ch5 PDF

5
Solvent Extraction Kinetics

PIER ROBERTO DANESI International Atomic Energy Agency,
Vienna, Austria
5.1 GENERAL PRINCIPLES

The kinetics of solvent extraction is a function of both the various chemical
reactions occurring in the system and the rates of diffusion of the various species
that control the chemistry of the extraction process.
5.1.1 Rate-Controlling Role

of the Chemical Reactions
The dependence of the kinetics on the chemical reactions is easily understood
by considering that the final products of any extraction process are usually in a
chemical state different from the initial unreacted species. This is true even for
the simple partition of neutral molecules between two immiscible liquid phases,
where the chemical change is in the solvation environment of the extracted
species. More drastic chemical changes take place in the extraction of a metal
cation from an aqueous solution by a chelating extractant dissolved in an organic
diluent. In the extraction some of the solvation water molecules can be removed
from the metal ion, and a new coordination compound, soluble in the organic
phase, is formed with the chelating group of the extractant. In addition, the
extracting reagent can undergo an acid dissociation reaction and, together with
the extractant-metal complex, can undergo changes in aggregation in the organic
phase. Consequently, whenever at least one of the chemical steps of the overall
reaction mechanism is slow enough, compared with the diffusion rate, the kinet-
ics of extraction would depend on the rate of the slow chemical reactions.
Copyright © 2004 by Taylor & Francis Group, LLC

For solvent extraction systems, we have two additional complications.
First, the chemical reactions can take place, at least in principle, in two bulk
phases, since we are dealing with two immiscible liquid layers. Second, the
chemical reactions can occur in the two-dimensional region called the liquid–
liquid interface, that separates the two immiscible liquids, or in a thin volume
region very close to it. When interfacial chemical reactions are important, the
situation is analogous to that describing the kinetics of the chemical reactions
encountered in heterogeneous catalysis and in some electrode processes. Al-
though a relatively large number of sophisticated techniques are available for
studying chemical reactions at solid–fluid interfaces, very few tools have been
developed to investigate chemical changes occurring at liquid–liquid interfaces.
Our knowledge of such interfacial reactions, therefore, is still limited and is
Downloaded by [Universite Laval] at 14:32 15 May 2016
based largely on indirect evidence and speculations.
5.1.2 Rate-Controlling Role

of the Diffusional Processes
To understand the dependence of the extraction kinetics on the rate of diffusion
of the various chemical species that participate in the extraction reaction, we
first have to distinguish between diffusion in the bulk phases and diffusion
through the thin layers adjacent to the interface. In most solvent extraction pro-
cesses of practical interest, both the aqueous and the organic phase are effi-
ciently stirred. It follows that transport of material from the bulk of the phases
up to a region very close to the interface can be considered instantaneous and
that diffusion in the bulk of the phases can be neglected. Nevertheless, diffu-
sional processes can still have an appreciable influence on the solvent extraction
kinetics. For example, even when the two phases are vigorously stirred, it is
possible to describe interfacial diffusion by assuming the existence of two stag-
nant thin layers of finite thickness located on the aqueous and organic side of
the interface. This model of the interface, often referred to as the two-film the-
ory [1,2], is extremely useful for describing extraction kinetics that are con-
trolled by diffusion occurring in proximity to the interface. The two-film model
is used throughout our treatment. Other models and theories of higher complex-
ity exist for describing diffusional transport in proximity to the interface. These
theories (penetration, surface renewal, boundary layer) are described in detail in
more advanced books [3,4].
The two films are schematically described in Fig. 5.1, in which the pres-
ence of an interfacially absorbed layer of extractant molecules is also shown. δo
and δw represent the thickness of the organic and aqueous films, respectively. In
these layers the liquid phases are considered completely stagnant (i.e., no move-
ment of the fluids takes place in spite of the mechanical energy that is dissipated
in the two-phase system to provoke mixing of the aqueous and organic phases).

Fig. 5.1 Interfacial diffusion films. δo and δw are the thickness of the organic and
aqueous films, respectively. The presence of an adsorbed layer of extractant molecules
at the interface is also shown.
These two thin liquid films, which are also called diffusion films, diffusion
layers, or Nernst films, have thicknesses that range between 10−2 and 10−4 cm
(in this chapter centimeter-gram-second (CGS) units are used, since most pub-
lished data on diffusion and extraction kinetics are reported in these units; com-
parison with literature values is, therefore, straightforward).
The description of the diffusion films as completely stagnant layers, having
definite and well-identified thicknesses, represents only a practical approxima-
tion useful for a simple mathematical description of interfacial diffusion. A

more realistic physical description should consider that, starting at a given dis-
tance from the liquid interface, the renewal of the organic and the aqueous
fluids becomes progressively less as the interface is approached.
The thickness of the diffusion films is, up to a certain limiting value, an
inverse function of the mechanical energy dissipated into the liquid system to
stir it. This means that δo and δw depend on the stirring rate of the two phases,
on the geometry of the solvent extraction equipment, and of the stirrers used, as
well as on the viscosity and density of the two liquids. For given liquid phases
and a fixed apparatus, the thickness of the diffusion films initially decreases
with increase in the rate of mixing. However, the decrease does not go beyond
a certain limiting value, and the thickness of the diffusion films never seems to
go down to zero, with limiting thickness of about 10−3 –10−4 cm eventually
reached with all systems. The exact limiting value depends on the specific physi-
cochemical properties of the liquids and the specific hydrodynamic conditions.
Even when dealing with the most efficiently stirred systems, the chemical spe-
cies that have to be transported to or away from the interface for the extraction
reaction to take place always have to diffuse through the diffusion films. The
time required for this often can be comparable or longer than that required for
the actual chemical changes in the extraction. As a consequence, diffusion
through these films controls the overall kinetics of extraction. Experience has
shown that film diffusion is the predominating rate-controlling factor in many
practical extraction processes. Although these diffusion processes occur through
very thin films, the films must be considered macroscopic, since the diffusion
distances involved exceed the molecular dimensions by several orders of magni-
tude.
5.2 DIFFUSION
Diffusion is that irreversible process by which matter spontaneously moves from
a region of higher concentration to one of lower concentration, leading to equal-
ization of concentrations within a single phase.
The laws of diffusion correlate the rate of flow of the diffusing substance
with the concentration gradient responsible for the flow. In general, in a multi-
component system, the process is described by as many diffusion equations as
the number of chemical species in the system. Moreover, the diffusion equations
are intercorrelated. Nevertheless, when dealing with solutions containing ex-
tractable species and extracting reagents at concentrations much lower than
those of the molecules of the solvents (dilute solutions), diffusion is sufficiently
well described by considering only those species for which the concentrations
appreciably change during the extraction reaction.
The diffusion flow, J, of a chemical species is defined as the amount of
matter of this species passing perpendicularly through a reference surface of

unit area during a unit time. The dimensions of J are those of mass per square
centimeter per second (mass cm−2 s−1). When the concentrations of the diffusing
species (c) are expressed in moles per cubic centimeter (mol cm−3), the correla-
tion between the flux across a unit reference area located perpendicularly to the
linear coordinate x (along which diffusion occurs) and the concentration gradi-
ent, ∂c /∂x, will be given by the Fick’s first law of diffusion, that is,
J = − D∂c / ∂x (5.1)
D is the diffusion coefficient of the species under consideration. Its dimen-
sions are given in cm2 s−1. Equation (5.1) implies that D is independent of con-
centration; it is a constant, characteristic of each diffusing species, in a given
medium, at constant temperature. This is only approximately true, but holds
sufficiently well for most diffusing species of interest in solvent extraction sys-
tems. The values of D for the majority of the extractable species fall in the
range 10−5 –10−6 cm2 s−1 in both water and organic solutions. When dealing with
viscous phases or bulky organic extractants, or with complexes that may un-
dergo polymerization reactions, the D values can drop to 10−7 or 10−8 cm2 s−1.
Equation (5.1) is extremely useful to evaluate the flux whenever the con-
centration gradient can be considered constant with time (i.e., when we can
assume a steady state). When a steady state cannot be assumed, the concentra-
tion change with time must also be considered. The non–steady-state diffusion
is expressed by Fick’s second law of diffusion:
∂c / ∂t = D∂ 2c / ∂x 2 (5.2)
Diffusion is a complex phenomenon. A complete physical description in-
volves conceptual and mathematical difficulties associated with the need to in-
volve theories of molecular interactions and to solve complicated differential
equations [3–6]. Here and in sections 5.8 and 5.9, we present only a simplified
picture of the diffusional processes, which is valid for limiting conditions. The
objective is to make the reader aware of the importance of this phenomenon in
connection with solvent extraction kinetics.
For steady-state diffusion occurring across flat and thin diffusion films,
only one dimension can be considered and Eq. (5.1) is greatly simplified. More-
over, by replacing differentials with finite increments and assuming a linear
concentration profile within the film of thickness δ, Eq. (5.1) becomes
J = − D dc / dx = − D(c2 − c1 ) / δ = (c2 − c1 ) / ∆ (5.3)
where
∆ = δ/ D (5.4)
is a diffusional parameter dependent on the thickness of the diffusion film and the
value of the diffusion constant of the diffusing species. The units of ∆ are cm−1 s.

Some simple considerations about the variation of concentrations with
time that can be expected in diffusion-controlled processes, can be made by
considering the imaginary one-liquid-phase system shown in Fig. 5.2. The imag-
inary system consists of two hypothetical well-stirred aqueous reservoirs of vol-
umes V1 and V2, containing different concentrations of the same solute, in con-
tact through a nonstirred aqueous liquid film across which the solute moves
only under the influence of molecular diffusion. Since c1 and c2 are allowed to
vary with time, diffusion through the film of thickness δ tends to equalize c1
and c2. This means that J, c1 and c2 are a function of time. If a linear concentra-
tion gradient is always assumed and the amount of diffusing matter inside the
film is negligible relative to that in the reservoirs, the flux equation J can be
integrated to yield a simple correlation between c1, c2 and time (t). By consider-
ing that the correlation between the flux and the concentration variations within
the reservoirs is
−V1 Q −1dc1 / dt + V2 Q −1 dc2 / dt = J (5.5)
where Q is the area of the diffusion film (equal on both sides), and, assuming,
for simplicity, that V1 = V2 = V, Eqs. (5.3) and (5.5) can be combined to give
VQ∆dc1 / dt = c1 − c2 (5.6)
For the initial conditions at t = 0, c2 = 0, and c1 = c , the integration of Eq.
0
1
(5.6) is straightforward. Separation of the variables leads to
Qdt / V ∆ = dc1 (c10 − 2c2 ) (5.7)
Fig. 5.2 Concentration profile of a solute diffusing across an unstirred liquid film of
thickness δ in contact with two well-stirred reservoirs. The same liquid phase is assumed
throughout this imaginary model system.

