Article

pubs.acs.org/OPRD

Liquid−Liquid Mass Transfer in an Oscillatory-Flow Mesoscale Coil

Reactor without Baffles
Sébastien S. Mongeon,† Dominique M. Roberge,*,‡ Michael Bittel,‡ Petteri Elsner,‡ and Arturo Macchi*,†
†
Centre for Catalysis Research and Innovation, Department of Chemical and Biological Engineering, University of Ottawa, K1N 6N5,
Ottawa, Canada
‡
Chemical Manufacturing Technologies, Lonza AG, CH-3930, Visp, Switzerland

ABSTRACT: Interphase mass transfer rates of an immiscible liquid−liquid system are investigated in an oscillatory-flow coil
reactor without baffles. A baffle-less system is chosen since much research has been completed in baffled tubes with little
performance comparison to baffle-less tubes, which offer lower operating and capital costs. In our experiments, interphase mass
transfer rates are evaluated via the two-phase alkaline hydrolysis of 4-nitrophenyl acetate (4-NPA) in a toluene−water biphasic
system. The mass transfer rate increased 7-fold at the maximum tested oscillation of 5000 oscillatory-flow Reynolds number. The
best application for the oscillatory-flow coil reactor is determined using a comparison to other liquid−liquid flow platforms in the
presented toolbox approach. The oscillatory-flow coil reactor becomes a clear complement to a plate microreactor for gaining
volume.

1. INTRODUCTION baffled reactors (meso-OBR)6,7 are currently being used to

A microreactor has become an essential tool in the develop- perform mixing of liquid−liquid systems. These meso-OBR
reactors have diameters smaller than 15 mm.6
ment of flow processes in the fine chemical and pharmaceutical
Further research is needed in the area to determine the
industry by enabling small flow rates and minimizing reagents.1
effect of flow oscillation on mass transfer in baffle-less tubes; to
A plate microreactor is useful for reactions that are transport-
determine the transition point in flow regime8,9 and its effect on
limited requiring rapid mass and heat transfer rates. A plate
mass transfer; and to obtain information for mesoscale reac-
with a proper structure is selected according to the phases of a
tors at oscillatory-flow Reynolds number in the turbulent-flow
reaction: homogeneous, liquid−liquid, or gas−liquid.1 In such regime, typically in the range between 1000 and 10 000, since
case, the cost per heat exchange area is attractive. However, current studies in meso-OBR have not gone above 316.6 This
upon scale-up, it is not the most appropriate reactor when a information will allow determining the best applications for
larger volume is required where the cost per volume may the reactor when comparing other flow platforms such as plate
become prohibitive. In such case, a metallic tube/coil, from its microreactors and meso-OBR in a reactor selection toolbox
simpler manufacturing technique, becomes the reactor of choice approach.1
for flow technologies in which a volume gain is required. A
tubular heat exchanger, a tube filled with static mixers/baffles,
and a long packed column are all variations of tube/coil and its 2. FLOW CHARACTERIZATION
simplest generic form remains the most useful for: resistance to Pulsatile and oscillatory-flow are respectively defined as when
fouling, ease of cleaning and operation, and ultimately cost. This the net-flow is greater than the amplitude of the oscillating
article will demonstrate that by a simple oscillation technique the component (i.e., the flow is always positive) and when the net-
coil reactor can become quite versatile enabling liquid−liquid flow is lesser than the amplitude of the oscillating component
reactions. (i.e., a negative or reverse-flow component is present at a
In continuous flow, mixing is categorized into passive and certain point in the cycle). Presented in Figure 1, the relation-
active mixing.1,2 Passive mixing uses the energy input of the ship between the net-flow and superimposed pulsatile/
flow, i.e., pump or source of pressure, as the single source of oscillatory-flow is broken down into five cases.9,10 The focus
mixing energy. The challenge in passive mixing comes from the of this research is on the reverse pulsatile/asymmetrical oscillatory-
compromise between selecting a production rate and obtaining flow.
the desired mixing as both properties are related to the same The steady unidirectional flow has a constant net positive
energy input. Active mixing uses an additional energy source velocity with zero amplitude pulse, i.e., flow without pulsation.
to create mixing, e.g., pressure disturbance, electrokinetics, The pulsing unidirectional flow has a net velocity larger than
acoustic, and radiation, among others.2 The use of active mixing the pulsation amplitude, which means that the fluid always
allows the independent manipulation of mixing intensity and move in the forward direction, cycling through acceleration and
residence time. This decoupling provides better control over deceleration. The start and stop flow is the transition between
the system to attain the desired mixing result at a chosen pulsatile and oscillatory flow; it has a net flow equal to the
production rate.
Many types of mixing structures/arrangement in plate Received: November 2, 2015
microreactors2−5 and mesoscale or millimeter scale oscillatory Published: March 15, 2016

