Acceleration of Pd-Catalyzed Amide N-Arylations Using Cocatalytic Metal Triflates: Substrate Scope and Mechanistic Study

Article abstract of DOI:10.1021/acscatal.7b01317

The Pd/xantphos-catalyzed cross-coupling of amides and aryl halides is accelerated by cocatalytic metal triflate additives. A survey of nitrogen nucleophiles reveals improved yields for a variety of N-aryl amide products when Al(OTf)₃ is employed as a catalytic additive, with some exceptions. Initial rates of catalysis indicate that the Lewis acid acceleration is more pronounced when bromobenzene (PhBr) is used in comparison with iodobenzene (PhI). The observation of an aryl halide dependence on rate and various qualitative kinetic experiments are consistent with a mechanism in which ligand exchange of halide for amide ("transmetalation") is turnover limiting. The mechanism may be different depending on whether PhBr or PhI is used as a coupling partner. Oxidative addition complexes (xantphos)Pd(Ph)(X) (X = Br, I; xantphos = 4,5-bis(diphenylphosphino)-9,9-dimethylxanthene), likely intermediates in catalysis, have been prepared; their differing interactions with Yb(OTf)₃ in solution resemble the halide dependence of the catalytic mechanism, which we propose originates from a reversible Lewis acid mediated halide abstraction during catalysis.

Full text of DOI:10.1021/acscatal.7b01317

Research Article
pubs.acs.org/acscatalysis
Acceleration of Pd-Catalyzed Amide N‑Arylations Using Cocatalytic
Metal Triflates: Substrate Scope and Mechanistic Study
Joseph Becica and Graham E. Dobereiner*
Department of Chemistry, Temple University, Philadelphia, Pennsylvania 19122, United States
S
* Supporting Information
ABSTRACT: The Pd/xantphos-catalyzed cross-coupling of amides and aryl
halides is accelerated by cocatalytic metal triflate additives. A survey of nitrogen
nucleophiles reveals improved yields for a variety of N-aryl amide products when
Al(OTf)3 is employed as a catalytic additive, with some exceptions. Initial rates of
catalysis indicate that the Lewis acid acceleration is more pronounced when
bromobenzene (PhBr) is used in comparison with iodobenzene (PhI). The
observation of an aryl halide dependence on rate and various qualitative kinetic
experiments are consistent with a mechanism in which ligand exchange of halide for amide (“transmetalation”) is turnover
limiting. The mechanism may be different depending on whether PhBr or PhI is used as a coupling partner. Oxidative addition
complexes (xantphos)Pd(Ph)(X) (X = Br, I; xantphos = 4,5-bis(diphenylphosphino)-9,9-dimethylxanthene), likely intermediates
in catalysis, have been prepared; their differing interactions with Yb(OTf)3 in solution resemble the halide dependence of the
catalytic mechanism, which we propose originates from a reversible Lewis acid mediated halide abstraction during catalysis.
KEYWORDS: palladium, cross coupling, amides, aryl halides, metal triflates, Lewis acids, transmetalation
INTRODUCTION
■
Pd-catalyzed C−N bond formation has had a major effect on
organic synthesis, and efforts to improve Pd couplings have
continued to yield increasingly valuable methods for synthetic
chemists.1−4 The coupling of primary and secondary amide,
sulfonamide, carbamate, and urea nucleophiles has found
important applications in the small- and large-scale synthesis
of bioactive molecules including drug candidates, medicinally
relevant heterocycles, and natural products, as well as valuable
products for materials science (photoswitches and fluorescent
probes, among others).5 Few methodologies broadly enable the
coupling of amide nucleophiles,6 and there is interest in the
development of more efficient protocols for couplings of these
substrates, given the prevalence of amide linkages in biological
and functional molecules.7
Advances in amide N-arylation have been largely made
possible by the design of novel phosphine ligands with special
properties, including wide bite angle bisphosphines8−11 and
sterically hindered12−15 or electron-deficient6,16 phosphines.
