Kinetic analysis of aqueous-phase Pd-catalyzed, Cu-free direct arylation of terminal alkynes using a hydrophilic ligand

Article abstract of DOI:10.1021/op4001274

The engineering of reaction mixtures that ensure the solubility of inorganic salt byproducts without compromising the reactivity is a grand challenge for continuous flow manufacturing in upstream pharmaceuticals and fine chemicals process development. Aqueous cross-coupling reactions are possible solutions. We report the study of an aqueous phase Pd-catalyzed Cu-free direct arylation of an alkyne using a hydrophilic ligand towards understanding the role of water on the cross-coupling kinetics. Kinetic analyses of theoretically estimated molar flux rates and the measured reaction kinetics reveal a transition from mass transfer to kinetically controlled deprotonation and carbopalladation mechanisms. Interfacial contact areas of immiscible aqueous-organic phases control the crossover from the mass-transfer-limited to the reaction-rate-limited regime. Highly agitated batch reactors and multiphase capillary flow reactors are needed to overcome the mass transport limitations and, thus, discover the transition from the apparent reaction kinetics to the true reaction kinetics. Comparison of previously reported Density Functional Theory calculations with experimentally measured activation energies, ranging from 14.8 to 20.0 kcal/mol, elucidates the theoretical possibility of designing the aqueous phase C-C cross-coupling reaction with similar reactivity as purely organic reactions. Although ambiguity remains concerning which reaction step is rate-determining in either the deprotonation or the carbopalladation mechanism, our discovery undergirds that one mechanism or the other could dominate in aqueous designed direct arylations.

Full text of DOI:10.1021/op4001274

Article
pubs.acs.org/OPRD
Kinetic Analysis of Aqueous-Phase Pd-Catalyzed, Cu-Free Direct
Arylation of Terminal Alkynes Using a Hydrophilic Ligand
Ria C. Domier,†,§ Jane N. Moore,‡,§ Kevin H. Shaughnessy,*,‡ and Ryan L. Hartman*,†
†Department of Chemical and Biological Engineering, The University of Alabama, Box 870203, Tuscaloosa, Alabama 35401, United
States
‡Department of Chemistry, The University of Alabama, Box 870336, Tuscaloosa, Alabama 35401, United States
S
* Supporting Information
ABSTRACT: The engineering of reaction mixtures that ensure the solubility of inorganic salt byproducts without compromising
the reactivity is a grand challenge for continuous flow manufacturing in upstream pharmaceuticals and fine chemicals process
development. Aqueous cross-coupling reactions are possible solutions. We report the study of an aqueous phase Pd-catalyzed Cu-
free direct arylation of an alkyne using a hydrophilic ligand towards understanding the role of water on the cross-coupling
kinetics. Kinetic analyses of theoretically estimated molar flux rates and the measured reaction kinetics reveal a transition from
mass transfer to kinetically controlled deprotonation and carbopalladation mechanisms. Interfacial contact areas of immiscible
aqueous−organic phases control the crossover from the mass-transfer-limited to the reaction-rate-limited regime. Highly agitated
batch reactors and multiphase capillary flow reactors are needed to overcome the mass transport limitations and, thus, discover
the transition from the apparent reaction kinetics to the true reaction kinetics. Comparison of previously reported Density
Functional Theory calculations with experimentally measured activation energies, ranging from 14.8 to 20.0 kcal/mol, elucidates
the theoretical possibility of designing the aqueous phase C−C cross-coupling reaction with similar reactivity as purely organic
reactions. Although ambiguity remains concerning which reaction step is rate-determining in either the deprotonation or the
carbopalladation mechanism, our discovery undergirds that one mechanism or the other could dominate in aqueous designed
direct arylations.
chemicals.7 Significant advances have been made in the design
of ligand-promoted, palladium-catalyzed cross-coupling reac-
INTRODUCTION
■
Water as a reaction solvent is potentially a transformative
approach to greener and safer synthetic chemistry,1 both in the
tions that provide high activities with a range of substrates
including inexpensive, widely available, but challenging, organic
laboratory and in production scenarios. Water is an attractive
solvent because it is cheap, nonflammable, and environmentally
chloride and sulfonate substrates.8 Despite the advances, large-
scale application of these methods is still limited by undesirably
benign. In addition, the use of water-soluble catalysts in
low TON or TOF values. Separation of the homogeneous
aqueous-biphasic systems presents the opportunity for catalyst
recovery.1c Catalyst recovery and its reuse are important
palladium catalyst from the organic product stream, particularly
to the levels required for pharmaceutical applications, also
because of the cost and the scarcity of the most catalytically
useful metals (Pd, Pt, Rh, Ru). Residual trace metal impurities
are stringently regulated in pharmaceuticals. Separation of
homogeneous catalysts from organic product streams can prove
challenging.2 Water can, in some cases, provide new or
improved reactivity in catalytic processes due to its unique
solvent properties.3 Furthermore, the precipitation of insoluble
inorganic salt byproducts, problematic in continuous flow
reactions in organic solvents,4 could potentially be dealt with
using water as a reaction solvent.
Palladium-catalyzed cross-coupling reactions of organic
halides with organoboron compounds (Suzuki coupling),
organostannanes (Stille coupling), alkenes (Heck coupling),
and alkynes (Sonogashira coupling), as well as the many
variations on these reactions, are indispensible methods to
construct carbon−carbon bonds.5 The generality and functional
group tolerance of these reactions allow them be utilized in the
synthesis of complex natural product targets.6 Their operational
simplicity and wide scope have led to their adoption in the fine
chemical industry in both development- and process-level
synthesis of active pharmaceutical ingredients and other fine
presents a major challenge.2
Palladium-catalyzed cross-coupling in aqueous-biphasic sol-
vents has a long history dating back to the initial development
of the Suzuki coupling of aryl halides and organoboronic acids.9
Casalnuovo pioneered the use of water-soluble phosphine
ligands to constrain the palladium catalyst in the aqueous phase
of a water/organic reaction mixture.10 Since this initial report,
catalysts derived from a wide range of water-soluble ligands
have been used to promote palladium catalyzed couplings in
water/organic solvent mixtures.1c By using a water-soluble
catalyst species, the catalyst containing an aqueous phase can be
readily separated from the hydrophobic product phase.
Furthermore, the catalyst solution can potentially be recycled
for use in subsequent reaction cycles. In addition to the
Special Issue: Engineering Contributions to Chemical Process
Development
Received: May 21, 2013
Published: September 3, 2013
© 2013 American Chemical Society
1262
dx.doi.org/10.1021/op4001274 | Org. Process Res. Dev. 2013, 17, 1262−1271

Organic Process Research & Development
Article
separation advantages offered by water, the unique properties of
water compared to organic solvents can result in enhancements
in reaction rates or changes in reactivity trends. Breslow first
noted the ability of water to accelerate reactions of hydrophobic
molecules.11 Recently, there has been a resurgence of interest in
the role of water in accelerating organic reactions.3,12
REACTION MECHANISM AND KINETICS
■
The Pd-catalyzed copper-free direct arylation of phenyl-
acetylene with 4-bromobenzotrifluoride is shown in Scheme
1. Water-soluble alkylphosphines, such as DTBPPS (3-(di-(tert-
butyl)phosphonium)propylsulfonate), are effective ligands in
the aqueous-phase Pd-catalyzed Suzuki, Sonogashira, and Heck
coupling reactions.5e−k,m,v,21 Sterically demanding and strongly
electron donating ligands have been shown to provide the most
effective catalysts for a wide range of Pd- and Rh-catalyzed
cross-coupling reactions. In addition, they have similar steric
and electronic properties to the ligands commonly used in the
direct arylation of C−H bonds.
