Reevaluation of the mechanism of the amination of aryl halides catalyzed by BINAP-ligated palladium complexes

Article abstract of DOI:10.1021/ja045533c

Two previous mechanistic studies of the amination of aryl halides catalyzed by palladium complexes of 1,1′-binaphthalene-2,2′- diylbis(diphenylphosphine) (BINAP) are reexamined by the authors of both studies. This current work includes a detailed study of the identity of the BINAP-ligated palladium complexes present in reactions of amines with aryl halides and rate measurements of these catalytic reactions initiated with pure precatalysts and precatalysts generated in situ from [Pd₂(dba) ₃] and BINAP. This work reveals errors in both previous studies, and we describe our current state of understanding of the mechanism of this synthetically important transformation. ³¹P NMR spectroscopy shows that several palladium(O) species are present in the catalytic system when the catalyst is generated in situ from [Pd₂-(dba)₃] and BINAP, and that at least two of these complexes generate catalytic intermediates. Further, these spectroscopic studies and accompanying kinetic data demonstrate that an apparent positive order in the concentration of amine during reactions of secondary amines is best attributed to catalyst decomposition. Kinetic studies with isolated precatalysts show that the rates of the catalytic reactions are independent of the identity and the concentration of amine, and studies with catalysts generated in situ show that the rates of these reactions are independent of the concentration of amine. Further, reactions catalyzed by [Pd(BINAP)₂] with added BINAP are found to be first-order in bromoarene and inverse first-order in ligand, in contrast to previous work indicating zero-order kinetics in both. These data, as well as a correlation between the decay of bromobenzene in the catalytic reaction and the predicted decay of bromobenzene from rate constants of studies on stoichiometric oxidative addition, are consistent with a catalytic process in which oxidative addition of the bromoarene occurs to [Pd(BINAP)] prior to coordination of amine and in which [Pd(BINAP)₂], which generates [Pd(BINAP)] by dissociation of BINAP, lies off the cycle. By this mechanism, the amine and base react with [Pd(BINAP)(Ar)(Br)] to form an arylpalladium amido complex, and reductive elimination from this amido complex forms the arylamine.

Full text of DOI:10.1021/ja045533c

Published on Web 03/01/2006
Reevaluation of the Mechanism of the Amination of Aryl
Halides Catalyzed by BINAP-Ligated Palladium Complexes
Shashank Shekhar, Per Ryberg, and John F. Hartwig*
Department of Chemistry, Yale UniVersity, P.O. Box 208107,
New HaVen, Connecticut 06520-8107
Jinu S. Mathew and Donna G. Blackmond*
Department of Chemistry, Imperial College, London SW7 2AZ, United Kingdom
Eric R. Strieter and Stephen L. Buchwald*
Department of Chemistry, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139
Received July 25, 2004; Revised Manuscript Received December 16, 2005;
E-mail: John.Hartwig@yale.edu; D.Blackmond@imperial.ac.uk; sbuchwal@mit.edu
Abstract: Two previous mechanistic studies of the amination of aryl halides catalyzed by palladium
complexes of 1,1′-binaphthalene-2,2′-diylbis(diphenylphosphine) (BINAP) are reexamined by the authors
of both studies. This current work includes a detailed study of the identity of the BINAP-ligated palladium
complexes present in reactions of amines with aryl halides and rate measurements of these catalytic
reactions initiated with pure precatalysts and precatalysts generated in situ from [Pd2(dba)3] and BINAP.
This work reveals errors in both previous studies, and we describe our current state of understanding of
the mechanism of this synthetically important transformation. 31P NMR spectroscopy shows that several
palladium(0) species are present in the catalytic system when the catalyst is generated in situ from [Pd2-
(dba)3] and BINAP, and that at least two of these complexes generate catalytic intermediates. Further,
these spectroscopic studies and accompanying kinetic data demonstrate that an apparent positive order
in the concentration of amine during reactions of secondary amines is best attributed to catalyst
decomposition. Kinetic studies with isolated precatalysts show that the rates of the catalytic reactions are
independent of the identity and the concentration of amine, and studies with catalysts generated in situ
show that the rates of these reactions are independent of the concentration of amine. Further, reactions
catalyzed by [Pd(BINAP)2] with added BINAP are found to be first-order in bromoarene and inverse first-
order in ligand, in contrast to previous work indicating zero-order kinetics in both. These data, as well as
a correlation between the decay of bromobenzene in the catalytic reaction and the predicted decay of
bromobenzene from rate constants of studies on stoichiometric oxidative addition, are consistent with a
catalytic process in which oxidative addition of the bromoarene occurs to [Pd(BINAP)] prior to coordination
of amine and in which [Pd(BINAP)2], which generates [Pd(BINAP)] by dissociation of BINAP, lies off the
cycle. By this mechanism, the amine and base react with [Pd(BINAP)(Ar)(Br)] to form an arylpalladium
amido complex, and reductive elimination from this amido complex forms the arylamine.
Introduction
these errors in data and conclusions and describe our current
state of understanding of the mechanism of this synthetically
important transformation.