© 2016 American Chemical Society 733 DOI: 10.1021/acs.oprd.5b00356

Org. Process Res. Dev. 2016, 20, 733−741
Organic Process Research & Development Article

scale and fluid component systems through the ratio of the

inertial to the viscous forces. When used in conjunction with a
pulsatile or an oscillatory component, it is referred to as either
the mean,12−14 average/time-average,8,15 or net/net-flow16−19
Reynolds number. The term net-flow Reynolds number is chosen
in this article due to its wider usage in literature.
The Reynolds number has also been applied to the pulsatile-
flow component in multiple pulsatile and oscillatory systems.
It is referred to as the pulsating Reynolds number20 or the
oscillating/oscillatory-flow Reynolds number, Reo.11,16,21−23
The term oscillatory-flow Reynolds number is chosen for its
current wide usage in the literature; however, both consist of
the same definition shown in eq 2. The velocity term is adapted
from the net-flow Reynolds number. The net velocity, un, is
replaced by the oscillatory velocity, uo. The oscillation ampli-
tude (center-to-peak), xo, which represents the length in the
coil that a fluid packet travels during the oscillation, is trans-
ferred into a velocity using the angular frequency, ω. To obtain
a single value from the constantly varying oscillating velocity,
the oscillatory maximum velocity, uo|max, is used. The maximum
velocity is attained at every cycle when the sinusoidal compo-
nent is equal to 1. Equations 4, 5, and 6 show the calculations
Figure 1. Types of pulsatile-flow.9,10 describing the oscillatory velocity.
uo = xoω sin(ωt ) (4)
pulsation amplitude. The name “start and stop” describes the
fluid coming to a stop at each cycle before starting its forward uo|max = xoω (5)
movement. The asymmetrical oscillatory flow has a net flow
between zero and the oscillation amplitude; the fluid velocity is ω = 2πf (6)
negative for a certain period of the oscillation, hence the reverse The third parameter to describe an oscillatory-flow is the
pulsatile-flow name. The symmetrical oscillatory-flow has no Womersley number, Wo, shown in eq 3. Introduced in 1955, it
net-flow; the only flow component present is the oscillation. evaluates the ratio of the transient inertial force of the oscilla-
Symmetrical oscillatory-flow is the only type not considered to tion to the viscous forces.24 When the Womersley number is
be in the continuous-flow category, it is an oscillatory batch larger than 10, the dominating forces are the oscillatory inertial
process. forces.25 Lower values indicate that the viscous forces are
From dimensional analysis, an oscillatory-flow can be fully dominant and that the pressure and flow are synchronized.24,26
described with the three following dimensionless parameters:11 In all of the three previously described dimensionless param-
net-flow Reynolds number: eters, the properties of the continuous phase, i.e., water, in
which generated eddies can breakup the dispersed phase, are
ρund tube
Ren = used to represent the liquid−liquid system. For the sake of
μ (1) conciseness and ease of reading, the mention of “continuous
phase” with each term and in subscript will be omitted in the
oscillatory-flow Reynolds number:
text.
ρuod tube
Reo = 3. EXPERIMENTAL TECHNIQUES
μ (2)
The selected reaction to perform the mass transfer tests is the
Womersley number: alkaline hydrolysis of 4-nitrophenyl acetate (4-NPA), shown
d tube ωρ in Figure 2. It is currently used as a chemical method for
Wo = interphase mass transfer rate evaluation.4,27,28 The reported
2 μ (3)
apparent pseudo first-order intrinsic rate constant for a
Using the dimensionless numbers Ren, Reo, and Wo, instead concentration of 0.02 mol/L NaOH is 0.280 s−1,29 which is
of standard units numbers un, uo, and f allows comparison of the already orders of magnitude greater than the expected overall
results between systems of different dimensions. The net-flow volumetric interphase mass transfer coefficient and thus
Reynolds number, Ren shown in eq 1, enables comparison of accomplishes the requirement that reaction kinetics are not

Figure 2. Alkaline hydrolysis of 4-nitrophenyl acetate.4

734 DOI: 10.1021/acs.oprd.5b00356

Org. Process Res. Dev. 2016, 20, 733−741
Organic Process Research & Development Article

rate-limiting for conversion. The stoichiometric ratio of

the reaction is 2:1 ratio; however, to ensure a pseudo-first
order reaction, a 9:1 ratio with the 4-NPA (0.05 mol/L) is
chosen.
Toluene is used as the organic phase in the system for its
relatively high interfacial tension with water, creating a very
low-miscible biphasic system. Earlier studies have shown that
this system can require residence/reaction times of several
minutes making it a suitable reaction for reactors of larger
volume such as a coil.4 The properties of the toluene and 0.5 M
of sodium hydroxide in water at 25 °C are shown in Table 1.

