You are on page 1of 29

13

Hydrodynamic Techniques for


Investigating Reaction Kinetics at
Liquid^Liquid Interfaces: Historical
Overview and Recent Developments
CHRISTOPHER J. SLEVIN, PATRICK R. UNWIN, and
JIE ZHANG Departme nt of Chemistry, University of Warwick, Coventry,
England

I. INTRODUCTION

As discussed throughout this book, reactions that occur at the interface between two
immiscible liquids are common in a wide range of areas. For example, industrially impor-
tant processes, such as solvent extraction [1,2] and phase transfer catalysis [3], rely on
optimizing reactions at liquid–liquid interfaces. In addition, immiscible liquid–liquid inter-
faces can be considered as useful analogs of biomembranes [4–6] for the experimental
investigation of cell membrane transfer processes, including drug delivery systems [7].
On a more general level, interest in the fundamental aspects of charge transfer processes
at the interface between two immiscible electrolyte solutions (ITIES) has been increasing
recently [8,9]. This research has led to developments in areas such as analysis with ampero-
metric ion-selective electrodes [10], as well as in understanding the fundamentals of charge
transfer mechanisms at liquid–liquid interfaces [8,11].
Many of the areas highlighted require an understanding of the kinetics and mechan-
isms of interfacial processes. An important aspect of progress in this field has been the
introduction of new techniques which are able to provide increasingly improved kinetic
and mechanistic insight. The investigation of the kinetics of reactions that occur at liquid–
liquid interfaces requires a number of factors to be considered. As depicted in the sche-
matic in Fig. 1, the overall rate may be limited by the transport of reactants, intermediates,
or products to and from the interface, interfacial processes, reactions in solution (not
shown), or by a combination of factors. In order to accurately measure interfacial reaction
rates, it is necessary to either operate under conditions where the overall rate is limited
only by the interfacial (or solution) chemical reaction, i.e., ensure that the transport step is
relatively fast and nonlimiting, or to make the transport step well-defined and calculable,
so that its effect can be accounted for when interpreting rate data [3,12,13].
For the study of reactions at immiscible liquid–liquid interfaces, achieving well-
defined contact between the two phases, with known interfacial area, under conditions
where variable and high rates of mass transport are attainable, is nontrivial.

325

Copyright © 2001 Marcel Dekker, Inc.


326 Slevin et al.

FIG. 1 Simplified schematic illustrating the types of steps involved in an interfacial reaction at a
liquid–liquid interface.

Complications introduced when a liquid–liquid interface, rather than a solid–liquid inter-


face, is studied include the possibility that transport on both sides of the interface, in each
of the phases, may need to be considered, and achieving a stable interface of known area is
much more difficult.
The general criteria for an experimental investigation of the kinetics of reactions at
liquid–liquid interfaces may be summarized as follows: known interfacial area and well-
defined interfacial contact are essential; controlled, variable, and calculable mass transport
rates are required to allow the transport and interfacial kinetic contributions to the overall
rate to be quantified; direct interfacial contact is preferred, since the use of a membrane to
support the interface adds further resistances to the overall rate of the reaction [14,15]; a
renewable interface is useful, as the accumulation of products at the interface is possible.
Finally, direct measurements of reactive fluxes at the interface of interest are desirable.
Many of the electrochemical techniques described in this book fulfill all of these
criteria. By using an external potential to drive a charge transfer process (electron or ion
transfer), mass transport (typically by diffusion) is well-defined and calculable, and the
current provides a direct measurement of the interfacial reaction rate [8]. However, there is
a whole class of spontaneous reactions, which do not involve net interfacial charge trans-
fer, where these criteria are more difficult to implement. For this type of process, hydro-
dynamic techniques become important, where mass transport is controlled by convection
as well as diffusion.
In this chapter, we describe some of the more widely used and successful kinetic
techniques involving controlled hydrodynamics. We briefly discuss the nature of mass
transport associated with each method, and assess the attributes and drawbacks. While
the application of hydrodynamic methods to liquid–liquid interfaces has largely involved
the study of spontaneous processes, several of these methods can be used to investigate
electrochemical processes at polarized ITIES; we consider these applications when appro-
priate. We aim to provide an historical overview of the field, but since some of the older
techniques have been reviewed extensively [2,3,13], we emphasize the most recent devel-
opments and applications.

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 327

II. THE LEWIS CELL


A. Methodology
Some of the earliest attempts to address the difficulties associated with making kinetic
measurements at immiscible liquid–liquid interfaces were made by Lewis [16,17] using the
stirred cell design illustrated in Fig. 2. The Lewis cell employs direct contact between the
two immiscible liquids, and reaction rates are evaluated by measuring concentration
changes in the bulk of one of the two phases, usually by a batch extraction technique.
The rate of change of concentration, dc=dt, is related to the interfacial reaction flux, j, by
dc jAint
¼ ð1Þ
dt V
where Aint is the interfacial area and V is the volume of the phase in which the concentra-
tion change is monitored. Stirrers are employed in each phase to generate convective
transport on both sides of the interface. This reduces the contribution of diffusion to
the overall kinetics, and it is generally assumed that the reaction rate can be evaluated
by neglecting diffusion. This, of course, places limitations on the range of kinetics that can
be investigated, as discussed shortly.
The basic cell employs stirrers rotating at the same speed to minimize surface break-
up, resulting in a relatively flat interface of determinable area. Although the rate of mass
transport to the interface can be enhanced by increasing the rate of stirring, the hydro-
dynamics are relatively ill-defined and transport rates cannot be calculated absolutely.
This technique therefore relies on the determination of interfacial reaction kinetics from
the plateau region of a plot of reaction rate vs. stirring rate, with the assumption that the
interfacial kinetics can be outrun by generating sufficiently high mass transport rates. At
the maximum attainable stirring rate, the size of the concentration boundary layer (or
diffusion layer) is in fact still relatively large, since high stirring rates cannot be used, as
convective effects close to the interface would cause excessive break-up of the interface. In
practice, the minimum diffusion layer thickness, D , is ca. 100 m, introducing a mass
transport resistance, kT ðcm s1 ), which may be evaluated using
D
kT ¼ ð2Þ
D

FIG. 2 Schematic of the basic Lewis cell design.

Copyright © 2001 Marcel Dekker, Inc.


328 Slevin et al.

where D is the diffusion coefficient of the reactant of interest. For typical values of D ¼
105 cm2 s1 and D ¼ 100 m, the mass transport resistance is of the order of 103 cm s1 .
This limits the technique to rather slow kinetics, with first-order interfacial rate constants,
ki  kT .
Another complication, caused by stirring the two phases at the same speed, occurs
when the two solutions have different viscosities, which is common for immiscible liquids.
The key fluid flow parameter is the Reynolds number, Re, which is the ratio of inertial to
viscous forces in the solution, as indicated by
vl
Re ¼ ð3Þ

where v is the fluid velocity, l is a characteristic length, and  is the kinematic viscosity of
the fluid. When the two phases have different viscosities, the Reynolds numbers are
different for each side of the interface.
Developments in the Lewis cell were made to overcome some of the problems
associated with the initial design. With the introduction of screens close to the interface,
on either side [18], stirring could be carried out at different speeds in each phase, allowing
the same Reynolds number to be attained on both sides of the interface. This modification
to the cell also allowed more vigorous stirring and efficient mixing in the bulk region
without disturbances to the interface, thereby reducing the size of the stagnant diffusion
layer and increasing the kinetic range of the technique.
A further useful development in the design was the introduction of a flow loop [19],
eliminating the need for large volumes of solution, and enhancing the batch sampling
procedures. Additional advances included the use of a porous gauze [20] to support the
interface with stirring close to the gauze employed to enhance mass transport. However,
even with these improvements, the main problems associated with using this type of cell
remain, primarily that mass transport is poorly characterized and rather slow. The tech-
nique is essentially limited to the measurement of first-order interfacial rate constants, ki <
105 cm s1 [3].

B. Applications
Although the Lewis cell was introduced over 50 years ago, and has several drawbacks, it is
still used widely to study liquid–liquid interfacial kinetics, due to its simplicity and the
adaptable nature of the experimental setup. For example, it was used recently to study the
hydrolysis kinetics of n-butyl acetate in the presence of a phase transfer catalyst [21].
Modeling of the system involved solving mass balance equations for coupled mass transfer
and reactions for all of the species involved. Further recent applications of modified Lewis
cells have focused on stripping–extraction kinetics [22–24], uncatalyzed hydrolysis [25,26],
and partitioning kinetics [27].

