microchannels
responsible for enhanced mass
transfer.

art ic l e i nf o a b s t r a c t

Article history: Microreactors offer a unique platform for chemical syntheses and have been applied to numerous re-
Received 2 May 2016 action types including nitrations, fluorinations and hydrogenations. A key feature of falling film micro-
Received in revised form reactors is the comparably large specific surface area they afford compared to conventional reactors. The
20 July 2016
enhanced heat and mass transfer characteristics can be exploited for rapid and exothermic reactions.
Accepted 3 August 2016
Available online 4 August 2016
Adequate understanding of the mass transfer processes occurring within microchannels is necessary for
proper reactor design and optimization. In the current study the influence of reaction plate orientation
Keywords: and gas flowrate on liquid-side mass transfer coefficient was investigated via CO2 absorption experi-
Microreactor ments. Lower plate angles resulted in lower liquid-side mass transfer coefficients. At higher film velo-
Mass transfer
cities the rate of mass transfer was greater. The experimentally determined mass transfer coefficients
CO2 absorption
were at least twice as high as those predicted either by film or penetration theory. The enhancement in
Inclined plate
mass transfer is suggested to be due to cellular convection in the microchannels. For inclined reaction
plates, increasing the gas flowrate had a positive effect on the mass transfer characteristics due to in-
duced fluctuations of the gas–liquid interface.
& 2016 Elsevier Ltd. All rights reserved.

1. Introduction down a microstructured reaction plate under the action of gravity

Microreactors are reactors with reaction channels of the order
of micrometers, at which diffusion is the dominant mixing me-
chanism rather than turbulent eddies. The interest in this tech-
nology has been focused on fine and speciality chemical manu-
facture (Fletcher et al., 2002). The falling film microreactor (FFMR)
is the most popular type of gas–liquid contactor in this class of
reactors. Under normal operating conditions, the liquid flows
n
Corresponding author.
E-mail address: lokhat@ukzn.ac.za (D. Lokhat).
and spreads to form an expanded thin film in each microchannel.
The gas flows over the liquid film contained within the micro-
channels. This design allows for specific gas–liquid interfacial areas
of up 20,000 m2/m3 to be achieved, which is noticeably higher
than other gas–liquid contacting reactors (Hessel et al., 2000; Yue
et al., 2007; Ziegenbalg et al., 2010). The high specific areas made
available by the reactor design allow for high heat and mass
transfer rates to be achieved which make FFMRs an efficient tool
for studying the kinetics of fast and highly exothermic gas–liquid
reactions. Researchers have exploited the high heat and mass
transfer rates of FFMRs for reactions such as catalysed