Table 1. Properties of Toluene and Water Used in the

Biphasic Liquid−Liquid System Figure 4. Cross-section schematic of the coil reactor.
solvent ρ (kg/m3) μ (mPa·s) σ with water (mN/m)
toluene30 862 0.552 35.4 Table 2. Dimensions of the Oscillatory-Flow Coil Reactor
0.5 M NaOH in H2O31 1020 1.124 - and Pulsator
details units values
The overview of the experimental setup is presented in dtube mm 4.57
Figure 3. The equi-massic feed flow is monitored by two Vcoil mL 800
Endress Hauser Promass 80A coriolis mass-flow meters. These dcoil mm 295
are relayed in a feedback control loop to two Fuji HYM-08 dpiston mm 14.0
metering pumps. The pulsatile-flow is generated by the LEWA f piston/max Hz 3.6
LDB1 metering pump. The pulsator has a head diameter of xpiston/max mm 7.5
14 mm with an independently adjustable length and frequency. Apiston/Atube 9.4
The maximum length and frequency of the piston are 15 mm xcoil/max mm 70.4
and 3.6 Hz, respectively. In order to translate the piston ampli- Zcoil m 48
tude to the pulsation length in the coil, the piston amplitude is
multiplied by 9.4 (the ratio of the piston area over the trans-
verse internal coil area). A blind flange is installed on the inlet Figure 4, and a summary of the key parameters of the pulsator
of the pump, while the outlet is connected with a T junction to and coil is shown in Table 2.
the reactant feed line going to the reactor. No average positive Following the coil reactor, the back pressure is regulated with
flow is thus created by the pulsator pump; the same fluid is a restriction device and nitrogen (N2) gas addition before the
pushed back and forth into the pulsator. depressurisation cyclone. The backpressure is applied in order
The 800 mL coil reactor is composed of a 48 m long tube to avoid cavitation with the oscillatory-flow. The fluid is then
with an inner diameter (ID) of 4.57 mm (1/4″ OD tube coil). depressurised to atmospheric pressure with the use of the
The tube is coiled with a 295 mm diameter (center-to-center of cyclone system. The built-in-house cyclone has a diameter
tube) and a spacing of 10 mm between each turn of the coil. of 0.15 m and height of 0.30 m. The reaction samples are
This spacing is present for an increased heat transfer capacity collected and allowed to settle for a short duration of 15 s for
(not evaluated in this article). A coil schematic is presented in phase separation. An equal volume of 0.5 mL of aqueous and

Figure 3. Experimental setup.

735 DOI: 10.1021/acs.oprd.5b00356

Org. Process Res. Dev. 2016, 20, 733−741
Organic Process Research & Development Article

organic phase is pipetted into a quench solution of 72.8 w/w% by taking the coil tube reactor volume divided by the net
acetonitrile, 26.5 w/w% water, and 0.7 w/w% acetic acid. The volumetric flow rate. The conversion result at each sampling
acetic acid reacts with the NaOH in the aqueous phase to halt time is normalized with (divided by) the average of the last
the possibility of the hydrolysis 4-nitrophenyl acetate reaction. three sample points of their respective test run.
The molar ratio of acetic acid to NaOH was set to 6:1. The The added oscillation created backmixing as shown by the
conversion was determined using an HPLC calibration curve earlier rise after ∼0.5 normalized residence time in Figure 6;
made using the response surface method. The HPLC system however, steady state is reached after ∼1.3 normalized residence
is composed of the HP Agilent 1100 series HPLC with a time in all of the cases. The samples during the following experi-
250 mm × 4.6 mm Agilent Zorbax SB-C8 column to separate ments were thus taken at time after 1.5 normalized residence
the compounds. time in order to minimize reactant use and ensure a reliable
Time to reach steady state is evaluated for the coil to deter- conversion result. At each operating condition, three distinct
mine the appropriate sampling time after a step change in samples are collected at steady state with a minimum interval of
operating set point. The first tests shown in Figure 5 are 1 min between each one.
The chosen operating conditions evaluated are presented in
Table 3 with the upper limit of the oscillatory-flow portion. The

Table 3. Selected Operating Flow Conditions

operating oscillatory-flow (maximum
parameter units net-flow tested)
Q mL/min 63.3; 84.8; 105.5 1180
u m/s 0.065; 0.086; 1.20
0.109
Re 270; 357; 450 4960
Wo 9.9
f Hz 3.3
xpiston mm 7.0
xcoil mm 65.7

maximum oscillatory-flow Reynolds number is chosen to be an

Figure 5. Time to steady state for the hydrolysis reaction without
order of magnitude greater than that of the net-flow Reynolds
oscillation for multiple net-flow rates. number creating the desired oscillatory dominated reverse
pulsatile flow system. The Womersley number of the con-
tinuous phase is set to reach as close to 10 in order to have the
performed using multiple flow rates without oscillation. The
oscillatory inertial forces dominate the viscous forces.25 The
second test shown in Figure 6 is performed using a piston
frequency of the piston is operated at a maximum of 90%
capacity (3.3 Hz) and the amplitude at 93% capacity (7.0 mm).
In order to avoid cavitation during pressure oscillations caused
by the oscillatory-flow, the system is kept under backpressure at
the outlet of the coil reactor using N2 gas and a line restriction
device.