III. CONSTANT INTERFACIAL CELL WITH LAMINAR FLOW

Controlled contact between two immiscible liquids has also been achieved by flowing one
liquid along a solid support submerged in the second phase [28,29]. Several different
arrangements have been used, although all are based on similar principles. For example,
a wetted wall column which offered liquid–liquid contact times of 0.5–10 s was used to
measure solute transfer rates [29].

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 329

A significant advance in this area was recently made by Li and coworkers [30,31],
who developed a laminar flow technique, that allowed the direct contact of two liquids
with better-defined mass transport compared to the Lewis cell. Laminar flow of the two
phases parallel to the interface was produced through the use of flow deflectors. By forcing
flow parallel to, rather than towards, the interface, it was proposed that the interface was
less likely to be disrupted. Reactions were followed by sampling changes in bulk solution
concentrations.
A schematic of the apparatus developed is shown in Fig. 3. Stirrers mix and push the
lighter and heavier phases in each compartment, with the maximum rotation speed gov-
erned by the need to maintain the interface steady. Flow deflectors ensure that the phases
are circulated in each chamber and that flow near the interface is laminar. The interfacial
plate (thickness 2 mm) is rectangular with a hole at its center. The distance from the
interface to the flow deflectors is less than 6 mm. The two phases are analyzed by with-
drawing small volumes via sampling holes.
The hydrodynamic and diffusion theory of this technique were obtained based on the
following assumptions: isothermal operation; constant densities; constant viscosities; con-
stant diffusivities; negligible diffusion along the interfacial plate direction; steady-state
operation; negligible flow in the y and z directions (see Fig. 4); and negligible stagnation
regions near the edges of the interfacial plates.
The thickness of the concentration boundary layer, D , was found to be governed by:
0 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1
1 96LDH 2 A
2D ¼ @9H 2  81H 4  ð4Þ
2 U

where H is the distance between the interface and flow deflectors, L is the length of
contacting interface in the flow direction, and U is the mean solution velocity.
The mass transfer equation applicable to the transport-limited extraction of a solute
from an aqueous solution to an organic phase (sink conditions), was derived:
  
V w cwb;0 3 Aint Dt
Tr ¼ 1  exp  ð5Þ
Aint ð1 þ Þ 2 V w D
where Tr is the mass transferred per unit area, is the partition coefficient of the solute
between the organic and aqueous phase, V w is the volume of aqueous phase, cw b;0 is the
initial concentration of the solute in the aqueous phase and t is time.
Equation (5) can be also written as
"  w #
1 cb 1 A k
f ðcb Þ ¼ ln
w
þ1 w  ¼  intw T ð þ 1Þt ð6Þ
cb;0 V

FIG. 3 Schematic of the constant interfacial cell with laminar flow.

Copyright © 2001 Marcel Dekker, Inc.


330 Slevin et al.

FIG. 4 The co-ordinate system for the constant interfacial cell.

where cwb denotes the bulk aqueous solute concentration at time, t. The mass transfer
coefficient is

3 D
kT ¼ ð7Þ
D þ D
2 w o

The superscripts o and w denote the organic and aqueous phases, respectively.
b Þ on t for a diffusion-controlled process.
Equation (6) suggests a linear dependence of f ðcw
This technique has been used to study the extraction kinetics of rare-earth elements
from an aqueous phase into heptane by 2-ethylhexyl phosphonic acid mono-2-ethylhexyl
ester (HEH/EHP) [31]. The linear dependence of f ðcw b Þ on t in Fig. 5 was considered to
indicate that the extraction of ErCl3 by HEH/EHP is a diffusion-controlled process.
The technique offers a known interfacial area under convective flow conditions that
are quite well-defined, with mass transport rates that are enhanced compared to the Lewis
cell and its analogs. However, in common with many other approaches, interfacial fluxes
must be determined indirectly from bulk solution measurements.

FIG. 5 Extraction of ErCl3 by HEH/EHP–heptane at pH 2.49 and 30 C. The experimental system


was characterized by the following parameters: L ¼ 6:0 cm, Aint ¼ 21:0 cm2 , H ¼ 0:6 cm,
u ¼ 3:4 cm s1 , V w ¼ 96 cm3 , and cw
b;0 ¼ 3  10
3
M. The organic phase initially contained 3:0
2
10 M HEH/EHP. (Reprinted from Ref. 31. Copyright 1998, Elsevier Science Ltd.)

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 331

IV. THE ROTATING DIFFUSION CELL


A. Principles
The rotating diffusion cell (RDC) [14] enables the study of liquid–liquid reaction kinetics
under conditions where interfacial hydrodynamics are well-defined and calculable. The
design of the RDC is based on the rotating disk electrode (RDE) [32], which has been
widely used for kinetic measurements at solid electrodes. The RDE consists of a disk-
shaped electrode sealed in an insulating cylinder rotated in solution, giving rise to well-
defined hydrodynamics, which have been calculated to give a complete description of the
flow field in the contacting solution [33]. The mass transfer coefficient for this device,
under transport-limited conditions, is given by [33]:

kT ¼ 1:554D2=3 1=6 W 1=2 ð8Þ

where W is the rotation frequency of the disk (Hz). The assumptions behind the derivation
of Eq. (8) are that flow is laminar and the Schmidt number, Sc ¼ =D, is in excess of 103 .
These assumptions generally hold for most practical conditions [34]. In practice, the rate
of mass transport to the disk can be controlled readily through the rotation speed. For
rotation speeds up to 100 Hz, which is experimentally accessible for a solid electrode, the
corresponding mass transport parameter, for typical D and  values of 1  105 cm2 s1
and 1  102 cm2 s1 , respectively, is variable up to 1:5 102 cm s1 . This corresponds
to a diffusion layer (concentration boundary layer) thickness of ca.7 m.
The RDC, depicted schematically in Fig. 6, operates with the interface supported at
a thin porous membrane between the inner and outer compartments of the cell. A number
of different arrangements are possible, with a common configuration involving aqueous
solutions in both the inner and outer compartments, with the organic phase impregnated
within the membrane. Abbreviated to w/o/w, this results in two aqueous–organic phase
interfaces. Alternatively, w/o/o, w/w/o, and o/o/o arrangements, etc., have also been used
[15]. The membrane is rotated in the solution and the hydrodynamic profiles of the
rotating disk are established on both sides of the membrane. Liquid–liquid contact is
achieved with a known area by treating the membrane with a solution which blocks the
pores, except in a small circular section in the center, which is untreated. Interfacial
reactions are usually studied by measuring changes in the bulk concentrations of reacting

FIG. 6 Schematic of the rotating diffusion cell. The reaction is usually followed by sampling the
bulk solution of the outer phase using a suitable analytical technique.

Copyright © 2001 Marcel Dekker, Inc.