Upon integration and insertion of Eq. (5.4), one obtains
ln{(2c1 − c10 ) / c10 } = −2 Q Dt / V δ (5.8)
Two conclusions follow from Eq. (5.8):
1. For t = ∞ , (2c1 − c10) = 0, that is
c1 (at t = ∞) = c2 (at t = ∞) = c10 / 2. (5.9)
This means that at equilibrium the concentrations in the two reservoirs are
equal.
2. The condition of equal concentrations is approached with a rate that is di-
rectly proportional to the contact area (Q) and the diffusion coefficient (D),
and inversely proportional to the volume of the reservoir (V) and the thick-
ness of the diffusion layer (δ).
Although these conclusions are intuitive and trivial and are based on sev-
eral crude approximations introduced into an imaginary model system consisting
of only one liquid phase, the simple mathematical treatment developed helps
the reader better appreciate the fundamental role that film diffusion plays in
controlling the rate of solvent extraction. When two different liquid phases
(aqueous and organic) are in contact, two liquid films can be assumed to exist
on the two opposite sides of the interface, and diffusion through both must be
considered. The presence of two phases at which species can be unevenly dis-
tributed, and with interfacial chemical reactions of varying degree of complex-
ity, will in fact appear in the extraction rate equations only as modifications of
this simple diffusion equation.
5.3 RATE LAWS AND MECHANISMS

5.3.1 Rate Laws and Reaction Mechanisms
For a general treatment of the various rate laws, the reader is referred to several
specialized books existing on the subject [7,8]. However, care must be exercised
in extrapolating the information obtained from the study of reactions in solutions
to solvent extraction systems. Solvent extraction deals with two solutions in
contact and in the presence of the interface, which has to be crossed by a solute
during the extraction, a fact that requires special consideration.
In this section, we briefly summarize the fundamental concepts that are
useful for better understanding those specific solvent extraction rates and mech-
anisms, which are presented in some detail in subsequent sections.
The immediate result of a kinetic study is a rate law. For a general reaction
with the stoichiometric equation
xX + yY = zZ + wW (5.10)

the rate is expressed by
−d [X] −d [Y] + d [ Z] + d [ W]
, , or (5.11)
dt dt dt dt
where t indicates time and the square brackets concentrations. The rate of the
reaction is given as function of the composition of the system:
rate = − d [X]/dt = f {[X], [Y], [Z], [W]} (5.12)
The first goal of any kinetic study is to devise experiments that establish
the algebraic form of the rate law and to evaluate the rate constants. Rate laws
of the form
−d [X]/dt = k [X]x [Y]y [ Z]z [ W]w
(5.13)
are relatively easy to treat mathematically when x, y, z, and w are small inte-
gers.
Rate laws can be derived by measuring concentration variations as func-
tion of time or the initial rates as function of the initial concentrations. Unfortu-
nately, there is no general method for finding the rate law and the reaction order
from measurements of concentration vs. time or other types of measurements.
Usually, a trial-and-error procedure is used, based upon intelligent guesses.
Once a rate law has been defined and the rate constants evaluated, the
next step is to correlate it with the most likely mechanism of the extraction
reaction.
Since the experimental kinetic data refer to a reaction rate and how this is
affected by variables, such as concentration, temperature, nature of the solvent,
presence of other solutes, structural variations of the reactants, and so forth, the
assignment of a mechanism is always only indirectly derived from primary data.
Therefore, it is not surprising that more than one mechanism has often been
proposed to explain the same rate law and that reaction mechanisms, which
were once consistent with all experimental information available on a system,
have later on been considered erroneous and have been disregarded, or drasti-
cally modified, as long as new experimental evidence was accumulated. In gen-
eral, the stoichiometry of the reaction, even when this is a simple one, cannot
be directly related with its mechanism, and when the reaction occurs through a
series of elementary steps, the possibility that the experimental rate law may be
interpreted in terms of alternative mechanism increases. Therefore, to resolve
ambiguities as much as possible, one must use all the physicochemical informa-
tion available on the system. Particularly useful here is information on the struc-
tural relations between the reactants, the intermediate, and the reaction products.
Table 5.1 presents the mathematical description of selected very simple
rate laws.
Irreversible first-order reactions [Eq. (5.14) in Table 5.1] depend only on

Table 5.1 Common Rate Expressions
Irreversible first-order reaction
Reaction X + other reactants (OR) → products (P) (5.14a)
Rate expression −d[X]/dt = k[X] (5.14b)
Solution [X] = [X]0e−kt (5.14c)
Conditions OR = const.; reverse rate: zero; [X] = [X]0 at t = 0
Reversible first-order reactions
k1
Reaction X + OR Y + OP (other products) (5.15a)
k2
Rate expression −d[Xl/dt = k1[X] − k2[Y] (5.15b)

−kt
Solution [X] = [X]0(k1 + k2)−1 (k2 + k1 e ) (5.15c)
Conditions k = k1 + k2; OR and OP = const.; at time t = 0, (5.15d)
[Y] = 0 and [X] = [X]0

Series first-order reactions
k1 k2
Reaction X→Y→Z (5.16a)
Rate expression d[Y]/dt = k1[X]0 e − k t − k2 [ Y]
1
(5.16b)
Solution [Y] = [X]0 k1 (k2 − k1 ) −1 (e − k t − e − k t )
1 2
(5.16c)
Conditions no Y or Z initially present; [X] = [X]0 at t = 0
the concentration of a reactant X, and the rate of the reverse reaction, rate
(reverse), is always equal to zero. Simple distribution processes between immis-
cible liquid phases of noncharged components characterized by very large val-
ues of the partition constant can be formally treated as first-order irreversible
reactions.
Reversible first-order reactions are given by Eq. (5.15) of Table 5.1.
Equation (5.15c) can be rewritten in a more elegant form by considering that at
equilibrium (t = ∞ ) the net rate is equal to zero, that is
k1[X]eq = k2 [Y]eq (5.17)
where subscript eq indicates equilibrium concentrations. The equilibrium con-
stant of the reaction is
K = k1 / k2 = [Y]eq /[X]eq (5.18)
Equation (5.15c) can then be reformulated as
ln{([X] − [X]eq )([X]0 − [X eq ]) −1} = −(k1 + k2 )t (5.19)
Equation (5.19) shows that the rate law is still first-order, provided the
quantity ([X] − [X]eq) is used instead of [X]. A plot of ln([X] − [X]eq) vs. t will
then be a straight line of slope -(k1 + k2). The individual rate constants of the
reaction can still be evaluated from the slope of such a plot, providing the

equilibrium constant [Eq. (5.18)] is also available. Many distribution processes
between immiscible liquid phases of noncharged species, as well as extraction
of metal ions performed at very low metal concentrations, can be treated as
first-order reversible reactions when the value of the equilibrium (partition) con-
stant is not very high.
Equations (5.16) of Table 5.1 refer to series first-order reactions. Of inter-
est for the solvent extraction kinetics is a special case arising when the concen-
tration of the intermediate, [Y], may be considered essentially constant (i.e.,
d[Y]/dt = 0). This approximation, called the stationary state or steady-state ap-
proximation, is particularly good when the intermediate is very reactive and
present at very small concentrations. This situation is often met when the inter-
mediate [Y] is an interfacially adsorbed species. One then obtains
[Y] = [X]0e − k1 t k1 / k2 (5.20)

and
[ Z] = [X]0 {1 − (1 + k1 / k2 )e − k1 t } (5.21)
k2 Ⰷ k1 means a very reactive intermediate present at very low concentra-
tion; t Ⰷ 1/k2 indicates that the induction period of the reaction has been passed,
that is, the initial phase of the reaction, when the rate of formation of Z is slow,
is immeasurably short in comparison with the time response of the measuring
technique.
5.4 SOME KINETICS AND MECHANISMS

WITH COMPLEX IONS
5.4.1 Introduction
In the solvent extraction of metal ions, the composition of the coordination
sphere often changes, either because of the formation of complexes between the
metal ions and a complexing reagent, preferentially soluble in an organic phase,
or because of the replacement of a ligand in the aqueous phase metal complex
with another more lipophilic one in the organic phase. Complexes that rather
readily exchange ligands or water molecules with other ligands are termed la-
bile. Inert complexes do not exhibit ligand exchange or do so very slowly. The
terms labile and inert refer to the velocity of the reactions and should not be
confused with the terms stable and unstable, which refer to the thermodynamic
tendency of species to exist under equilibrium conditions. Although, in many
cases, thermodynamic stability parallels the inertness of a complex, there are
many examples of complex ions that are thermodynamically stable, but kinetic-
ally labile, or the reverse.
Kinetic studies of ligand exchange reactions seek to elucidate the mecha-

nisms of such reactions. In the following section, information on the mech-
anisms that control solvent exchange and complex formation is reported. Al-
though the information has been derived from studies of one-phase liquid
systems, much of it can be safely extrapolated to solvent extraction systems
whenever interfacial film diffusion processes can be neglected as the rate-
determining processes.
5.4.2 Mechanisms
Solvent exchange and complex formation are special cases of nucleophilic sub-
stitution reactions.
Generally, when a coordination bond is broken in a metal ion complex,
the electron pair responsible for the bond accompanies the leaving group (com-
monly, the organic ligand). It follows that reactions involving solvent exchange
and complex formation have as reactive centers electron-deficient groups. There-
fore, they can all be considered as nucleophilic substitutions (in symbols SN).
The basic classification of nucleophilic substitutions is founded on the
consideration that when a new metal complex is formed through the breaking
of a coordination bond with the first ligand (or water) and the formation of a
new coordination bond with the second ligand, the rupture and formation of the
two bonds can occur to a greater or lesser extent in a synchronous manner.
When the rupture and the formation of the bonds occur in a synchronous way,
the mechanism is called substitution nucleophilic bimolecular (in symbols SN2).
On the other extreme, when the rupture of the first bond precedes the formation
of the new one, the mechanism is called substitution nucleophilic unimolecular
(in symbols SN1). Mechanisms SN2 and SN1 are only limiting cases, and an
entire range of intermediate situations exists.
Inorganic chemists prefer a slightly more detailed classification and subdi-
vide the range of possible mechanisms into the following four groups.
1. D mechanism. This is the limiting dissociative mechanism, and a transient
inter-mediate of reduced coordination number is formed. The intermediate
persists long enough to discriminate between potential nucleophiles in its
vicinity. Here we are dealing with an SN1 limiting process, since the dissoci-
ation of the metal-ligand bond fully anticipates the formation of the new
bond. The vacancy in the coordination shell that occurs as a result of the
dissociation is then taken by the new ligand.
2. Id mechanism. This is the dissociative interchange mechanism and is similar
to the previous one in the sense that dissociation is still the major rate-
controlling factor. Therefore, we are still dealing with an SN1 process. Nev-
ertheless, differently from the D mechanism, no experimental proof exists
that an intermediate of lower coordination is formed. The mechanism in-
volves a fast outer sphere association between the initial complex and the

incoming nucleophile (the new ligand), and when the complex metal ion
and the incoming ligand have opposite charges, this outer sphere associate
is an ion pair. As a result of the formation of this new outer sphere associ-
ate, the new group is now suitably placed to enter the primary coordination
sphere of the metal ion as soon as the outgoing group has left.
3. Ia mechanism. This is the associative interchange mechanism and is similar
to the Id mechanism, in the sense that here, also, no proof exists that an
intermediate of different coordination is formed. However, differently from
Id, we have significant interaction between the incoming group and the
metal ion in the transition state. In this instance, the process is partially of
the SN2 type.
4. A mechanism. This is the pure associative mechanism, and here we are
dealing with a process entirely of the SN2 type. The rate-determining step
of the mechanism is the association between the complex metal ion and the
entering ligand, leading to the formation of an intermediate of increased
coordination number that can be experimentally identified. This mechanism
is often operative in ligand-displacement reactions occurring with planar
tetracoordinated complexes.
In practice, there is a continuous gradation of mechanisms from D,
through Id and Ia to A, depending on the extent of interaction between the metal
cation and the incoming group in the transition intermediate. Moreover, the
diagnosis of the mechanism is not a straightforward and unambiguous process.
A variety of methods have to be used, depending on circumstances, since an
approach based only on rate-law consideration can easily lead to false conclu-
sions.
5.4.3 Water Exchange and Complex Formation

from Aquoions
The rate at which solvent molecules are exchanged between the primary solva-
tion shell of a cation and the bulk solvent is of primary importance in the kinet-
ics of complex formation from aquocations. In both water exchange and com-
plex formation, a solvent molecule in the solvated cation is replaced with a new
molecule (another water molecule or a ligand). Therefore, strong correlations
exist between the kinetics and mechanisms of the two types of reactions.
The basic mechanism for the formation of complexes from aquocations
[9] is based on the following observations:
For a given metal ion the rates and the activation parameters for complex
formation are similar to those for water exchange, with the complex formation
rate constants usually about a factor of 10 lower than those for water exchange,
For a given metal ion the rates show little or no dependence on the
identity of the ligand.