4. RESULTS AND DSCUSSION

4.1. Conversion of the 4-Nitrophenyl Acetate Hydrol-
ysis. The conversion for the alkaline hydrolysis of 4-NPA, η, is
obtained at a net-flow rate of 80 g/min and a fixed frequency of
2.9 Hz. As shown in Figure 7, a 60 percentage point increase in
conversion is obtained through a piston amplitude increase
from 0.0 mm to 7.0 mm. The equivalent amplitude (center-to-
peak) in the coil is 0.0 mm to 65.7 mm.
Typically, an increase of 8.5 percentage point in conversion is
observed with the addition of the depressurisation cyclone sys-
Figure 6. Time to steady state for the hydrolysis reaction with tem. The gain in conversion will affect positively the reported
oscillation (net-flow rate of 100 g/min, piston amplitude of 2.8 mm, interphase mass transfer rate for the coil; however, the depres-
and frequency of 3.3 Hz).
surization component is an integral part of the system as the
coil needs to operate under backpressure to avoid cavitation
frequency of 3.3 Hz and amplitude of 2.8 mm at a net-flow rate during pulsation. As such, the reported values of interphase
of 100 g/min. The solvents without reactants are filled into the mass transfer are of the overall reactor system rather than
coil in order to prime the system. The feeds are then changed strictly the coil component.
to include the reactants while the chosen operating parameters 4.2. Interphase Mass Transfer Rate in a Toluene−
are set. This is considered to be the starting time, i.e., t = 0. The Water System. The overall volumetric mass transfer coeffi-
normalized residence times are calculated using the sample cient of the organic phase, Korga, shown in eq 7 is calculated
collection time at the outlet of the system divided by the system using a reactive extraction mass balance.32 It is used to com-
residence time at the operating conditions, which is estimated pare various reactor systems due to its normalized volume.4
736 DOI: 10.1021/acs.oprd.5b00356
Org. Process Res. Dev. 2016, 20, 733−741
Organic Process Research & Development Article

Figure 7. Hydrolysis conversion with varying piston amplitude (net-flow rate of 80 g/min and frequency of 2.9 Hz).

Figure 8. Overall volumetric mass transfer coefficient as a function of the oscillatory-flow Reynolds number (net-flow rate of 80 g/min and frequency
of 2.9 Hz).

The assumption of the plug flow model is made. The equi- slope increase indicates a change in flow regime. We propose
massic feed in the experiments gives a constant volumetric that the flow is transitioning to a drop flow regime.
organic phase fraction, φorg, of 0.537, with the assumption that In the drop flow regime, the Sauter mean diameter can be
there is no slip velocity between the phases. The oscillation estimated using eq 8, described and used with the same
amplitude and frequency are put into dimensionless form using biphasic system and reaction by Plouffe et al.34 The equation is
the oscillatory-flow Reynolds number. derived from a stagnant spheres model with a fully developed
concentration profile model and can be used since the reaction
φorg can be modeled as a fast pseudo-first order reactive extraction.32
Korga = − ln(1 − η) The estimation is used to validate that a droplet diameter
τ (7)
smaller than that of the coil is obtained after transitioning to
The overall volumetric mass transfer coefficient as a function drop flow. The values used for the parameters are given in
of Reynolds number is shown in Figure 8. The frequency is Table 4.
kept constant at 2.9 Hz, while the pulsator amplitude is varied
from 0.0 mm to 7.0. A 7-fold (6.7 times) increase in mass HADAd ⎛ 12φd DAck rxnC Bc ⎞
d32 = ⎜⎜ 1 + − 1⎟⎟
transfer rate is obtained from the no oscillation case (2.30 × DAck rxnC Bc ⎝ Kca HADAd ⎠ (8)
10−4 s−1) to the oscillatory-flow Reynolds number of 4976 at a
piston displacement of 7.0 mm (1.53 × 10−3 s−1). At the transition to drop flow, a droplet size diameter of
The overall volumetric mass transfer coefficient stays 1.7 mm is estimated. This diameter represents roughly a third
unchanged from an oscillatory-flow Reynolds number of 0 to of the 4.57 mm coil diameter. At the maximum oscillatory-
355. In this region, the flow type is a pulsing unidirectional or flow Reynolds number, a droplet size diameter of 0.6 mm is
very close to a start and stop-flow. When flow reversal occurs at obtained. From experience in microreactor plates, these values
an oscillatory-flow Reynolds number of 355, the interphase are far enough from the coil diameter to avoid interaction with
mass transfer rate begins to increase. This is in agreement with the wall to obtain for example slug flow. The relatively small
the mass transfer rate being dominated by the net-flow until droplet diameter corroborates the flow transitioning to a drop
flow reversal occurs.33 It is also noticed that a distinct increase flow regime.
in the slope of the mass transfer coefficient occurs at a Reynolds 4.3. Amplitude versus Frequency. In order to determine
number of ∼2670. From our experience and literature,4 this if the amplitude and frequency term are equal in mixing
737 DOI: 10.1021/acs.oprd.5b00356
Org. Process Res. Dev. 2016, 20, 733−741
Organic Process Research & Development Article