332 Slevin et al.

species, as described above for the Lewis cell, although interfacial fluxes can be measured
directly by modifying the technique with the addition of a ring electrode, as described
later.
The kinetics of the reaction must be calculated from the measured flux, taking into
account transport to and from the disk-shaped contacting area and through the mem-
brane. The flux of a target species may be quantified according to the following equations;
here we consider a simple solute transfer reaction from an organic donor phase to an
aqueous receptor phase in the o/o/w configuration. Firstly, the flux across the diffusion
layer towards the membrane on the donor side is given by
j ¼ Do ðcob  coi Þ=oD ð9Þ
where Do is the diffusion coefficient of the solute in the oil phase, coi
and cob
are the
concentrations of the solute in the organic donor phase at the membrane surface, and
in the bulk of the organic solution, respectively. The diffusion layer thickness for the
hydrodynamic rotating disk arrangement, D , is given by
gD ¼ 0:643D1=3 1=6 W 1=2 ð10Þ
where g denotes phase o or w. This applies to both donor and receptor compartments. The
flux through the membrane is given by
j ¼ Do ðcom  cw
m Þ=lm ð11Þ
where is the membrane porosity, lm is the membrane thickness, and comand cw
mare the
concentrations of the solute within the membrane, specifically, where the membrane meets
the organic donor and water receptor, respectively. For this particular arrangement, coi
and com are equal, since there is no transfer resistance on the donor compartment side. The
interfacial transfer reaction is controlled by
j ¼ ðkow cw
m  kwo ci Þ
w
ð12Þ
where cw
iis the corresponding concentration at the other side of this interface. The rate
constants kwo and kow are for transfer across the organic–water interface from the water to
the organic and the organic to the water phases respectively. Finally, the diffusion away
from the interface into the receptor phase may be evaluated from
j ¼ Dw ðcw
i  cb Þ=D
w w
ð13Þ
where cw
bis the bulk concentration of the solute in the water receptor phase.
Given that Pow ¼ kwo =kow is the solute partition coefficient, then the expression for
the total flux for this particular case can be written as
cob  Pow cw P  o l 1
b
¼ ow D;w þ D þ m þ ð14Þ
j Dw Do Do kow
Combining Eq. (10) and (14) yields
!
cob  Pow cw 1=2 o
1=6
Pow 1=6
w l 1
b
¼ 0:643W þ þ m þ ð15Þ
j 2=3
Do Dw2=3 Do kow

Analysis of data obtained from the RDC usually involves plotting the reciprocal of
the measured flux, j 1 , against W 1=2 , and extrapolating to infinite rotation speed
(W 1=2 ! 0), such that the diffusion layer thicknesses tend to zero and diffusion in the
solution on either side of the membrane is discounted. The intercept on the j 1 axis then
gives a measure of the flux, governed by the interfacial kinetics and the transport resistance

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 333

[for example, the last two terms on the right-hand side of Eq. (15)]. The membrane transfer
resistance must be subtracted in order to obtain the true interfacial resistance.

B. Drawbacks and Improvements


In contrast to the RDE, the range of rotation speeds used in the RDC is rather limited.
The upper limit is around 6–8 Hz, while Eq. (8) breaks down below approximately 1–2 Hz,
where the hydrodynamic boundary layer,

H ¼ 1:4ð=WÞ1=2 ð16Þ
becomes comparable with the rotating disk radius [35]. Consequently, rate data are
obtained over a very narrow range of mass transport coefficients, the maximum kT avail-
able being ca. 4  103 cm s1 . Significant extrapolation of reciprocal flux data to W 1=2
! 0 has been suggested to be an unreliable way to determine interfacial kinetics [15].
A modification of the RDC design, based on the ring–disk arrangement of the RDE
[36], incorporated an arc electrode [37,38] deposited on the surface of the membrane
around the untreated area. This facilitated the electrochemical detection of species reacting
at the interface at short times following the reaction. This method was used to study the
solvent extraction of cupric ions, which were detected by reduction to copper metal at the
arc electrode. The resulting current flow was related to the interfacial flux at the mem-
brane.
There are several drawbacks to the RDC that need to be emphasized. First, the fact
that the interface must be supported adds a considerable resistance to the transport of
species, which is in addition to that from the concentration boundary layers on both sides
of the membrane. This limits the range of kinetics that can be studied. Second, in practical
applications, blocking of the membrane can be problematic for some reactions. Third,
measurements are generally made in the bulk of the solution and not at the interface
although, as mentioned above, for certain processes it is possible to measure fluxes via
a ring or an arc electrode.

C. Applications
The majority of RDC studies have concentrated on the measurement of solute transfer
resistances, in particular, focusing on their relevance as model systems for drug transfer
across skin [14,39–41]. In these studies, isopropyl myristate is commonly used as a solvent,
since it is considered to serve as a model compound for skin lipids. However, it has since
been reported that the true interfacial kinetics cannot be resolved with the RDC due to the
severe mass transport limitations inherent in the technique [15]. The RDC has also been
used to study more complicated interfacial processes such as kinetics in a microemulsion
system [42], where one of the compartments contains an emulsion.
A comprehensive study of the complex interfacial processes involved in the solvent
extraction of cupric ion by oxime ligands represents one of the most detailed and success-
ful studies carried out with the RDC [37,38]. Recently, the technique was also used to
study the transfer of tetrabutylammonium cations [43] and the kinetics of partitioning of
compounds between octanol and water [44]. In the latter study, Fisk and coworkers
investigated the rates of partitioning of 23 compounds from octanol to an aqueous
phase. The RDC arrangement used most frequently in this work is of the o/o/w type.
So according to Eq. (15), Pow and kow can be calculated from the gradient and intercept of

Copyright © 2001 Marcel Dekker, Inc.


334 Slevin et al.

FIG. 7 Analysis of RDC data for the partition kinetics of cyanazine between octanol and water.

a plot of c 0 =j vs. W 1=2 , where c 0 ¼ cob  Pow cw


b . Figure 7 illustrates some typical results for
cyanazine transfer from octanol to the aqueous phase.

V. THE LIQUID JET RECYCLE REACTOR (LJRR)

Developed by Freeman and Tavlarides [45,46], and based on the liquid jet technique
[47,48], the LJRR provides a method of measuring liquid–liquid reaction kinetics with
direct contact, known interfacial area, renewable interface, and reasonably defined hydro-
dynamics. This method operates by employing an aqueous liquid jet in a concurrent,
coaxially flowing organic solution, shown schematically in Fig. 8.
The aqueous solution flows from the jet nozzle to a receiving capillary with no
overflow into the outer stream, resulting in short contact times of around 0.05 s.
Analysis is implemented by flowing the outer organic phase continually through a closed
loop and monitoring concentration changes spectrophotometrically. The apparatus used
by Freeman and Tavlarides employed capillaries with internal diameters of 2 mm, and the

FIG. 8 Schematic diagram of the liquid jet recycle reactor (LJRR).

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 335

jet length was 3.54 cm. The aqueous jet was arranged vertically, and a gravity fed flow
system was used. This approach enabled measurements to be made on a clean interface
with a known area (measurable by photography). The following equations of mass trans-
fer from the jet are based on certain assumptions, namely, isothermal operation; constant
densities, viscosities, and diffusivities; no ionization or homogeneous reactions; small
penetration depth; negligible curvature; negligible axial diffusion; and steady-state opera-
tion:

@co @c @2 c
u þ v o ¼ Do 2o y>0 ð17Þ
@x @y @y
@cw @c @2 c
u þ v w ¼ Dw 2w y<0 ð18Þ
@x @y @y

where u is the velocity in the x direction and v is the velocity in the y direction. The co-
ordinate system for the liquid–liquid jet is shown in Fig. 9.
The main complication with this technique is that the mass transfer analysis is
nontrivial. For example, the change of velocity as a function of the distance down the
jet was not taken into account in modeling the system. The LJRR has been used to study
the diffusivities of benzene and toluene in water [41] and cupric ion extraction [42,49].

VI. MOVING-DROP TECHNIQUES

The moving-drop method [2] employs a column of one liquid phase through which drops
of a second liquid either rise or fall. The drops are produced at a nozzle situated at one end
of the column and collected at the other end. The contact time and size of the drop are
measurable. Three regimes of mass transport need to be considered: drop formation, free
rise (or fall) and drop coalescence. The solution in the liquid column phase or drop phase
(after contact) may be analyzed to determine the total mass transferred, which may be
related to the interfacial reaction only after mass transfer rates have been determined.

FIG. 9 Co-ordinate system for the LJRR.

Copyright © 2001 Marcel Dekker, Inc.


336 Slevin et al.

The hydrodynamics of the moving drop are difficult to calculate, particularly the
flow characteristics within the droplet itself. However, this technique is still used widely,
because it is a simple and straightforward method. It was recently applied to study the
stripping–extraction kinetics of Mn(II) in an aqueous–kerosene system [50,51]. The effect
of anionic surfactants on the kinetics of extraction of lactic acid from an aqueous phase by
Alamine 336 in a toluene phase was also studied by this technique [52].
A growing-drop method has been reported [53] for measuring interfacial liquid–
liquid reactions, in which mass transport to the growing drop was considered to be
well-defined and calculable. This approach was applied to study the kinetics of the solvent
extraction of cupric ions by complexing ligands.