However, weakly nucleophilic amides and hindered amides
(such as secondary acyclic amides) remain challenging
substrates for conventional N-arylation methods. Pd catalysts
featuring electron-deficient ligands enable efficient coupling of
challenging amide substrates, but only with a limited scope of
electrophiles.6 Therefore, more generalized protocols for Pd
amide couplings continue to be sought.
Figure 1. Proposed mechanism of Pd-catalyzed cross-coupling of
amides and aryl halides.
C−N bond-forming reductive elimination from D generates the
N-arylated amide product. Bidentate or sterically hindered
phosphines have been shown to prevent catalytically inactive κ2
coordination of the amidate ligand (E),22 and the chelating 4,5-
The mechanism of amide N-arylation17−21 (Figure 1) is
believed to proceed via oxidative addition of an aryl halide or
sulfonate (PhX) to Pd0 (A). Next, ligand exchange of the X
ligand for an amidate ligandcommonly referred to as
“transmetalation”involves amide coordination to B and
base-mediated dehydrohalogenation to generate D. Subsequent
Received: April 24, 2017
Revised: July 19, 2017
© XXXX American Chemical Society
5862
DOI: 10.1021/acscatal.7b01317
ACS Catal. 2017, 7, 5862−5870

ACS Catalysis
Research Article
bis(diphenylphosphino)-9,9-dimethylxanthene (xantphos) is
commonly employed for this class of reactions.23,24
evaluate possible Lewis acid effects in catalysis. Initial
observations in our laboratory indicated that Lewis acid
acceleration is general for Pd-catalyzed amide N-arylation,
including heteroaromatic and homocyclic aryl halide substrates.
We discovered that the yields within 2 h of reaction time for the
coupling of bromobenzene (PhBr) or iodobenzene (PhI) with
pyrrolidin-2-one (1) to generate N-phenylpyrrolidin-2-one (2)
using 2 mol % Pd/xantphos catalyst and K3PO4 base in
refluxing toluene (PhMe) are improved by the addition of 5
mol % of Lewis acids.31 A screen of additives yielded optimized
reaction conditions (Scheme 1).
The turnover-limiting step of amide N-arylation is often
transmetalation,6,13 most likely due to the weak binding of
amides to Pd. In contrast to ligand effects, amide coordination
to Pd could instead be encouraged by increasing the
electrophilicity of key intermediates in situ through secondary
coordination sphere interactions.25 Related effects shown in
several notable reports demonstrate that Lewis acid additives
enable challenging transition-metal-mediated reactions, partic-
ularly reductive elimination (Figure 2).26−29 Nolan and
Scheme 1. Initial Screen of Lewis Acids in the Pd-Catalyzed
a
Coupling of Aryl Halides and Pyrrolidin-2-one
a
Reaction conditions: 1.0 equiv of pyrrolidin-2-one (0.2 M [1]), 1.1
equiv of aryl halide, 1.2−1.5 equiv of K3PO4, 2 mol % of Pd(dba)2, 2
Figure 2. Key examples of Lewis acid enabled reductive elimination
reactions.26−28
mol % of xantphos, 5 mol % of M(OTf)3, 2 mL of anhydrous toluene,
1
110 °C, 2 h. The yield was quantified by H NMR of the amide
product versus an internal standard (1,3,5-trimethoxybenzene).
Moloy26 reported that adducts of a Lewis acid to PdII cyanide
complexes accelerate the reductive elimination of nitriles, while
Hartwig and Shen27 reported that adducts of Lewis acids to PdII
amidate complexes accelerate the reductive elimination of N-
heteroarylamides. In both cases, the rate of reductive
elimination from Pd increases with the electrophilicity of the
Lewis acid, where adduct formation apparently encourages
reductive elimination by forming a more electrophilic metal
intermediate. Similarly, Bergman and Tilley28,29 demonstrated
that remote binding of a Lewis acid dramatically increases the
rate of reductive elimination of electron-deficient biaryls from
PtII. Efforts in our laboratory have focused on understanding
how these stoichiometric effects, among others, may benefit the
catalytic coupling of amides and aryl halides or heteroaryl
halides, similar to the Lewis acid effects identified by Hartwig
and Shen.27 The Hartwig−Shen study identified a stoichio-
metric Lewis acid induced reductive elimination reaction, as
well as a Lewis acid cocatalytic effect in amide N-
heteroarylation reactions using a Pd/xantphos catalyst. Several
heteroaryl halides were coupled with lactams, benzamides, and
secondary sulfonamides in significantly higher yields when
triethylboron (BEt3) was added to the reaction mixture.