Palladium complexes of 1,1′-binaphthalene-2,2′-diylbis-
(diphenylphosphine) (BINAP) and other bisphosphines are
catalysts for the amination of aryl halides.1,2 This reaction has
become a widely used practical synthetic method for C-N bond
formation and has been the subject of intensive mechanistic
analysis. In 2000 and 2002, two papers were published in this
journal describing kinetic and mechanistic studies on the
(BINAP)Pd-catalyzed amination of bromobenzene. Each of
these papers contained errors. In the current article we correct
In 2000,3 a series of synthetic, spectroscopic, and kinetic
studies on the oxidative addition of aryl bromides to [Pd-
(BINAP)2] and [Pd(DPPF)2] and on the catalytic amination of
bromoarenes catalyzed by these two compounds were reported.
These data indicated that the catalytic process occurred by a
pathway (Scheme 1) in which oxidative addition of the
bromoarene occurred to [Pd(BINAP)], which was generated by
dissociation of ligand from [Pd(BINAP)2]. Subsequent work
indicated that a minor pathway involving displacement of a
(1) Hartwig, J. F. In Handbook of Organopalladium Chemistry for Organic
Synthesis; Negishi, E. I., Ed.; Wiley-Interscience: New York, 2002; Vol.
1, p 1051.
(3) Alcazar-Roman, L. M.; Hartwig, J. F.; Rheingold, A. L.; Liable-Sands, L.
(2) Muci, A. R.; Buchwald, S. L. Top. Curr. Chem. 2002, 219, 131.
M.; Guzei, I. A. J. Am. Chem. Soc. 2000, 122, 4618.
9
3584
J. AM. CHEM. SOC. 2006, 128, 3584-3591
10.1021/ja045533c CCC: $33.50 © 2006 American Chemical Society

Amination of Aryl Halides Catalyzed by Pd−BINAP Complexes
A R T I C L E S
Scheme 1
Scheme 3
Scheme 4
of studies that provide further insight into the mechanism of
this important catalytic process. These studies have led to
mechanistic conclusions that modify both reaction schemes.
Results and Discussion
1. Identification of the Major Catalyst Components and
the Rates of Reactions with Pure Precatalysts. A. Identity
of the Palladium Complexes in the Catalytic Solutions. To
interpret rate data on a catalytic process and to draw mechanisms
that correspond to these rate data, information should be gained
on the identities of the species present in the catalytic process.
In some cases, these observed species may be true intermediates,
but in other cases they lead to intermediates by reversible or
irreversible processes. The kinetic data reported in 2002 were
collected on reactions with a catalyst generated in situ (1-IS)
by combining BINAP and Pd2(dba)3 and incubating at 60 °C
with alkoxide for 1 h (Scheme 3). In some other experiments,
BINAP, Pd2(dba)3, base, and amine were incubated prior to
addition of the bromoarene. No information on the palladium
complexes in solution was provided. Thus, these and related
reactions with various combinations of amine, base, and aryl
halide were monitored by 31P NMR spectroscopy to identify
the major palladium complexes in solution.
The reactions of compounds generated from BINAP and Pd2-
(dba)3 with amine and base are summarized in Scheme 4. We
first describe transformations starting with pure palladium
phosphine complexes. The combination of BINAP and Pd2(dba)3
alone generates free dba and [Pd(BINAP)(dba)] (1),6 which was
the Pd(0)-dba complex included in the mechanisms of the 2002
report.6 Purified complex 1 reacts with NaOtAm (Am ) CMe2-
Et) and N-methylpiperazine or with NaOtAm and octylamine
at 60 or 65 °C to form the new dinuclear complex [Pd-
(BINAP)]2(dba) (2). This complex presumably forms because
the amine adds to dba by a Michael addition under the basic
conditions.7 The dinuclear palladium complex was isolated in
85% yield and characterized by spectroscopy and X-ray dif-
fraction (see Supporting Information).
Scheme 2
partially dissociated κ1-BINAP in [Pd(κ2-BINAP)(κ1-BINAP)]
by bromoarene prior to oxidative addition occurred in parallel.4
In this mechanism, the complex [Pd(BINAP)2] was proposed
to lie on the catalytic cycle because the rate behavior had been
measured to be zero order in ArBr and added ligand.
In 2002, Blackmond, Buchwald and co-workers5 reported
studies of in situ catalyst activation in amination reactions
initiated with mixtures of Pd2(dba)3 and BINAP and with pure
[Pd(BINAP)2] as catalyst precursor. Kinetic modeling of reac-
tions catalyzed by mixtures of Pd2(dba)3 and BINAP supported
the proposal, shown in Scheme 2, that rate-limiting oxidative
addition of ArX occurs to an amine-bound Pd complex and that
oxidative addition to [Pd(BINAP)(amine)] is faster than oxida-
tive addition to [Pd(BINAP)]. This work also proposed a role
of the amine substrate in catalyst activation using BINAP and
Pd2(dba)3 mixtures in which the amine displaced dba to form
the palladium(0) amine complex [Pd(BINAP)(amine)]. In ad-
dition, these studies proposed an alternate interpretation of the
results of ref 3, suggesting that the apparent zero-order kinetics
in [ArBr] could be rationalized by a slow activation of a
precatalyst complex. These studies implied that [Pd(BINAP)2]
lay off of the catalytic cycle and fed [Pd(BINAP)] into the cycle
by ligand dissociation.