VII. TECHNIQUES BASED ON DISPERSED LIQUID–LIQUID SYSTEMS

Stirred suspensions of droplets have proven to be a popular approach for studying the
kinetics of liquid–liquid reactions [54–57]. The basic principle is that one liquid phase takes
the form of droplets in the other phase when two immiscible liquids are dispersed. The
droplet size can be controlled by changing the agitator speed. For droplets with a diameter
< 0:15 cm the inside of the drop is essentially stagnant [54], so that mass transfer to the
inside surface of the droplet occurs only by diffusion. In many cases, this technique can
lack the necessary control over both the interfacial area and the transport step for deter-
mination of fundamental interfacial processes [3], but is still of some value as it reproduces
conditions in industrial reactors.
A significant advance was made in this field by Watarai and Freiser [58], who
developed a high-speed automatic system for solvent extraction kinetic studies. The extrac-
tion vessel was a 200 mL Morton flask fitted with a high speed stirrer (0–20,000 rpm) and a
teflon phase separator. The mass transport rates generated with this approach were con-
sidered to be sufficiently high to effectively outrun the kinetics of the chemical processes of
interest. With the aid of the separator, the bulk organic phase was cleanly separated from
a fine dispersion of the two phases in the flask, circulated through a spectrophotometric
flow cell, and returned to the reaction vessel.
The absorbance data enabled the determination of extraction rate constants. For a
pseudo-first-order reaction, the following equation describes the extraction process:
 
AF  A0
ln ¼ kobs t ð19Þ
AF  At
where A0 , At , and AF correspond to absorbance values for the extracted solute at t ¼ 0, t,
and equilibrium, respectively, and kobs is the rate constant observed experimentally. The
extraction of nickel (II) by dithizone in a chloroform phase was studied using this tech-
nique by recording the absorbance for nickel dithizonate. A typical result is shown in Fig.
10. A first-order rate constant of 1:42  102 s1 was obtained. To relate this to an inter-
facial flux, one needs to know the area/volume ratio in the system. This can be obtained,
but requires that the experimental system is ‘‘calibrated’’ by studying a process with
known interfacial kinetics.
Extraction rates of zinc (II) and nickel (II) with ethyldithizone, butyldithizone, or
hexyldithizone in an organic phase (chloroform, CCl4 , n-heptane, or benzene) showed a
first-order dependence on the ligand and metal ion concentration and an inverse-first-
order dependence on the proton concentration. The results were explained by chelate
formation in the interfacial region [59]. The effects of stirring on the distribution equili-

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 337

FIG. 10 Extraction of nickel (II) by dithizone in chloroform. (a) Absorbance of the nickel (II)
dithizone complex in chloroform vs. time. (b) Kinetic analysis in terms of Eq. (19).

brium of n-alkyl-substituted dithizones in several aqueous–organic (chloroform, CCl4 , n-


heptane, or benzene) systems were also reported [60].

VIII. CENTRIFUGAL LIQUID MEMBRANE (CLM) METHOD

The CLM method is a new technique, developed by Nagatani and Watarai [61]. This
method produces a stable, ultrathin two-phase liquid membrane by the centrifugal force
due to the rotation of a cylindrical cell, using the arrangement shown in Fig. 11. The inner
diameter and inner height of the cylindrical cell were 19 and 29 mm, respectively. The
rotation speed was controlled in the range 6000–7500 rpm. The summation of the
absorption spectra of both interfacial and bulk organic phase species was measured in
the direction perpendicular to the rotation axis with a diode array spectrophotometer.
Mass transfer for this technique was examined by studying the protonation and
aggregation kinetics of 5,10,15,20-tetraphenylporphyrin (H2 TTP) at the dodecane–aqu-
eous interface [61]. The rate law for the diffusion-controlled protonation of H2 TTP at the
interface was derived.

Copyright © 2001 Marcel Dekker, Inc.


338 Slevin et al.

FIG. 11 Schematic of the centrifugal liquid membrane apparatus. Typically, the organic phase,
0:150 cm3 , and the aqueous phase, 0:250 cm3 , are spread as an inner liquid membrane of 79 m film
thickness and an outer liquid membrane of 145 m thickness, respectively. (Reprinted from Ref. 61.
Copyright 1998, American Chemical Society.)

The demetalation kinetics of ZnTTP by an acidic aqueous phase have also been
reported [61]. In this study, ZnTTP was considered to adsorb at the interface producing
Zn2þ and free base porphyrin by proton attack. The demetalation kinetics of ZnTTP were
analyzed as a pseudo-first-order reaction, because the proton concentration in the aqueous
phase was in large excess. The rate law was found to be described by
d½ZnTTPT A
 ¼ kobs  int ð20Þ
dt Vo
where kobs ðs1 Þ is the observed rate constant,  is the surface excess concentration (mol
cm2 ) of ZnTTP and [ZnTTPT is the total or initial concentration of ZnTTP dissolved in
the bulk organic phase. Typical first-order kinetic profiles derived from these studies are
shown in Fig. 12. The dependence of the observed demetalation rate constant of ZnTTP
on the proton concentration in the aqueous phase is shown in Fig. 13. The results indicate
that the observed demetalation rate constant has a first-order dependence on the proton
concentration. Based on this observation, the authors suggested that the rate-determining
step was the first attack of protons on ZnTTP at the interface.
A further interesting application of CLM involves the fluorescence quenching reac-
tion between (5,10,15,20-tetraphenylporphyrinato) zinc (II) and methylviologen at a dode-
cane–water interface. [62] This study demonstrated that the quenching reaction could
occur only in the presence of an anionic surfactant, sodium dodecyl sulfate (SDS). The
quenching efficiency depended on the concentration of SDS in the aqueous phase with a
maximum value of 13.5%

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 339

FIG. 12 First-order kinetic profiles observed from CLM measurements of the diprotonation of
H2 TTP (a) and the demetalation of ZnTPP (b). The initial concentrations of H2 TPP and ZnTPP
were 1:71  105 and 8:9  106 mol dm3 , respectively. The concentration of hydrochloric acid was
2:0 mol dm3 . (Reprinted from Ref. 61. Copyright 1998, American Chemical Society.)

IX. ELECTROLYTE DROPPING METHOD

The electrolyte dropping electrode [63] method, introduced in 1976, and subsequently used
in conjunction with the four-electrode potentiostat [64], is a hydrodynamic technique,
offering controlled convective transport. In essence, this approach is identical to the
dropping mercury electrode [65]; however, the drop consists of a flowing electrolyte liquid
phase which forms a polarized ITIES with an immiscible continuous (receptor) phase. In

FIG. 13 Logarithmic plot of the observed demetalation rate constants of ZnTPP vs. acidity func-
tion of hydrochloric acid, at 298 K. The slope of the straight line is close to unity. The concentration
of ZnTPP was 8:9  106 mol dm3 , and the ionic strength was maintained at 2.0 mol dm3 .
(Reprinted from Ref. 61. Copyright 1998, American Chemical Society.)

Copyright © 2001 Marcel Dekker, Inc.


340 Slevin et al.

this approach, the interface is continually refreshed, minimizing the possible problems of
interfacial blocking (by reactants, intermediates, and products). A fast miniaturized ver-
sion of this technique has recently been reported [66].
The electrolyte dropping electrode has found particular application in the study of
ion transfer at the polarized ITIES, with an emphasis on analysis. A range of species have
been detected amperometrically by measuring the transport-limited current:
i ¼ nFD1=2 q2=3 t1=6
d cb ð21Þ
where  is the Ilkovich constant, n is the charge number of the ion transferred, F is
Faraday’s constant, cb is the bulk concentration of the species of interest, q is the volume
flow rate, and td is the drop time. Many aqueous ions which are not accessible to ampero-
metric detection by conventional electrolysis may be measured through the application of
this technique, with suitable complexing ligands in the organic phase. A review focusing on
the analytical applications of the method has appeared [67].
A comprehensive and systematic investigation of the fundamental factors in ion
transfer measurements in the electrolyte dropping electrode arrangement has been carried
out in order to develop a quantitative method for analysis [68]. It was reported that the
limiting currents were proportional to the concentration of monovalent ions such as Csþ ,
tetramethylammonium ion, ClO     
4 , IO4 , I , Br , ReO4 , and BF4 in the concentration
5 3
range between 10 and 10 M in aqueous solution, when 1,2-dichloroethane (DCE) was
used as the organic phase.
A hanging electrolyte drop has also been applied to determine ionic species in solu-
tion using differential-pulse-stripping voltammetry procedures [69]. Particular emphasis
was given to assessing the selectivity and sensitivity of the method. The technique of
current-scan polarography has also been applied in the study of electron-transfer [70]
and coupled electron-transfer–ion-transfer [71,72] reactions at the ITIES in this config-
uration.