The design of aqueous phase catalyzed cross-couplings with
process design in mind motivates a closer look at the laboratory
techniques commonly employed for their discovery. Classical
batch syntheses mediated by hydrophilic ligands rely on
efficiency in heterogeneous reaction mixtures where molecular
diffusion potentially influences the reaction rates. The extent of
mass transfer limitations must be considered to properly
evaluate intrinsic kinetics when reactions occur at liquid−liquid
interfaces. It is readily known that molecular flux to and from
bulk fluids in gas−liquid and liquid−liquid systems influences
the reaction kinetics with the interfacial contact areas being
principally important.13 The same principles hold true for batch
and continuous flow reactors with the characteristic diffusion
length scales being of utmost importance. Extensive knowledge
is available on mesoscale to full production batch mixing
principles,14 and numerous elegant works outline the mixing
principles in continuous laminar flow reactors.14h,15 In either
case, properly assessing the role of mass transport limitations
on the experimentally measured kinetics requires analyses of
reaction time scales relative to molecular diffusion time scales,16
which is made possible through first principle calculations. The
existence of mass transport limitations does not necessarily
constrain synthetic methodology discovery. However, process
chemistry requires intrinsic knowledge to accurately scale-up
from the laboratory to production.17 Flow chemistry has been a
topic of intense research4b,c,5u,17c,f,18 for these reasons and many
other benefits related to green chemistry and engineering.19
Aqueous derived synthetic methodologies1,5j,18j,20 have the
potential to expand the scope of continuous process
intensification. However, a first principle understanding of
laboratory-scale mass transport phenomena must accompany
the synthesis techniques to ensure the accurate estimation of
kinetics.
Scheme 1. Pd-catalyzed copper-free cross-coupling between
terminal alkyne and 4-bromobenzotrifluoride using water-
soluble DTBPPS
Two different mechanisms, deprotonation and carbopallada-
tion,5s,22 have been proposed for the palladium-catalyzed
copper-free cross-coupling between a terminal alkyne and an
aryl halide. The two catalytic cycles, shown in Figures 1a and b,
both undergo the reversible oxidative addition of the aryl halide
to the catalytically active Pd0LX species 1. Large concentrations
of aryl halide and alkyne relative to other free species drive
either catalytic cycle in the forward direction, which establishes
reaction conditions far from equilibrium. The resulting
In the present work, we examine the palladium-catalyzed
synthesis of diarylacetylenes using hydrophilic ligand with the
goal of evaluating the influence of mass transfer on the kinetics.
Molar flux and interfacial contact areas potentially control the
kinetics in multiphase aqueous−organic reactions instead of
intrinsic mechanisms. Capillary flow reactors are applied to
estimate the molar flux rates in multiphase classical batch
reactors, and in turn the reaction space necessary for our
laboratory-scale kinetics discovery is identified. We report that
highly agitated batch reactors and multiphase capillary flow
reactors provide sufficient interfacial contact areas to identify
the crossover from the mass-transfer-limited to the reaction-
rate-limited regime (i.e., the transition from the apparent
reaction kinetics to the true kinetics). The subsequent
comparison of mass transfer limited models to intrinsic kinetic
models, derived from first principles and based on the
deprotonation and the carbopalladation catalytic cycles, reveals
the transformative approach of engineering water as a reaction
solvent in fine chemicals synthesis and manufacturing. Our
study using capillary flow and batch reactors establishes the
groundwork for the innovation of reactors and hydrophilic
ligands for aqueous phase catalyzed cross-couplings.
Figure 1. Proposed deprotonation (a) and carbopalladation (b)
mechanisms for the palladium-catalyzed copper-free cross-coupling
between a terminal alkyne and an aryl halide.
1263
dx.doi.org/10.1021/op4001274 | Org. Process Res. Dev. 2013, 17, 1262−1271

Organic Process Research & Development
Article
oxidative addition adduct, species 2, then undergoes reversible
coordination to the alkyne, forming alkyne−PdII complex 3. In
the deprotonation mechanism, the base then removes a proton
from the alkyne−PdII complex 3, and as a result forms the
vinyl−PdII complex 4. The cross-coupled product R−Ar then
undergoes dissociation, restoring the catalytically active Pd0LX
species 1. Reversible cationic and anionic alternatives for the
deprotonation mechanism have been proposed. In the
carbopalladation mechanism, however, the intramolecular
rearrangement (5) of the alkyne−PdII complex 3 takes place
through migratory insertion of the aryl group Ar to the terminal
alkyne. The product R−Ar is then expelled from the vinyl−PdII
complex, complex 4′, via reductive elimination to restore the
catalytically active Pd0LX species 1. The concentration of the
base in either mechanism will have a larger impact on reductive
elimination when the base is strong. Cationic and ionic versions
of the deprotonation mechanism rely on the solution ionic
strength, which makes aqueous reaction mixtures interestingly
complex. The mechanisms postulated in Figure 1 further
support that base-to-ligand interactions could influence the
catalytic cycles, which motivates an even deeper molecular
understanding of hydrophilic ligand design.
liquid−liquid segmented flow was established, and the reactions
were quenched with 1.0 M HCl to ensure accurate analyses
offline with gas chromatography. In batch, samples were
acquired using syringes and immediately diluted with ethyl
acetate followed by their analyses with gas chromatography.
Batch experiments at 50, 70, and 80 °C were conducted over a
range of stirrer rates (from 176 to 1100 rpm). Analyses of
aliquots taken at 1 to 5 min intervals over 30 min in batch
reactors and aliquots collected for 1, 3, and 5 min residence
times yielded the substrate concentrations as a function of the
reaction time (i.e., molar flow rates of less than 25.0 μmol/
min).
Monitoring the substrate concentrations as functions of the
reaction times yielded interesting kinetic observations. One
observes in Figures 3, S1a, and S1b that increasing the reaction
Although significant contributions have been published, there
remains ambiguity in the literature towards understanding
which of the two catalytic cycles controls the copper-free
Sonogashira reaction.5s,22 Both theoretical and experimental
works summarize the mechanistic studies on the deprotonation
and the carbopalladation mechanisms.5s,22 Density Functional
Theory (DFT) calculations by Garcia-Melchor et al.22a
conclude a high-energy barrier for the carbopalladation
mechanism, indicating that the deprotonation mechanism
operates under the reaction space they investigated. However,
the mechanistic complexity necessitates understanding specific
reaction scenarios, as the multiple reaction pathways exhibit
competitive rates as functions of the solvent, ligands, substrates,
base, and, as our results will demonstrate, the reactor design.
Figure 3. Aryl bromide concentrations as functions of their reaction
times at 50, 70, and 80 °C and 1100 rpm. Aryl alkyne (Figure S1a) and
base (Figure S1b) concentrations are available in the Supporting
Information.
temperatures from 50 to 80 °C at 1100 rpm increased the rates
of disappearance of the aryl bromide, alkyne, and the base.