When complex 1 was generated in situ from [Pd2(dba)3] and
BINAP in the presence of amine and base, the Pd(0) species
was different. The results of these reactions are shown in
Scheme 5, and the 31P NMR spectra of these mixtures are shown
in Figure 1. When [Pd2(dba)3], BINAP, base, and N-methylpip-
erazine were mixed at once and stirred at 60 °C, a mixture of
dimer 2 and [Pd(BINAP)2] (3) was formed in a 0.7:1.0 ratio
(5) Singh, U. K.; Strieter, E. R.; Blackmond, D. G.; Buchwald, S. L. J. Am.
Chem. Soc. 2002, 124, 14104.
(6) Amatore, C.; Broeker, G.; Jutand, A.; Khalil, F. J. Am. Chem. Soc. 1997,
119, 5176.
(7) The reaction of free dba with NaOtBu and either the primary or secondary
amine rapidly consumes dba at room temperature and generates several
products.
To address the differences between these mechanistic propos-
als, the authors of both previous papers report herein a series
(4) Alcazar-Roman, L. M.; Hartwig, J. F. Organometallics 2002, 21, 491.
9
J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006 3585

A R T I C L E S
Shekhar et al.
a procedure in the 2002 report. These experiments showed that
the ratio of 2 to 3 was relatively constant (2:3) after the first
portion of bromoarene had reacted. However, the amount of
free BINAP decreased initially, signals for 2 and 3 decreased
in intensity, and broad or poorly resolved resonances in the
region of coordinated BINAP grew in intensity.
These data lead to several conclusions about the palladium
complexes in solution. The major palladium(0) species is [Pd-
(BINAP)2] (3); significant amounts of dinuclear complex [Pd-
(BINAP)]2(dba) (2) are present; the ratios of 2 to 3 depend on
the conditions by which the catalyst is generated; and complexes
2 and 3 are decomposing over the time of the catalytic reactions
of secondary amines. The catalyst decomposition and depen-
dence of the ratio of 2 to 3 on the type of amine are important
for understanding the rate data. The identification of 2 and 3 as
the major species allows one to propose a mechanism for the
catalytic process that begins with a precatalyst defined by
spectroscopic data.
Figure 1. (Top) 31P NMR spectra obtained during the reaction of
3-bromoanisole (0.13 M) with N-ethylpiperazine (0.93 M) catalyzed by Pd2-
(dba)3 (5.0 mM) and BINAP (10 mM) (1-IS) at 60 °C. (Bottom) 31P NMR
spectra during the same reaction with octylamine.
Scheme 5
B. Relative Rates of Reactions Initiated with Pure Forms
of the Observed Palladium Complexes in the Catalytic
Solutions. To determine the contributions of the different Pd-
(0) species to the rate of the catalytic process, the relative rates
of the catalytic aminations initiated with isolated [Pd(BINAP)-
(dba)] (1), dinuclear 2, and bis-BINAP complex 3 were studied.
Because the reactions of PhBr and 3-bromoanisole contained
the same three Pd(0) species, because the σ value of a 3-methoxy
group is only 0.1, and because the concentration of 3-bromoani-
sole is easily followed by 1H NMR spectroscopy, these
comparisons were conducted with 3-bromoanisole instead of
PhBr. However, reactions of PhBr catalyzed by 1-IS were
monitored by GC to ensure that the conclusions about induction
periods and relative rates of reactions of primary and secondary
amines would be the same for reactions of PhBr and 3-bro-
moanisole.
Scheme 6
after 1 h, and in a 1:1 ratio after 3 h. When [Pd2(dba)3], BINAP,
base, and octylamine, instead of N-methylpiperazine, were
mixed at once and stirred at 60 °C, a mixture of dimer 2 and
monomer 3 was formed in a 0.2:1.0 ratio after 1 h, and in a
0.3:1.0 ratio after 3 h. Although the precise ratios of 2 to 3
varied from experiment to experiment, the amount of bis-BINAP
3 was always larger in reactions of primary amines than in
reactions of secondary amines.
The reactions of some of these complexes occurred with
induction periods, and reactions of both amines initiated with
mononuclear dba complex 1 occurred with long induction
periods. The rates of reaction of the different catalysts were
measured after the induction period by the procedure in the 2002
report, in which sequential portions of aryl bromide were added
to a mixture of the other reaction components. The decay of
Most of the rate measurements by reaction calorimetry were
conducted by incubating [Pd2(dba)3], BINAP, and NaOtAm
prior to the addition of amine and aryl halide. The compositions
of the solutions generated from heating [Pd2(dba)3], BINAP,
and NaOtAm for 1 h were different from that just shown to
form from heating of these three species with amine.5 Scheme
6 shows a summary of the palladium complexes formed from
combining [Pd2(dba)3], BINAP, and NaOtAm. After 1 h of
incubation of [Pd2(dba)3], BINAP, and NaOtAm at 60 °C, the
three complexes [Pd(BINAP)2], [Pd(BINAP)(dba)], and [Pd-
(BINAP)]2(dba), as well as free BINAP, are all present.