X. MICROELECTROCHEMICAL MEASUREMENTS AT EXPANDING


DROPLETS (MEMED)
A. Methodology and Theory
In the majority of methods described thus far, the interfacial kinetics are deduced by
measuring concentration changes in the bulk of the solution rather than at the interface,
where the reaction occurs. This introduces a time lag, limiting the resolution of the mea-
surement in the determination of interfacial kinetics. A more direct approach is to identify
the interfacial flux. This can be achieved in the electrolyte dropping electrode, via the
current flow, but this method is only applicable to net charge-transfer processes at exter-
nally polarized interfaces.
We have developed a new approach, for measuring reaction rates at liquid–liquid
interfaces using ultramicroelectrodes (UMEs). Termed microelectrochemical measure-
ments at expanding droplets (MEMED), the technique overcomes many problems inher-
ent in previous approaches used to investigate spontaneous reactions at liquid–liquid
interfaces. In MEMED, liquid–liquid contact is established by forming drops of one liquid
from a capillary submerged in a second immiscible liquid. The internal diameter of the
capillary is typically in the range 100–200 m. The feeder solution, either an organic or an
aqueous phase, flows into the receptor solution at a constant rate, such that drops form at

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 341

the capillary tip, grow and eventually detach, periodically in a well-defined manner. This
approach is similar to the dropping mercury electrode (DME) [65] and electrolyte drop-
ping electrode [63,64,67–70], discussed above, and conveys the same advantage by provid-
ing a constantly renewable, clean interface.
In MEMED, the interfacial reaction is monitored with a stationary UME, posi-
tioned directly beneath or above (depending on the relative densities of the two liquids)
the expanding drop. The electrode operates in either an amperometric or potentiometric
mode, to measure local changes in concentration in the receptor phase, at the probe tip, as
the drop approaches the electrode. The probe penetrates, and measures directly the devel-
oping concentration profile adjacent to the drop surface in the receptor phase, due to the
two-phase reaction. By solving the convective-diffusion equation for this particular con-
figuration, with appropriate boundary conditions, theoretical concentration profiles can
be generated for comparison with experiment. In this way, the nature of mass transport
and the interfacial reaction can be investigated quantitatively.
To ensure that the detector electrode used in MEMED is a noninvasive probe of the
concentration boundary layer that develops adjacent to the droplet, it is usually necessary
to employ a small-sized UME (less than 2 m diameter). This is essential for amperometric
detection protocols, although larger electrodes, up to 50 m across, can be employed in
potentiometric detection mode [73]. A key strength of the technique is that the electrode
measures directly the concentration profile of a target species involved in the reaction at
the interface, i.e., the spatial distribution of a product or reactant, on the receptor phase
side. The shape of this concentration profile is sensitive to the mass transport character-
istics for the growing drop, and to the interfacial reaction kinetics. A schematic of the
apparatus for MEMED is shown in Fig. 14.
Mass transfer to an expanding droplet has been considered previously for the DME
arrangement [33,74,75] and other growing liquid drop techniques [53,76]. At a simple level,

FIG. 14 Schematic diagram of the experimental arrangement for MEMED.

Copyright © 2001 Marcel Dekker, Inc.


342 Slevin et al.

convective diffusion to a growing drop may be treated by assuming a symmetrically


expanding sphere geometry:
!
@c @2 c 2 @c @c
¼D þ  vr ð22Þ
@t @r2 r @r @r

In Eq. (22), r is the spherical co-ordinate starting at the center of the drop. The variable vr
represents the convective velocity due to the moving surface of the expanding drop. Within
a reference framework centered on the moving drop surface, this is given by
 
q 1 1
vr ¼  ð23Þ
4 r2 r20

where r0 is the (time-dependent) drop radius. This equation assumes that the drop behaves
as a symmetrically expanding sphere, i.e., that the drop expands from a fixed center. While
this is not the case for a droplet growing from a capillary, since the center of the drop
moves, it has been found to be a reasonable approximation to the real system at the
relatively slow flow rates employed.
A second approach is to assume that the drop surface approaching the electrode is a
moving plane. This is appropriate, since the diffusion layer is almost always considerably
smaller than the size of the drop, for most of its lifetime under practical conditions. To a
good approximation, the convective effect close to the moving front is then calculated
based on velocities which are twice those determined from Eq. (23), in order to account for
the moving center of the drop. The convective-diffusion equation which describes this case
is given by

@c @2 c @c
¼ D 2  2vr ð24Þ
@t @r @r

A linear approximation for the velocity term, used to treat the DME problem [74], does
not work for MEMED, because the concentration boundary layers tend to be much larger
for MEMED due to the longer drop times employed.
The presence of the capillary can be ignored in the simulation of mass transport in
MEMED since, for most of its lifetime, the dimensions of the drop are much larger than
the capillary diameter. Moreover, the capillary is not expected to influence measurements
that are made (in a micrometer region) at the opposite side of the drop, given the time scale
of the measurements. For applications of MEMED hitherto, depletion or accumulation
inside the drop have been neglected for two reasons. First, the probe is positioned close to
the portion of the drop where the surface is constantly renewed (see Fig. 15). Internal
streamlines in a drop expanding from a capillary have been investigated experimentally
[77,78], and these observations suggest that fresh material is carried from the capillary to
the front face of the drop, directly opposite the capillary. Second, for the systems studied,
the concentrations of reactants have been carefully chosen such that neither depletion nor
accumulation inside the droplet would be important. This means that in solving the mass
transport problem for MEMED, it is only necessary to consider the domain r  r0 by
solution of either Eq. (22) or (24)
Equations (22) and (24) are readily cast into dimensionless form using the following
terms [74]:

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 343

FIG. 15 Schematic of the internal streamlines in a drop expanding from a capillary tip into a
second immiscible fluid phase.

r
R ¼ pffiffiffiffiffiffiffiffi ð25Þ
Dtd
t
¼ ð26Þ
td
c
C¼ ð27Þ
cb
rffiffiffiffi
t
Vr ¼ vr d ð28Þ
D

where cb is typically the initial or bulk concentration of the reactant in the receptor
solution, but may be the concentration at the surface of the drop (for processes involving
transport-limited transfer from the drop to the receptor phase). The variable td is the total
drop time from the beginning of growth to the point of contact with the probe electrode.
The resulting normalized equations can readily be solved subject to boundary conditions
describing the chemical processes of interest. For the expanding sphere and the moving
plane models, mass transport is described by Eqs. (29) and (30), respectively:

@c @2 c 2 @c @c
¼ þ  Vr ð29Þ
@ @R2 R @R @R
@c @2 c @c
¼  2Vr ð30Þ
@ @R2 @R
Models have been formulated to enable the simulation of the concentration vs. radial
distance profile as it develops with time, from which the time-dependent concentration vs.
distance, d, profile, observed at the probe, can be extracted for comparison with experi-
mental data. Models based on Eqs. (29) and (30) give similar results for conditions
encountered practically.

Copyright © 2001 Marcel Dekker, Inc.


344 Slevin et al.

Typical theoretical concentration profiles, observed at a probe electrode, for the


consumption of a receptor phase species in a first-order interfacial reaction are shown
in Fig. 16. The simulation involved solving Eq. (30) with appropriate boundary condi-
tions.
These results show that the reactant concentration becomes increasingly depleted
close to the drop surface as the rate constant increases, until a limiting case is reached,
where the reaction becomes transport controlled. MEMED can be used to measure quite a
wide range of rate constants, but is particularly sensitive to rate constants in the low–
medium range. In its present form, first-order rate constants higher than 0:01 cm s1
cannot readily be distinguished from a transport-controlled process, although this upper
limit can be increased slightly by the use of shorter drop times.
The nature of mass transport in MEMED has been confirmed with both ampero-
metric and potentiometric studies of bromine transfer from an aqueous phase to DCE [79].
Figure 17 shows typical amperometric data for this case, in which the DCE phase acts as a
sink for Br2 , and a depleted region of Br2 is measured adjacent to the droplet in the
aqueous phase. Video images are also provided, which correspond to particular times
during the amperometric transient: at position (3) the edge of the developing concentra-
tion boundary layer, around the drop, reaches the electrode; the concentration profile is
then mapped out between points (3) and (4). The measured current, i, can be related to the
local concentration, c, via

i ¼ 4nFDac ð31Þ

where n is the number of electrons transferred, F is Faraday’s constant, and a is the radius
of the UME.