However, Figures 4, S2a, and S2b illustrate that the stirring rate
and the nature of the reactor (batch or continuous) each have a
RESULTS AND DISCUSSION
■
Diaryl Alkyne Synthesis. Direct arylations of a terminal
alkyne were performed in batch reactors (4.0 mL vials) and in
0.1 mL continuous flow capillary reactors (PFA tubing 1/16″
O.D., 1.0 mm I.D.). In either case, a heated oil bath was used to
establish isothermal reaction conditions. Batch reactors were
agitated via magnetic stirrers (dimp/dflask = 0.77) whereas a
Harvard Apparatus PHD2000 syringe pump was used to inject
reagents as shown schematically in Figure 2. In flow, multiphase
Figure 4. Aryl bromide concentrations as functions of their reaction
times (at 80 °C) in 4.0 mL vials mixed at 176, 522, and 1100 rpm. The
concentrations exiting a capillary flow reactor (at 80 °C) were
measured at residence times, τ, of 1, 3, and 5 min. Aryl alkyne (Figure
S2a) and base (Figure S2b) concentrations are available in the
Supporting Information.
Figure 2. Continuous flow capillary reactor apparatus consisting of 1.0
mm I.D. PFA tubing (0.1 mL). Batch reactor experiments were
performed by replacing the capillary reactor system with a 4.0 mL vial
sealed via septum and agitated with magnetic stirrers.
1264
dx.doi.org/10.1021/op4001274 | Org. Process Res. Dev. 2013, 17, 1262−1271

Organic Process Research & Development
Article
profound impact on the reaction. Batch experiments (at 80 °C)
carried out at 176, 522, and 1100 rpm demonstrate that
increasing the stirrer rate increases the rate of disappearance of
the substrate. Stirrer rates of 1100 rpm represent the maximum
agitation achievable using a conventional magnetic stirrer plate.
The general observation is a preliminary indication of
inadequate shear between the immiscible aqueous and organic
fluids (i.e., the existence of mixing limitations) at lower stirring
rates of 176 and 522 rpm. The mixing of two immiscible liquids
in batch is complex with both convective and diffusive terms
being challenging to characterize. Low Reynolds number (Re <
10) flow, however, is well characterized, and it offers a
ubiquitous reactor platform to evaluate either mass transfer
limited or reaction rate limited synthesis.15h,i,16,23 Examination
of the substrate concentrations measured in the capillary flow
reactor demonstrates results comparable to batch experiments
at 1100 rpm, as shown in Figures 4, S2a, and S2b. Table 1
Table 1. Experimentally calculated diaryl alkyne yields for
batch and flow experiments
reaction time
(min)
diaryl alkyne yield
std.
b
a
b
entry
reactor T (°C)
(%)
dev.
1
1.0
5.0
10
batch
batch
batch
batch
batch
batch
batch
batch
batch
flow
50
50
50
70
70
70
80
80
80
80
80
80
7
37
64
13
70
87
12
81
90
5.5
17
5
7
2
3
10
2
4
1.0
5.0
10
5
10
4
6
7
1.0
5.0
10
6
8
10
2
9
Figure 5. Linearized aryl bromide mass transfer rates, given by
expressions (S3) and (S4), for the instantaneous reaction via either
mechanism, deprotonation, or carbopalladation. The data illustrate the
potential for mass transfer influences on either mechanism (a) for
varying fluid mechanics and (b) at different reaction temperatures.
10
11
12
1.0
3.0
5.0
0.3
0.4
4
flow
flow
36
a
b
Batch experiments performed at 1100 rpm. All yields calculated
from product concentrations measured in three samples. No
byproducts were measured.
On the other hand, the flow results of Figure 5a motivate a
more in-depth analysis that reveals the role of mass transfer.
Figure 5a approximates the measured value, kca = 0.1 min−1,
using expression (S4) for flow. Theoretically calculated kca
values are available, using expression (S2), by substituting the
liquid−liquid interfacial contact area in the milliscale flow
reactor, a = 90 cm2/cm3, for liquid slugs having equivalent
droplet diameters,24 di = 0.21 cm. Theoretical and experimental
values of the substrate diffusivity, predicted by the Wilke−
Chang correlation24 and Valencia and Gonzalez,25 are 2.3 ×
10−3 and 6.2 × 10−3 cm2/min at 80 °C, respectively.
Substituting the diffusivities into expression (S2) when Sh =
2 yields kca values of 1.9 and 5.2 min−1. The calculated mass
transfer coefficients being an order of magnitude greater than
our measured value both indicate a kinetically controlled direct
arylation. Strictly speaking, the corresponding molar flux rate is
an order of magnitude greater than the reaction rate. However,
best practices of performing kinetic measurements when the
mass transfer coefficient exceeds at least 1 of order magnitude
ensure the discovery of intrinsic mechanistic knowledge.
Reducing the flow reactor characteristic length from 1.0 mm
to microreactor dimensions (e.g., < 0.5 mm), hence an order of
magnitude increase in the interfacial liquid−liquid contact area,
is expected to eliminate any mass transfer influences on the
measured kinetics altogether. The batch results of Figure 5b (at
1100 rpm) further support that the kinetics are principally
additionally highlights the calculated yields for batch and flow
experiments. In all cases, however, comparisons of the physical
rate processes to the chemical rate processes are needed to
identify the mechanisms truly governing the complex reaction
system.
Instantaneous Reaction with Respect to Mass Trans-
fer. The mass transfer rates derived for instantaneous reaction
at the liquid−liquid interface elucidate the role of molecular
diffusion on the kinetics. Expressions (S3) and (S4) are shown
in Figures 5a and b via plotting the natural logarithm of the
substrate concentrations. One observes that in Figure 5a strong
correlations exist for stirrer rates of 176 and 522 rpm. Plausible
explanations are that the reactions were mass transfer limited
rather than kinetically controlled by either the deprotonation or
the carbopalladation catalytic cycles. The resulting mass transfer
coefficients, kca, are 7.0 × 10−3 and 1.0 × 10−2 min−1 for stirrer
rates of 176 and 522 rpm, respectively. Predicted initial
substrate concentrations of 0.418 and 0.420 M are in agreement
with the actual initial 4-bromobenzotrifluoride concentration of
0.424 M. These results undergird that, at stirrer rates of 176
and 522 rpm, the molecular flux of the substrate controls the
reaction rather than the deprotonation or the carbopalladation
mechanisms themselves.
1265
dx.doi.org/10.1021/op4001274 | Org. Process Res. Dev. 2013, 17, 1262−1271

Organic Process Research & Development
Article
Figure 6. Calculated and measured aryl bromide, aryl alkyne, and base concentrations as functions of their reaction times (at 50, 70, and 80 °C and
1100 rpm) for the deprotonation mechanism. Rate-determining step scenarios of (a) oxidative addition far from equilibrium (OA), (b) coordination
far from equilibrium (COR), (c) reductive elimination (RE), and (d) dissociation (DISS) were calculated. Each set of oxidative addition and
coordination calculations of the deprotonation mechanism match those for the carbopalladation mechanism. Calculated values for (MI) migratory
insertion also are reported for the carbopalladation mechanism. Near equilibrium simulation results for oxidative addition (OA, EQ) and
coordination (COR, EQ) steps are available in the Supporting Information as Figures S3a and S3b, respectively.
controlled by either the deprotonation or the carbopalladation
mechanism. Although reasonable linear correlations are
reported in these figures, the initial concentrations of 0.449
M (in flow at 80 °C), 0.445 M (at 50 °C and 1100 rpm), 0.516
M (at 70 °C and 1100 rpm), and 0.518 M (at 80 °C and 1100
rpm) are over predictions. Thus, the mass transfer coefficients
(for flow and a batch stirrer rate of 1100 rpm) estimated by
expressions (S3) and (S4) are actually modified first-order
kinetic rate constants (according to expression (S1)) that
inaccurately describe the cross-coupling. Our observations do
not necessarily rule out the existence of mass transfer
limitations when performing laboratory synthesis. Instead,
they indicate the potential for mass transfer influences when
mixing effects are neglected. More accurate evaluations of the
kinetics are next accomplished through intrinsic calculations of
the deprotonation and the carbopalladation mechanisms.