This means that, at least as a first approximation, the complex formation
mechanism can be described by the rapid equilibrium formation of an ion pair
(for oppositely charged species) or of an outer sphere association complex be-
tween the aquometal ion and the ligand, followed by a rate-determining inter-
change step (D or Id mechanism). Here, the ligand bonds to the complex cation
only after a water molecule has been released from the primary coordination
sphere. For a six-coordinated single-charged cation, the mechanism can be ex-
pressed as
M(H 2O)6+ + L− → (L− , M (H 2O)6+ ) → kM (H 2O)5 + H 2O
fast outer sphere slow
associate (5.22)
The rate law corresponding to such a mechanism is:

rate = {K ass k[M + ][L− ]}/{1 + K ass [L− ]} (5.23)
In Eqs. (5.22–5.23), coordination water molecules are omitted and Kass
indicates the ion pair (or outer sphere association) constant. In general Kass is a
small number (≤10) and Eq. (5.23) simplifies to
rate = K ass k[M + ][L− ] (5.24)
Rate Eq. (5.23) indicates that the process is second-order, even if the rate-
controlling step is unimolecular. Furthermore, since the constant Kass differs only
slightly from species to species, depending mostly on the charge density, only
a small dependence on the nature of the incoming ligand can be expected.
Once the value of Kass is estimated, either from theoretical considerations
or from separate experimental measurements, the rate constants of the water
exchange rate-controlling step can be calculated. The values of k are in good
agreement with the independently evaluated rate constants of the water ex-
change process:
M(H 2O)6+ + H 2O* → M(H 2O)*(H 2O)5 +H 2O (5.25)
In Eq. (5.25) H2O* represents a water molecule initially present outside the
coordination sphere of the metal ion, which, as a result of the exchange, has
entered the first coordination sphere. It follows that the degree of kinetic reactiv-
ity of aquometal ions with complexing agents parallels their kinetic lability to-
ward water exchange. Moreover, since the water exchange rate constants of
most metal ions are known, predictions on the rate of complex formation of
aquometal ions can be made.
Rates of water exchange at metal ions vary tremendously with the nature
of the cation. The range is from almost 1010 s−1, for Cu2+ or Cr2+, down to less
than 10−7 s−1 for Rh3+. The values for many metal aquocations are shown in Fig.

5.3 on a logarithmic scale. They have been obtained by nuclear magnetic reso-
nance (NMR) and sound absorption, for the rapidly exchanging ions, and by
isotopic labeling for the more inert ones.
Generally, in solvent extraction, for the simplest case of an organic com-
plexing reagent (ligand) reacting with an aquometal ion, when the water ex-
change rate constant is <102 s−1, the reaction can be considered slow enough that
the complex formation rate may compete with the rate of diffusion through the
interfacial films in controlling the overall extraction kinetics. In exceptional
cases, dealing with very efficiently stirred phases (in which the thickness of the
diffusion films may even be reduced to about 10−4 cm), complexation reactions
with rate constants as high as 106 s−1 can be rate limiting in solvent extraction.
In all other cases, the ions can be considered as reacting instantaneously relative
to the rate of film diffusion. The rate of diffusion also is the predominating rate-
controlling process when the aquocations are extracted into the organic phase
with their entire coordination sphere of water molecules. These considerations
do not hold when other slow chemical processes, such as hydrolysis reactions
of the metal cation, polymerizations in the organic phase of the extractant or
the metal complex, adsorption-desorption processes at the interphase, keto-enol
Fig. 5.3 Logarithm of water exchange constants (s−1) for aquometal ions at 25°C.
(From Ref. 26.)

tautomerizations, or others, control the chemistry of the solvent extraction sys-
tem. In these cases, it is possible that the slow chemical control predominates
over the faster diffusional process. In such systems, solvent extraction rate data
can still provide information on the rate and mechanism of the chemical pro-
cesses.
5.4.4 Ligand Displacement Reactions

For ligand displacement reactions occurring in octahedral complexes, very few
generalizations can be made, since the reaction mechanisms tend to be specific
to each chemical system. This is unfortunate because many solvent extraction
processes are indeed ligand displacement reactions of octahedral complexes in
which L is the extractant [see Eq. (5.22)]. However, it has been experimentally
observed, at least for aqueous-phase reactions, that the variation of the rates with
the identity of the ligand correlates well with the variation in the thermodynamic
stability of the complex. Therefore, whenever the complex is not extracted un-
changed into the organic phase, thermodynamically very stable complexes can
be expected to react slowly with the extractant. An example is the extraction of
trivalent lanthanide or actinide cations from aqueous solutions by diethylhexyl
phosphoric acid (HDEHP). If the aqueous medium contains only weakly com-
plexing ligands (Cl− or NO−3), the extraction kinetics are very fast. On the other
hand, in the presence of polyaminocarboxylic acids, such as EDTA that are
powerful complexing agents, the extraction reaction proceeds only slowly.
For planar tetracoordinated complexes, the mechanistic problem is more
straightforward. A large amount of experimental evidence on ligand substitution
reactions for complexes of transition metals with electronic configuration d8,
such as Pt(II), Pd(II), Au(III), Ni(II), and Rh(I), has been accumulated. The
transition state is a five-coordinated species having the structure of a trigonal
bipyramid on which the entering group lays on the trigonal plane. Ligand-substi-
tution reactions with planar tetracoordinated complexes are often slow in com-
parison with the rate of diffusion through the interfacial diffusional films. It
follows that the kinetics of solvent extraction of planar tetracoordinate com-
plexes in a well-stirred system are often entirely controlled by the rate of the
chemical process. Studies on the rate of extraction can then be directly used
to obtain information on the rate of the biphasic ligand-substitution reactions,
neglecting the complications from diffusional contributions.
5.5 THE LIQUID–LIQUID INTERFACE

As mentioned in section 5. 1, when the two immiscible aqueous and organic
phases are in contact and transfer of chemical species occurs from one phase to
the other, the diffusional transport in proximity of the interface can be described

by assuming the presence on either side of the interface of two stagnant films
(the diffusional films). Although these films are thin when compared with the
bulk phases, their thickness is still macroscopic (about 10−3 or 10−4 cm) in com-
parison with the molecular dimensions and the range of action of the molecular
and ionic forces. Thus, these films have a finite volume, and the concentrations
of chemical species within them must always be considered as volume concen-
trations. Moreover, the physicochemical properties of the liquids within these
films, such as density, viscosity, dielectric constant, charge distribution, are the
same as those of the bulk phases. However, the situation changes when the
distance from the interface approaches the order of magnitude of the molecular
dimensions and the range of action of the molecular and ionic forces. In this
region, generally defined as the interface, the physicochemical properties differ
from those of the bulk phases and, therefore, any chemical change occurring
herein is likely to be affected by the different environment. The difference will
be particularly enhanced by the presence on the interface of adsorbed layers of
polar or ionizable extractant molecules. The extractant, because of its simultane-
ous hydrophobic-hydrophilic (low water affinity–high water affinity) nature,
necessary to maintain a high solubility in low dielectric constant diluents, and
selective complexing power relative to water-soluble species, tends to orient
itself with the polar (or ionizable) groups facing the aqueous side of the inter-
face. The rest of the molecule, having a prevalent hydrophobic character, will
be directed instead toward the organic phase.
Unfortunately, little direct information is available on the physicochemical
properties of the interface, since real interfacial properties (dielectric constant,
viscosity, density, charge distribution) are difficult to measure, and the interpre-
tation of the limited results so far available on systems relevant to solvent ex-
traction are open to discussion. Interfacial tension measurements are, in this
respect, an exception and can be easily performed by several standard physico-
chemical techniques. Specialized treatises on surface chemistry provide an ex-
haustive description of the interfacial phenomena [10,11]. The interfacial ten-
sion, γ, is defined as that force per unit length that is required to increase the
contact surface of two immiscible liquids by 1 cm2. Its units, in the CGS system,
are dyne per centimeter (dyne cm−1). Adsorption of extractant molecules at the
interface lowers the interfacial tension and makes it easier to disperse one phase
into the other.
Interfacial tension studies are particularly important because they can pro-
vide useful information on the interfacial concentration of the extractant. The
simultaneous hydrophobic-hydrophilic nature of extracting reagents has the re-
sulting effect of maximizing the reagent affinity for the interfacial zone, at
which both the hydrophobic and hydrophilic parts of the molecules can mini-
mize their free energy of solution. Moreover, as previously mentioned, a prefer-
ential orientation of the extractant groups takes place at the interface. Conse-

quently, most solvent extraction reagents are interfacially adsorbed and produce
a lowering of the organic-aqueous interfacial tension. The extent of this adsorp-
tion is a function of the chemical nature and structure of the hydrophilic and
hydrophobic groups, of the extractant bulk concentration, and of the physico-
chemical properties of the diluent. The extractant interfacial concentration can
be extremely high, depending on the way molecules can pack themselves at the
interface, and even moderately strong surfactants can form an interface fully
covered with extractant molecules when their bulk concentration is as low as
10−3 M or less. It follows that, when an extractant is a strong surfactant and
exhibits low solubility in the water phase, the interfacial zone is a region where
a high probability exists that the reaction between an aqueous soluble species
and an organic soluble extracting reagent can take place. This is why a strong
surface activity of an extractant can sometimes lend support to extraction mech-

anisms involving interfacial chemical reactions.
Interfacial concentrations can be evaluated from interfacial tension mea-
surements by utilizing the Gibbs equation
d ∏ / d ln c = nikT (5.26)
In Eq. (5.26), Π is the interfacial pressure of the aqueous-organic system,
equal to (γ0 − γ) [i.e., to the difference between the interfacial tensions without
the extractant (γ0) and the extractant at concentration c (γ)], c is the bulk organic
concentration of the extractant, and ni is the number of adsorbed molecules of
the extractant at the interface. The shape of a typical Π vs. ln c curve is shown
in Fig. 5.4; ni can be evaluated from the value of the slopes of the curve at each
c. However, great care must be exercised when evaluating interfacial concentra-
tions from the slopes of the curves because Eq. (5.26) is only an ideal law, and
many systems do not conform to this ideal behavior, even when the solutions
are very dilute. Here, the proportionality constant between dΠ/d ln c and ni is
different from kT. Nevertheless, Eq. (5.26) can still be used to derive infor-
mation on the bulk organic concentration necessary to achieve an interface
completely saturated with extractant molecules (i.e., a constant interfacial con-
centration). According to Eq. (5.26), the occurrence of a constant interfacial
concentration is indicated by a constant slope in a Π vs. ln c plot. Therefore,
the value of c at which the plot Π vs. ln c becomes rectilinear can be taken as
the bulk concentration of the extractant required to fully saturate the interface.
Many extractants reach a constant interfacial concentration at bulk organic
concentrations far below the practical concentrations that are generally used to
perform extraction kinetic studies. This means that when writing a rate law for
an extraction mechanism that is based on interfacial chemical reactions, the
interfacial concentrations can often be incorporated into the apparent rate con-
stants. This leads to simplifications in the rate laws and to ambiguities in their
interpretation, which are discussed in later sections.

Fig. 5.4 Typical interfacial pressure (Π) vs. logarithm of bulk organic concentration
(log c) plotted for an extractant exhibiting surface-active properties.
It is interesting to observe that alkylammonium salts, alkylarylsulfonic

acids, hydroxyoximes, alkylphosphoric acids, and alkylhydroxamic acids, as
well as neutral extractants such as crown ethers and tributyl phosphate, all form
water-organic interfaces saturated with extractant molecules when their bulk
organic concentration is larger than 10−3 M.
The liquid–liquid interface is not a sharply defined surface, but rather an
ill-defined region the boundaries of which extend above and below the layer of
the interfacially adsorbed extractant molecules. Nevertheless, it is often conve-
nient to assume a mathematical dividing surface located where the physico-
chemical properties of the two-phase system experience the sharpest discontinu-
ity. This imaginary surface of zero thickness, which we refer to as the interface,
is useful for defining the concentrations of interfacially adsorbed species, which
can be expressed in moles per square centimeter (mol cm−2). On the other hand,

when in later sections we refer to interfacial concentrations, we mean volume
concentrations in that region which is immediately adjacent to this imaginary
dividing surface (i.e., at the extreme limit of the diffusional films).
The physical depth of the microscopic interfacial region can be estimated
to correspond to the distance over which interfacial molecular and ionic forces
exert their influence. Although molecules or ions experience no net forces in
the interior of the bulk phases, these forces become unbalanced as the ions or
molecules move toward the interface. On the aqueous side of the interfaces,
where monolayers of charged molecules or polar groups can be present, these
forces can be felt over several nanometers. On the organic side of the interface,
where van der Waals forces are mainly operative, the interfacial region generally
extends for tens of nanometers. The van der Waals forces decrease with the
seventh power of the intermolecular distances, so molecules experience essen-

tially symmetric forces, once they are a few molecular diameters away from the
interface.
Although an experimentally verified molecular model of the interface is
not yet available, a qualitative description of the interfacial zone can be at-
tempted on the basis of speculations that extrapolate to the liquid–liquid inter-
face results that are predicted from existing and verified physicochemical theo-
ries. Considering that extracting reagents can be present at the interface with
their polar or ionizable groups facing the water side, a more structured water
can be expected in the interfacial zone. The polar heads of the extractant tend
to polarize the water molecules close to the interface, producing a more struc-
tured and viscous interfacial water, as depicted in Fig. 5.5. This interfacial water,
having the possible depth of several molecular layers, is the result of the water
penetration of the hydrophilic groups of the extractant, which hydrogen bond
with the water molecules close to the interface. As a consequence, interfacial
water should represent a more viscous environment to diffusing species than
ordinary bulk water. The aqueous side of the interface can then be envisioned, at
the microscopic level, as a very particular medium that still retains the chemical
properties of bulk water, but for which the physical properties are much more
like those of a strongly structured solvent, such as glycerol [12]. This interfacial
zone should be characterized by lower fluidity, lower dielectric constant, higher
stiffness, and be more compact than bulk water. The extent of these effects is a
function of the surface-active properties of the extractant for a given organic
diluent–aqueous electrolyte solution system. They can slow ligand exchange
reactions, with octahedral aquoions occurring at the interface, since the more
viscous interfacial environment slows the associative component of the overall
interchange mechanism that controls this type of reactions. The associative step
is, by contrast, very fast and, therefore, not rate controlling for most ligand-
substitution reactions of octahedral aquoions occurring in bulk water.