FIG. 16 Theoretical first-order reaction approach curves for MEMED. Final drop
radius ¼ 0:05 cm, drop time ¼ 6:3 s, and D ¼ 7:0  106 cm2 s1 were used in the simulation.
From top to bottom the curves are for first-order interfacial rate constants of 1  105 , 1  104 ,
1  103 , and 1  102 cm s1 , and a transport-controlled process.

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 345

FIG. 17 (a) Typical current (i)–time (t) characteristics for bromine transfer from an aqueous phase
to an expanding DCE drop measured at a 1 m diameter Pt UME positioned beneath the growing
drop. The images (1)–(4) correspond to the points indicated on the transient. The final drop size in
(4) is 1.00 mm. Data are analyzed to produce (normalized) concentration vs. distance profiles, such
as that in (b), showing experiment (*) and theory for a transport-limited process (—).

B. Applications
MEMED has been used to study the hydrolysis of triphenylmethyl chloride (TPMCl) in a
two-phase DCE–aqueous arrangement [73]:
TPMClðDCEÞ þ H2 O ! TPMOHðDCEÞ þ Hþ 
ðwÞ þ ClðwÞ ð32Þ

The reaction was followed by the local measurement of chloride ions, at a potentio-
metric Ag/AgCl microelectrode probe, positioned in the aqueous receptor phase, as DCE
droplets containing TPMCl were grown (feeder phase). The reaction was shown unam-
biguously to occur interfacially, and was first-order in TPMCl with a hydrolysis rate
constant of 6:5  105 cm s1 . A typical concentration–distance profile determined in
these experiments is shown in Fig. 18.
MEMED has also been used to investigate the nature of coupled ion-transfer pro-
cesses involved in spontaneous electron transfer at ITIES [80]. In this application, a key
strength of MEMED is that all of the reactants and products involved in the reaction can
be measured, as shown in Figs. 19 and 20. The redox reaction studied involved the
oxidation of either ferrocene (Fc) or decamethylferrocene (DMFc) in a DCE phase
0
(denoted by FcDCE ) by either IrCl2 3
6 or FeðCNÞ6 in the aqueous phase (denoted by Oxw ):
0
0
FcDCE þ Oxw ! Fc þ þ Redw ð33Þ
The results in Figs. 19 and 20 clearly show that these reactions occur at a transport-
0
controlled rate. For example, the concentration of FcDCE is zero at the interface [Figs.
19(a) and 20(a)] with DCE as the receptor phase, and the IrCl2
6 interfacial concentration

Copyright © 2001 Marcel Dekker, Inc.


346 Slevin et al.

FIG. 18 Chloride concentration profile recorded by a microelectrode probe during the hydrolysis
of TPMCl at a DCE drop–aqueous interface (*). The concentration of TPMCl in the organic phase
was 50 mmol dm3 , the drop time from formation to contact with the probe was 4.80 s, and the final
drop radius was 0.55 mm. The solid lines represent theoretical time-dependent concentration profiles,
from top to bottom, generated for k ¼ 3:50  109 , 3:25  109 , and 3:00  109 mol cm2 s1 . A
value of 1:8  105 cm2 s1 was employed for the diffusion coefficient of chloride. (Reprinted from
Ref. 73. Copyright 1997, American Chemical Society.)

is zero with an aqueous receptor phase. The ability to probe all of the reactant and product
profiles is also of value in demonstrating that DMFcþ effectively remains in the DCE
phase under the experimental conditions, while Fcþ is lost extensively from DCE.
The recent applications of MEMED to electron-transfer reactions demonstrate that
it complements scanning electrochemical microscopy (SECM) [81,82] which is discussed in
detail in Chapter 12. MEMED can be used to measure rate constants at the lower scale of
those measurable by SECM, but in the region where the two techniques overlap, they
provide consistent information. In applications hitherto, the reduction of 7,7,8,8-tetracya-
noquinodimethane (TCNQ) in DCE by aqueous ferrocynide and the oxidation of DMFc
by aqueous oxidants have been studied. The purpose of these investigations was to deter-
mine the dependence of heterogeneous electron transfer rate constants on the interfacial
potential drop, which was controlled through the use of potential-determining ions. For a
single potential-determining ion system, the potential drop across an ITIES follows the
Nernst–Donnan equation [5,8],
 
RT a
o
¼ o
þ
w w o
ln o ð34Þ
zF aw
where w o
¼

is the potential drop across the water–oil interface, o
is the
w o w o

standard ion-transfer potential across the water–oil boundary, z is the charge number of
the potential-determining ion, R and T have their usual meanings, and ao and aw are the
activities of the potential-determining ion in the oil and aqueous phases, respectively.
Under experimental conditions where ao is fixed and the activity coefficient of the poten-
tial-determining ion (typically ClO4 ) in the aqueous phase is constant within the concen-
tration range of interest, Eq. (34) can be simplified to

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 347

FIG. 19 Normalized concentration profiles (solid lines) of the reactants and products in the DCE
(a) or aqueous (b) receptor phase for the reaction between Fc (DCE) and IrCl2 6 (aqueous) with
0.1 M ClO 4 in both DCE and the aqueous phase. In each case, the reactant concentration in the
receptor phase was 1 mM, with 10 mM reactant inside the droplet. Drop times and final sizes were
(a) 5.54 s and 0.96 mm, and (b) 6.32 s and 1.00 mm. The theoretical profiles (dashed lines) are for a
transport-controlled reaction, with no transfer of the product ions. (Reprinted from Ref. 80.
Copyright 1999, Royal Society of Chemistry.)

Copyright © 2001 Marcel Dekker, Inc.


348 Slevin et al.

FIG. 20 Normalized concentration profiles (solid lines) of the reactants and products in the DCE
(a) and aqueous (b) receptor phases for the reaction between DMFc (DCE) and IrCl2 6 (aqueous),
with 0.1 M ClO 4 in both DCE and the aqueous phase. In each case, the reactant concentration in the
receptor phase was 1 mM, with 10 mM reactant inside the droplets. Drop times and final sizes were
(a) 5.54 s and 0.96 mm, and (b) 6.32 s and 1.00 mm. The theoretical profiles (dashed lines) are for a
transport-controlled reaction, with no transfer of the product ions. (Reprinted from Ref. 80.
Copyright 1999, Royal Society of Chemistry.)

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 349

ow
¼ const  0:059 log½ClO
4 w ð35Þ
Previous SECM studies have suggested that a Butler–Volmer type approximation
could be used for the ITIES at low driving forces [83]. For a system where an ET reaction
occurs between an aqueous electron donor and an oxidant in the organic phase, the free
energy barrier is given by
 
G6¼ ¼ 0F E 0 þ ow
ð36Þ
where E 0 is the standard potential difference between the organic oxidant and aqueous
electron donor and  is the transfer coefficient. ðE 0 þ ow
Þ denotes the driving force for
the reaction. With a fixed concentration of oxidant in the organic phase, the dependence
of the ET rate on the driving force can be written as [83,84]
k ¼ const 0 expðG6¼ =RTÞ ð37Þ
where k is the bimolecular interfacial reaction rate constant. For a reaction between a
redox couple in each phase, the E 0 value is fixed and the combination of equations (35),
(36), and (37) yields
log k ¼ const 00  2:303 log½ClO
4 w ð38Þ
Thus, the log k vs. log½ClO
4 wdependence should be linear with a slope proportional to .
Typical results, shown in Fig. 21(a), demonstrate that the rate constant for the
reaction between TCNQ and aqueous FeðCNÞ4 6 increases with increasing driving force,
promoted by decreasing ½ClO 4 w as evidenced by the steeper FeðCNÞ6
4
concentration
4
profiles. Moreover, the Tafel plot obtained for ET between FeðCNÞ6 and TCNQ is linear
with an apparent measured  value of 0:31  0:02. In these studies, the concentration of
reactant in the droplet phase was always at least 10 times the concentration of the reactant
in the receptor phase, to ensure that depletion (and diffusional) effects within the droplet
were negligible.
Recently, MEMED was also used to investigate cupric ion extraction-stripping
kinetics in an aqueous–DCE system containing the oxime extractant ligand, Acorga P50
[85,86]. The stripping process showed a first-order dependence on both proton concentra-
tion in the aqueous phase and the complex concentration in the DCE phase. This infor-
mation was consistent with a reaction mechanism involving the first attack of protons on
the cupric ion complex as the rate limiting step, as found in recent studies of Zn2þ strip-
ping from acidic complexing ligands in an immiscible liquid–liquid system [61]. The strip-
ping rate constants measured by MEMED were comparable with those derived from
earlier SECM studies [87].