Instantaneous Mass Transfer with Respect to Reac-
tion. As we have seen, theoretical mass transfer coefficient
estimates are in agreement in that the measured kinetics are
likely controlled by their intrinsic mechanism at batch stirrer
rates of 1100 rpm and dimp/dflask = 0.77. Our quest to further
evaluate the kinetics of either the deprotonation or the
carbopalladation mechanism begins by assuming the oxidative
addition step to control the overall reaction. The relationships
k2/(k1[Pd]T) ≪ 1 and k2′/(k1′[Pd]T) ≪ 1 when the first step is
far from equilibrium and [PdnLX] ≈ [PdnLX−1] ≈ [Pd]T and [L]
≈ [L]T. The conditions are relevant during the early stages of
the reaction (e.g., reaction times <5 min) where the substrate
concentrations are high relative to those of the ligand and the
palladium intermediates. The resulting rate expressions for
oxidative addition have first-order dependence on each the aryl
bromide and the catalytically active Pd0LX species 1
concentration. Near equilibrium (i.e., for longer reaction
times), the concentrations of both the intermediates and the
free ligand become important. The numerical solutions to the
simplified coupled ODEs, for nonequilibrium conditions, are
shown in Figures 6a−d, S3a, and S3b (for deprotonation and
carbopalladation steps) and in Figure 7a and b (for
carbopalladation steps) as the calculated concentrations of
aryl bromide substrates as functions of their reaction time. The
calculated values span the temperatures investigated, and the
measured values also are included in the figure. Interestingly,
the forward reaction rate constants, k1 and k1′, are calculated to
be equivalent as shown in Table 2. Linearization of the
temperature dependence of the rate constants (i.e., the
Arrhenius expression) yields the apparent activation energies
near and far from equilibrium of 9.8 kcal/mol for the forward
oxidative addition of 4-bromobenzotrifluoride. Oxidative
addition is generally thought to have the lowest activation
energy with values reported in the same range for either
1266
dx.doi.org/10.1021/op4001274 | Org. Process Res. Dev. 2013, 17, 1262−1271

Organic Process Research & Development
Article
addition adduct, species 2, concentrations. Near equilibrium,
k3/k4 ≈ 1 and k3′/k4′ ≈ 1, and the rate approximation is first
order in each the alkyne, species 2, and the ligand. The
corresponding coupled ODEs were solved numerically, and
their solutions are again reported in Figures 6 and 7. The sum
of the squared differences between the measured and the
calculated concentrations (i.e., s2) yields the corresponding
kinetic parameters, k3 and k3′, reported in Table 2. Analyses of
s2 values in Table 2 and the results of Figures 6 and 7 reveal
excellent agreement between the measured and the calculated
differences when one considers the coordination step to
determine the kinetics, near and far from equilibrium. One
observes that the linearization of the temperature dependence
of the rate constants gives the apparent activation energies near
(10.0 kcal/mol) and far (9.7 kcal/mol) from equilibrium that
closely match the forward oxidative addition of 4-bromobenzo-
trifluoride. Our analogy of the coordination step, in either
mechanism, estimates the apparent activation energies of
similar magnitude to purely organic systems.
Similarly, the removal of a proton from the alkyne−PdII
complex 3, and the resultant vinyl−PdII complex 4 (i.e.,
reductive elimination in deprotonation), also was investigated
for its potential as the rate-determining step. Reductive
elimination has been postulated to follow either cationic or
anionic mechanisms. Here, we presume, as a close approx-
imation, irreversible first-order dependence on each the
alkyne−PdII complex 3, the free ligand, and the base
concentrations. In doing so, k5 values were calculated and the
corresponding apparent activation energies of reductive
elimination (9.6 kcal/mol) are reported in Table 2. Kinetic
parameter estimates also are reported in Table 2 and in Figure
6 for the final step of the deprotonation mechanism: the
dissociation of the product R−Ar and its restoration of the
catalytically active Pd0LX species 1. Values of 7.6 kcal/mol
indicate a lower apparent energy barrier of R−Ar dissociation
compared to the other reaction steps. Our combined theoretical
and experimental analysis of all reaction steps in the
deprotonation mechanism does not reveal a single, governing
reaction step over the temperature range of 50 to 80 °C.
Instead, our analogy supports that all the reaction steps of our
aqueous phase Pd-catalyzed Cu-free direct arylation of terminal
alkyne, when considering only the deprotonation mechanism,
have apparent activation energies of the same order of
magnitude as purely organic systems. Additionally, it would
appear that the product dissociation step is not rate-
determining.
Our observations based on the deprotonation mechanism are
in agreement with the carbopalladation mechanism. Table 2
and Figure 7 further outline the kinetic parameter calculations
for (i) the intramolecular rearrangement (5) of the alkyne−PdII
complex 3 through migratory insertion of the aryl group Ar to
the terminal alkyne; (ii) the formation of the vinyl−PdII
complex, complex 4′, via ligand association; and (iii) the
subsequent R−Ar product formation via reductive elimination.
Interestingly, the migratory insertion and ligand coordination
apparent activation energies of 7.6 and 7.4 kcal/mol are
indications that oxidative addition, coordination, or reductive
elimination (with apparent activation energies of ∼10 kcal/
mol) could be the rate-determining step in the carbopalladation
mechanism. The calculated apparent activation energies are
again of similar magnitude as purely organic systems, and our
kinetic parameter estimates accurately predict the measured
Figure 7. Calculated and measured base concentrations as functions of
their reaction times (at 50, 70, and 80 °C and 1100 rpm) for the
carbopalladation mechanism. Rate-determining step scenarios of (a)
ligand coordination (LC) and (b) reductive elimination (RE) were
calculated.
mechanism. Density Functional Theory calculations22a have
previously reported the energy barrier for oxidative addition in
purely organic systems to be low (in the range of 17.0 kcal/mol
for iodides) compared to energy barriers as high as 40.4 kcal/
mol for the successive reductive elimination of the complex 4′
(see Figure 1b) in the carbopalladation mechanism. Our
measured values of 9.8 kcal/mol elucidate that aqueous energy
barriers are of the same magnitude as those observed in purely
organic systems, assuming no quantum mechanical tunneling
(i.e., the activation energy is approximately equal to the energy
barrier). The sum of the squared differences between the
measured and the calculated aryl bromide concentrations (i.e.,
s2) used to extract the kinetic parameters show excellent
agreement by s2 values ≪1 in Table 2. The measured
concentrations compared to the calculated concentrations for
the subsequent reaction steps further validate our kinetic
assumptions.