Table 4. Parameter Values for the Sauter Mean Diameter the oscillatory-flow coil with the addition of baffle type orifice
Calculation inserts throughout the length of the tube. Steady research has
been completed on the OBR since the end of the 1980s20,22
parameter description value
and meso-OBR have also been studied since 2003.6 However,
DAc molecular diffusion coefficient of 4-NPA 6.32 × 10−10 m2 s−1 reported experiments in meso-OBR have used low continuous
in the continuous phase
CBc concentration of NaOH in the 0.5 mol L−1
phase oscillatory-flow Reynolds number of 36−316 in liquid−
continuous phase liquid systems of vegetable oil and methanol in the trans-
H organic/aqueous concentration 229.81 esterification to produce biodiesel,37−39 and no mass transfer
distribution coefficient rates are reported. As such, no clear comparison can be made
krxn second order reaction rate constant 14 m3 mol−1 s−1 with an OBR of similar dimensions for an immiscible liquid−
liquid system. Reported single phase liquid and biphasic gas−
liquid are presented below, followed by toluene−water mass
contribution, the evaluation of the mass transfer rate at a
constant product of amplitude and frequency is completed, i.e., transfer rates in microreactor plates.
a constant oscillatory-flow Reynolds number. Three frequencies Mackley and Ni reported a 10-fold increase in dispersion
are arbitrarily chosen to span the operating range of the pulsator; number for a water system in a 25 mm tube in the range of
1.5, 2.2, and 3.3 Hz. Three constant products of pulsator 100 to 1800 oscillatory-flow Reynolds number.40 Hewgill et al.
amplitude and frequency are chosen; 5.4, 9.1, and 10.9 mm·s−1 reported a 6-fold increase in mass transfer rates for an air−
which have an oscillatory-flow Reynolds number equivalent to water system using an oscillatory-flow Reynolds number of 0 to
1330, 2240, and 2670, respectively. The plunger amplitude is 7840 in a 26 mm tube diameter. No increase was noticed using
then calculated to obtain the desired velocity product at each an oscillation in a tube without baffles.41 Ni and Gao reported a
frequency. 5−6-fold increase in mass transfer rates for an air−water system
As shown in Figure 9, the overall mass transfer coefficient in a 50 mm and 100 mm diameter OBR. In the 50 mm
increases with higher frequency (Wo number) for the three diameter OBR, the minimum oscillation was 3 Hz and 4 mm,
constant velocity (Reo) curves. Thus, the frequency component and the maximum oscillation was 8 Hz at 12 mm. In the
shows a greater influence on the mass transfer rate than the 100 mm diameter OBR, the minimum oscillation was 2 Hz and
amplitude. Similar results for an air−water system in an OBR 5 mm, and the maximum oscillation was 6 Hz at 12 mm.35
are also reported in literature.35 To compare the obtained interphase mass transfer rate of the
It is noted that a higher pressure oscillation is experienced oscillatory-flow coil on an absolute value basis, reported results
when using a higher frequency vs larger amplitude at equal of a toluene−water system in two microreactor plates, the
oscillatory-flow Reynolds number. The transient inertial forces Sickle and the SZ FlowPlate by Ehrfeld Mikrotechnik BTS, are
over the viscous forces are increasing with the increment in used.4 The sickle is an obstacle-based micromixer, while the SZ
frequency and the higher pressure oscillation can be attributed is a curvature-based micromixer; these are shown in Figure 10
to inertia and relaminarisation effects.36 An increase in fre- and Figure 11. The dimensions at the contraction in width by
quency at a constant amplitude results in a greater mass transfer depth of the Sickle and SZ are, respectively, 0.20 × 0.50 mm
rate compared to a proportional increase in amplitude at a and 0.50 × 1.25 mm. The contraction is indicated by the arrow
constant frequency. It is also noticed at the higher Womersley in the figures. The reported results are shown alongside the
number of 10 (frequency of 3.3 Hz) that the increase in overall obtained oscillatory-flow coil results in Figure 12. The
mass transfer coefficient at a constant frequency is greater for secondary horizontal axis of the flow rate for the microreactor
an equivalent increase in oscillatory-flow Reynolds number than plates does not have a direct relationship with the primary
at the lower frequencies. This frequency is at the lower limit of horizontal axis of the Reynolds number used for the coil. Both
the oscillatory inertia dominant regime at the Womersley axes are chosen to represent the operating range of each sys-
number of 10, previously mentioned in the flow character- tem. The comparison of the reactor systems is also completed
ization section.25 using the average energy dissipation rate.
4.4. Reactor Performance Comparison. A clear choice The average energy dissipation rate for the system, shown in
for comparison is the OBR. It has all of the flow components of eqs 9 and 10 and graphically in Figure 13, is estimated using the

Figure 9. Overall volumetric mass transfer coefficient as a function of the Womersley number for three iso-oscillatory-flow Reynolds number curves.