The reactions catalyzed by the material formed from [Pd2-
(dba)3], BINAP, and NaOtAm were also monitored by 31P NMR
spectroscopy. In the catalytic reactions of N-methylpiperazine
with either PhBr (as reported in 2002) or 3-bromoanisole with
complex 1 generated in situ, both dimer 2 and bis-BINAP 3
were present. 31P NMR spectra were obtained of four sequential
reactions. The first was initiated with a 7:1 ratio of amine to
bromoarene, and the three subsequent reactions were run after
addition of sequential portions of additional ArBr. This proce-
dure of addition of multiple sequential portions of ArBr mimics
1
aryl bromide in each sequential reaction was monitored by H
NMR spectroscopy.
The rates of the reactions of N-methylpiperazine with
3-bromoanisole, conducted with isolated 1, 2, and 3 as precata-
lyst, obtained by 1H NMR spectroscopy at 50 °C are illustrated
Figure 2. Comparison of the rates of reaction of N-methylpiperazine (0.93
M) with 3-bromoanisole catalyzed by [Pd(BINAP)(dba)] (1) (10 mM), [Pd-
(BINAP)]2(dba)] (2) (10 mM), and [Pd(BINAP)2] (3) (10 mM) at 50 °C.
The decay of the first portion of the bromoarene is shown for reactions
initiated with catalysts 1 and 2, and the decay of the first and third portions
of bromoarene are shown for reactions initiated with catalyst 3.
9
3586 J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006

Amination of Aryl Halides Catalyzed by Pd−BINAP Complexes
A R T I C L E S
Figure 3. Comparison of the decay of 3-bromoanisole during the reactions
of 3-bromoanisole (0.13 M) with N-methylpiperazine (0.93 M) and
octylamine (0.93 M) catalyzed by 10 mM 1-IS at 60 °C. This plot
corresponds to the reaction of the first portion of aryl bromide in the
experiments in which sequential portions of ArBr are added.
by the curves in Figure 2. Clearly, the relative reactivities of
the precatalysts are 1 , 3 < 2. The reaction initiated with 1
occurs with a long induction period, and the initial rate with
this catalyst was too slow to measure. The reaction initiated
with 2 occurred without an induction period and with an initial
rate that was (1.1 ( 0.1) × 10-4 M s-1 at 50 °C. The reaction
of these same reagents initiated with bis-BINAP complex 3 in
the absence of added BINAP occurred with a small induction
period (vide infra).8 By NMR spectroscopy, the third and fourth
portions of 3-bromoanisole reacted with the same decay profile.
Whether the initial rate of reaction of the first portion of aryl
bromide ((3.5 ( 0.2) × 10-5 M s-1) or the third portion of
ArBr ((5.2 ( 0.2) × 10-5 M s-1) is used for comparison, the
initial rate of reactions with [Pd(BINAP)2] as precatalyst was
smaller than that of reactions with 2 as precatalyst and larger
than that of reactions with 1 as precatalyst. Thus, the long
induction period during reactions catalyzed by 1-IS results from
the transformation of 1 to 2, 1 to 3, or 1 to a combination of 2
and 3.
Figure 4. Comparison of the decay of ArX during sequential reactions of
PhBr (0.13 M) with N-methylpiperazine (0.93 M) and octylamine (0.93
M) catalyzed by 10 mM 1-IS at 60 °C. Influence of preincubation time on
the rate of the first reaction in the sequence. (Inset) Comparison of second
reactions in the same sequence.
shows that the first sequential reaction of N-methylpiperazine
with PhBr using 1-IS is significantly faster when the incubation
time is increased by just 10 min. This large difference in catalyst
activation presumably arises because the combination of amine
and base serves as a better nucleophile for Michael addition to
the dba than either base or amine alone. Figure 4 also shows
that the length of the induction period does not affect the rate
behavior of the catalyst once it attains steady state. If, however,
the mixture of Pd2(dba)3 and BINAP is incubated with NaOtAm
and the reaction is initiated by addition of amine and bro-
moarene, then less of the more reactive dinuclear 2 is generated
than if the two catalyst components are incubated with alkoxide
and amine together. Thus, the procedure for incubation affects
the absolute measured rates of reaction.
2. Evaluation of the Induction Periods of Reactions
1
Initiated with 1-IS by H NMR Spectroscopy. Because the
data in section 1 showed that the distribution of Pd(0) complexes
depended on subtle differences in the conditions of catalyst
incubation, differences in the induction period were tested by
monitoring the reactions of octylamine and N-methylpiperazine
3. Evaluation of Concentration Dependences for Amine
and PhBr in Reactions Catalyzed by a Mixture of Pd2dba3
and BINAP. The 2002 study by Blackmond, Buchwald, and
co-workers5 reported concentration dependences of aryl halide
and primary or secondary amine in amination reactions employ-
ing catalysts generated in situ by combining BINAP and Pd2-
(dba)3 with alkoxide base and amine prior to addition of the
bromoarene. Consecutive reactions were monitored using reac-
tion calorimetry by injecting equal aliquots of PhBr successively
into the reaction mixture containing the catalyst and an excess
of amine and base. After the initial induction period in the first
reaction in the sequence, decreasing reaction rates for each
consecutive reaction in the series were observed for secondary
amines but not for primary amines.