The coordination of a terminal alkyne with the oxidative
addition adduct, species 2, is generally thought to have a higher
activation barrier than oxidative addition in purely organic
systems. Ratios of k4/k3 ≪ 1 and k4′/k3′ ≪ 1 exist when the
forward coordination step determines the reaction rate (i.e., far
from equilibrium) for either catalytic cycle. Under these
circumstances the reaction rate approximation from either
mechanism, similarly to oxidative addition, is expected to be
first order with respect to each the alkyne and the oxidative
1267
dx.doi.org/10.1021/op4001274 | Org. Process Res. Dev. 2013, 17, 1262−1271

Organic Process Research & Development
Article
Table 2. Kinetic parameters estimated for both the deprotonation and the carbopalladation mechanisms with the reaction steps,
oxidative addition, coordination, ligand coordination, migratory insertion, reductive elimination, and product dissociation, each
a
having been calculated separately
kinetic
fitted
EApp
b
c
c
d
e
entry mechanism
reaction step
oxidative addition
T (°C)
parameter
value
units
ki,app (min−1
)
(kcal/mol)
s2 = ∑Ni=1(ci,m − ci,c)2
b
1
2
3
4
DM, CM
DM, CM
DM, CM
DM, CM
80
70
50
80
k1, k1′
k1, k1′
k1, k1′
k1, k1′
57.5
M−1 min−1
M−1 min−1
M−1 min−1
M−1 min−1
0.616
0.405
0.168
0.629
9.8
0.001
0.090
0.006
0.001
b
b
b
oxidative addition
oxidative addition
37.8
15.7
58.7
oxidative addition, near
equilibrium
9.8
b
b
5
6
DM, CM
DM, CM
oxidative addition, near
equilibrium
70
50
k1, k1′
k1, k1′
37.8
15.9
M−1 min−1
M−1 min−1
0.405
0.170
0.091
0.006
oxidative addition, near
equilibrium
b
b
b
b
b
7
DM, CM
DM, CM
DM, CM
DM, CM
DM, CM
coordination
80
70
50
80
70
50
80
70
50
80
k3, k3′
k3, k3′
k3, k3′
k3, k3′
k3, k3′
k3, k3′
k5
58.8
37.9
15.6
58.8
37.9
16.1
M−1 min−1
M−1 min−1
M−1 min−1
M−1 min−1
M−1 min−1
M−1 min−1
M−2 min−1
M−2 min−1
M−2 min−1
min−1
0.630
0.406
0.167
0.630
0.406
0.172
4.30
10.0
9.7
0.302
0.108
0.008
0.309
0.111
0.008
0.300
0.108
0.007
0.234
8
coordination
9
coordination
10
11
12
13
14
15
16
coordination, near equilibrium
coordination, near equilibrium
coordination, near equilibrium
reductive elimination
reductive elimination
reductive elimination
b
DM, CM
DM
5210
9.6
DM
k5
3500
1460
2.89
DM
k5
1.21
b
b
b
DM, CM
dissociation/migratory
insertion
k6, k5′
16.9
63.4
7.6
17
18
DM, CM
DM, CM
dissociation/migratory
insertion
70
50
k6, k5′
k6, k5′
13.1
6.3
min−1
min−1
49.1
23.6
0.073
0.003
dissociation/migratory
insertion
19
20
21
22
23
24
CM
CM
CM
CM
CM
CM
ligand coordination
ligand coordination
ligand coordination
reductive elimination
reductive elimination
reductive elimination
80
70
50
80
70
50
k6′
k6′
k6′
k7′
k7′
k7′
1540
M−1 min−1
M−1 min−1
M−1 min−1
M−1 min−1
M−1 min−1
M−1 min−1
30.9
23.7
7.4
0.232
0.066
0.003
0.300
0.108
0.007
1180
583
11.7
58.3
1.17
9.8
37
0.743
0.315
15.7
a
b
The sums of the squared differences between the calculated and the measured concentrations are also reported. Rate-determining step
c
d
approximates equivalent kinetic parameters for both mechanisms. Calculation assumes the reaction step is rate-determining. Kinetic parameter
e
values calculated using the nonlinear least-squares method. Apparent activation energy values estimated by plotting ln(ki) as functions of 1/T (K−1).
Linearized correlations of R2 > 0.99 for all scenarios.
concentrations, as shown in Figure 7 and by the s2 values of
Table 2.
As a consequence, the smallest reaction rate constant ki,app =
0.616 min−1 at 80 °C, as shown in Table 2. Our capillary flow
calculations of the mass transfer coefficient, kca, indicate values
on the order of 0.1 min−1. Therefore, one must also consider
mixed mass transfer and reaction rate limitations to properly
evaluate the activation barriers of the cross-coupling kinetics.
Mixed Mass Transfer and Reaction Rate Limited
Regime. When the molar flux rate approaches the same
order of magnitude as the reaction rate there exists the
potential for mixed mass transfer and kinetically limited
regimes. Thus, the measured rate constants are expected to
be less than their true values. The measured, true activation
energy in classical heterogeneous catalysis is derived as twice
that of the apparent activation energy (i.e., ET = 2EApp).13,26
The rule applies when spherical liquid droplets with diameters,
di, contain film resistance, and both chemical reaction and
molecular diffusion have Arrhenius temperature dependencies.
Further analysis of the apparent values of Table 2 offers
additional scientific insights on the kinetics in the mixed mass
transfer and intrinsically limited regime. From Table 2, our true
activation energies of the oxidative addition of the aryl bromide
(ET = 19.6 kcal/mol) and the coordination of the alkyne (ET =
19.5 kcal/mol), for either deprotonation or carbopalladation,
are in excellent agreement with recent DFT calculated values of
Revisiting the kinetic expressions of (S5) and (S7) offers
fundamental insight into the catalytically active species’
transition states. Our initial condition of the palladium mole
balance is an approximation [PdnLX] ≈ [PdnLX−1] ≈ [Pd]T, but
it offers fundamental insight into the activation energies of the
steps in either mechanism. Using the aforementioned
assumption, the fitted kinetic parameters of Table 2 do not
accurately describe the measured kinetics when fully construct-
ing and solving the ODEs of expressions (S5) and (S7) both
near and far from equilibrium, an indication that the
concentration of catalytic intermediates are a fraction of the
total palladium concentration. Accordingly, the corresponding
coupled ODEs were solved numerically for both mechanisms
by varying the fraction of the catalytically active palladium
species, ϕ, using expressions (S9), (S10), and (S11). Varying ϕ
values to minimize s2 values for expressions (S5) and (S7)
reveals the fraction of the total palladium in the form of any
catalytically active intermediate under far equilibrium con-
ditions for either mechanism. Thus, the apparent reaction rate
constants of Table 2 are recalculated using expressions (S12)
and (S13) for scenarios when one does not assume our initial
condition that all of the palladium participates in the reaction.
1268
dx.doi.org/10.1021/op4001274 | Org. Process Res. Dev. 2013, 17, 1262−1271

Organic Process Research & Development
Article
17.0 and 23.0 kcal/mol for iodides, respectively.22a Density
Functional Theory predictions of the highest activation barrier,
the rate-determining step of reductive elimination, for the
cationic (ET = 14.0 kcal/mol) and the anionic (ET = 13.1 kcal/
mol) deprotonation mechanisms are less than the measured
value of 19.2 kcal/mol from our aqueous phase cross-
coupling.22a Differences of 5.2 and 6.1 kcal/mol are
theoretically the result of the hydrophilic nature of the ligand.
Our measured activation energy of the diaryl product
dissociation in the deprotonation mechanism (ET = 15.1
kcal/mol) is 3.6 kcal/mol greater than DFT calculated
values,22a supporting that the hydrophobic nature of the
product could inhibit it from being expelled into the mixed
aqueous−organic solvent.
Further analysis of the carbopalladation mechanism in the
mixed mass transfer and reaction-controlled regime fosters a
deeper understanding of its kinetics. The energy barrier of the
intramolecular rearrangement (5) of the alkyne−PdII complex
3 via migratory insertion of the aryl group Ar to the terminal
alkyne has been calculated to be 14.2 kcal/mol.22a Our
measured value of 15.1 kcal/mol confirms the DFT predictions.