738 DOI: 10.1021/acs.oprd.5b00356

Org. Process Res. Dev. 2016, 20, 733−741
Organic Process Research & Development Article

The mass transfer performance of a plate microreactor and

an oscillatory-flow coil reactor become similar when the flow
enters the drop flow regime. Moreover, for the oscillatory-flow
reactor, this is likely the transition between laminar and
turbulent flow as depicted in Figure 8. Prior to drop flow
regime, the oscillatory-flow coil reactor underperforms relative
to the plate microreactor. This is most likely due to its intrinsic
milli-structure where the beneficial effect from the micro-
structure (large phase interfacial area due to geometry) is
reduced. The mass transfer performance of a plate microreactor
and an oscillatory-flow coil reactor becomes similar once the
flow regime enters drop flow. Within its typical operating range
(see Figure 12), a microreactor will provide greater Korga at
Figure 10. SZ micromixer structure. corresponding energy dissipation rates making it a more
suitable reactor for very fast reactions (Type A, see section 4.5
for reaction types). This is also applicable to heat transfer as
microreactors have a greater intrinsic area (m2 m−3). On the
other hand, the oscillatory-flow reactor provides significantly
greater volume at a reduced energy dissipation rate and Korga
but can still operate in the same flow regime (drop flow) as the
plate microreactor. In such case, the oscillatory-flow coil reactor
becomes a clear complement to a plate microreactor for gaining
volume, which naturally leads to an association with slower
reactions. It can thus be used jointly (often Type B reactions)
or independently (often Type C reactions) to the plate
microreactor (Table 5).
4.5. Integration to the Microreactor Toolbox. The
microreactor toolbox1 previously proposed by this group and
Figure 11. Sickle micromixer structure. shown in Table 5 uses reaction rates and phase types to deter-
mine the appropriate reactor to select. The three reactions
rates, presented as type A, B, and C, are ordered respectively
sum of the net-flow4 and the oscillatory energy dissipation rate from the very fast type A (seconds) to the slow type C (several
component.42 The referenced oscillatory energy dissipation rate minutes).43−45
is divided by the density due to the original units being The previously proposed usage of the pulsated/oscillatory-
reported in W m−3 instead of the W kg−1, and the phase angle flow coil for type B and C in liquid−liquid system holds true;
(δ) is considered to be zero (we assumed pressure and velocity
however, as previously demonstrated through mass transfer rate
to be in-phase). The average energy dissipation rate for the
performance, there is a clear overlap between the performance
referenced microreactor plates is calculated using the net-flow
average energy dissipation rate shown in the first term of eq 10. of a plate and coil that enables a smooth transition between
both technologies or even a hybrid use (first a plate for when
ε ̅ = εn̅ + εo̅ (9) the reaction requires elevated transport rates, then a coil for the
volume and residence time gain). Not shown in the toolbox is
ΔPn 1 ΔPo cos(δ)xoω
ε̅ = + the ease of adaptability in volumes and relative low cost of
ρτ 2 Zρ (10) production for the coil compared to microreactor plates.

Figure 12. Oscillatory-flow coil overall mass transfer coefficient comparison with other microreactors (microreactor plate results from Plouffe et al.
2015).4

739 DOI: 10.1021/acs.oprd.5b00356

Org. Process Res. Dev. 2016, 20, 733−741
Organic Process Research & Development Article

Figure 13. Oscillatory-flow coil overall mass transfer coefficient microreactor comparison as a function of the average energy dissipation rate
(microreactor plate results from Plouffe et al. 2015).4

Table 5. Microreactor Toolbox: Flow Modules for Reaction Notes

Types1 The authors declare no competing financial interest.
rates/
phases
Type A
homogeneous liquid−liquid
plate SZ plate LL
gas−liquid
plate Venturi with TG
solid−
liquid
CSTR
■ ACKNOWLEDGMENTS
A special mention is made to Lonza AG for its financial
plate HEART mixing units packed bed contribution and laboratory use and the laboratory technician
Type B plate SZ plate LL plate Venturi with TG CSTR Anton Pfammatter, the CREATE program in Continuous Flow
coil plate HEART mixing units packed bed Science (NSERC), for the financial support to promote
continuous-flow research