In addition, the form of the rate profile reported in 2002 was
different for reactions of secondary and primary amines, with
primary amines exhibiting slightly less than overall first-order
kinetics compared to apparent first-order kinetics for secondary
amines. The data used in ref 5 for kinetic modeling of the
mechanism in Scheme 2, comprising the second, third, and
fourth reactions in the sequence for n-hexylamine and N-
methylpiperazine, have recently been reproduced at both MIT
and Imperial College.9 This different kinetic behavior of the
1
by H NMR spectroscopy, in parallel with the same batch of
catalyst. [Pd2(dba)3] was incubated with BINAP and NaOtAm
for 1 h, and the resulting solution was split into two portions.
To one portion was added octylamine, and to the other portion
was added N-methylpiperazine. To both reactions were added
four sequential portions of 3-bromoanisole, and the decay of
1
each portion of 3-bromoanisole was monitored by H and 31P
NMR spectroscopy in parallel in two instruments.
Instead of the reaction of the primary amine occurring with
a much shorter induction period than the reaction of the
secondary amine or with no induction period at all, as reported
in 2002, the reactions of the two amines with the same batch
of catalyst occurred with induction periods of nearly identical
length. An overlap comparison of the decay of the reactions of
octylamine and N-methylpiperazine is shown in Figure 3.
Further rate measurements by reaction calorimetry after
incubation of a mixture of the catalyst components with both
the amine and the base show that the activation behavior is
extremely sensitive to the conditions of incubation. Figure 4
(8) Rosner, T.; Pfaltz, A.; Blackmond, D. G. J. Am. Chem. Soc. 2001, 123,
(9) Data obtained in 2002 and in 2005 from MIT and Imperial College agree
within 5%.
4621.
9
J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006 3587

A R T I C L E S
Shekhar et al.
Figure 5. Comparison of the rates of reaction of 3-bromoanisole with
octylamine (0.93 M) or N-methylpiperazine (0.93 M) catalyzed by 10 mM
2 at 50 °C (top) and by 10 mM 3 with 2 mM added BINAP at 70 °C
(bottom). The reaction of octylamine is shown with red diamonds, and the
reaction of N-methylpiperazine is shown with blue squares.
Figure 6. Comparison of the rates of reaction of 3-bromoanisole (0.13 M)
catalyzed by 10 mM 2 at 50 °C (top) and by 10 mM 3 with 2 mM added
BINAP at 70 °C (bottom) with different concentrations of amine. The
reaction with 0.45 M amine is shown with blue squares, and the reaction
with 1.0 M amine is shown with red squares.
aryl halide in the presence of the two different amines along
with the apparent dependence of rate on the concentration of
secondary amine led to the suggestion that the amine must be
involved in the mechanism prior to the rate-limiting step.
This proposal was recently reevaluated by the Yale authors
with a set of new experiments that probed the dependence of
rate on the identity and concentration of the amine. Data were
sought that would reveal whether this difference in rate was
due to chemistry that occurred directly on the catalytic cycle
or from chemistry that occurred off of the catalytic cycle. To
study this, the rates of the reactions of primary and secondary
amines initiated with pure 2 and 3 as precatalyst were measured,
in this case by 1H NMR spectroscopy. If the difference in rate
were due to chemistry on the catalytic cycle, then the relative
rates of reactions of the two amines with these pure complexes
should mimic the relative rates of reactions of the two amines
with the catalyst generated in situ.
Figure 7. Decay of 3-bromoanisole during the sequential reactions of
3-bromoanisole with N-methylpiperazine catalyzed by 5.0 mM Pd2(dba)3
+ 10 mM BINAP at 60 °C. The initial concentration of N-methylpiperazine
was 0.93 M, and 0.13 M 3-bromoanisole was added five times. Legend:
blue, decay of the first portion of 3-bromoanisole; red, decay of the second
portion of 3-bromoanisole; green, decay of the third portion of 3-bromoani-
sole; black, decay of the fourth portion of 3-bromoanisole; pink, decay of
the fifth portion of 3-bromoanisole after addition of amine to generate a
concentration of N-methylpiperazine of 1.1 M.
The data on the reactions of 3-bromoanisole initiated with
pure 2 and with pure 3 contrasted with the reactions of PhBr
initiated with 1-IS. The decay of 3-bromoanisole during
reactions with octylamine was indistinguishable from that during
reactions with N-methylpiperazine when catalyzed by pure 2.
These data are shown at the top of Figure 5. These reactions
were conducted with 10 mM dimeric 2 as catalyst (2 mol %
overall, 8 mol % based on each portion of bromoarene).
Likewise, the decay of 3-bromoanisole during reactions with
octylamine was indistinguishable from that during reactions with
N-methylpiperazine when catalyzed by pure 3. These data are
shown at the bottom of Figure 5. These reactions were conducted
with 10 mM 3 as catalyst (2 mol % overall, 8 mol % based on
each portion of bromoarene) in the presence of 2 mM added
BINAP (0.4 mol % overall, 1.6 mol % based on each portion
of ArBr).
the reactions catalyzed by pure 2 and 3, the decays of aryl
bromide from reactions of 3-bromoanisole with 0.45 and 1.0
M concentrations of N-methylpiperazine in the presence of
NaOtAm catalyzed by 10 mM 2 (Figure 6, top), or 10 mM 3
with 2 mM added BINAP (Figure 6, bottom), were measured
at 50 °C. As shown clearly by the two plots in Figure 6, the
reactions with the 2-fold different concentrations of amine were
indistinguishable.