We report here that metal triflate additives accelerate amide
N-arylation using a Pd/xantphos catalyst. The acceleration
effect is general for a wide variety of N-aryl amide products. On
the basis of a mechanistic study, Lewis acids appear to
accelerate the reaction by lowering the barriers to turnover-
limiting transmetalation.
A range of nitrogen nucleophiles were evaluated, including
primary benzamides, primary aliphatic amides, secondary
anilides, secondary acyclic amides, carbamates, and indoles
(Chart 1). The following sections describe the yield enhance-
ments observed upon adding Lewis acids to N-arylation
reactions of various amide substrates. Al(OTf)3 was primarily
employed as the Lewis acid, being the most inexpensive metal
triflate and effective in accelerating the formation of 2 (Scheme
1). In general, adding 5 mol % of Al(OTf)3 provided significant
increases in reaction yield (20 examples, as much as a 45% yield
increase). In our survey of amides, the Pd to Lewis acid ratio
was not individually optimized for each substrate. K3PO4 was
found to be a suitable base for the coupling of primary
benzamides or lactams, while reactions of other substrates
required Cs2CO3. In select cases where the addition of 5 mol %
of Al(OTf)3 to a 2 mol % Pd/xantphos catalyst mixture resulted
in a decrease in yield, an increase in yield could be observed
when a lower 1:1 Al:Pd ratio was employed. This is in concert
with our mechanistic findings: significantly higher Lewis acid
concentrations result in a reduced rate of catalysis (vide infra).
Cyclic Amides. Despite the reduced nucleophilicity of the
cyclic carbamate oxazolidin-2-one in comparison to the
analogous lactam 1, this substrate (and variants) have been
successfully employed in Pd-catalyzed N-arylation reactions
with various ligands and a range of electrophiles.22,32,33 Under
our conditions, the reaction of oxazolidin-2-one and PhBr with
2 mol % of Pd/xantphos and 5 mol % of Al(OTf)3 generates N-
phenyloxazolidin-2-one (3) in quantitative yielda 41%
increase in yield versus the reaction without added Lewis
acid. The performance of our catalyst without Lewis acid (58%
yield) is similar to that of reported catalyst systems.32
RESULTS AND DISCUSSION
■
Optimization of Catalysis and Substrate Scope. To
yield the broadest and most accessible synthetic protocol, we
chose widely available phenyl halides, the relatively inexpensive
xantphos ligand, and moisture-tolerant metal triflate salts30 to
The reaction of 3-bromopyridine with oxazolidin-2-one
generated N-3-pyridyloxazolidin-2-one (5) in reduced yield
5863
DOI: 10.1021/acscatal.7b01317
ACS Catal. 2017, 7, 5862−5870

ACS Catalysis
Research Article
f
Chart 1. Survey of Lewis Acid Effect in Pd-Catalyzed N-Arylation with Various Nitrogen Nucleophiles
a
b
c
d
K3PO4 was used rather than Cs2CO3. Isolated yield. 5 mol % of Pd(dba)2 and 5 mol % of xantphos were used. 5 mol % of Pd(dba)2, 5 mol % of
e
f
xantphos, and 5 mol % of Al(OTf)3 were used. 2.5 equiv of aryl halide and 3 equiv of Cs2CO3 were used. Reaction conditions: 1.0 equiv of amide,
1.1 equiv of aryl halide, 1.2−1.5 equiv of Cs2CO3, 2 mol % of Pd(dba)2, 2 mol % of xantphos, 5 mol % of M(OTf)3, 2 mL of anhydrous toluene, 110
°C, 18 h. Unless otherwise noted, the yield was quantified by 1H NMR of the amide product versus an internal standard (1,3,5-trimethoxybenzene).