Fig. 5.5 Simplified picture of the interfacial structuring effect caused by an ionizable
extractant adsorbed at the organic-water interface.
5.6 RATE-CONTROLLING EXTRACTION REGIMES

5.6.1 Definition of the Extraction Regimes
Section 5.1 describes how, in a stirred system, solvent extraction kinetics can
be controlled only by slow chemical reactions or only by diffusion through the
interfacial films. An intermediate situation can also occur whereby both the rates

of chemical reactions and of interfacial film diffusion simultaneously control
the rate of extraction.
When one or more of the chemical reactions is sufficiently slow in com-
parison with the rate of diffusion to and away from the interface of the various
species taking part in an extraction reaction, such that diffusion can be consid-
ered instantaneous, the solvent extraction kinetics occur in a kinetic regime. In
this case, the extraction rate can be entirely described in terms of chemical
reactions. This situation may occur either when the system is very efficiently
stirred and when one or more of the chemical reactions proceeds slowly, or
when the chemical reactions are moderately fast, but the diffusion coefficients
of the transported species are very high and the thickness of the two diffusion
films is close to zero. In practice the latter situation never occurs, as diffusion
coefficients in liquids generally do not exceed 10−4 cm2 s−1, and the depth of the
diffusion films apparently is never less than 10−4 cm.
If, however, all chemical reactions in the biphasic system are very fast in
comparison with the diffusional processes, the solvent extraction kinetics are
said to occur in a diffusional regime. In this situation, the rate of extraction can
be described simply in terms of interfacial film diffusion. Chemical engineers
define this as “mass transfer with instantaneous chemical reaction.” A diffu-
sional regime may control the extraction rate also in the presence of relatively
slow chemical reactions whenever the degree of stirring of the two phases is so
poor that thick diffusional films can be assumed on the two sides of the inter-
face. When both chemical reactions and film diffusion processes occur at rates
that are comparable, the solvent extraction kinetics are said to take place in a
mixed diffusional-kinetic regime, which in engineering, is often referred to as
“mass transfer with slow chemical reactions.” This is the most complicated case,
since the rate of extraction must be described in terms of both diffusional pro-
cesses and chemical reactions, and a complete mathematical description can be
obtained only by simultaneously solving the differential equations of diffusion
and those of chemical kinetics.
The unambiguous identification of the extraction rate regime (diffusional,
kinetic, or mixed) is difficult from both the experimental and theoretical viewpoints
[12,13]. Experimental difficulties exist because a large set of different experimental
information, obtained in self-consistent conditions and over a very broad range of
several chemical and physical variables, is needed. Unless simplifying assump-
tions can be used, frequently the differential equations have no analytical solu-
tions, and boundary conditions have to be determined by specific experiments.
5.6.2 Identification of the Extraction Regimes

The experimental identification of the regime that controls the extraction kinet-
ics is, in general, a problem that cannot be solved by reference to only one set

of measurements. On the contrary, in some systems, a definite situation cannot
be obtained even when the rate of extraction is studied as a function of both
hydrodynamic parameters (viscosity and density of the liquids, geometry of the
apparatus, geometry of stirrers, or stirring rate) and concentrations of the chemi-
cal species involved in the extraction reaction. This difficulty is because some-
times extraction rates may show the same dependence on hydrodynamic and
concentration parameters, even though the processes responsible for the rate are
quite different. For a correct hypothesis on the type of regime that controls the
extraction kinetics, it is necessary to supplement kinetic investigations with
other information concerning the biphasic system. This may be the interfacial
tension, the solubility of the extractant in the aqueous phase, the composition
of the species in solution, and so forth.
The criteria that are most often used to distinguish between a diffusional
regime and a kinetic regime are
1. Comparison of the heat transfer and the mass transfer coefficients
2. Evaluation of the activation energy of extraction
3. The reference substance method
4. Dependency of the rate of extraction on the rate of stirring of the two
phases.
Criteria 1–3 are difficult to apply and have been used only occasionally
to evaluate the extraction regime.
1. According to this criterion, when, in the same apparatus, the same depen-
dence of the heat transfer coefficient and the mass transfer coefficient on
the stirring rate of the phases is observed, the conclusion can be reached
that the extraction occurs in a diffusional regime.
2. The rationale behind the use of this method is that, when the rate is con-
trolled by a chemical reaction, the activation energy is generally higher than
that expected for a diffusion-controlled process (diffusion coefficients vary
only slightly with temperature). However, this criterion is not always very
meaningful, since many chemical reactions occurring in solvent extraction
processes exhibit activation energies of only a few kilocalories per mole,
i.e., have the same order of magnitude as those of diffusional processes.
3. The reference substance method is based on the addition to the solution,
containing the species for which the transfer rate is going to be investigated,
of another inert component for which the rate of extraction is known to be
controlled only by diffusion. By following the simultaneous transfer of the
species of interest and of the reference component as function of the hydro-
dynamic conditions in the extraction apparatus, a diffusional regime will be
indicated by a similar functional dependence, whereas a kinetic regime is
indicated by a sharply different one.
4. This criterion is the simplest, as proved by its widespread use. A typical

curve of rate of extraction vs. stirring rate is shown in Fig. 5.6. Such curves
are generally obtained both when constant interfacial-area-stirred cells
(Lewis cells) and vigorously mixed flasks are used (see section 5.10 for a
description of the various techniques). In general, a process occurring under
the influence of diffusional contributions is characterized by an increase of
the rate of extraction, as long as the stirring rate of the two phases is
increased (see Fig. 5.6, zone A). When, on the other hand, the rate of ex-
traction is independent of the stirring rate (see Fig. 5.6, zone B), it is some-
times possible to assume that the extraction process occurs in a kinetic
regime.
The rationale behind this criterion is that an increase in stirring rate produces
a decrease in thickness of the diffusion films. Moreover, since at low stirring
rates the relationship between stirring rate and the inverse thickness of the
diffusion films (at least in a constant interfacial-area-stirred cell) is approxi-
mately linear, the first part of such a plot will usually approximate a straight
line. In any case, the rate of extraction will increase with the rate of stirring,
as long as a process is totally or partially diffusion controlled. When, eventu-
ally, the thickness of the diffusion films is reduced to zero, only chemical
reactions can be rate controlling, and the rate of extraction becomes indepen-
dent of the stirring rate. Unfortunately, this kind of reasoning can lead to erro-
Fig. 5.6 Typical extraction rate vs. stirring rate for constant interfacial–area-stirred cell.

neous conclusions with both constant interfacial-area-stirred cells and vigor-
ously stirred flasks.
In the case of constant interfacial–area-stirred cells, although zone A is
certainly an indication that the process is controlled by diffusional processes,
the opposite is not true for zone B. In fact, in spite of the increased stirring rate,
it may happen that the thickness of the diffusion films never decreases below
a sufficiently low value to make diffusion so fast that it can be completely
neglected relative to the rate of the chemical reactions. This effect, some-
times called “slip effect,” depends on the specific hydrodynamic conditions
of the apparatus in which the extraction takes place and simulates a kinetic re-
gime.
As far as vigorously stirred vessels are concerned, strong ambiguities are also
faced when this criterion is applied. Unless the overall interfacial area is known
and taken into consideration, an increase in the rate of extraction can take place
with the increase in stirring rate when the system is in a kinetic regime deter-
mined by interfacial chemical reactions. The increase observed in zone A indi-
cates here an increase in the number of droplets of the dispersed phase (propor-
tional to the overall interfacial area) and not a decrease in the thickness of the
diffusion film. Moreover, the plateau region of zone B does not necessarily
prove that the extraction occurs in a kinetic regime. Although at high stirring
rates the number of droplets of the dispersed phase eventually becomes con-
stant, since the rate of drop formation equals the rate of drop coalescence, lack
of internal circulation and poor mixing can occur inside the droplets of the
dispersed phase. This is particularly true with small droplets and in the pres-
ence of extractants that are strong surfactants. Therefore, here also, a plateau
region may simulate a diffusion-controlled regime. It is then apparent that crite-
rion 4 can also lead to ambiguous or misleading conclusions.
Finally, it has to be emphasized that both the hydrodynamic parameters
and the concentrations of the species involved in the extraction reaction simulta-
neously determine whether the extraction regime is of kinetic, diffusional, or
mixed diffusional-kinetic type. It, therefore, is not surprising that different in-
vestigators, who studied the same chemical solvent extraction system in differ-
ent hydrodynamic and concentration conditions, may have interpreted their re-
sults in terms of completely different extraction regimes.
5.7 KINETIC REGIME

In a kinetic regime system, the kinetics of solvent extraction can be described
in terms of chemical reactions occurring in the bulk phases or at the interface.
The number of possible mechanisms is, in principle, very large, and only the
specific chemical composition of the system determines the controlling mecha-
nism. Nevertheless, some generalizations are possible on considerations based

on the nature of the reagents involved in the extraction reaction. Since, in many
metal extractions, ligand substitution reactions take place, rate laws similar to
those for complexation reactions in solution may be expected. During most ex-
traction processes, coordinated water molecules or ligands are substituted in part
or wholly by molecules of a more organophilic ligand (the extractant) or of the
organic diluent. When the extractant shows little solubility in the aqueous phase
and is a strong surfactant, the ligand substitution reaction may take place at the
interface. The highest concentration of extractant molecules is at the interface,
so the probability of reaction is higher than in the bulk phases.
In this section, we describe three simple cases of rates and mechanisms
that have been found suitable for the interpretation of extraction kinetic pro-
cesses in kinetic regimes. These simple cases deal with the extraction reaction
of a monovalent metal cation M+ (solvation water molecules are omitted in the

notation) with a weakly acidic solvent extraction reagent, BH. The overall ex-
traction reaction is
M + + BH(org) → MB(org) + H + (5.27)
with equilibrium extraction constant, Kex equal to
K ex = [MB]eq [H + ]eq /[M + ]eq [BH]eq (5.28)
Ideal behavior of all solute species will be assumed.
Case 1: The rate-determining step of the extraction reaction is the aque-
ous phase complex formation between the metal ion and the anion of the extract-
ing reagent. Even if at very low concentration, BH will always be present in
the aqueous phase because of its solubility in water. The rate-determining step
of the extraction reaction is as follows:
k1
M + + B− ↔ MB (slow) (5.29)
k−1
with a reaction rate expression:

rate = −d [M + ]/ dt = k1[M + ][B− ] − k−1[MB] (5.30)
To derive an expression for the rate in terms of easily measurable quanti-
ties (initial concentrations), the following (instantaneously established) equilib-
ria have to be included (cf. Chap. 4):
1. BH ↔ BH; K DB = [BH]/ [BH] (5.31)
−
2. BH ↔ B +H ; K a = [B ][H ]/[BH]
+ + +
(5.32)
3. MB ↔ MB(org); K DM = [MB]/[MB] (5.33)
By substituting Eqs. (5.31–5.33) into Eq. (5.30), we obtain the rate ex-
pression