To test whether the decrease in rate measured for reactions
of secondary amines catalyzed by the mixture of [Pd2(dba)3]
and BINAP was due to a dependence of the rate on the
concentration of amine or other factors that retard catalytic
reactions, such as catalyst decay, product inhibition, or change
in medium, the rates of reactions of five portions of ArBr were
measured, but an amount of N-methylpiperazine was added to
the fifth portion of aryl bromide to increase the concentration
of amine 2-fold. These reactions are shown in Figure 7. The
decay of the first portion of bromoarene (shown in blue) is slow
because of the induction period discussed in section 2. Most
important for assessing the dependence of the rate on the
concentration of amine, the decay of ArBr after the concentration
of amine was increased 2-fold (shown in pink) is inconsistent
4. Evaluation of the Dependence of the Rate on the
Concentration of Amine. If the dependence of the rate of the
actual catalytic cycle depended on the concentration of amine,
then a dependence of the rate on the concentration of amine
should be observed during reactions initiated with the pure
catalyst precursors 2 and 3. To determine the order in amine of
9
3588 J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006

Amination of Aryl Halides Catalyzed by Pd−BINAP Complexes
A R T I C L E S
with a positive order in amine. This fifth portion of 3-bro-
moanisole did not decay with a faster rate than the fourth
portion. The same lack of an increase in rate was observed if
the additional N-methylpiperazine was added along with the
third portion of aryl bromide. These experimental results have
been confirmed by Blackmond, Buchwald, and co-workers.
These experiments show conclusively that the amination
reaction does not exhibit positive order kinetics in secondary
amine concentration. Instead, the apparent positive order in
amine can be attributed to catalyst deactivation occurring during
the reactions of secondary amines that were modeled. This
catalyst deactivation is supported by the 31P NMR spectroscopic
evidence provided in section 2, showing that the decrease in
concentration of 2 and 3 is more significant for secondary
amines compared to primary amines in similar consecutive
reaction sequences.
5. Evaluation of Concentration Dependences on PhBr and
BINAP in Reactions Catalyzed by [Pd(BINAP)2]. The data
shown in Figure 3 for reactions conducted with pure Pd-
(BINAP)2 in the absence of added ligand corroborate similar
observations reported by Blackmond, Buchwald, and co-workers
in ref 5. That report suggested that increasing rate could be
attributed to an increase in the active catalyst concentration
within the cycle over time. The slow step in the reaction
sequence catalyzed by [Pd(BINAP)2] in the absence of added
BINAP was suggested to be the introduction of [Pd(BINAP)]
from the off-cycle [Pd(BINAP)2] complex, in analogy to an
earlier example of false zero-order kinetics in Heck coupling.8
The anomalous zero-order kinetics shown at early stages of the
reaction of Figure 3 for the [Pd(BINAP)2] system with no added
ligand may thus be rationalized by an increase in the concentra-
tion of active catalyst at early stages of the reaction.
Figure 8. Reaction progress kinetic profiles for reactions of PhBr with
n-hexylamine at 60 °C after the induction period. (a) Reactions using Pd2-
dba3/BINAP mixtures at 0.014 M Pd or Pd(BINAP)2 at 0.01 M Pd with
0.002 M BINAP added: [PhBr]0 ≈ 0.13-0.15 M; [amine]0 ≈ 0.85 M;
[NaOtAm]0 ) 0.85 M. Data are plotted as rate vs fraction conversion of
PhBr. Magenta squares and blue circles, reactions using Pd2dba3/BINAP
mixtures from MIT laboratories in 2002 and 2005, respectively; green
triangles, reaction using Pd(BINAP)2 from Imperial College laboratories
in 2005. (b) Reactions carried out at same [“excess”] ) 0.70 M. Blue circles,
[PhBr]0 ) 0.31 M, [amine]0 ) 1.01 M, [NaOtAm]0 ) 1 M; magenta squares,
[PhBr]0 ) 0.15 M, [amine]0 ) 0.85 M, [NaOtAm]0 ) 0.85 M. Data from
Imperial College in 2005, plotted as reaction rate vs [PhBr]. Note that
reaction progress is from right to left in panel (b).
When the [Pd(BINAP)2]-catalyzed reaction is carried out with
2 mol % added BINAP, however, the induction period is not
observed, and the system obeys steady-state kinetics that are
zero order in the concentration of amine and close to first-order
in the concentration of ArX for both primary and secondary
amines, as was shown in Figures 5 and 6. This confirms that
the reactions show no dependence on the concentration or
identity of the amine for the [Pd(BINAP)2] catalyst system, just
as was found above for the Pd2dba3/BINAP system. This
observation of the absence of an induction period under these
conditions contrasts with the induction period described in 2002
by Buchwald, Blackmond, and co-workers that has not been
reproduced.
more reactive 2 appears to compensate for the lower concentra-
tion of 3 generated with the 1:1 ratio of metal to ligand. This
higher reactivity of 2 is illustrated by comparison of the rates
of reactions catalyzed by the two complexes at 50 and 70 °C in
Figure 6. The identical form of these rate curves is consistent
with a reaction through a common intermediate, regardless of
the provenance of the Pd precursor.