The bold bond indicates the C−N bond(s) formed.
when a Lewis acid was used, in contrast to increased yields in
the N-heteroarylation of 1 with 3-bromopyridine. Consistent
with Hartwig and Shen’s study of amide N-heteroarylation,27
increased yields of N-3-pyridylpyrrolidin-2-one (7, 81% yield
with 5 mol % Sc(OTf)3, 39% yield increase) are observed.
4-Methylbenzamide. Various protocols have been devel-
oped to selectively N-arylate primary benzamides using Pd
catalysts.15,21,23,34,35 Under our conditions, the N-arylation of 4-
methylbenzamide proceeded to a 50% yield of N-phenyl-4-
methylbenzamide (4) using 2 mol % Pd/xantphos and
improved to 64% of 4 using 2 mol % of Pd/xantphos and 5
mol % of Al(OTf)3. A larger increase in yield was observed in
the case of coupling 4-methylbenzamide with 3-bromopyridine.
Addition of Al(OTf)3 afforded 5 in 94% yield (44% yield
increase), again consistent with the Lewis acid acceleration
observed by Hartwig and Shen.27
N-Methyl-p-toluenesulfonamide. Buchwald reports that the
reaction of N-methyl-p-toluenesulfonamide is successfully
coupled with the electron-rich 3,5-dimethylbromobenzene
using 1 mol % of Pd/xantphos in refluxing PhMe (0.125 M)
with Cs2CO3 base with long reaction times (38 h).24 In our
hands, the reaction of N-methyl-p-toluenesulfonamide with the
less electron rich PhBr yields N-methyl-N-phenyl-p-toluene-
sulfonamide (7, 8% yield) using 2 mol % of Pd/xantphos in
refluxing PhMe (0.2 M) with Cs2CO3. At different catalyst and
Lewis acid loadings (5 mol % of Pd/xantphos/Al(OTf)3), 8
was formed in 72% yield (20% increase with respect to 5 mol %
Pd/xantphos). Additionally, N-methyl-N-3-pyridyl-p-toluene-
sulfonamide was formed in 54% yield (5 mol % Pd/
xantphos/Al(OTf)3, 29% yield increase with respect to 5 mol
% Pd/xantphos).
α-4-Fluorophenylacetamide. Carbonyl substrates with α
hydrogens are susceptible to α-arylation,36 yet primary α-
phenylacetamide substrates have been employed in Pd coupling
cascade reactions to prepare naphthyridinones and quinoli-
nones under weakly basic conditions.37 Here, N-phenyl-α-4-
fluorophenylacetamide (10) and N-3-pyridyl-α-4-fluoropheny-
lacetamide (11) required a higher catalyst loading (5% of Pd/
xantphos) to achieve good yields. Yields of 10 and 11 were
comparable when Al(OTf)3 was employed (61% and 64%,
respectively). Indeed, the N,α-diarylation product (11b) was
observed as a byproduct in reactions which generated 11 (∼2%
yield) and the formation of 11b was greater (10%) in the
presence of 5 mol % of Al(OTf)3 (Scheme 2). The analogous
N,α-diarylation product was not observed in reactions which
generated 10, nor were N,N-diarylation products.
5864
DOI: 10.1021/acscatal.7b01317
ACS Catal. 2017, 7, 5862−5870

ACS Catalysis
Research Article
diarylation (12c), and N,N,α-triarylation (12d) products (the
minor products could not be isolated), and the yields of these
products were determined by 19F NMR.38 When this reaction
was repeated with 5 mol % of Al(OTf)3, 12 was selectively
formed (72% yield). The reaction of levetiracetam, a primary
aliphatic amide substrate with a hindered CH bond α to the
amide carbonyl, with 3-bromopyridine selectively produced the
mono-N-arylation product 13 (66% yield, 2 mol % of Pd/
xantphos and 5 mol % of Al(OTf)3, 22% yield increase).