Remarkably, the hydrophilic ligand coordination step exhibits a
true activation energy of 14.8 kcal/mol, which is greater than
the value of 6.4 kcal/mol calculated by DFT.22a The difference
is possibly the outcome of competing hydrogen-bonding effects
during the ligand’s stereochemical coordination. Recent DFT
calculations of the carbopalladation mechanism reveal its
energy barrier for reductive elimination (i.e., the rate-
determining step) in the purely organic cross-coupling to be
40.4 kcal/mol, implying that the deprotonation mechanism
exhibits an overall lower energy catalytic cycle. The measured
true activation energy of reductive elimination in our
hydrophilic cross-coupling (ET = 19.7 kcal/mol) is 20.7 kcal/
mol less than the DFT calculated value.22a Consequently, either
the deprotonation or the carbopalladation mechanism could
dominate in the aqueous phase cross-coupling, which supports
that water plays a vital role in the catalysis.
into new 15 mL vials sealed under a nitrogen atmosphere. The
two solutions were then used to carry out reactions in batch
and flow reactors.
General Procedure for Batch Experiments. 1 mL each
of the acetonitrile and aqueous solutions was transferred into
three sealed 4.0 mL vials containing magnetic stir bars. The 4.0
mL vials had each been sealed with a screwcap and PTFE
coated septums under a nitrogen atmosphere. The three vials
were then tied together and carefully centered in an oil bath (80
1 °C) on a stir plate. Aliquots of 25 μL were taken from each
vial at intervals of 1 to 30 min and diluted into 1.0 mL of ethyl
acetate in ambient air. Gas chromatography (Varian 3800) was
used to analyze each aliquot to determine the progression of
the reaction. No byproducts were detected in any experiments.
The procedure was repeated at 70 1 °C and 50 1 °C) and
stir speeds of 2 (176 rpm), 5 (522 rpm), and 10 (1100 rpm).
General Procedure for Flow Experiments. Reagents
were transferred to glass syringes (5.0 mL, SGE) and delivered
to a 0.1 mL PFA capillary reactor (1/16″ O.D., 1.0 mm I.D.)
with a Harvard Apparatus PHD2000 syringe pump. All fluidic
connections were made using 1/4″-28 PTFE fittings and PFA
tubing (1/16″ O.D., 1.0 mm I.D., VWR). The capillary reactor
was coiled and submerged into a well mixed oil bath (80
1
°C) on a stir plate. The bath temperature was monitored via a
thermocouple and maintained with a Waage immersion heater
controlled by a J-KEM Scientific Gemini PID controller. Upon
exiting the reactor, the mixture was introduced with 1.0 M HCl
to quench the reaction before it was collected in a GC vial
sealed with a septum. Three experiments were performed each
at three different residence times of 1, 3, and 5 min. Gas
chromatography (Varian 3800) was used to analyze each
aliquot to determine the progression of the reaction. No
byproducts were detected in any experiments.
Procedure for Evaluation of Kinetic Parameters. Raw
data from synthesis experiments were fit using nonlinear
regression for each rate-determining step separately. The
coupled ODEs were then solved numerically by the Runge−
Kutta−Fehlberg28 and the Rosenbrock methods29 using
Polymath 6.1,30 with the resulting solutions fit using the
nonlinear least-squares method, and the kinetic parameters for
each rate-determining scenario identified.
EXPERIMENTAL SECTION
■
Reagent Preparation and Analytical. Bis(acetonitrile)-
palladium dichloride was synthesized from palladium chloride
by refluxing in acetonitrile for several hours, then precipitating
with cold diethyl ether, and filtering to yield a yellow-orange
powder.27 3-(Di-tert-butylphosphonium)propane sulfonate
(DTBPPS) was synthesized according to a previously reported
procedure.5j Di-tert-butylphosphine was obtained from FMC,
Lithium Division. All other materials were purchased from
VWR and used without further purification. The 4.0 mL vials
were 13 mm in diameter, and the magnetic stir bars were 10
mm × 3.0 mm × 4.0 mm. The stir plate used was a Corning
PC-420 stirrer/hot plate. All reaction vials were assembled in a
nitrogen-filled drybox (MBraun Labmaster 130).
CONCLUSION
■
The phase behaviors of aqueous Pd-catalyzed cross-couplings
using hydrophilic ligand are key to their study in the laboratory
and their scale-up to production systems. Synthetic organic
chemistry and chemical reaction engineering points-of-views
must be considered in parallel to discover the most efficient
routes to useful synthesis products. Otherwise, molar flux and
interfacial contact areas potentially control the kinetics instead
of either postulated catalytic cycle, deprotonation or
carbopalladation. Synthesis in isothermal batch reactors at
low stirring rates supports the fact that mass transfer indeed
influences the multiphase reaction, whereas isothermal multi-
phase flow reactors offer intrinsic kinetic insights. Multiphase
capillary reactors offer the unique opportunity to elucidate the
role of mass transfer in isothermal batch reactors. Our analysis
of mass transfer coefficients measured in flow and their
comparison to theoretical predictions motivates the need for
special considerations when performing kinetic studies on
multiphase reactions with classical batch techniques. The
present work lays the groundwork for studies on heat transfer
influences on the aqueous biphasic reaction kinetics.
In the drybox, PdCl2(CH3CN)2 (41.5 mg, 0.160 mmol) was
measured into a 15 mL vial and sealed with a septum. DTBPPS
(48.0 mg, 0.160 mmol) and CsOH·H2O (1.47 g, 8.75 mmol)
were measured into a separate 15 mL vial and sealed with a
septum. Both vials were removed from the drybox, and the first
vial was charged with 4-bromobenzotrifluoride (1.12 mL, 8.00
mmol), phenylacetylene (0.960 mL, 8.75 mmol), mesitylene
(0.800 mL, 5.74 mmol), and 8 mL of nitrogen flushed
acetonitrile. The second vial was charged with 8 mL of nitrogen
flushed deionized water. Both vials were allowed to sit for 1 h
before filtering through a nitrogen flushed PTFE syringe filter
1269
dx.doi.org/10.1021/op4001274 | Org. Process Res. Dev. 2013, 17, 1262−1271

Organic Process Research & Development
Article
Our discovery that the aqueous Pd-catalyzed Cu-free direct
arylation of terminal alkyne with 4-bromobenzotrifluoride
exhibits activation energies in excellent agreement with DFT
calculated values reported for purely organic systems establishes
the groundwork for exciting possibilities. Additional work on
the modified ionic deprotonation mechanism will aid in
defining the molecular interactions of water that potentially
influence either the deprotonation or the carbopalladation
mechanism. Computational techniques in collaboration with
precise experimental control of the reaction conditions (e.g.,
the use of microreactors) could potentially resolve ambiguity
revolving around either the deprotonation or the carbopalla-
dation mechanisms. Our discoveries thus far support that either
mechanism could control the reaction kinetics in aqueous phase
cross-couplings.
Time-dependent changes in both the ligand and the
palladium intermediate concentrations, smaller in magnitude
relative to the substrates, are expected to have a profound
impact on the near-equilibrium kinetics. Moreover, the kinetics
of formation of the catalytically active palladium species could
ultimately control the overall reaction when the rates of the
individual steps in either catalytic cycle are fast, relatively
speaking (e.g., approaching flash chemistry). The palladium and
the hydrophilic ligand lifecycles themselves, especially in
multiphase reactions, are interesting pathways that demand
the attention of laboratory techniques that offer an even deeper
molecular understanding of the science. Innovations of
microreactors with online analytics have the potential to reveal
the true reaction mechanism, carbopalladation and/or depro-
tonation, and the rate-determining step(s) over the entire
kinetically controlled reaction space of aqueous phase cross-
couplings. The study of well-characterized multiphase micro-
fluidics, where reaction-rate-limited conditions exist, is expected
to provide a first principle understanding of the science of water
and its potential use in the continuous processing of fine
chemicals.