http://dx.doi.org/10.1016/j.ces.2016.08.002
0009-2509/& 2016 Elsevier Ltd. All rights reserved.
 D. Lokhat et al. / Chemical Engineering Science 155 (2016) 38–44
Nomenclature
AL liquid cross sectional area in the microchannel, m2
Cout liquid phase concentration of CO2 in the reactor outlet,
mol
dm 3
Cn saturated concentration of CO2, mol
dm 3
CHCl concentration of HCl, mol
dm 3
D CO2 diffusivity in the liquid phase, m2
s 1
Fo Fourier number
jL mean liquid velocity, m
s 1
kL liquid-side mass transfer coefficient, m
s 1
kL, pen penetration model liquid-side mass transfer coeffi-
cient, m
s 1
kL, film film model liquid-side mass transfer coefficient,
m
s 1
Larc arc length of parabola, m
Lch length of microchannel, m
Nch number of channels
hydrogenation (Yeong et al., 2003), direct fluorination of aromatics
(Jähnisch et al., 2000) and direct synthesis of hydrogen peroxide
(Inoue et al., 2007).
The measurement of gas–liquid mass transfer in a falling film
micro reactor has formed the basis for a number of published
studies (Zanfir et al., 2005; Zhang et al., 2009). The reactive ab-
sorption of carbon dioxide (CO2) and sulphur dioxide (SO2) in
aqueous sodium hydroxide have both been used effectively to
characterize mass transfer efficiency in microreactors. For the
measurement of liquid phase mass transfer coefficients, CO2 ab-
sorption is preferred. The CO2 reacts rapidly with the NaOH at the
gas–liquid interface and forms sodium carbonate. The analysis of
liquid reaction products can be done simply by titration and the
reactants are inexpensive and non-toxic. For the measurement of
gas phase mass transfer coefficient, the SO2 method is preferred.
Very low concentrations of the reactive gas-phase species must be
used for measurement of the gas-phase mass transfer coefficient,
to ensure that the process is limited by the gas phase mass transfer
(Commenge et al., 2011). An SO2 gas analyser is much more sen-
sitive and selective at this concentration range compared to CO2
sensors which may encounter interference from the nitrogen
diluent.
Applications of these methods are illustrated in the literature.
Zanfir et al. (2005) carried out an experimental and modelling
study of the reactive absorption of carbon dioxide in a falling film
microreactor. A two-dimensional model was formulated to re-
present the flow through the reactor and the experimental results
were compared to the model prediction in terms of carbon dioxide
conversion. The model gave good agreement with the experi-
mental data at low inlet NaOH concentration. Moreover, the
parametric analysis of the model indicated that the major rate
limitation was on the liquid side and that the CO2 was consumed
within a short distance from the gas–liquid interface. Zhang et al.
(2009) studied the hydrodynamics and mass transfer of gas–liquid
flow in a FFMR. Once again the CO2 absorption method was em-
ployed, with titration of the liquid phase product to determine the
efficiency of the process. Al-Rawashdeh et al. (2008) developed a
three-dimensional computational fluid dynamic model of a falling
film microreactor based on realistic channel geometry profiles. The
model was validated experimentally by the absorption of CO2 in
NaOH aqueous solution. The model allowed the authors to in-
vestigate the effects of channel fabrication precision, liquid flow
distribution, gas chamber height, and hydrophilic-hydrophobic
plate material on the conversion of CO2. It was found that
39
QL liquid volumetric flowrate, m3
s 1
Re Reynold number
t collection time, s
Wch width of microchannel, m
We Weber number
VHCl volume of HCl dispensed between titration endpoints,
dm3
x width dimension of microchannel
y depth dimension of microchannel
Greek symbols
δL thickness of falling film, m
νL kinematic viscosity of liquid, m2
s 1
θ angle of reaction plate, °
sL surface tension of liquid, N
m 1
ρL density of liquid, kg
m 3
fabrication imprecisions and liquid maldistribution lowered the
reaction conversion by approximately 2%. However, none of these
previous studies have considered the effect of reactor orientation
on the rate of mass transfer. The reactor orientation is considered
an additional operating parameter that can be used to adjust the
performance of a FFMR. In particular it has been suggested that
changes to the reactor orientation can be used to modify liquid
residence times for slower reactions (Hessel et al., 2008). There is a
need therefore to understand the effect of reactor orientation on
other operational characteristics of the apparatus, particularly the
rate of mass transfer.
In this work CO2 absorption measurements were used to
quantify the effect of the degree of inclination of the reaction plate
and gas flowrate in a FFMR on the liquid-side mass transfer
coefficient. For the specific geometry of the microchannels used in
this study, approximations for the shape and length of the liquid
menisci were used to compute the mass transfer coefficients.
These results were compared to the case of a flat interface and the
relative differences were analysed.
2. Experimental
The FFMR used in this study is a well-known and well-studied
microstructured reactor which was invented by the Institut für
Mikrotechnik Mainz GmbH and has successfully shown its ability
to maintain thin liquid films of up to 100 mm in thickness (Al-
Rawashdeh et al., 2008). The FFMR consisted of a copper cooling