Secondary Acyclic Amides. To the best of our knowledge,
only Pd catalysts featuring the ligand JackiePhos are broadly
competent for N-arylation of secondary acyclic amides.6
Coupling of secondary acyclic amides is otherwise sluggish
unless strongly activated aryl halides (such as 4-cyanobromo-
benzene and 4-nitroiodobenzene) are used.24,39,40 In the
absence of these features, Buchwald reports the N-arylation
of N-methylacetamide with PhBr, generating N-methyl-N-
phenylacetamide (14, 73% yield) over 37 h using 4 mol % of
Pd/xantphos with 1,4-dioxane solvent (0.25 M) and Cs2CO3
base.24 Under our conditions, 14 is produced in 48% yield over
18 h (5 mol % of Pd/xantphos/Al(OTf)3, PhMe solventa
22% yield increase with respect to 5 mol % of Pd/xantphos).
The reaction of N-methylacetamide with methyl 4-bromoben-
zoate generates 15 in 89% yield (2 mol % of Pd/xantphos and 5
Scheme 2. N-Arylation and α-Arylation Products Observed
in Reactions with α-4-Fluorophenylacetamide
The reaction of α-4-fluorophenylacetamide with 2.5 equiv of
methyl 4-bromobenzoate using 2 mol % of Pd/xantphos, on the
other hand, produced a roughly 2:1:1:1 distribution of four
products, the major product of which is the N-monoarylated
product 12 (41% yield; Scheme 2). The minor products are
tentatively identified as the N,α-diarylation (12b), N,N-
Figure 3. Comparison of selected catalytic reaction profiles: (A) N-arylation of 1 as a function of aryl halide; (B) N-arylation (PhBr) of 1 as a
function of Lewis acid (1.5 mol %) or NBu4Br (1 equiv); (C) N-arylation (PhI) of 1 as a function of Lewis acid; (D) N-arylation (PhBr) of 1 as a
function of Lewis acid concentration. Trend lines representing initial rates for selected trials are superimposed.
5865
DOI: 10.1021/acscatal.7b01317
ACS Catal. 2017, 7, 5862−5870

ACS Catalysis
Research Article
mol % of Al(OTf)3, 45% yield increase with respect to 2 mol %
of Pd/xantphos) while the reaction of N-methylacetamide with
3-bromopyridine generated 16 in 54% yield (5 mol % of Pd/
xantphos/Al(OTf)3, 18% yield increase with respect to 5 mol %
of Pd/xantphos). For 16, adding a metal triflate allows for
reduced Pd loadings (and comparable conversion) in
comparison to the Lewis acid free reactions. 16 is generated
in 34% yield with 2 mol % of Pd/xantphos and 5 mol % of
Al(OTf)3roughly the same yield as when the reaction is
performed with 5 mol % of Pd/xantphos and no Lewis acid.
The reaction of N-butylpropionamide and N-phenylpropiona-
mide (larger secondary acyclic amides) with methyl 4-
bromobenzoate, forming 17 and 18, shows a decrease in yield
when 5 mol % of Al(OTf)3 was employed. No N-arylation
product was observed in the reaction of N-butylpropionamide
or N-phenylpropionamide with PhBr.
suggested by Hartwig and Shen.27 When pyridyl halide
electrophiles are used, a Lewis acid/base adduct of the pyridyl
nitrogen in LPd(Py)(NR2) is possible (as in Figure 2b);
however, such an adduct is not possible with phenyl halide
electrophiles. This suggests that the triflate salt has various
potential roles in the catalyst system. With unsubstituted aryl
halides, turnover-limiting transmetalation in amide N-arylation
via Pd has been invoked,6,13 and our data are consistent with a
turnover-limiting transmetalation step. The universal Pd
intermediate involved in reductive elimination for electrophiles
PhBr, PhI, and PhOTf(xant)Pd(Ph)(NR2)should in
principle undergo a unimolecular reductive elimination at the
same rate regardless of which electrophile were used; however,
the differences in rate behavior in Figure 3A suggest a halide
influence in the turnover-limiting step.13 Further, 31P NMR
spectra of the catalytic reaction mixtures reveal (xant)Pd(Ph)-
(X) (22, X = Br; 23, X = I) to be the major phosphorus-
containing component of the resting state mixture of
complexes. Finally, a mechanism in which oxidative addition
of electrophile is rate-limiting for this catalyst system is possible
but would be inconsistent with extensive literature prece-
dent.4,6,13,16,23,24,27
Melatonin, a Substrate with Multiple Functional Groups.