REFERENCES
■
(1) (a) Dunn, P. Water as a green solvent for pharmaceutical
applications. In Handbook of Green Chemistry - Green Solvents;
Crabtree, R. H., Leitner, W., Jessop, P. G., Li, C.-J., Wasserscheid,
P., Stark, A., Eds.; John Wiley & Sons: Weinheim, 2010; Vol. 5, pp
363−383. (b) Hailes, H. C. Org. Process Res. Dev. 2007, 11 (1), 114−
120. (c) Shaughnessy, K. H. Chem. Rev. 2009, 109 (2), 643−710.
(d) Shaughnessy, K. H.; DeVasher, R. B. Curr. Org. Chem. 2005, 9 (7),
585−604.
(2) Garrett, C. E.; Prasad, K. Adv. Synth. Catal. 2004, 346, 889−900.
(3) Butler, R. N.; Coyne, A. G. Chem. Rev. 2010, 110, 6302−6337.
(4) (a) Hartman, R. L. Org. Process Res. Dev. 2012, 16 (5), 870−887.
(b) Hartman, R. L.; et al. Org. Process Res. Dev. 2010, 14 (6), 1347−
1357. (c) Noel, T.; et al. Chem. Sci. 2011, 2 (2), 287−290.
(5) (a) Johansson Seechurn, C. C. C.; et al. Angew. Chem., Int. Ed.
́
́
2012, 51, 5062−5085. (b) Molnar, A. Palladium-Catalyzed Coupling
Reactions; Wiley-VCH: 2013; p 692. (c) Negishi, E.-i. Angew. Chem.,
Int. Ed. 2011, 50 (30), 6738−6764. (d) Suzuki, A. Angew. Chem., Int.
Ed. 2011, 50 (30), 6722−6737. (e) Shaughnessy, K. H.; Booth, R. S.
Org. Lett. 2001, 3 (17), 2757−2759. (f) DeVasher, R. B.; et al. J. Org.
Chem. 2004, 69 (23), 7919−7927. (g) Moore, L. R.; Shaughnessy, K.
H. Org. Lett. 2004, 6 (2), 225−228. (h) Huang, R.; Shaughnessy, K. H.
Chem. Commun. 2005, 35, 4484−4486. (i) Huang, R.; Shaughnessy, K.
H. Organometallics 2006, 25 (17), 4105−4112. (j) Brown, W. S.; et al.
Synthesis 2008, 12, 1965−1970. (k) Moore, L. R.; et al. Organometallics
2008, 27 (4), 576−593. (l) Alonso, F.; et al. Tetrahedron 2005, 61
(50), 11771−11835. (m) Anderson, K. W.; Buchwald, S. L. Angew.
Chem., Int. Ed. 2005, 44 (38), 6173−6177. (n) Beletskaya, I. P.;
Cheprakov, A. V. Chem. Rev. 2000, 100 (8), 3009−3066. (o) Cabri,
W.; Candiani, I. Acc. Chem. Res. 1995, 28 (1), 2−7. (p) Cabri, W.; et al.
J. Org. Chem. 1992, 57 (13), 3558−3563. (q) de Vries, J. G. Can. J.
Chem.-Rev. Can. Chim. 2001, 79 (5−6), 1086−1092. (r) Dounay, A.
B.; Overman, L. E. Chem. Rev. 2003, 103 (8), 2945−2963.
(s) Ljungdahl, T.; et al. Organometallics 2008, 27 (11), 2490−2498.
(t) Marion, N.; Nolan, S. P. Acc. Chem. Res. 2008, 41 (11), 1440−
1449. (u) Mennecke, K.; et al. Synthesis 2008, No. 10, 1589−1599.
(v) Moore, L. R.; Shaughnessy, K. H. Org. Lett. 2004, 6 (2), 225−228.
(6) Nicolaou, K. C.; et al. Angew. Chem., Int. Ed. 2005, 44 (29),
4442−4489.
(7) (a) Magano, J.; Dunetz, J. Chem. Rev. 2011, 111, 2177−2250.
(b) Torborg, C.; Beller, M. Adv. Synth. Catal. 2009, 351, 3027−3043.
(8) (a) Fu, G. C. Acc. Chem. Res. 2008, 41 (11), 1555−1565.
(b) Lundgren, R. J.; Stradiotto, M. Chem.Eur. J. 2012, 18, 9758−
9769. (c) Valente, C.; et al. Angew. Chem., Int. Ed. 2012, 51, 3314−
3332. (d) So, C. M.; Kwong, F. Y. Chem. Soc. Rev. 2011, 40, 4963−
4972.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental details. This material is available free of charge via
the Internet at http://pubs.acs.org.
(9) Miyaura, N.; et al. J. Am. Chem. Soc. 1985, 107, 972−980.
(10) Casalnuovo, A. L.; Calabrese, J. C. J. Am. Chem. Soc. 1990, 112,
4324−4330.
AUTHOR INFORMATION
■
(11) Rideout, D. C.; Breslow, R. J. Am. Chem. Soc. 1980, 102 (26),
7816−7817.
Corresponding Authors
*First corresponding author. E-mail: rhartman@eng.ua.edu.
Tel: +1 (205) 348-1696. Fax: +1 (205) 348-7558.
*Second corresponding author. E-mail: kshaughn@bama.ua.
edu. Tel: +1 (205) 348-4435. Fax: +1 (205) 348-9104.
(12) Chanda, A.; Fokin, V. V. Chem. Rev. 2009, 109 (2), 725−748.
(13) Levenspiel, O. Chemical Reaction Engineering, 3rd ed.; John
Wiley & Sons: New York, NY, 1999; p 578.
(14) (a) Alvarez, M. M.; et al. AIChE J. 2002, 48 (10), 2135−2148.
(b) Alvarez-Hernandez, M. M.; et al. Chem. Eng. Sci. 2002, 57 (17),
3749−3753. (c) Baldyga, J.; Bourne, J. R. Chem. Eng. Sci. 1992, 47 (8),
1839−1848. (d) Fournier, M. C.; et al. Chem. Eng. Sci. 1996, 51 (22),
5053−5064. (e) Ottino, J. M. Phys. Fluids 2010, 22 (2), 12. (f) Ottino,
J. M. Chem. Eng. Sci. 1994, 49 (24A), 4005−4027. (g) Ottino, J. M.;
et al. AIChE J. 1981, 27 (4), 565−577. (h) Ottino, J. M.; Wiggins, S.
Philos. Trans. R. Soc. London, Ser. A 2004, 362 (1818), 923−935.
(i) Paul, E. L., et al. Handbook of Industrial Mixing: Science and Practice;
John Wiley & Sons, Inc.: Hoboken, NJ, 2004; p 1385. (j) Post, T.
Chem. Eng. Prog. 2010, 106 (3), 25−32. (k) Zalc, J. M.; et al. AIChE J.
2002, 48 (10), 2124−2134.
Author Contributions
§These authors contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the National Science Foundation (EEC-1062705;
CBET-1264630; CHE-1058984) for partial financial support of
this work; FMC, Lithium Division for donation of di-tert-
butylphosphine; and Johnson-Matthey for donation of
palladium salts.