−1
rate = k1K a / K DB [M + ][BH][H + ]−1 − k−1K DM
−1
[MB] (5.34)
In the right-hand side of Eq. (5.34), the first term represents the forward
rate of extraction and the second term the reverse rate of extraction.
The values of the apparent rate constants of the extraction reaction,
−1 −1
k1K a K DB and k−1K DM (5.35)
permit the evaluation of the rate constants of the aqueous complex formation
reaction only if KDB, KDM, and Ka are known. Here, comparisons can be made
with literature values of the rate constants for reaction in aqueous solutions
between the same metal ion and aqueous soluble ligands containing the same
complexing groups as the extracting reagent. Agreement between the values
supports, but does not prove, the validity of the proposed mechanism.
The apparent values of the rate constants of the solvent extraction reaction
are usually evaluated by measuring the rate of extraction of M+ as function of
[BH] (at [H+] constant), of [H+] (at [BH] constant), and of [MB] (at [H+] and
[BH] constant). The experimental conditions are usually chosen in such a way
that the reaction can be assumed pseudo-first-order for [M+]. The apparent rate
constants are evaluated from the slope of the straight lines obtained by plotting
ln ([M + ] − [M + ]eq )/([M + ]0 − [M + ]eq ) vs. t (5.36)
Subscript 0 indicates initial concentrations in the aqueous phase.
It must be observed that when this mechanism holds, the rate of the extrac-
tion reaction is independent of the interfacial area, Q, and the volume of the
phases, V. The expected logarithmic dependency of the forward rate of extrac-
tion on the specific interfacial area (S = Q/V), the organic concentration of the
extracting reagent and the aqueous acidity, is shown in Fig. 5.7, case 1.
Case 2: The rate-determining step of the extraction reaction is the inter-
facial formation of the complex between the metal ion and the interfacially
adsorbed extracting reagent. Here, the rate-determining step of the extraction
reaction can be written as
k2
M + + B− (ad) ↔ MB(ad) (slow) (5.37)
k−2
where (ad) indicates species adsorbed at the liquid–liquid interface. The rate of
the reaction that follows is
rate = −d [M + ]/ dt = k2 S w [M + ][B− ]ad − k2 So [MB]ad (5.38)
where Sw and So represent the specific interfacial areas [i.e., the ratios between
the interfacial area Q (cm2) and the volumes of the aqueous (Vw) and organic
(Vo) phase, respectively]. Throughout the following treatments we will always
assume, for simplicity, that Vw = Vo = V, and

Fig. 5.7 Logarithm of the forward extraction rate (normalized to the extractable metal
concentration) vs. log interfacial area (a), log [BH], and log [H+] for some exemplary
kinetic regimes.

Q / V = S = SW = So (5.39)
To derive a rate of extraction in terms of easily measurable quantities, we
express the concentration of the reagent adsorbed at the liquid–liquid interface
as function of its bulk organic concentration. This can be done by utilizing the
Langmuir’s adsorption law: that is,
[BH]ad = (α 2 [BH]/ γ 2 ) /(1 + [BH]/ γ 2 ) (5.40)
In Eq. (5.40), α2 and γ2 are Langmuir adsorption constants. Their values

are characteristic of each specific system. We can then distinguish two regions
of adsorption. One defined as ideal, where 1 Ⰷ [BH]/ γ2, yielding
[BH]ad = (α 2 [BH]/ γ 2 ) = α[BH] (5.41)
and one of complete interface saturation with extractant molecules, where 1 Ⰷ

[BH]/ γ2, yielding
[BH]ad = α 2 (5.42)
With extractants exhibiting strong surface active properties, Eq. (5.42)
holds true in the entire concentration range, which is generally of interest in
practical studies, and the interface becomes fully saturated with extractant mole-
cules when their bulk concentration is as low as 10−3 M.
The following, instantaneously established, equilibria then have to be con-
sidered:
1. BH(org) ↔ BH(ad) (5.43)
adsorption of the extractant molecules at the interface described by Langmuir’s
law [see Eq. (5.40)]
2. BH(ad) ↔ B− (ad ) + H + ; K a* = [B− ]ad [H + ][BH]ad
−1
(5.44)
interfacial dissociation of the extractant (involving no interfacial adsorption of
the dissociated protons)
3. MB(ad) + BH(org) ↔ MB(org)+BH(ad);
K ex0 = ([MB][BH]ad ) /([MB][BH] (5.45)
fast replacement at the interface of the interfacially adsorbed metal complex
with bulk molecules of the extractant. Equations (5.40), (5.44), (5.45) can be
inserted into Eq. (5.38) to obtain:
rate = S k2α 2 K a*[M + ][H + ]−1 − S k−2 K 0 ex −1 α 2 [MB]−1[BH] (5.46)
which holds for a fully saturated interface, and
rate = S k2 α K a*[M + ][BH][H + ]−1 − S k−2 K 0 ex −1 α[MB] (5.47)

which holds for ideal adsorption.
The equilibrium constant of the extraction reaction [see Eq. (5.28)] is
equal to
K ex = k2 K a* K ex0 k−−21 (5.48)
for both ideal adsorption and for a saturated interface. The expression for Kex
is derived from either Eq. (5.46) or Eq. (5.47), since at equilibrium, rate (for-
ward) = rate (reverse).
Equations (5.34) and (5.47) have the same dependence on the concentra-
tion variables. The only difference is that for the interfacial reaction the apparent
rate constants are directly proportional to the ratio S = Q/V. This dependency
can be experimentally verified by measuring the apparent rate of extraction as
a function of the interfacial area and the volume of the phases. A plot of the
apparent rate constant of the forward rate of extraction vs. S must yield a straight
line through the origin of the axes when case 2 holds.
Equation (5.46) shows a further difference relative to Eq. (5.34). Equation
(5.46) indicates a zero reaction order relative to [BH] in the forward extraction
rate, reflecting the complete saturation of the interface with the extracting re-
agent.
The determination of the true rate constants, k1 and k−2, is generally more
difficult than for case 1, since information on α, α2, K*a and K0ex can be difficult
to obtain.
Case 3: There are two interfacial rate-determining steps, consisting of:
(1) formation of an interfacial complex between the interfacially adsorbed mole-
cules of the extractant and the metal ion; and (2) transfer of the interfacial
complex from the interface to the bulk organic phase and simultaneous replace-
ment of the interfacial vacancy with bulk organic molecules of the extractant.
For this mechanism, we distinguish two possibilities. The first (case 3.1) de-
scribes the reaction with the dissociated anion of the extracting reagent, B−(ad).
The second (case 3.2) describes the reation with the undissociated extractant,
BH(ad).
Case 3.1: The first mechanism is represented by the following equa-
tions:
k3
M + + B− (ad) ↔ MB(ad) (slow) (5.49)
k−3
k4
MB(ad) + BH(org) ↔ BH(ad) + MB(org) (slow) (5.50)
k−4
The rate equations holding for the two slow steps are:
rate 1 = S k3 [M + ][B− ]ad − S k−3 [MB]ad (5.51)

rate 2 = S k4 [MB]ad [BH] − S k−4 [BH]ad [MB] (5.52)
When the metal concentration is sufficiently low and the concentration of
the interfacially adsorbed metal complex is consequently low, for the stationary
condition at [MB]ad
rate 1 = rate 2 (5.53)
Equation (5.53) can then be solved for [MB], after having introduced the substi-
tution:
[B− ]ad = K a* [BH]ad [H + ]−1 (5.54)
and considering that, for a fully saturated interface [BH]ad = α2. The expression
for [MB]ad found in this way is

[MB]ad = {k3 K a* α 2 [M + ][H + ]−1 + k−4 α 2 [MB]}/{k−3 + k4 [BH]} (5.55)
Substituting the value of [MB]ad into either Eq. (5.51) or Eq. (5.52) gives
rate = {S k3k4 K a*α 2 [BH][M + ][H + ]−1}{k−3 + k4 [BH]}−1
− {S k−3k−4α 2 [MB]}{k−3 + k4 [BH]}−1 (5.56)
An expression for the equilibrium constant of the extraction reaction can be
obtained from Eq. (5.56) by considering that, at equilibrium, one can set rate
(forward) = rate (reverse). Thus
K ex = K 3 k4 K a* k−−31 k−−41 (5.57)
Comparison of Eqs. (5.34), (5.47), and (5.56) shows that if [BH] is suffi-
ciently small to allow introduction of the approximation k4 [BH] Ⰶ k−3 into the
denominator of Eq. (5.56) (i.e., for a low and restricted concentration range of
the extractant), the three rate equations have exactly the same dependence on
[M+], [BH], and [H+], although the extraction mechanisms are characterized by
different rate-determining steps:
1. A homogeneous reaction in the aqueous phase with the anion of the extract-
ant [see Eq. (5.34)].
2. An interfacial reaction with the anion of the extractant, which is ideally
adsorbed at the interface [see Eq. (5.47)].
3. Two sequential interfacial reactions: the first one being the reaction with
the anion of the extractant that saturates the interface; the second one being
the slow desorption of the interfacial complex from the interface [see Eq.
(5.56)].
The difference among these mechanisms can be seen only by measuring
the dependence of the rate of extraction on the specific interfacial area, S, and
by using the broadest possible concentration range of the reactants. If a depen-

dence on S exists, case 1 can be excluded. If in addition the extraction rate is
initially first order and then becomes zero order when [BH] increases, case 2
can also be ruled out.
Case 3.2: When the interfacially reactive species are the undissociated
molecules of the extractant adsorbed at the interface [i.e., the first rate-determin-
ing step of the two-step mechanism is the reaction between the metal ion M+
and HB(ad)], the following equations will hold:
M + + BH(ad) ↔ MB(ad) + H + (slow) (5.58)
MB(ad) + BH(org) ↔ BH(ad) + MB(org) (slow) (5.59)

The rate equations for the slow steps [see Eqs. (5.58) and (5.59)] are
rate 1 = k3* S [M + ][BH]ad − k−*3 S [MB]ad [H + ] (5.60)
rate 2 = k4* S [MB]ad [BH] − k−*4 S [BH]ad [MB] (5.61)

By assuming that the stationary condition, (rate 1 = rate 2), can be applied
and by introducing the same substitutions as those used to derive Eq. (5.56), we
obtain (for a fully saturated interface):
[MB]ad = {α 2 k3* [M + ] − α 2 k−4 [MB]}{k−*3 [H + ] + k4 [BH]}−1 (5.62)
The insertion of Eq. (5.62) into Eq. (5.61) gives
rate = rate (forward) − rate (reverse) =
={S k3*k4α 2 [BH][M + ]}{k−*3[H + ] + k4 [BH]}−1
−{S k−*3k−4α 2 [MB][H + ]}{k−*3 + k4 [BH]}−1 (5.63)
Equation (5.63) has the same dependence on [M+], [BH], and [H+] as Eqs.
(5.34) and (5.47) only when the simplification k4[BH] Ⰶ k−3 [H+] can be intro-
duced in the denominator.
Figure 5.7 shows the logarithmic dependencies expected between the nor-
malized rates of forward extraction, rate (forward)/[M+], and S, [BH] and [H+],
for selected cases.
Although the three described cases represent rates and mechanisms that
have often been used in describing solvent extraction systems that were claimed
to occur in the absence of diffusional contributions, we have to emphasize that
many other mechanisms and rates are possible [12]. However, the three forego-
ing cases are useful in showing the type of ambiguities often faced when trying
to assign a reaction mechanism on the basis of an experimentally determined
solvent extraction rate law. These ambiguities are only partly reduced by study-
ing the extraction rate in a very broad concentration range of the reagents and
by varying S. Much additional information on the species present and on the

physicochemical properties of the system may be necessary to identify and sup-
port a mechanism. For example, an appreciable water solubility and weak ad-
sorption at the interface of the extracting reagent generally favor cases like 1.
Alternatively, highly water-insoluble and a strongly interfacially adsorbed che-
lating extractant increase the possibility that cases 2 and 3 may describe the
mechanism of extraction.
The simplified chemical system (M+, BH, B−) used here to derive the
equations of cases 1–3, although useful to shown similarities, differences, and
ambiguities that can be encountered when dealing with the rates and mecha-
nisms of solvent extraction processes, only approximately describes the very
complex and diverse chemical systems that have been reported in the literature.
With minor changes and modifications, it is not difficult to derive other equa-
tions suitable for describing many other more complex situations. For example,
with the help of case 1, the equations describing the kinetic of extraction of
divalent metal ions, such as Zn2+, Ni2+ and Co2+, by dithizone and its derivatives
can be derived. Minor modifications also have to be introduced in cases 2 and
3 to obtain the equations that describe the extraction of metal ions, such as Cu2+
and Fe3+, by hydroxyoximes; Fe3+ by hydroxamic acids; and trivalent lanthanide
and actinide cations by alkylphosphoric acids. Other chemical systems in which
neutral or negatively charged metal species are extracted by neutral extractants
or by alkylammonium salts have been described by mechanisms involving aque-
ous-phase complexation or interfacial reactions (one-step or two-step mecha-
nisms) as rate-determining steps. Examples are the extraction of UO2(NO3)2 by
tributyl phosphate (TBP), or FeCl−4 and Pu(NO3)4 by trialkylammonium salts.
The equations reported in cases 1–3 can be easily modified to adapt them to the
specific chemical features of the reagents involved [13]. In most instances, these
modifications involve writing the correct charge for the extracted species (in-
cluding zero and negative charges), eliminating the acid dissociation step of the
extractant whenever this is appropriate, including additional equilibrium equa-
tions, and additional simple rate-determining steps. However, examples of com-
pletely different mechanisms and rates, which in no way can be related to cases
1–3, are abundant in the literature.
5.8 DIFFUSIONAL REGIME