Figure 8b reveals the important point that the catalytic cycle
exhibits a steady state in catalyst concentration under the
conditions of these reactions.10 As described in a recent review
of reaction progress kinetic analysis,11 when “overlay” is
demonstrated as in Figure 8b between kinetic plots of rate vs
[ArX] from two experiments carried out at the same values of
[“excess”] ([“excess”] ) [amine]0 - [ArX]0), the absence of
an increase or decrease of [catalyst] within the cycle is
established.
Reaction calorimetric studies of reactions catalyzed by a
mixture of Pd2dba3 and BINAP, incubated with amine and base
prior to addition of aryl bromide and of [Pd(BINAP)2] under
conditions similar to the reactions shown in Figures 5 and 6,
illustrate a positive dependence on the concentration of ArBr
and reveal (1) that the reaction of primary amines with the two
catalysts gives a similar rate profile and (2) that the reaction
catalyzed by [Pd(BINAP)2] plus BINAP fortuitously occurs at
about the same rate per Pd added to the system as the reaction
catalyzed by the [Pd2(dba)3]/BINAP mixture (Figure 8a).
Although the 1:1 stoichiometry of metal to ligand allows for
only half the concentration of active catalyst to form, the
spectroscopic studies reported in the first section of this paper
show that a mixture of complexes 2 and 3 is formed from the
mixture of [Pd2(dba)3] and BINAP. Thus, the presence of the
The dependence of the rate on the concentration of aryl halide
revealed in Figures 6-8 contradicts the order in aryl halide from
a mechanism in which [Pd(BINAP)2] lies on the reaction
pathway. Rate constants determined in the studies on the
stoichiometric oxidative addition12 may be used to predict the
(10) If a set of reactions similar to those in Figure 6 had been carried out with
Pd2dba3/BINAP, a lack of overlay between the two curves could have
confirmed that catalyst deactivation, rather than positive order kinetics in
[secondary amine], accounted for the rate profiles observed in Figure 1b
of ref 5.
(11) Blackmond, D. G. Angew. Chem., Int. Ed. 2005, 44, 4302.
(12) These data were obtained as part of concurrent work on the oxidative
addition of bromoarenes to [Pd(BINAP)2] in the presence and in the absence
of amine: Shekhar, S.; Ryberg, P.; Hartwig, J. F. Org. Lett. 2006, 8, 851.
9
J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006 3589

A R T I C L E S
Shekhar et al.
Scheme 7
Figure 9. Decay of 3-bromoanisole during the catalytic reaction of
3-bromoanisole with N-methylpiperazine catalyzed by 10 mM [Pd(BINAP)2]
in the presence of 2 mM BINAP, predicted by the rate constants obtained
from the study of stoichiometric oxidative addition of 3-bromoanisole to
[Pd(BINAP)2] at 70 °C.
the amine adds bromoarene to generate a five-coordinate Pd-
(II) product or the product amine is displaced associatively by
BINAP from the 16-electron complex [Pd(BINAP)(product)]
to regenerate [Pd(BINAP)2]. The rate equation for the mecha-
nism in Scheme 7 is provided in eq 1. This equation was derived
k1k2
-d[ArX]
dt
)
[ArBr][Pd]
(1)
k-1[L]
using the steady-state approximation and simplifications based
on experimental knowledge of the relative rate constants of
different steps from studies of stoichiometric oxidative addition
and reactivity of arylpalladium halide and arylpalladium amido
complexes. The derivation of this equation and the approxima-
tions to simplify the full rate equation are provided in Supporting
Information. The steady-state approximation is valid in this case
because no intermediates accumulate, as determined by 31P
NMR spectroscopy. Further, numerical simulation of the pal-
ladium species in the catalytic system using known and
estimated rate constants for the individual steps that lead to a
close fit to the experimental decay of bromoarene shows that
the palladium complexes within the cycle by this mechanism
would not accumulate. These equations demonstrate that a
reaction by the pathway in Scheme 7 will be zero-order in
amine, first-order in bromoarene, and inverse first-order in added
ligand. The equilibrium relationship between 3 and PdL
manifested by the rate law for Scheme 7 holds for any case
where a reversible “dead-end” step is linked to a steady-state
catalytic cycle.13 Equation 1 indicates that equilibrium may be
established between a species off the cycle, such as 3, and a
species within the catalytic cycle, such as PdL, at concentrations
that would provide a non-equilibrium steady-state concentration
of PdL in an identical stoichiometric reaction sequence.
Figure 10. Dependence of the rate of the reaction of 0.10-2.7 M
bromobenzene with 13 mM hexylamine catalyzed by 1.26 mM 3 with 2.3
mM BINAP at 50 °C on the concentration of PhBr (top) and of the rate of
the reaction of 2.0 M bromobenzene with 13 mM hexylamine catalyzed by
1.26 mM 3 with 2.3-9.2 mM BINAP on the concentration of BINAP
(bottom).
decay of bromoarene by a pathway with [Pd(BINAP)2] on the
reaction pathway under similar conditions. Figure 9 shows that
the catalytic reaction is much faster than predicted for this
mechanism from the rate constants of the study of oxidative
addition and exhibits positive order kinetics at concentrations
of [ArX] and [BINAP] where the mechanism with [L2Pd] on
the cycle should begin to be saturated in [ArX].