plate, a stainless steel reaction plate with etched microchannels
and a viewing window, fabricated from glass, which enclosed the
gas headspace. The device was fabricated from 316Ti stainless steel
and was equipped with a microstructured panel containing 32
microchannels of dimensions 641 mm  257 mm  66.4 mm. The
experimental system is shown in Fig. 1.
Carbon dioxide gas (Afrox,499%) was supplied in a standard
cylinder. The feed gas flowrate was controlled using a thermal
mass flow controller (Alicat). Inlet gas flowrates were fixed at ei-
ther 50 cm3
min 1 and 80 cm3
min 1. Deionized water (con-
ductivity 0.07 μS) was mixed with ethylene glycol (Riedel-de
Haen, 499%) to produce a 12 wt% solution and was used as the
liquid feed. This ensured satisfactory wetting of the channel sur-
face due to a lower surface tension of the liquid. In all experiments
the liquid flow into the microreactor was controlled using a HPLC
pump. The feed pump supplies liquid to a header. The liquid
40 D. Lokhat et al. / Chemical Engineering Science 155 (2016) 38–44
Fig. 1. Experimental setup for CO2 absorption experiments.
overflows from the header through slits into the microchannels,
falling under gravity to the lower collection plate from where it is
withdrawn. The withdrawal pump is a peristaltic-type pump op-
erating at a flowrate higher than the inlet pump to ensure that
there is no accumulation of fluid in the channels. Since the liquid
flow is gravity driven the high flow withdrawal pump does not
completely empty the channel. The temperature in the micro-
reactor (298 K) was maintained using an external water bath cir-
culator. In a typical experiment, the system was initially flooded
with liquid. The withdrawal pump was then activated and the inlet
flow was set at the test value using the HPLC pump until a uniform
liquid film was formed within each microchannel. The gas was
admitted and the outlet liquid was directed to a beaker containing
a solution of NaOH (0.1 mol
dm 3). The CO2 absorbed by the
water was converted to Na2CO3. The experiment was allowed to
proceed for 20 min. The contents of the beaker was then titrated
against a 0.1 mol
dm 3 solution of HCl. The outlet liquid phase
concentration of CO2 was then given by:
VHCl CHCl
Cout =
Q Lt
where VHCl is the volume of HCl dispensed between the titration
endpoints (representing hydroxide and carbonate species) and t is
the collection time (Zhang et al., 2009). Special wooden cradles at
60° and 75° were fabricated to hold the reactor unit for the in-
clined plate experiments. Each experiment was repeated three
times and average results are presented for each set of conditions.
3. Results and discussion
Falling liquid films may be described through the use of three
progressive regimes, based on the Reynolds number of the flow. In
the original work on the subject by Nusselt (1916) such flow was
restricted to flat plates and correction factors are introduced for
structured plates. The Reynolds number is defined as four times
the volumetric flow rate per unit width of channel divided by the
kinematic viscosity.
4Q L
ReL =
Wch νL (2)
For Reynolds numbers up to 25 the liquid flow is characterized
by a thin film with a smooth, flat interface and is referred to as the
smooth laminar regime. Between Reynolds numbers of 25-1000
(laminar flow with rippling surface) the flow is characterized by a
wavy interface with fluctuating film thickness. Turbulent flow in
falling liquid films is usually encountered above a Reynolds
number of 1000. Flow patterns in FFMRs can be characterized as
either corner rivulet flow, falling film with dry patches or complete
falling film flow. The latter is the most appropriate for gas–liquid
mass transfer and reaction. For complete falling film flow char-
acterized by a continuous liquid film without dry patches, laminar
flow conditions do prevail according to Nusselt theory. The validity
of these regimes have been confirmed experimentally (Commenge
et al., 2011; Zhang et al., 2009).
The first study of falling liquid films was conducted by Nusselt
(1916). He assumed a steady state system, a flat interfacial region
and rectilinear flow (i.e. all streamlines are straight) and con-
structed a shell momentum balance over a thin layer of the fluid.
(1)
In the limit, the thickness of this layer was allowed to approach
zero in order to generate a differential equation which could be
integrated to represent the momentum flux. The differential
equation was solved to give the velocity profile of the fluid. The
boundary conditions used to solve these equations for falling li-
quid films were: (i) the velocity of the fluid at the wall must be
zero relative to the velocity of the wall and (ii) the momentum
flux, or shear stress, and velocity gradient must be zero at the li-
quid-gas interface. By following the steps indicated above for a
falling film, equations describing the velocity profile, wall shear
stress and film thickness were derived. The flow through micro-
sized channels is characterized by a very low Reynolds number
and is generally considered laminar. Commenge et al. (2011) state
that the film thickness in a FFMR can be related to the liquid
flowrate, liquid properties and geometric parameters according to
Nusselt condensation theory:
3Q L ν L
δL = 3
Nch Wch g (3)
 D. Lokhat et al. / Chemical Engineering Science 155 (2016) 38–44 41