Let us now consider the systems in a diffusional regime. Given the simplified
picture of the interface described by the two-film theory (see Fig. 5.1), we can
now visualize that any species going from one phase to the other, crossing the
liquid–liquid interface, will encounter a total resistance R, which is the sum of
three separate resistances. Two of these resistances are of a diffusional nature
and depend on the fact that diffusion through the stagnant interfacial films can
be a slow process. These resistances are indicated in Fig. 5.1 as Rw and Ro. Rw

includes the diffusional contributions on the water side of the interface, assumed
to occur entirely through the stagnant film of thickness, δw. Ro includes the
contributions on the organic side, where the thickness of the stagnant film is δo.
The resistance associated with the crossing of the interface is indicated as Ri.
As a consequence, the total transfer resistance, R, is
R = Rw + Ri + Ro (5.64)
This equation holds at the steady state. In a diffusional regime and in ab-
sence of rigid interfacial films, Ri is generally negligible relative to Rw and Ro.
Consider two simple cases of extraction processes in which kinetics are
controlled by interfacial film diffusion (the solutions are always considered
stirred). The two cases are treated with the simplifying assumptions introduced
in section 2 (i.e., steady-state and linear concentration gradients throughout the

diffusional films).
Case 4: The interfacial partition between the two phases of unchanged
species is fast. The rate is controlled by the diffusion to and away from the
interface of the partitioning species. In the absence of an interfacial resistance,
the partition equilibrium of A between the aqueous and organic phase, occurring
at the interface, can be always considered as an instantaneous process. Here, A
is any species, neutral or charged, organic or inorganic. This instantaneous parti-
tion process (interfacial equilibrium) is characterized by a value of the partition
coefficient equal to that measured when the two phases are at equilibrium.
The following equations describe the extraction process:
A ↔ A(org) (5.65)
with an extraction equilibrium constant KDA, reached at t = ∞ ;
A i ↔ A i (org) (fast) (5.66)
interfacial equilibrium, holding at any time (subscript i indicates species in con-
tact with the interface); thus
K DA = [A]eq /[A]eq = [A]i /[A]i (5.67)
The symbol [ ]i indicates concentrations at the extreme limit of the diffu-
sional film (i.e., volume concentrations in the region in direct contact or very
close to the liquid-liquid interface). Because of the fast nature of the distribution
reaction, local equilibrium always holds at the interface.
The diffusional process can be treated by applying the first Fick’s diffu-
sion law to the two diffusional processes occurring on both sides of inter face,
under the simplifying assumptions of steady-state and linear concentration gra-
dients shown in Fig. 5.8. The diffusional fluxes through the aqueous, Jw and
organic, Jo, diffusion films will then be
J w = − DA ∂ [A]/ ∂x = − DA ([A]i − [A]) / δ w (5.68)

Fig. 5.8 Linear concentration profiles for the partition of A between an aqueous and
an organic phase.
J o = − DA ∂ [A]/ ∂x = − DA ([A] − [A]i ) / δo (5.69)

DA and DĀ are the diffusion coefficients of the distributing species in the
aqueous and organic layers, respectively. By considering that, at the steady state,
Jw = Jo = J and by setting
∆ w = δ w / DA , ∆ o = δo / DA (5.70)
Equations. (5.68) and (5.69) can be solved for the interfacial concentrations to
yield
[A]i = [A] − J ∆ w (5.71)
[A]i = J ∆ o + [A] (5.72)

By inserting Eqs. (5.71) and (5.72) into Eq. (5.67), solving for J, and
considering the correlation between flux and rate [see Eq. (5.5)], we obtain
rate = −d [A]/ dt = rate (forward) − rate (reverse)=
=S K DA [A]/( K DA ∆w + ∆ o ) − S[A]/( K DA ∆ w + ∆ o ) (5.73)
The diffusion-controlled, extraction kinetics of A, therefore, can be de-
scribed as a pseudo-first-order rate process with apparent rate constants
k1 = K DA [A]/( K DA ∆ w + ∆ o ) and k−1 = 1 /( K DA ∆ w + ∆ o ) (5.74)

Rydberg_5063-2_Ch05_R3_02-09-04 11:11:10
The rate equation is indistinguishable from that of an extraction process

occurring in a kinetic regime, which is controlled by a slow, interfacial partition
reaction:
k1
A ↔ A(org) (slow) (5.75)
k−1
The diffusion-controlled process [see Eqs. (5.66), (5.68), and (5.69)] can
be experimentally differentiated from the process occurring in the kinetic regime
[see Eq. (5.75)] only by measuring the variations of kl and k−1 with δw and δo .
Otherwise, the identical rate laws will not permit one to distinguish between the
two mechanisms.
Equation (5.73) can be integrated to obtain
ln ([A]o − [A]eq ) /([A] − [A]eq ) = S ( K DA + 1)t /

( K DA ∆ w + ∆ o ) = S (k1 + k−1 )t (5.76)
Equations (5.74) and (5.76) show that when KDA is very high (i.e., the
partition of A is all in favor of the organic phase):
k1 ~ DA / δ w and k−1 ~ 0
and the extraction rate is controlled only by the aqueous-phase diffusional resis-
tance. On the other hand, when KDA is very small, it is
k1 ~ 0 and k−1 ~ DA / δo
and the rate is only controlled by the organic-phase diffusional resistance.
Case 5: A fast reaction between the metal cation and the undissociated
extracting reagent occurs at the interface (or in proximity of the interface). The
rate is controlled by the diffusion to and away from the interface of the species
taking part in the reaction.
The overall extraction reaction (extraction stoichiometry) is represented
by the equation:
M + + BH(org) = MB(org) + H + (5.77)
with equilibrium extraction constant Kex chemical reaction occurring at or near
the interface is considered to be always at equilibrium, although bulk phases
reach this condition only at the end of the extraction process. The condition of
interfacial (local) equilibrium is expressed by
M i+ + BH i (org) ↔ MBi (org) + H i+ (fast) (5.78)
with interface equilibrium constant Ki, which alternatively can be written as the
sum of the two equilibria.
M i+ + Bi− (org) ↔ MBi (org), K1 = [MB]i /[M + ]i [B− ]i (5.79)

BH i (org) ↔ H i+ + Bi− (org), K 2 = [B− ]i [H + ]i /[BH]i (5.80)
with
K ex = K1K 2 = ([MB]eq [H + ]eq ) /([BH]eq [M + ]eq )
= ([MB]i [H + ]i ) /([BH]i [M + ]i ) = K i (5.81)
Following the previous line of reasoning, the rate of extraction can be
obtained by considering the diffusional fluxes through the diffusion films of M+,
H+, MB, and BH, expressed through Fick’s law, and introducing the chemical
reaction [see Eq. (5.78)] as boundary condition to the differential equations
describing diffusion through the aqueous and organic diffusion films. Under the
usual simplifying assumptions of steady-state and linear concentration gradients

and referring to Fig. 5.9, where the concentration gradients of the four species
M+, H+, BH, and MB are schematically indicated, the rate of extraction of M+
can be derived. Equations (5.82a–d) describe the diffusional fluxes of M+, H+,
BH, MB through the stagnant films:
Fig. 5.9 Linear concentration profiles for the extraction of M+ by BH(org) in presence
of an interfacial chemical reaction.

J M+ = − DM+ ∂ [M + ]/ ∂x = − DA+ ([M + ]i − [M + ]) / δ w (5.82a )
J H+ = − DH+ ∂ [H + ] / ∂x = − DH+ ([H + ]i − [H + ]) / δ w (5.82b)
J MB = − DMB∂ [MB]/ ∂x = − DAB ([MB] − [MB]i ) / δo (5.82c)
J BH = − DBH ∂ [BH]/ ∂x = − DBH ([BH] − [BH]i ) / δo (5.82d)
Furthermore, the four fluxes are not independent and, at the steady state,
are related by the condition:
J M+ = J MB = − J H+ = − J BH (5.83)
+
indicating that, although M and MB diffuse in the positive direction of the x
axis, H+ and BH diffuse in the opposite one. Moreover,
−d [M + ]/ dt V / Q = J = J M+ (5.84)
By solving Eqs. (5.82a–d) for the interfacial concentrations [ ]i after tak-
ing into account the correlation between the fluxes [see Eqs. (5.83) and (5.84)]
and introducing the simplification that the diffusion coefficients depend only on
the nature of the phase, that is,
∆ w = δ w / DM+ = δ w / DH+ (5.85a )
∆ o = δo / DMB = δo / DBH (5.85b)
we obtain
[M + ]i = [M + ] − J ∆ w (5.86a )
+ +
[H ]i = [H ] + J ∆ w (5.86b)
[MB]i = [MB] + J ∆ o (5..86c)
[BH]i = [BH] − J ∆ o (5.86d)
Equations (5.86a–d) can then be inserted into Eq. (5.81) to yield
K ex = [MB] + J ∆ o )([H + ] + J ∆ w ) /([M + ] − J ∆ w )(BH] − J ∆ o ) (5.87)
Equation (5.87) can be solved for J to derive an expression for J (and the
extraction rate) as function of the four concentration variables in the bulk
phases, the equilibrium constant of the extraction reaction, and the two diffu-
sional parameters ∆w and ∆o.
A simple analytical solution for J (and the extraction rate) can be derived
from Eq. (5.87) by neglecting the term in J2 (very low fluxes) and isolating J:
rate = rate (forward) − rate (reverse) = −d [M + ]/ dt = JS =
= {S K ex [M + ][BH]}{∆ w ([MB] + K ex [BH]) + ∆ o ([H + ] + K ex [M + ])}−1 −
− {S [MB][H + ]}{∆ w ([MB] + K ex [BH]) + ∆ o ([H + ] + K ex [M + ])}−1 (5.88)