XI. CONCLUSIONS AND SUGGESTIONS

This review has highlighted the wide range of techniques, which have been used to inves-
tigate reaction kinetics at liquid–liquid interfaces. While significant progress has been
made in the last few years, in particular, many of the techniques currently in use have a
number of drawbacks, and there is considerable scope for the introduction of further
techniques which can match the criteria proposed in Section I.
Of the methodologies considered, the Lewis cell, employs direct contact of the two
liquids, but does not have well-defined hydrodynamics. The constant interface cell with
laminar flow has better-defined hydrodynamics, but the interfacial flux is not measured

Copyright © 2001 Marcel Dekker, Inc.


350 Slevin et al.

FIG. 21 (a) Normalized concentration profiles (solid lines) of FeðCNÞ4 6 (1 mM) in the aqueous
receptor phase, obtained using MEMED technology. The reaction involved the reduction of 10 mM
TCNQ in DCE (feeder phase) by FeðCNÞ4 6 . In all measurements, the DCE phase contained 0.1 M
tetra-n-hexylammonium perchlorate. The drop time and final size were 9.2 s and 0.99 mm. From
bottom to top, the solid experimental curves are shown for ½ClO 4 w ¼ 0:01, 0.025, 0.1, and 0.25 M,
while the corresponding dashed curves (also bottom to top) are for the theoretical behavior for
k ¼ 0:22, 0.09, 0.04, and 0.02 cm s1 M1 . (b) Dependence of k on ClO 4 concentration in the aqueous
phase; the data are those from (a).

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 351

directly. The RDC introduces the added transport resistance of a supported interface in an
effort to obtain well-defined hydrodynamics. The LJRR possesses both direct contact and
reasonably defined hydrodynamics, but mathematical modeling of the flowing solutions is
complex, and the method is difficult to use experimentally. The centrifugal liquid mem-
brane method is a promising new approach, and it will be interesting to see how this
develops with further applications. However, as with all of these methods, the measure-
ment of concentration changes occurs in the bulk of solution and not at the interface
where the reaction occurs.
MEMED meets all of the criteria listed in Section I, for the investigation of liquid–
liquid interfacial kinetics, but is limited in the range of rate constants that can be deter-
mined. While SECM, discussed in Chapter 12, enhances the kinetic domain that can be
measured with ultramicroelectrodes, there are many spontaneous reactions to which
SECM cannot be applied.
The development of hydrodynamic techniques which allow the direct measurement
of interfacial fluxes and interfacial concentrations is likely to be a key trend of future work
in this area. Suitable detectors for local interfacial or near-interfacial measurements
include spectroscopic probes, such as total internal reflection fluorometry [88–90], surface
second-harmonic generation [91], probe beam deflection [92], and spatially resolved UV-
visible absorption spectroscopy [93]. Additionally, building on the ideas in MEMED,
submicrometer or nanometer scale electrodes may prove to be relatively noninvasive
probes of interfacial concentrations in other hydrodynamic systems. The construction
and application of electrodes of this size is now becoming more widespread and general
[94–96].

ACKNOWLEDGMENTS

We wish to thank Dr. John Atherton (Avecia, Huddersfield) for helpful discussions on
liquid–liquid reactions over the past few years. The support of our research in this area by
the EPSRC, Avecia (formerly Zeneca Specialities), and the University of Warwick is
gratefully acknowledged.

REFERENCES

1. H. Freiser. Chem. Rev. 88:611 (1988).


2. P. R. Danesi and R. Chiarizia. CRC Crit. Rev. Anal. Chem. 10:1 (1980).
3. J. H. Atherton. Res. Chem. Kinet. 2:193 (1994).
4. R. B. Gennis, Biomembranes: Molecular Structure and Function, Springer-Verlag, New York,
1989.
5. A. G. Volkov, D. W. Deamer, D. L. Tanelian, and V. S. Markin, Liquid Interfaces in Chemistry
and Biology, Wiley, New York, 1998.
6. J. Koryta. Electrochim. Acta 24:293 (1979).
7. K. Arai, F. Kusu, and K. Takamura, in Liquid–Liquid Interfaces: Theory and Methods, (A. G.
Volkov and D. W. Deamer, eds.), CRC Press, Boca Raton, 1996, pp. 375–400.
8. H. H. Girault and D. J. Schiffrin, in Electroanalytical Chemistry (A. J. Bard, ed.), vol. 15,
Marcel Dekker, New York, 1989, pp. 1–141.
9. P. Vanýsek. Electrochim. Acta 40:2841 (1995).
10. M. Senda and Y. Yamamoto, in Liquid–Liquid Interfaces: Theory and Methods (A. G. Volkov
and D. W. Deamer, eds.), CRC Press, Boca Raton, 1996, pp. 277–294.

Copyright © 2001 Marcel Dekker, Inc.