gδL2 channel grooves were thus assumed to have an elliptical shape

jL = and the liquid meniscus the form of a parabolic arc (cf. Fig. 2).
3ν L (4)
Based on the locus of the parabolic vertex and the radii of the
Zhang et al. (2009) also propose the use of these equations to ellipse the following functions were developed:
relate the liquid flowrate, mean velocity and film thickness. For
⎡ ⎤
exactly the same type of reactor as that used in this study, Al- ⎢ Dch − δL ⎥ ⎛ W ⎞2
Rawashdeh et al. (2008) state that the liquid flow is solely driven y1 = ⎢ 2⎥
× ⎜ x − ch ⎟ + δL
⎝ 2 ⎠
by gravity and found that Eqs. (3) and (4), developed for a free (
⎢⎣ Wch − Wch
2 ) ⎥⎦
(6)
falling film over a flat plate, are also valid for flow in the micro-
channels at liquid flowrates below 10 ml
min 1.
⎡
( x − ) ⎤⎥⎥
0.5
Wch 2
Liquid films in microchannels can exhibit morphologies other ⎢ 2 2
than the flat profile. Due to the action of capillary forces, the y2 = ⎢ 1 − × Dch + Dch
Wch 2
surface of the liquid film in microchannels usually takes the form ⎢⎣ ( ) ⎥⎦
2 (7)
of a flowing meniscus rather than a completely flat film (Zhang
et al., 2010). The shape of the film depends on the wetting ability In Fig. 2, y1 is the function representing the parabolic meniscus
of the liquid and aspect ratio of the channels (Al-Rawashdeh et al., and y2 is the function representing the elliptical channel. The area
2008). Under favourable conditions (low contact angle, low liquid in between these two curves (the liquid flow area) is given by:
flow-rate and proper aspect ratio) the liquid will form a meniscus Wch
that is pinned at the edges of the groove. The Weber number AL = ∫0 ( y1 − y2 ) dx
(8)
characterizes the ratio of liquid kinetic energy to interfacial en-
ergy. The volumetric flowrate through the microchannel is:

ρL δL jL2 Q L = jL AL Nch (9)

We =
σL
3
At a sufficiently low Weber number ( o10 ), the interfacial
energy dominates and a meniscus in the shape of a concave cir-
cular arc can be assumed (Al-Rawashdeh et al., 2008). The mi-
croreactor used in this study was equipped with an etched reac-
tion plate consisting of a number of rounded microgrooves. The
Weber number for each experiment (cf. Table 1) was calculated
based on the fluid properties and the film thickness and velocity
from Eqs. 3 and 4. In all cases the Weber number was below the
critical value and a curved meniscus was extant. The film thickness
and liquid velocity were then determined by an alternative
method based on approximate geometries of the channel and film
meniscus. Anastasiou et al. (2013) carried out a systematic in-
vestigation of the liquid phase hydrodynamics in FFMRs. The effect
of various fluid parameters as well as geometrical characteristics
and inclination angle of the conduit on the geometry of the liquid
phase was studied. For all cases it was found that the shape of the
meniscus could be described accurately using a quadratic equation
representing a parabolic arc. A true circular arc is not possible
since the meniscus is pinned to the outer edges of the channel and
the width of the channel is much greater than the depth. The
(5) where the value of jL can be obtained from Eq. 3. For those ex-
periments that were conducted at plate orientations other than
90°, the angular component of the gravitational acceleration was
used to correct the velocity:
sin θ × gδL2
jL =
3ν L (10)
At a completely horizontal configuration the liquid velocity
becomes zero since the liquid from the pump merely fills the
header without overflow into the microchannels. Eventually when
enough liquid has accumulated in the header it will begin to
spread into the microchannels and the gas cavity. This is an
anomalous situation that is not encountered during normal op-
eration of the device, i.e. the basis for operation of the FFMR is that
there is some inclination which induces flow through the micro-
channels after overflow from the header.
The film thickness is not known a priori. For each of the three
volumetric flowrates used in this study, the film thickness is found
from the simultaneous solution of Eqs. (6)–(10). That is, the vo-
lumetric flowrate governs the thickness of the film and the velo-
city. The results are presented in Table 1. Fig. 3 shows the liquid

Table 1
Experimental results for gas–liquid mass transfer experiments a.