These inconsistencies led the Yale group to repeat the
measurements of the order of the catalytic reaction in bro-
moarene and added ligand from ref 3. Inspection of the primary
data and laboratory records makes clear that an error had been
made when conducting the experiments that led to observation
of the same reaction rate in experiments stated to have varied
concentrations of ligand and bromobenzene. The set of experi-
ments by the experimentalist in the current work (Figure 10)
clearly shows that the reaction is first-order in bromoarene and
inverse first-order in ligand. This is contrary to the previous
report, where the kinetic data implied that the rate of the
oxidative addition process and the rate of the catalytic cycle
both depended only on the rate of dissociation of BINAP.
6. Mechanistic Conclusions. A pathway is shown in Scheme
7 in which the bromoarene reacts with [Pd(BINAP)], and [Pd-
(BINAP)] is generated by dissociation of the product amine,
rather than the pathways in which a palladium(0) complex of
Several lines of data indicate that the major pathway for the
catalytic process is the mechanism in Scheme 7. First, the zero-
order dependence of the rate on the concentration of amine,
first-order dependence of the observed rate constants on the
concentration of bromoarene, and inverse dependence on the
concentration of ligand measured most recently are consistent
with the rate equation for reaction by the mechanism in Scheme
7, but are inconsistent with the rate equations for mechanisms
(13) Cornish-Bowden, A. In Fundamentals of Enzyme Kinetics; Portland Press:
London, 1995; p 88.
9
3590 J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006

Amination of Aryl Halides Catalyzed by Pd−BINAP Complexes
A R T I C L E S
in Schemes 1 and 2. Second, the observation of an induction
period during reactions conducted with [Pd(BINAP)2] as catalyst
in the absence of any added ligand makes it unlikely that [Pd-
(BINAP)2] lies directly on the reaction pathway. Third, the
predicted decay of bromoarene by the mechanism in Scheme 7
from the rate constants measured independently by studying the
stoichiometric oxidative addition12 accounts for a majority of
the rate constant of the observed catalytic process. These data
provide strong evidence that a mechanism in which bromoarene
reacts with [Pd(BINAP)], and [Pd(BINAP)] is generated by [Pd-
(BINAP)2] lying off the cycle, is the major pathway. One could
also consider that the difference between predicted and observed
rates leaves open the potential that one or more additional minor
pathways contribute to the total observed rate constant or that
the rate constants measured in the medium of the oxidative
addition process are slightly different from those for the same
reactions in the medium of the catalytic system that contains
base as well as alcohol and salt byproducts.
Most generally, these studies show the importance of com-
bining rate data on a catalytic system with both spectroscopic
measurements and kinetic data on stoichiometric reactions that
provide information about the mechanism of the individual steps
of the catalytic cycle. A comparison of the saturation in kobs at
high bromoarene concentration in the stoichiometric oxidative
addition and the first-order dependence on bromoarene in the
catalytic pathway demonstrates that the reaction orders of linear
and cyclic reaction sequences can differ when they include a
common reversible step that lies off the catalytic pathway. The
deepest insight into the mechanism of a catalytic process is
obtained by a series of techniques, including catalytic rate
measurements, methods that identify the components present
in the catalytic system, and quantitative and qualitative measure-
ments of rates of stoichiometric reactions under wide ranges of
concentrations.
Acknowledgment. J.F.H. thanks the NIH (GM-55382) for
support and Johnson-Matthey for palladium. S.L.B. thanks the
NIH (GM-58160) for funding. D.G.B. thanks Pfizer Global
Research for an unrestricted research grant. All authors thank
Prof. David B. Collum (Cornell) and an anonymous referee for
invaluable input.
Summary
Studies of the identity of the palladium(0) species in the
reactions of aryl halides with amines catalyzed by a complex
generated in situ, and studies of the rates of reactions catalyzed
by pure palladium components in this system, have shown that
the palladium-catalyzed amination occurs by a mechanism in
which the bromoarene undergoes oxidative addition prior to
reaction with amine (Scheme 7), and not by a pathway in which
amine coordinates to the palladium(0) prior to oxidative addition,
as had been postulated in Scheme 2. At the same time, new
data on the order of the reactions of [Pd(BINAP)2] with added
BINAP in the concentration of ArBr and BINAP show that [Pd-
(BINAP)2] lies off the catalytic cycle, and not directly on the
cycle, as had been postulated in Scheme 1.3
Note Added after ASAP Publication. After this paper was
published ASAP on March 1, 2006, corrections were made to
Scheme 7 and in the next-to-last sentence in the last paragraph
of section 2 of the Results. The corrected version was published
ASAP March 6, 2006.
Supporting Information Available: Experimental procedures
and raw kinetic data (PDF, CIF). This material is available free
of charge via the Internet at http://pubs.acs.org.
JA045533C
9
J. AM. CHEM. SOC. VOL. 128, NO. 11, 2006 3591