(15) (a) Aoki, N.; Mae, K. Chem. Eng. J. 2006, 118 (3), 189−197.
(b) Bottausci, F.; et al. Lab Chip 2007, 7 (3), 396−398. (c) Chen, D.
L.; et al. J. Am. Chem. Soc. 2005, 127 (27), 9672−9673. (d) Ehrfeld,
1270
dx.doi.org/10.1021/op4001274 | Org. Process Res. Dev. 2013, 17, 1262−1271

Organic Process Research & Development
Article
W.; et al. Ind. Eng. Chem. Res. 1999, 38 (3), 1075−1082. (e) Falk, L.;
Commenge, J. M. Chem. Eng. Sci. 2010, 65 (1), 405−411.
(f) Guichardon, P.; Falk, L. Chem. Eng. Sci. 2000, 55 (19), 4233−
4243. (g) Guichardon, P.; et al. Chem. Eng. Sci. 2000, 55 (19), 4245−
4253. (h) Gunther, A.; Jensen, K. F. Lab Chip 2006, 6 (12), 1487−
1503. (i) Gunther, A.; et al. Langmuir 2005, 21 (4), 1547−1555.
(j) Hessel, V.; et al. AIChE J. 2003, 49 (3), 566−577. (k) Hessel, V.;
et al. Chem. Eng. Sci. 2005, 60 (8−9), 2479−2501. (l) Hu, G. Q.; Li, D.
Q. Chem. Eng. Sci. 2007, 62 (13), 3443−3454. (m) Kockmann, N.;
et al. Sens. Actuators, B 2006, 117 (2), 495−508. (n) Lu, L. H.; et al. J.
Microelectromech. Syst. 2002, 11 (5), 462−469. (o) Paik, P.; et al. Lab
Chip 2003, 3 (4), 253−259. (p) Panic, S.; et al. Chem. Eng. J. 2004,
101 (1−3), 409−419. (q) Song, H.; et al. Appl. Phys. Lett. 2003, 83
(22), 4664−4666.
(29) Press, W. H., et al. Numerical Recipes, 2nd ed.; Cambridge
University Press: Cambridge, 1992.
(30) Shacham, M., et al. Polymath 6.1, Numerical Solutions for
Scientific and Engineering Mathematical Problems; Polymath Software:
Willimantic, CT, 2006.
(16) Hartman, R. L.; et al. Angew. Chem., Int. Ed. 2011, 50, 7502−
7519.
(17) (a) Calabrese, G. S.; Pissavini, S. AIChE J. 2011, 57 (4), 828−
834. (b) Kockmann, N.; et al. Chem. Eng. J. 2011, 167 (2−3), 718−
726. (c) Kockmann, N.; et al. Chem.Eur. J. 2008, 14 (25), 7470−
7477. (d) Roberge, D. M.; et al. Chem. Eng. Technol. 2005, 28 (3),
318−323. (e) Roberge, D. M.; et al. Org. Process Res. Dev. 2008, 12 (5),
905−910. (f) Hessel, V. Chem. Eng. Technol. 2009, 32 (11), 1655−
1681. (g) Chevalier, B.; Schmidt, F. Chim. Oggi-Chem. Today 2008, 26
(5), 6−7. (h) Pennemann, H.; et al. Org. Process Res. Dev. 2004, 8 (3),
422−439. (i) Caygill, G.; et al. Org. Process Res. Dev. 2006, 10 (3),
539−552. (j) Taghavi-Moghadam, S.; et al. Org. Process Res. Dev. 2001,
5 (6), 652−658. (k) Renken, A.; et al. Chem. Eng. Process. 2007, 46
(9), 840−845.
(18) (a) Baumann, M.; et al. Mol. Diversity 2011, 15 (3), 613−630.
(b) Baxendale, I. R.; et al. Chem. Commun. 2006, 24, 2566−2568.
(c) Bogdan, A. R.; et al. Angew. Chem., Int. Ed. 2009, 48 (45), 8547−
8550. (d) Carter, C. F.; et al. Org. Process Res. Dev. 2010, 14 (2), 393−
404. (e) Qian, Z. Z.; et al. Synlett 2010, No. 4, 505−508. (f) Riva, E.
Tetrahedron 2010, 66 (17), 3242−3247. (g) Shu, W.; et al. Angew.
Chem., Int. Ed. 2011, 50 (45), 10665−10669. (h) Hartman, R. L.; et al.
Angew. Chem., Int. Ed. 2010, 49 (5), 899−903. (i) Li, P. F.; Buchwald,
S. L. Angew. Chem., Int. Ed. 2011, 50 (28), 6396−6400. (j) Naber, J. R.;
Buchwald, S. L. Angew. Chem., Int. Ed. 2010, 49 (49), 9469−9474.
(k) Noel, T.; Buchwald, S. L. Chem. Soc. Rev. 2011, 40 (10), 5010−
5029. (l) Bedore, M. W.; et al. Org. Process Res. Dev. 2010, 14 (2),
432−440.
(19) (a) Haswell, S. J.; Watts, P. Green Chem. 2003, 5 (2), 240−249.
(b) Mason, B. P.; et al. Chem. Rev. 2007, 107 (6), 2300−2318.
(20) (a) Fors, B. P.; et al. Org. Lett. 2008, 10 (16), 3505−3508.
(b) Huang, X.; et al. J. Am. Chem. Soc. 2003, 125 (22), 6653−6655.
(21) Huang, R.; Shaughnessy, K. H. Chem. Commun. 2005, No. 35,
4484−4486.
(22) (a) Garcia-Melchor, M.; et al. ACS Catal. 2012, 2 (1), 135−144.
(b) Amatore, C.; et al. Eur. J. Org. Chem. 2004, 2, 366−371.
(c) Tougerti, A.; et al. Chem.Eur. J. 2007, 13 (2), 666−676.
(23) (a) Gunther, A.; et al. Lab Chip 2004, 4 (4), 278−286.
(b) Hartman, R. L.; Jensen, K. F. Lab Chip 2009, 9, 2495−2507.
(c) Losey, M. W.; et al. J. Microelectromech. Syst. 2002, 11 (6), 709−
717. (d) Losey, M. W.; et al. Ind. Eng. Chem. Res. 2001, 40 (12),
2555−2562. (e) Yen, B. K. H.; et al. Angew. Chem., Int. Ed. 2005, 44
(34), 5447−5451.
(24) Geankoplis, C. J. Transport Processes and Unit Operations, 3rd
ed.; Prentice Hall PTR: Englewood Cliffs, NJ, 1993.
(25) Valencia, D. P.; Gonzalez, F. J. Electrochem. Commun. 2011, 13
(2), 129−132.
(26) Fogler, H. S. Elements of Chemical Reaction Engineering, 4th ed.;
Prentice Hall: Upper Saddle River, NJ, 2005.
(27) Wimmer, F. L., et al. (2,2′-Bipyridine)Dichloropalladium(II)
and [N′-(2-Diaminoethyl)-1,2-Ethanediamine]Chloropalladium(II)
Chloride. In Inorganic Syntheses; John Wiley & Sons, Inc.: 2007; pp
185−187.
(28) Forsythe, G. E., et al. Computer Methods for Mathematical
Computation; Prentice Hall PTR: Englewood Cliffs, NJ, 1977.
1271
dx.doi.org/10.1021/op4001274 | Org. Process Res. Dev. 2013, 17, 1262−1271