Experiment 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Plate angle / ° 90 90 90 90 90 90 75 75 75 75 75 75 60 60 60 60 60 60
QL /cm3
min 1 3 4 5 3 4 5 3 4 5 3 4 5 3 4 5 3 4 5
QG /cm3
min 1 50 50 50 80 80 80 50 50 50 80 80 80 50 50 50 80 80 80

Flat interface
jL / m
s 1,  102 2.60 3.15 3.66 2.60 3.15 3.66 2.51 3.04 3.53 2.51 3.04 3.53 2.25 2.73 3.17 2.25 2.73 3.17
δL / m,  104 0.94 1.03 1.11 0.94 1.03 1.11 0.94 1.03 1.11 0.94 1.03 1.11 0.94 1.03 1.11 0.94 1.03 1.11
We,  104 2.35 3.79 5.50 2.35 3.79 5.50 2.19 3.54 5.13 2.19 3.54 5.13 1.76 2.84 4.13 1.76 2.84 4.13
Fo 0.48 0.32 0.24 0.48 0.32 0.24 0.49 0.34 0.25 0.49 0.34 0.25 0.55 0.38 0.28 0.55 0.38 0.28
kL / m
s 1,  105 5.73 6.41 6.65 5.76 6.48 6.75 5.20 5.86 6.09 5.46 6.12 6.45 5.13 5.56 5.87 5.26 5.89 6.09

Curved interface
jL / m
s 1,  103 7.16 9.70 12.3 7.16 9.70 12.3 7.17 9.72 12.3 7.17 9.72 12.3 7.22 9.79 12.4 7.22 9.79 12.4
δL / m,  105 4.92 5.72 6.44 4.92 5.72 6.44 5.01 5.83 6.57 5.01 5.83 6.57 5.30 6.18 6.96 5.30 6.18 6.96
Larc / m,  104 7.91 7.81 7.72 7.91 7.81 7.72 7.90 7.80 7.71 7.90 7.80 7.71 7.86 7.75 7.66 7.86 7.75 7.66
Fo 6.29 3.43 2.13 6.29 3.43 2.13 6.05 3.30 2.05 6.05 3.30 2.05 5.36 2.91 1.81 5.36 2.91 1.81
kL / m
s 1,  105 4.64 5.26 5.52 4.67 5.32 5.60 4.22 4.81 5.07 4.43 5.03 5.36 4.18 4.60 4.91 4.29 4.87 5.10
kL, pen / m
s 1,  105 1.84 2.14 2.41 1.84 2.14 2.41 1.84 2.14 2.41 1.84 2.14 2.41 1.84 2.15 2.42 1.84 2.15 2.42
kL, film / m
s 1,  105 3.34 2.87 2.54 3.34 2.87 2.54 3.28 2.81 2.50 3.28 2.81 2.50 3.09 2.65 2.36 3.09 2.65 2.36

a 7 1
Average deviation on calculated mass transfer coefficients from repeated experiments was 72  10 m
s
42 D. Lokhat et al. / Chemical Engineering Science 155 (2016) 38–44

Fig. 2. Approximations of channel and film geometries.

cross sectional profile in the channel at each of the three flowrates where L arc is the arc length of the parabola representing the top of
for the vertically oriented plate. In all cases the calculated film the liquid film:
thickness for the curved interface was lower than that for the flat Wch
interface. The film profiles obtained using these approximations L arc = ∫0 1 + y1 ′(x)2
(12)
compare favourably with the results of Al Rawashdeh et al. (2008).
The liquid-side mass transfer coefficient was calculated based Thus,
on the method developed by Zhang et al. (2009). A material bal-
QL ⎛ C* ⎞
ance over an elementary shell of the falling film for a flat interface kL = ln ⎜ ⎟
L arc × L ch ⎝ C * − Cout ⎠ (13)
yields the following:

QL ⎛ C* ⎞ Calculations showed that on average there is a 22% difference in

kL = ln ⎜ ⎟ the mass transfer coefficients obtained using the flat interface and
Wch L ch ⎝ C − Cout ⎠
* (11)
curved interface approximations.
For a curved meniscus the interfacial area is given by L arc × L ch , Fig. 4 shows the calculated liquid side mass transfer coefficients