■
static mixer
pulsated coil coil pressure
Type C static mixer static mixer coil pressure pulsated
NOMENCLATURE
coil
General Symbols
coil pulsated coil
a Specific interfacial area [m−1]
A Cross-sectional area [m2]
There are still research areas in need of further investigation
Cij Concentration of molecule ‘‘i” in phase ‘‘j” [mol m−3]
such as
d Diameter [m]
• the impact of net-flow rate on overall volumetric mass d32 Sauter mean diameter [m]
transfer coefficient; Dij Molecular diffusion coefficient of molecule ‘‘i” in the
• the effect of scale on mass transfer rate; the influence of phase ‘‘j” [m2 s−1]
higher oscillatory-flow Reynolds number using smaller f Frequency of oscillation [Hz]
diameter tubes; H Organic/aqueous concentration distribution coefficient
• the system operability with solid presence and/or fixed [-]
bed catalyst incorporation; ID Inner diameter [m]
• the overall volumetric mass transfer coefficient for krxn second order reaction rate constant [m3 mol−1 s−1]
different organic phase; K Overall mass transfer coefficient [m s−1]
• the heat transfer capacity. Kca Overall volumetric mass transfer coefficient of the
continuous phase [s−1]
5. CONCLUSIONS Korga Overall volumetric mass transfer coefficient of the
The use of a baffle-less oscillatory-flow mesoreactor proved to organic phase [s−1]
be successful in increasing the mass transfer rate of a low OD Outer diameter [m]
miscibility biphasic liquid−liquid system of toluene and water. P Pressure [Pa]
A 7-fold increase in interphase mass transfer rate is obtained Q Volumetric flow rate [m3 s−1]
compared to the unidirectional-flow base case, i.e., without Re Reynolds number [−]
oscillation. A proportional increase in oscillation frequency t Time [s]
compared to amplitude results in a higher mass transfer rate. u Superficial velocity [m s−1]
Mass transfer rates reach the lower end of the plate microreactor V Volume [m3]
performance. Thus, a baffle-less oscillatory-flow mesoreactor can Wo Womersley number [−]
be used in low miscibility liquid−liquid systems as a complement Z Coil length [m]
to the plate type microreactor and typically to gain volume for Greek Symbols
longer reaction times. δ Phase angle between pressure and velocity (time between

■ AUTHOR INFORMATION
Corresponding Authors
peaks multiplied by the angular frequency) [rad]
ε̅ Average energy dissipation rate [W kg−1]
η Conversion of 4-NPA [%]
*E-mail: [email protected]. μ Dynamic viscosity [Pa s]
*E-mail: [email protected]. ρ Density [kg m−3]
740 DOI: 10.1021/acs.oprd.5b00356
Org. Process Res. Dev. 2016, 20, 733−741
Organic Process Research & Development Article

σ Interfacial tension [N m−1] (27) Ahmed, B.; Barrow, D.; Wirth, T. Adv. Synth. Catal. 2006, 348,
τ Residence time [s] 1043−1048.
φ Volumetric phase fraction [-] (28) Ahmed-Omer, B.; Barrow, D.; Wirth, T. Chem. Eng. J. 2008,
x Centre-to-peak amplitude [m] 135, 280−283.
(29) Parker, V. D. J. Phys. Org. Chem. 2006, 19, 714−724.
ω Angular frequency [rad s−1] (30) Backes, H. M.; Jing Jun, M.; Bender, E.; Maurer, G. Chem. Eng.
Indices Sci. 1990, 45, 275−286.
A 4-nitrophenyl acetate (31) Alexandrov, A. A. In 14th International Conference on the
aq Aqueous phase Properties of Water and Steam; Kyoto, 2005; pp 86−90.
(32) Levenspiel, O. Chemical Reaction Engineering, 3rd ed.; Wiley:
B Sodium hydroxide base New York, 1999.
c Continuous phase (33) Howell, J. A.; Finnigan, S. M. In Effective Industrial Membrane
coil Coil properties of the oscillatory-flow coil reactor Processes: Benefits and Opportunities; Turner, E., Ed.; Springer:
d Dispersed phase Netherlands, 1991; pp 49−60.
max Maximum (34) Plouffe, P.; Roberge, D. M.; Sieber, J.; Bittel, M.; Macchi, A.
n Net-flow component Chem. Eng. J. 2016, 285, 605−615.
o Oscillatory-flow component (35) Ni, X.; Gao, S. Chem. Eng. J. Biochem. Eng. J. 1996, 63, 157−166.
org Organic phase (36) Ribas, F. A.; Deschamps, C. J. In International Compressor
piston Piston of the pulsator Engineering Conference; Purdue University: West Lafayette, 2004; pp
tot Total of continuous and dispersed phase 1−8.
(37) Phan, A. N.; Harvey, A. P.; Eze, V. Chem. Eng. Technol. 2012, 35,
tube Tube properties of the oscillatory-flow coil reactor 1214−1220.