352 Slevin et al.

11. A. G. Volkov and D. W. Deamer (eds.), Liquid–Liquid Interfaces: Theory and Methods, CRC
Press, Boca Raton, 1996.
12. P. R. Unwin. Faraday Trans. 94:3183 (1998).
13. G. J. Hanna and R. D. Noble. Chem. Rev. 85:583 (1985).
14. W. J. Albery, J. F. Burke, E. B. Leffler, and J. Hadgraft. J. Chem. Soc. Faraday Trans. I
72:1618 (1976).
15. D. E. Leahy and A. R. Wait. J. Pharm. Sci. 75:1157 (1986).
16. J. B. Lewis. Chem. Eng. Sci. 3:248 (1954).
17. J. B. Lewis. Chem. Eng. Sci. 3:260 (1954).
18. W. Nitsch and K. Hillekamp. Chem. Ztg. 96:254 (1972).
19. P. R. Danesi, C. Cianetti, E. P. Horowitz, and H. Diamond. Sep. Sci. Tech. 17:961 (1982).
20. M. Bhaduri, C. Hanson, M. A. Hughes, and R. J. Whewell, Int. Solv. Extr. Conf (Proc.), 1983,
pp. 293–294.
21. S. Asai, H. Nakamura, and Y. Furuichi. AIChE J 38:397 (1992).
22. S. I. Eldessouky, J. A. Daoud, and H. F. Aly. Radiochim. Acta 85:79 (1999).
23. Z. M. Zhou, W. C. Wang, J. S. Wang, H. F. Du, and M. L. Ye. J. Radioanal. Nucl. Chem.
231:167 (1998).
24. E. Combes, C. Sella, D. Bauer, and J. L. Sabot. Hydrometal 46:1 (1997).
25. C. Albasi, J. P. Riba, I. Sokolovska, and V. Bales. J. Chem. Technol. Biotechnol. 69:329 (1997).
26. Y. Miyake, M. Ohkubo, and M. Teramoto. Biotechnol. Bioeng. 38:30 (1991).
27. J. A. Daoud, E. H. Borai, and H. F. Aly. J. Radioanal. Nucl. Chem. 181:165 (1994).
28. C. A. P. Bakker, F. H. Fentener van Vlissingen, and W. J. Beek. Chem. Eng. Sci. 22:1349
(1967).
29. N. G. Maroudas and H. Sawistowski. Chem. Eng. Sci. 19:919 (1964).
30. Z. Zheng and D. Li. Int. Solv. Extr. Conf. (Proc.), 1996, pp. 171–176.
31. Z. Zheng, J. Lu, D. Li, and G. Ma. Chem. Eng. Sci. 53:2327 (1998).
32. V. Yu Filinovskii and Yu V. Pleskov, in Comprehensive Treatise of Electrochemistry, vol. 9 (E.
Yeager, J.O.’M. Bockris, B. E. Conway, and S. Sarangapani, eds.), Plenum Press, New York,
1984.
33. V. G. Levich, Physiochemical Hydrodynamics, Prentice-Hall, Englewood Cliffs, NJ, 1962.
34. J. V. Macpherson and P. R. Unwin. Prog. React. Kinet. 20:185 (1995).
35. A. J. Bard and L. R. Faulkner, Electrochemical Methods, John Wiley, New York, 1980.
36. W. J. Albery and M. L. Hitchman, Ring-Disc Electrodes, Oxford University Press, Oxford,
1971.
37. W. J. Albery, R. A. Choudhery, and P. R. Fisk. Faraday Discuss. Chem. Soc. 77:53 (1984).
38. W. J. Albery and R. A. Choudhery. J. Phys. Chem. 92:1142 (1988).
39. W. J. Albery and J. Hadgraft. J. Pharm. Pharmacol. 31:65 (1979).
40. R. H. Guy, T. P. Aquino, and D. H. Honda. J. Phys. Chem. 86:280 (1982).
41. R. H. Guy, T. R. Aquino, and D. H. Honda. J. Phys. Chem. 86:2861 (1982).
42. W. J. Albery, R. A. Choudhery, N. Z. Atay, and B. H. Robinson. J. Chem. Soc. Faraday
Trans. I 83:2407 (1987).
43. J. A. Manzanares, R. Lahtinen, B. Quinn, K. Kontturi, and D. J. Schiffrin. Electrochim. Acta
44:59 (1998).
44. P. R. Fisk, M. G. Ford, and P. Watson, in Comprehensive Chemical Kinetics, (R. G. Compton
and G. Hancock, eds.), vol. 37, Elsevier, Amsterdam, 1999, pp. 161–194.
45. R. W. Freeman and L. L. Tavlarides. Chem. Eng. Sci. 35:559 (1980).
46. R. W. Freeman and L. L. Tavlarides. Chem. Eng. Sci. 37:1547 (1982).
47. H. Yu and G. F. Scheele. Int. J. Multiphase Flow 2:153 (1975).
48. P. Gonspodinov, S. Radev, and S. Penchev. Int. J. Multiphase Flow 5:87 (1979).
49. L. T. Huang and R. W. Freeman, Abstracts of AIChE Spring National Meeting, Anaheim, CA,
May 1984, p. 64d.
50. R. K. Biswas, M. A. Habib, and M. F. Bari. Hydrometal. 46:349 (1997).
51. R. K. Biswas, M. A. Habib, and M. F. Bari. Hydrometal. 42:399 (1996).

Copyright © 2001 Marcel Dekker, Inc.


Hydrodynamic Techniques 353

52. V. Thom, B. Gutierrez, C. Pazos, and J. Coca. J. Dispersion Sci. Technol. 17:407 (1996).
53. G. L. Bauer. Ph.D. Thesis, University of Wisconsin, 1975.
54. S. Nagata, Mixing: Principles and Applications, Halsted Press, 1975.
55. F. M. Menger. J. Am. Chem. Soc. 92:5965 (1970).
56. M. Halpern, Y. Sasson, and M. Rabinovitz. J. Org. Chem. 49:2011 (1984).
57. R. Solaro, S. D’Antone, and E. Chiellini. J. Org. Chem. 45:4179 (1980).
58. H. Watarai, L. Cunningham, and H. Freiser. Anal. Chem. 54:2390 (1982).
59. H. Watarai and H. Freiser. J. Am. Chem. Soc. 105:189 (1983).
60. H. Watarai and H. Freiser. J. Am. Chem. Soc. 105:191 (1983).
61. H. Nagatani and H. Watarai. Anal. Chem. 70:28060 (1998).
62. H. Nagatani and H. Watarai. Chem. Lett. 701 (1999).
63. J. Koryta, P. Vanýsek, and M. Brezina. J. Electroanal. Chem. 67:263 (1976).
64. Z. Samec, V. Marecek, J. Weber, and D. Homolka. J. Electroanal. Chem. 99:385 (1979).
65. J. Heyrovsky and P. Zuman, Practical Polarography, Academic Press, New York, 1961.
66. A. Baars, K. Aoki, and J. Watanabe. J. Electroanal. Chem. 464:128 (1999).
67. E. Wang and Z. Sun. Trends Anal. Chem. 7:99 (1988).
68. S. Kihara, M. Suzuki, K. Maeda, K. Ogura, S. Umetani, M.Matsui, and Z. Yoshida. Anal.
Chem. 58:2954 (1986).
69. V. Marecek and Z. Samec. Anal. Chim. Acta 151:265 (1983).
70. S. Kihara, M. Suzuki, K. Maeda, K. Ogura, M. Matsui, and Z. Yoshida. J. Electroanal. Chem.
271:107 (1989).
71. H. Ohde, K. Maeda, Y. Yoshida, and S. Kihara. Electrochim. Acta 44:23 (1998).
72. K. Maeda, M. Nishihara, H. Ohde, and S. Kihara. Anal. Sci. 14:85 (1998).
73. C. J. Slevin and P. R. Unwin. Langmuir 13:4799 (1997).
74. D. Britz, Digital Simulation in Electrochemistry, 2nd edn., Springer-Verlag, New York, 1988.
75. S. Pons, B. Speiser, J. F. McAleer and P. P. Schmidt. Electrochim. Acta 27:1711 (1982).
76. A. T. Popovich, R. E. Jervis, and O. Trass. Chem. Eng. Sci. 19:357 (1964).
77. R. Clift, J. P. Grace, and M. E. Webber, Bubbles, Drops and Particles, Academic Press, New
York, 1978.
78. D. K. Guha, F. Vyarawalla, and F. De. Can. J. Chem. Eng. 65:448 (1987).
79. C. J. Slevin and P. R. Unwin. Langmuir 15:7361 (1999).
80. J. Zhang, C. J. Slevin, and P. R. Unwin. Chem. Commun. 1501–1502 (1999).
81. J. Zhang and P. R. Unwin. J. Phys. Chem. B 104:2341 (2000).
82. J. Zhang, A. L. Barker, and P. R. Unwin. J. Electroanal. Chem. 483:95 (2000).
83. M. Tsionsky, A. J. Bard, and M. V. Mirkin. J. Phys. Chem. 100:17881 (1996).
84. R. A. Marcus. J. Phys. Chem. 94:4152 (1990). Addendum, J. Phys. Chem. 94:7742 (1990).
85. J. Zhang and P. R. Unwin. Phys. Chem. Chem. Phys. 2:1267 (2000).
86. D. Chapman, C. J. Slevin, J. Zhang, and P. R. Unwin, manuscript in preparation.
87. A. L. Barker, J. V. Macpherson, C. J. Slevin, and P. R. Unwin. J. Phys. Chem. B 103:7260
(1998).
88. M. Fujiwara, S. Tsukahara, and H. Watarai. Phys. Chem. Chem. Phys. 1:2949 (1999).
89. R. A. W. Dryfe, Z. Ding, R. G. Wellington, P. F. Brevet, A. M. Kuznetzov, and H. H. Girault.
J. Phys. Chem. A 101:2519 (1997).
90. H. Watarai and F. Funaki. Langmuir 12:6717 (1996).
91. D. A. Higgins and R. M. Corn. J. Phys. Chem. 97:489 (1993).
92. M. C. Barbero, O. H. Miras, and R. Kötz. J. Electrochem. Soc. 138:669–672 (1991).
93. A. Deputy, H. P. Wu, and R. L. McCreey. J. Phys. Chem. 94:3620 (1990).
94. M. V. Mirkin, F. R. F. Fan, and A. J. Bard. Science 257:364 (1992).
95. B. D. Pendley and H. D. Abruna. Anal. Chem. 62:782 (1990).
96. C. J. Slevin, N. J. Gray, J. V. Macpherson, M. A. Webb, and P. R. Unwin. Electrochem.
Commun. 1:282 (1999).

Copyright © 2001 Marcel Dekker, Inc.

You might also like