In the limiting condition of very low metal ion concentrations, the terms con-
taining [M+] and [MB] can be neglected in the denominator of Eq. (5.88). This
means that [BH] and [H+], in practice, are constant with time, and their initial
and equilibrium concentrations are essentially the same. The rate equation then
becomes
rate = {S K ex [M + ][BH]eq }{∆ w K ex [BH]eq + ∆ o [H + ]eq }−1 −
− {S [MB][H + ]eq }{∆ w K ex [BH]eq + ∆ o [H + ]eq }−1 (5.89)
By dividing the numerators and denominators in Eq. (5.89) by [H+]eq and
setting
K ex* = K ex [BH]eq /[H + ]eq (5.90)

where K*e x is the conditional equilibrium constant of equilibrium [see Eq. (5.77)]
at [H+] and [BH] constant, Eq. (5.89) becomes identical with Eq. (5.73) when
KDA is replaced by Kex. As expected, the rate of extraction of species present at
high dilution in a system of fixed composition can be formally treated as the
simple distribution of an unchanged species between two immiscible phases.
The comparison of Eq. (5.89) with Eq. (5.63) also indicates that the rate laws
obtained for a system in a kinetic regime controlled by slow two-step interfa-
cial chemical reactions [see Eqs. (5.58) and (5.59)] and for a system in diffu-
sional regime controlled by slow film diffusion processes coupled to an in-
stantaneous reaction occurring at the interface (or in the proximity of the
interface) [see Eqs. (5.82a–d) and (5.78)], have the same functional depen-
dence on the concentration variables when low fluxes and low metal concen-
trations are involved. Alternatively, this result can be formulated by stating
that, because of diffusional contributions to the extraction rate, a fast extrac-
tion reaction occurring with a simple stoichiometry can mimic a two-step
interfacial chemical reaction occurring in absence of diffusional contributions
[14,15]. This type of ambiguity, in the past, has led solvent extraction chem-
ists (including this author) to sometimes erroneously derive extraction mech-
anisms invoking a series of two slow interfacial chemical reactions as rate
determining steps in systems for which the two phases were insufficiently
stirred. However, a distinction between the two cases can occasionally be
made if it is possible to conduct kinetic experiments in which the values of
δw and δo are varied. When the rate is independent of δw and δo, Eq. (5.63)
rather than Eq. (5.89) of this section describes the rate of extraction. Never-
theless, since the film thicknesses never seem to go to zero, the ambiguity
may not be resolved.
Film diffusion usually is slower than the rate of many ligand-substitution

reactions. Therefore, when rate laws, such as Eqs. (5.63) or (5.89), are found
for the extraction kinetics of metal species, preference should, in general, be

given to an interpretation in which the determining step of the extraction reac-
tion is film diffusion.
When the fast reactions occurring in the system have stoichiometries dif-
ferent from the simple one shown by Eq. (5.78), analytical solutions of the
diffusion equations are difficult to obtain. Nevertheless, numerical solutions can
be obtained by iterative routines, and the results are conceptually similar to
those described. The additional complications introduced by non–steady-state
diffusion and nonlinear concentration gradients can be similarly handled.
5.9 MIXED DIFFUSIONAL–KINETIC REGIME

A mixed diffusional–kinetic regime occurs when one or more of the chemical
steps of the reaction mechanism proceed with a velocity that is comparable with
that of the diffusional processes through the interfacial films. The other two
regimes are thus limiting cases of this more general one. Up to a certain limit,
the diffusional processes in an isothermal system can be accelerated by stirring,
which reduces the thickness of the diffusion films, and the rate of chemical
reactions can be increased by increasing the concentrations of the reactants. The
range of experimental conditions that determine the nature of the extraction
regime is specific for each chemical system and depends on the hydrodynamic
conditions of the apparatus used for the extraction.
The analytical description of a mixed regime is rather complicated, as all
difficulties and ambiguities associated with both the kinetic and diffusional re-
gimes may be present. To fully describe the extraction kinetics, it is necessary
to solve simultaneously the equations of diffusion and of chemical kinetics.
Furthermore, the slow chemical reactions can be of two types: homogeneous,
wherein the chemical changes occur in the entire volume of a phase or in a thin
layer, generally not too far from the interface; and heterogeneous, wherein the
chemical changes occur at the interface. The description of a diffusion regime
coupled to a homogeneous chemical reaction is particularly complicated, since
the differential equations describing the rate of formation of new chemical spe-
cies have to be solved simultaneously with the system of differential equations
describing the diffusional processes. On the other hand, when the slow chemical
reactions are heterogeneous (i.e., interfacial), the rates of the chemical reactions
appear in the boundary conditions of the differential diffusion equations.
A mathematically simple case, that occurs frequently in solvent extraction
systems, in which the extracting reagent exhibits very low water solubility and
is strongly adsorbed at the liquid interface, is illustrated. Even here, the interpre-
tation of experimental extraction kinetic data occurring in a mixed extraction
regime usually requires detailed information on the boundary conditions of the
diffusion equations (i.e., on the rate at which the chemical species appear and
disappear at the interface).

Thus, the chemical kinetics at the interface must be known from separate
experiments carried out in a pure kinetic regime. In fact, although the problem
has some similarities with that of a diffusional regime associated with a fast
interfacial chemical reaction, for which the equilibrium law was set up as an
interfacial boundary condition to the diffusion of the species, here, the further
complication exists that the interfacial chemical rate laws cannot be a priori
known and have to be themselves derived from kinetic experiments. This diffi-
culty can sometimes be avoided by assuming that the interfacial rate laws are
simply related to the reaction stoichiometry or by trying to derive information
on the interfacial reactions through fitting, to experimental data obtained in a
mixed regime, analytical solutions of differential equations that take into ac-
count both diffusion and chemical reactions. Unfortunately, both of these proce-
dures can lead to erroneous interpretations, as in only a few cases can the rate
laws be correctly derived from stoichiometric considerations, and when too
many variables (e.g., rate constants, rate laws, and diffusional parameters) are
simultaneously adjusted to fit experimental data, many alternative models can
usually satisfy the same equations. Therefore, it is important that the boundary
condition of the diffusion equations (i.e., the interfacial rate laws) be derived
by separate, suitable experiments.
We now treat, in some detail, under the usual assumptions of steady-state
diffusion and linear concentration gradients through the diffusion films, only
one simple case of mixed regime. The case deals with a situation that is an
extension of case 4 described in section 5.8. Here the additional complication
of a slow interfacial reaction is introduced.
Case 6: The interfacial partition reaction between the two phases of un-
charged species is slow. The rate is controlled both by the slow partition reaction
and by diffusion to and away from the interface of the partitioning species.
Although the extraction process is still described by Eq. (5.65), the slow
interfacial chemical reaction and the corresponding interfacial flux, Ji, are
k1
A i ↔ A i (org) (5.91)
k−1
J i = k1[A]i − k−1 [A]i (5.92)

Following the same derivation as for case 4 of section 5.8 and referring
again to Fig. (5.8) (here, however, the concentrations [A]i and [A]i are at equi-
librium only at the end of the extraction process), we obtain for the interfacial
concentrations, equations identical with Eqs. (5.71) and (5.72). By substituting
them into Eq. (5.92), keeping into account that at the steady state
Ji = J w = Jo = J (5.93)
and solving for J, we obtain

rate = −d [A]/ dt = rate (forward) − rate (reverse) =
= S k1[A](k1∆ w + k−1 ∆ o + 1) −1 − S k−1[A](k1 ∆ w + k−1∆ o + 1) −1 (5.94)
A comparison between Eq. (5.94) and Eq. (5.73) indicates that the pres-
ence of a slow interfacial chemical reaction shows up as an additional term in
the denominator of the rate laws. Nevertheless, the same functional dependence
on S, [A], and [A] is exhibited by Eq. (5.73) and Eq. (5.94). This means that
the rate law does not allow choice between partition kinetics controlled only by
diffusion, or only by a slow partition reaction, or by a combination of the two.
When the rate constants are large numbers (fast reactions), both the numera-
tors and the denominators of Eq. (5.94) can be divided by k−1. Since k1/k−1 =
KDA and (k−1)−1 can be neglected relative to the other terms in the sum (KDA ∆w +
∆o + k−1), an equation identical with Eq. (5.73) is obtained for which only dif-
fusion controls the partition rate. At the other extreme, when interfacial film
diffusion is very fast, and the reactions are very slow, we can set 1 Ⰷ (k1∆w +
k−l ∆o) and the rate equation becomes equal to
rate = S k1 [A] − S k−1 [A] (5.95)
that is, the partition rate is controlled only by the slow chemical reaction.
5.10 EXPERIMENTAL TECHNIQUES

Several experimental techniques are available for measuring the kinetics of sol-
vent extraction. They differ in the different efficiency of stirring which can be
achieved in the two phases, and in the control of the hydrodynamic conditions
in proximity to the interface and in the interfacial area. Most of these techniques
can be grouped into five categories.
1. Highly stirred vessels: Those techniques in which the two phases are so
highly stirred that droplets of one phase are dispersed in a continuum of the
second phase (the dispersing phase).
2. Constant interfacial-area-stirred cells: Those techniques in which the two
phases can be stirred to a variable extent, while a known and constant inter-
facial area is maintained.
3. Rotating diffusion cells: These techniques bring the two phases in contact
at the surface of a rotating membrane filter having well characterized hydro-
dynamics.
4. Moving drops: These techniques measure the extraction that takes place
when a single droplet of one phase travels along a vertical tube filled with
the second phase.
5. Short-time phase contact methods: These techniques do not stir the two
phases, and the contact time between them is very short.

5.10.1 Highly Stirred Vessels
Different versions of highly stirred vessels are used in kinetic studies. The com-
mon feature is that samples are removed from the extraction apparatus at dif-
ferent times to measure concentrations. The extraction occurs under con-
ditions similar to those that are met in practical technological or analytical
extractions.
Figure 5.10 shows an advanced version of this technique [16]. The appara-
tus incorporates a continuous monitoring of the rate and permits instantaneous
data analyses that are accomplished by the use of a Teflon phase separator and
an online microcomputer. The device allows measurement of extraction rates
with half-lives as short as 10 s.
When the two phases are so highly stirred that one phase is dispersed into
the other in the form of minute droplets, it is generally difficult to control the
interfacial area through which the extraction takes place. This constitutes the
major drawback to the use of this technique to derive unequivocal information
Fig. 5.10 Computer-assisted extraction kinetics-measuring apparatus for highly stirred

phases: (A) high-speed stirrer; (B) stirrer shaft; (C) sample inlet; (D) Teflon stirring bar;
(E) Teflon phase separator; (F) water bath; (G) flow-cell; (H) spectrophotometer; (I)
peristaltic pump; (J) chart recorder; (K) A/D converter; (L) clock; (M) minicomputer;
(N) dual-floppy disk drive; (O) printer, (P) plotter. (From Ref. 16.)

on the extraction regime. Typical drop diameters are 200 µm, that is, about
10–20 times larger than the claimed minimum thickness of the diffusion films
usually reached in well-stirred biphasic systems.
5.10.2 Constant Interfacial-Area-Stirred Cells

In this technique, the two phases are stirred while the aqueous and the organic
layers are positioned one on top of the other and the transfer occurs through the
contact area [17–19]. The interfacial area, which coincides with the geometric
horizontal section at which the contact between the two phases is established,
is always kept constant, whereas the degree of stirring of the two phases is
varied over the widest possible range. The major limitations of the technique is
that the motion of the liquids in proximity of the interface and the interfacial
turbulence are not well defined and vary with the equipment. Consequently,
poorly defined diffusion films result. Nevertheless, experience has shown that
by using well-designed equipment, reproducible data can be obtained, and film
diffusion can be treated quantitatively by using the two-film theory. With these
cells, ultimate film thicknesses of about 10−3 cm can be reached. Figure 5.11
shows a version of a constant interfacial-area-stirred cell.
Several different versions of extraction cells are used. The main difference
between the various cells lies in the presence or absence of internal baffles that
modify the internal forced convection, in the shape and dimensions of the stir-
rers that drive the forced convection, and in the presence of grids (or screens)
in proximity to the interface that help to stabilize the interface when the two
Fig. 5.11 Constant interfacial–area-stirred cell (“Lewis-type cell”) for liquid–liquid

extraction kinetic measurements.

phases are stirred. In presence of these baffles and grids, the range of variation
of the stirring speed of the phases can be quite wide. The grids (or screens)
transform most of the translation kinetic energy of the liquids into turbulent
energy, preventing the rippling of the interface. The cells allow an independent
change of the revolution number of the stirrers in the two phases. In this way,
the different influence of the density and viscosity of the two phases can be
taken into account, and the organic and aqueous phase can be stirred in such a
way to have similar fluid dynamic conditions.
5.10.3 The Rotating Diffusion Cell

In this cell, the aqueous and the organic phase are brought in contact at the
surface of a microporous membrane filter, the pores of which are filled with the
organic phase [20]. The microporous filter is attached to a hollow cylinder filled
with the same organic phase. The cylinder is dipped into the aqueous phase and
is rotated by a motor. A scheme of the rotating diffusion cell is presented
in Fig. 5.12. The cell has a filter with well-defined hydrodynamics on both
Fig. 5.12 Cross section of the rotating diffusion cell: B, bearing; BA, internal baffle;
FM, filter mount; I, inner compartment; L, lid with holes; M, membrane; MA, mounting
rod; O, outer compartment; P, pulley block; S, hollow rod; T, thermostated beaker. (From
Ref. 20.)