■ REFERENCES
(1) Plouffe, P.; Macchi, A.; Roberge, D. M. Org. Process Res. Dev.
(38) Zheng, M.; Skelton, R. L.; Mackley, M. R. Process Saf. Environ.
Prot. 2007, 85, 365−371.
(39) Phan, A. N.; Harvey, A. P.; Rawcliffe, M. Fuel Process. Technol.
2011, 92, 1560−1567.
2014, 18, 1286−1294. (40) Mackley, M. R.; Ni, X. Chem. Eng. Sci. 1991, 46, 3139−3151.
(2) Nguyen, N.-T.; Wu, Z. J. Micromech. Microeng. 2005, 15, R1− (41) Hewgill, M. R.; Mackley, M. R.; Pandit, A. B.; Pannu, S. S. Chem.
R16. Eng. Sci. 1993, 48, 799−809.
(3) Plouffe, P.; Macchi, A.; Donaldson, A. A. Heat Transfer Eng. 2013, (42) Mackley, M. R.; Stonestreet, P. Chem. Eng. Sci. 1995, 50, 2211−
34, 159−168. 2224.
(4) Plouffe, P. Micro-Reactor Design for Fast Liquid- Liquid Reactions; (43) Roberge, D. M.; Gottsponer, M.; Eyholzer, M.; Kockmann, N.
University of Ottawa, 2015. Chem. Today 2009, 27, 8−11.
(5) Kashid, M. N.; Renken, A.; Kiwi-Minsker, L. Chem. Eng. Sci. (44) Roberge, D. M.; Ducry, L.; Bieler, N.; Cretton, P.;
2011, 66, 3876−3897. Zimmermann, B. Chem. Eng. Technol. 2005, 28, 318−323.
(6) McDonough, J. R.; Phan, A. N.; Harvey, A. P. Chem. Eng. J. 2015, (45) Kockmann, N.; Gottsponer, M.; Roberge, D. M. Chem. Eng. J.
265, 110−121. 2011, 167, 718−726.
(7) Reis, N.; Harvey, A. P. Trans IChemE 2005, 357−371.
(8) Ö zdinç Ç arpinlioǧlu, M.; Yaşar Gündoǧdu, M. Flow Meas.
Instrum. 2001, 12, 163−174.
(9) Nabavi, M.; Siddiqui, K. Meas. Sci. Technol. 2010, 21, 042002.
(10) Southard, J. An Introduction to Fluid Motions, Sediment Transport,
and Current-Generated Sedimentary Structures; Boston 2006, 184−200.
(11) Peacock, J.; Jones, T.; Tock, C.; Lutz, R. Exp. Fluids 1998, 24,
1−9.
(12) Trip, R.; Kuik, D. J.; Westerweel, J.; Poelma, C. Phys. Fluids
2012, 24.
(13) Gebreegziabher, T.; Sparrow, E. M.; Abraham, J. P.; Ayorinde,
E.; Singh, T. Numer. Heat Transfer, Part A 2011, 60, 811−826.
(14) Einav, S.; Sokolov, M. J. Biomech. Eng. 1993, 115, 404−411.
(15) Trabelsi, F. Pulsatile Flow in a Conical Tube; University of
Ottawa, 1993.
(16) Ni, X. J. Chem. Technol. Biotechnol. 1995, 64, 165−174.
(17) Smith, K. B.; Mackley, M. R. Chem. Eng. Res. Des. 2006, 84,
1001−1011.
(18) Palma, M.; Giudici, R. Chem. Eng. J. 2003, 94, 189−198.
(19) Lobry, E.; Lasuye, T.; Gourdon, C.; Xuereb, C. Chem. Eng. J.
2015, 259, 505−518.
(20) Ni, X.; Gough, P. Chem. Eng. Sci. 1997, 52, 3209−3212.
(21) Ni, X.; Mackley, M. R. Chem. Eng. J. Biochem. Eng. J. 1993, 52,
107−114.
(22) Ni, X.; Mackley, M. R.; Harvey, A. P.; Stonestreet, P.; Baird, M.
H. I.; Rama Rao, N. V. Trans IChemE 2003, 81, 373−383.
(23) Xia, Q. Enhancement of Liquids Mixing Using Active Pulsation in
the Laminar Flow Regime; University of Manchester, 2012.
(24) Womersley, J. R. J. Physiol. 1955, 127, 553−563.
(25) Hussein, A. K. M. F.; Settler, J. C. J. Fluid Mech. 1986, 170,
169−197.
(26) McDonald, D. A. Blood flow in arteries, 2nd ed.; Arnold, E., Ed.;
London, 1974.

741 DOI: 10.1021/acs.oprd.5b00356

Org. Process Res. Dev. 2016, 20, 733−741