Fig. 3. Liquid cross sectional profile in the channel for flowrates of (a) 3 cm3
min 1
, (b) 4 cm3
min 1
and (c) 5 cm3
min 1
. Reaction plate oriented at 90°.
 D. Lokhat et al. / Chemical Engineering Science 155 (2016) 38–44
Fig. 4. Liquid-side mass transfer coefficient as a function of liquid film velocity, all
plate orientations.
at various liquid velocities, gas flowrates and plate angles. A de-
crease in the plate angle gave lower mass transfer coefficients.
There was also a distinct difference in the mass transfer coeffi-
cients at high and low gas flowrates for 60° and 75° angles. When
the reactor was oriented vertically there was no significant dif-
ference in the mass transfer coefficients obtained at the two gas
velocities.
Several mass transfer models have been proposed to describe
the absorption of gas into a liquid film (King, 1966). Zhang et al.
(2009) consider the film and penetration theories of interfacial
mass transfer for theoretical validation of the experimental results
specifically due to the presence of a laminar flow regime. The film
theory assumes a stagnant film exists near the interface. In the
penetration theory, the interfacial region is imagined to be a very
thick film continuously generated by flow, which may be laminar.
For both of these theories turbulence in the bulk is not a requisite.
The surface renewal theory on the other hand requires both a thick
interfacial film region and a large well mixed bulk region. The
latter theory was therefore not considered in the analysis of the
experimental results.
The Fourier number, the ratio of liquid residence time to dif-
fusion time, can be used to determine the applicability of a par-
ticular model.
L ch D
Fo =
jL δL2
If the Fourier number is much less than 1, then the penetration
model is applicable and the mass transfer in the falling film is far
from equilibrium. On the other hand if the Fourier number is much
greater than 1, then sufficient time is available to reach equili-
brium and the film model is applicable (Zhang et al., 2009). The
Fourier number was calculated for each data point using the flat
interface and curved interface approximations. For the former, all
Fourier numbers were lower than 1 whereas for the realistic liquid
profiles the Fourier numbers were greater than 1 implying that the
film model was applicable. Mass transfer coefficients were pre-
dicted based on each of these fundamental models:
1.5jL D
kL, pen = 2
πL ch
D metries based on elliptic and parabolic functions have yielded film
kL, film =
δL
The results of these calculations are presented in Table 1 and
illustrated in Fig. 5 for the vertically oriented reaction plate.
Both models underpredicted the mass transfer coefficient by at
least half the experimental value. Interestingly, the experimental
43
Fig. 5. Liquid-side mass transfer coefficient as a function of liquid film velocity for
vertically oriented reaction plate.
trend showed more similarity to the penetration model predic-
tions than the film model. Similar results were obtained in the
literature for CO2 absorption in a microreactor containing rectan-
gular microchannels (Zhang et al., 2009). Zhang et al. (2009)
suggested that the discrepancy in the calculated and predicted
mass transfer coefficients was due to the presence of a wavy gas–
liquid interface and a protracted liquid residence time as a result of
strong side-wall shear stress in the microchannels. Sobieszuk et al.
(2010) on the other hand presented evidence of mass transfer
enhancement through cellular convection (Marangoni effect) in
the microchannels of a falling film microreactor. This cellular
convection phenomenon gives behaviour similar to the motion of
turbulent eddies that is proposed to be the principle mechanism of
interphase mass transfer underlying the penetration theory, but in
a much more intensified manner. This may be the reason why the
experimental data shows an increase in mass transfer coefficient
as the liquid velocity increases. The enhancement in mass transfer
due to cellular convections appears to be highest when the reac-
tion plate is vertically oriented. As the plate angle is reduced the
effect is less pronounced, resulting in a lower mass transfer
coefficient.
As shown by Zhang et al. (2009) the effect of gas flowrate on
liquid-side mass transfer coefficient is not significant for a verti-
cally oriented plate (cf. Fig. 5). However the results of this study
also show that at lower plate angles distinctly higher mass transfer
(14) coefficients are obtained at higher gas flowrates (cf. Fig. 4). This
can only be connected to the disturbance of the liquid interface as