sides. This allows description of the diffusional processes of the reactants and
products of the extraction reactions on the aqueous and the organic side of the
interface.
The basic assumption is that the rotating filter creates a laminar flow field that
can be completely described mathematically. The thickness of the diffusion
boundary layer (δ) is calculated as a function of the rotational speed (ω), vis-
cosity, density, and diffusion coefficient (D). The thickness is expressed by the
Levich equation, originally derived for electrochemical reactions occurring at
a rotating disk electrode:
δ = 0.643 ω−1 / 2 D1 / 3ν1 / 6 (5.96)
ν is the kinematic viscosity. The organic phase inside the membrane is con-
sidered at rest, and the membrane thickness and porosity control the diffusion
rate. The major drawbacks of this technique are difficulties in well characteriz-
ing the thickness and porosity of the membrane filter, in utilizing really thin
filters (less than 10−3 cm), and that, although the hydrodynamics inside the
rotating cylinder can be assimilated to those of a rotating disk, no studies have
conclusively shown that the flows inside and outside the cylinder are the same.
A more recent version of this cell [21] makes use of a rotating capillary,
containing a gelified aqueous phase. This eliminates the microporous membrane
filter.
5.10.4 Moving Drops

In the moving drop technique (also described in Chapters 7 and 9), a drop of
the organic or aqueous phase is produced at the end of a vertical column filled
with the other phase. The drop travels along the tube, during which extraction
occurs across the drop surface. By measuring the time of traveling, the drop
size, and from the volume of collected drops, it is possible to evaluate the rate
of extraction (see Chapter 9 for a detailed discussion of drop behavior and mass
transfer).
The technique can be applied to drops that move both upward and downward,
the choice depending on the higher or lower density of the drops relative to
the continuum liquid that fills the column. Although the technique allows one
to control both drop size and interfacial area, there are some disadvantages that
impose the use of great care both in the treatment of the experimental data and
in the design of the apparatus. When the extraction is very rapid, much of the
extraction can take place during drop formation. Further extraction can also
take place in the stagnant pool where the drops are collected. On the other
hand, when the extraction is very slow, very long columns are required to
assure sufficient drop residence times. The hydrodynamic control of the sin-
gle drops can also be very difficult because of lack of internal circulation and
the presence of drop oscillations. When the drops are quite small and the

extractant adsorbs on the drop surface, the internal circulation can be absent
or strongly reduced, and it is difficult to establish whether the degree of
turbulence of the two phases in the moving drop method can ever be enough
to ensure the measurement of a transfer rate in a kinetic regime or, on the
opposite side, is so low that the extraction rate is always occurring in a diffu-
sional regime.
5.10.5 Short-Time Phase-Contact Method

This method, which has been developed and extensively utilized in Russia
[22,23], deals with nonstirred liquid phases and is based on measurements of
the electrical conductivity of the aqueous phase and of its changes during extrac-
tion. In the experimental device, the two phases are initially kept separated
inside a thermostated cell and then instantaneously contacted through a known
contact area. At this moment, the concentration variation in one phase is re-
corded as a function of time, generally using conductometry.
The aqueous phase is usually contained in a metal capillary. The capillary is
pushed into the lower part of the cell, containing the organic phase, at time
zero of the extraction. The method examines the reaction during very short
time intervals (from 0.05–1 s) and is relatively simple, as the two phases are
not stirred. The consumption of the reactants is very small (about 20 mm3 per
experiment) and permits one to make many determinations in a relatively short
time (about 10 kinetic curves per hour). Unfortunately, the method proves inap-
plicable for concentrated aqueous solutions and when many species are in-
volved in the extraction reaction.
This technique is based on a non–steady-state approach to diffusion. The
non–steady-state diffusion is not treated in our simplified approach to extraction
kinetics; the interested reader is referred to Yagodin and Tarasov [24] and Tara-
sov and Yagodin [25].
SYMBOLS
Parameter Denoted by First shown, in text order
Aqueous phase No symbol, or index w General

Organic phase Symbols org, index o (not to be con- Eq. (5.27)
fused with zero 0) or a bar above
the variable, constant, etc.
J Diffusion flow, mass cm−2 s−1 Eq. (5.1)
D Diffusion coefficient, cm−2 s−1 Eq. (5.1)
c Concentration, mol cm3 Eq. (5.1), Fig. 5.2
x Diffusion coordinate, cm Eq. (5.1)

∂ Derivative Eq. (5.1)

d Also used for derivative Eq. (5.3)
δ Thickness of organic δo or aqueous Eq. (5.3), Figs. 5.1 and 5.2
film δw at interface
∆ Diffusional parameter, =δ/D, cm−1 s Eq. (5.4)
R Film resistance, transfer resistance, for Fig. 5.1, Eq. (5.64)
indexed phase
V Volume of indexed reservoir Eq. (5.5), Fig. 5.2
Q Area of diffusion film, cm2 Eq. (5.5)
γ Interfacial tension, dynes cm−1 Section 5.5
γ2
Langmuir adsorption constant Eq. (5.40)

Π Inerfacial pressure, dynes cm−1 Eq. (5.26)
k Boltzmann constant Eq. (5.26)
T Temperature, kelvin Eq. (5.26)
t Time, seconds, s Eq. (5.5)
index 0 at beginning of experiment Eq. (5.36)
index eq at equilibrium Eq. (5.28), (5.36)
X, Y, Z, W Abbreviations for chemical com- Eq. (5.10)
pounds
k Rate constant; k−1 rate constant of re- Eq. (5.13), (5.29)
verse reaction
M+ Metal cation (central atom in complex Eq. (5.22), (5.27)
MB, or ML)
L− Ligand anion (ligand in complex ML) Eq. (5.22)
BH Extractant Eq. (5.27), (5.32)
B− (Organic) anion (ligand in complex Eq. (5.29)
MB)
K Equilibrium constant (index D for dis- Eq. (5.18)
tribution)
index ass for association Eq. (5.23)
index ex for (overall) extraction Eq. (5.28)
index DB, distribution constant of Eq. (5.31)
extractant BH
index a, acid dissociation constant Eq. (5.32)
of HB
index DM distribution constant of Eq. (5.33)
complex MB
index MB complex formation Chapter 3
constant of MB
ad Index for species adsorbed at the inter- Eq. (5.37)
face
A Any solute distributing between two Eq. (5.65)
phases

α2 Langmuir adsorption coefficient Eq. (5.40)

S Specific interfacial area, cm−1 Eq. (5.38)
REFERENCES
1. Whitman, W. G. Chem. Metal Eng., 1923, 29(4):146.
2. Whitman, W. G.; Davies, D. S. Ind. Eng. Chem., 1924, 16:1233.
3. Bird, R. B.; Stewart, W. E.; Lightfoot, E. N. Transport Phenomena, John Wiley
and Sons, New York, 1960.
4. Cussler, E. L. Diffusion, Mass Transfer in Fluid Systems, Cambridge University

Press, Cambridge, 1984.
5. Astarita, G. Mass Transfer with Chemical Reaction, Elsevier, Amsterdam, 1967.
6. Frank-Kamenetskii, D. A. Diffusion and Heat Transfer in Chemical Kinetics, Ple-
num Press, New York, 1969.
7. Frost, A. A.; Pearson, R. G. Kinetics and Mechnanism, Wiley-Interscience, New
York, 1956.
8. Moelwin-Hughes, E. A. The Kinetics of Reactions in Solution, Oxford University
Press, London, 1950.
9. Eigen, M.; Wilkins, R. W. Mechanisms of Inorganic Reactions, Vol. 49, Adv.
Chem. Ser., American Chemical Society, Washington, DC, 1965.
10. Adamson, A. W. Physical Chemistry of Surfaces, John Wiley and Sons, New York,
1982.
11. Davies, J. T.; Rideal, E. K. Interfacial Phenomena, Academic Press, New York,
1961.
12. Danesi, P. R.; Chiarizia, R.; Vandergrift, G. F. J. Phys. Chem., 1980, 84:3455.
13. Danesi, P. R.; Chiarizia, R. The Kinetics of Metal Solvent Extraction, Crit. Rev.
Anal. Chem., 1980, 10:1.
14. Danesi, P. R. Solv. Extr. Ion Exch. 1984, 2:29.
15. Danesi, P. R.; Vandegrift, G. E.; Horwitz, P.; Chiarizia, R. J. Phys. Chem., 1980,
84:3582.
16. Watarai, H.; Cunningham, L.; Freiser, H. Anal. Chem., 1982, 54:2390.
17. Lewis, J. B. Chem. Eng. Sci., 1954, 3:218.
18. Nitsch, W.; Kahn, G. Ger. Chem. Eng. Engl. Trans., 1980, 3:96.
19. Danesi, P. R.; Cianetti, C.; Horwitz, E. P.; Diamond, H. Sep. Sci. Technol. 1982,
17.961.
20. Albery, W. J.; Burke, J. E; Lefier, E. B.; Hadgraft, J. J. Chem. Soc. Faraday Trans.,
1976, 72:1618.
21. Simonin, J. P.; Musikas, C.; Soualhia, E.; Turq, P. J. Chim. Phys., 1987, 84:525.
22. Tarasov, V. V.; Yagodin, G. A.; Kizim, N. F. Dokl. Akad. Nauk. SSR, 1971, 45:
2517.
23. Tarasov, V. V.; Yagodin, G. A. VINITI, Khim. Teknol. Neorganich. Khim., 1974, 4.
24. Yagodin, G. A.; Tarasov, V. V. Proc. Int. Solv. Extr. Conf., ISEC 71, Vol. 1,

Society of Chemical Industry, (Gregory, J. G.; Evans, B; Weston, P.C. Eds.), Lon-
don, p. 888, 1971.
25. Tarasov, V. V.; Yagodin, G. A. Interfacial Phenomena in Solvent E. Ytraction. In
Ion Exchange Solvent Extraction, Vol. 10, Ch. 4 (Marinsky, J. A. and Marcus, Y.
Eds.), Marcel Dekker, New York, pp. 141–237, 1988.
26. Eigen, M. Pure Appl. Chem., 1963, 6:97.

ch5 PDF

Uploaded by

Copyright:

Available Formats

ch5 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ch5 PDF

Uploaded by

Copyright:

Available Formats

5

Solvent Extraction Kinetics

5.1 GENERAL PRINCIPLES

5.1.1 Rate-Controlling Role

Copyright © 2004 by Taylor & Francis Group, LLC

based largely on indirect evidence and speculations.

5.1.2 Rate-Controlling Role

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

5.3 RATE LAWS AND MECHANISMS

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

Rate expression −d[Xl/dt = k1[X] − k2[Y] (5.15b)

[Y] = 0 and [X] = [X]0

Copyright © 2004 by Taylor & Francis Group, LLC

[Y] = [X]0e − k1 t k1 / k2 (5.20)

5.4 SOME KINETICS AND MECHANISMS

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

5.4.3 Water Exchange and Complex Formation

Copyright © 2004 by Taylor & Francis Group, LLC

The rate law corresponding to such a mechanism is:

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

5.4.4 Ligand Displacement Reactions

5.5 THE LIQUID–LIQUID INTERFACE

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

surface activity of an extractant can sometimes lend support to extraction mech-

Copyright © 2004 by Taylor & Francis Group, LLC

It is interesting to observe that alkylammonium salts, alkylarylsulfonic

Copyright © 2004 by Taylor & Francis Group, LLC

seventh power of the intermolecular distances, so molecules experience essen-

Copyright © 2004 by Taylor & Francis Group, LLC

5.6 RATE-CONTROLLING EXTRACTION REGIMES

Copyright © 2004 by Taylor & Francis Group, LLC

5.6.2 Identification of the Extraction Regimes

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

5.7 KINETIC REGIME

Copyright © 2004 by Taylor & Francis Group, LLC

of a monovalent metal cation M+ (solvation water molecules are omitted in the

with a reaction rate expression:

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

In Eq. (5.40), α2 and γ2 are Langmuir adsorption constants. Their values

and one of complete interface saturation with extractant molecules, where 1 Ⰷ

Copyright © 2004 by Taylor & Francis Group, LLC

Copyright © 2004 by Taylor & Francis Group, LLC

for [MB]ad found in this way is

Copyright © 2004 by Taylor & Francis Group, LLC

MB(ad) + BH(org) ↔ BH(ad) + MB(org) (slow) (5.59)

rate 1 = k3* S [M + ][BH]ad − k−*3 S [MB]ad [H + ] (5.60)

rate 2 = k4* S [MB]ad [BH] − k−*4 S [BH]ad [MB] (5.61)