induced by the higher gas flowrate and hence an increase in the
wavy region of the falling film (Zhang et al., 2009; Dukler, 1977).
4. Conclusions
CO2 absorption experiments in an FFMR have shown that the
angle of orientation of the reaction plate has a direct influence on
the rate of mass transfer, i.e. lower plate angles give lower liquid-
side mass transfer coefficients. Mass transfer appears to proceed
according to the penetration model but at significantly enhanced
rates. These observations appear to be the result of cellular con-
vection which is more pronounced in the vertically oriented re-
(15)
action plate. Approximations for the channel and meniscus geo-
thicknesses and velocities 50% lower than the flat interface ap-
(16)
proximation. These values compare favourably with literature (Al-
Rawashdeh et al., 2008).
The effect of gas flowrate on liquid-side mass transfer coeffi-
cient is negligible at 90° plate angle but a significant difference is
observed over a narrow range of gas flowrates for 75° and 60°
44 D. Lokhat et al. / Chemical Engineering Science 155 (2016) 38–44
plate angles. Increased gas flow at lower plate angles has a
stronger effect on the hydrodynamics of the liquid phase, inducing
rippling at the gas–liquid interface and extending the wavy in-
terface region.
Acknowledgement
This work was supported by the South African Research Chairs
Initiative of the Department of Science and Technology and the
National Research Foundation. This work is based on the research
supported in part by the National Research Foundation of South
Africa for the grant, Unique Grant No. 99339.
References
Anastasiou, A.D., Gavriilidis, A., Mouza, A.A., 2013. Study of the hydrodynamic
characteristics of a free flowing liquid film in open inclined microchannels.
Chem. Eng. Sci. 101, 744–754.
Al-Rawashdeh, M., Hessel, V., Lob, P., Mevissen, K., Schonfeld, F., 2008. Pseudo 3-D
simulation of a falling film microreactor based on realistic channel and film
profiles. Chem. Eng. Sci. 63, 5149–5159.
Commenge, J.M., Obein, T., Framboisier, X., Rode, S., Pitiot, P., Matlosz, M., 2011. Gas-
phase mass-transfer measurements in a falling-film microreactor. Chem. Eng.
Sci. 66, 1212–1218.
Dukler, A.E., 1977. The role of waves in two-phase flow: some new understandings.
Chem. Eng. Ed. 11, 108–117.
Fletcher, P.D.I., Haswell, S.J., Pombo-Villar, E., Warrington, B.H., Watts, P., Wong, Y.F.,
Zhang, X., 2002. Microreactors: principles and applications in organic synthesis.
Tetrahedron 58, 4735–4757.
Hessel, V., Ehrfeld, W., Herweck, T., Haverkamp, V., Lowe, H., Schiewe, J., Wille, C.,
Kern, T., Lutz, N., 2000. Gas/liquid microreactors: hydrodynamics and mass
transfer. In: Proceedings of the Fourth International Conference on Micro-
reaction Technology, Atlanta, USA, 5–9 March, pp. 174–186.
Hessel, V., Löb, P., Löwe, H., 2008. Gas-liquid reactions. In: Wirth, T. (Ed.), Micro-
reactors in Organic Synthesis and Catalysis. Wiley-VCH, Weinheim, p. 142.
Jähnisch, K., Baerns, M., Hessel, V., Ehrfeld, W., Haverkamp, V., Loewe, H., Wille, C.,
Guber, A., 2000. Direct fluorination of toluene using elemental fluorine in gas/
liquid microreactors. J. Fluor. Chem. 105, 117–128.
King, C.J., 1966. Turbulent liquid phase mass transfer at free gas-liquid interface.
Ind. Eng. Chem. Fund. 5, 1–8.
Nusselt, W., 1916. Die oberflachenkondensation des wasserdamphes. VDI For-
chungscheft 60, 541.
Sobieszuk, P., Pohorecki, R., Cygański, P., Kraut, M., Olschewski, F., 2010. Marangoni
effect in a falling film microreactor. Chem. Eng. Sci. 164, 10–15.
Yeong, K.K., Gavriilidis, A., Zapf, R., Hessel, V., 2003. Catalyst preparation and de-
activation issues for nitrobenzene hydrogenation in a falling film reactor. Catal.
Today 81, 641–651.
Yue, J., Chen, G., Yuan, Q., Luo, L., Gonthier, Y., 2007. Hydrodynamics and mass
transfer characteristics in gas–liquid flow through a rectangular microchannel.
Chem. Eng. Sci. 62, 2096–2108.
Zanfir, M., Gavriilidis, A., Wille, C., Hessel, V., 2005. Carbon dioxide absorption in a
falling film microstructured reactor: experiments and modeling. Ind. Eng.
Chem. Res. 44, 1742–1751.
Zhang, H., Chen, G., Yue, J., Yuan, Q., 2009. Hydrodynamics and mass transfer of
gas–liquid flow in a falling film microreactor. AIChE J. 55, 1110–1120.
Zhang, H., Yue, J., Chen, G., Yuan, Q., 2010. Flow pattern and break-up of liquid film
in single-channel falling film microreactors. Chem. Eng. J. 163, 126–132.
Ziegenbalg, D., Löb, P., Al-Rawashdeh, M.M., Kralisch, D., Hessel, V., Schönfeld, F.,
2010. Use of ‘smart interfaces’ to improve the liquid-sided mass transport in a
falling film microreactor. Chem. Eng. Sci. 65, 3557–3566.