Melatonin has two secondary nitrogens: an indole and a
secondary acyclic amide site, both of which could potentially
undergo N-arylation. We note that indole N-arylation via Pd or
Cu is another area of cross-coupling with practical limitations.15
In our case, the reaction of melatonin with PhBr results in
selective N-arylation at the indole nitrogen (19) and proceeds
to higher yield in the presence of 5 mol % of Al(OTf)3 (60%
yield, 30% increase), indicating that Lewis acid additives can
accelerate coupling rates for nitrogen nucleophiles beyond
amide substrates. No N-arylation products at the amide site on
The catalysis was then repeated with metal triflate additives
(Figure 3B) using a 1:1 Lewis acid:Pd ratio. Initial rates were
indeed increased in the reaction of PhBr with 1. The most
dramatic increase in reactivity was observed upon the addition
of 1.5 mol % of Sc(OTf)3 (ν = (36 21) × 10−4 M min−1) and
1.5 mol % Yb(OTf)3 (ν = (36 6.4) × 10−4 M min−1) (Table
1).41 In contrast, the addition of metal triflates did not produce
1
melatonin were observed by H NMR when using PhBr, even
with excess PhBr (3 equiv) or after long reaction times (48 h).
However, the reaction of melatonin with the more electrophilic
methyl 4-bromobenzoate (2.5 equiv) resulted in diarylation at
both nitrogen positions. Without Lewis acid, the indole N-
arylation product (20) and the N,N′-diarylation product (21)
are formed in roughly 1:1 yield (20, 42%; 21, 38%). However,
when the reaction is performed with 5 mol % of Al(OTf)3,
moderate selectivity for 20 is observed (20, 67%; 21, 27%).
In summary, we observe that metal triflate additives can
improve yields and selectivity for various N-arylation reactions
or allow for reduced Pd loadings. Use of a heteroaryl halide is
not a requirement for an apparent accelerating effect, although
various N-3-pyridylamide products have particularly significant
increases in yield when Al(OTf)3 is employed. The influence of
Al(OTf)3 is general for a variety of substrates, but additives
reduce the overall reaction yield in some cases.
Table 1. Observed Initial Rates of Catalysis for Pd-Catalyzed
a
Coupling of Aryl Halides and 1
entry
aryl halide
Lewis acid
initial rate (10−4 M min−1
)
1
PhBr
PhI
N/A
14 1.3
9.5 2.4
4.7 1.6
18 5.5
36 21
32 21
36 6.4
26 4.5
6.2 2.6
12 4.4
16 3.0
2
N/A
3
PhOTf
PhBr
PhBr
PhBr
PhBr
PhBr
PhI
N/A
4
In(OTf)3
Sc(OTf)3
Al(OTf)3
Yb(OTf)3
Zn(OTf)2
In(OTf)3
Zn(OTf)2
Yb(OTf)3
5
6
7
8
9
10
11
PhI
PhI
Mechanistic Investigation. To quantify the effect of Lewis
acids, we first performed kinetic measurements of cross-
coupling under various conditions. Given the difference in
reactivity between PhBr and PhI in our system (represented in
Scheme 1), we first determined the relative initial rates of
catalysis in the absence of Lewis acids using PhBr, PhI, and
phenyl triflate (PhOTf) as substrates (Figure 3a). The reaction
of 1 and PhBr (1.1 equiv) using Pd(dba)2 (1.5 mol %),
xantphos (1.5 mol %), and K3PO4 (1.2 equiv) in PhMe (5.5
mL) proceeded to completion in approximately 9 h at 110 °C.
The initial reaction profile shows that the concentration of the
amide product 2 increases linearly with time, consistent with
pseudo-zero-order kinetics with an observed initial rate of (14
1.3) × 10−4 M min−1. The analogous reactions employing
PhI and PhOTf are slower (ν = (9.5 2.4) × 10−4 and (4.7
1.6) × 10−4 M min−1, respectively) and begin to decrease in
rate within 1 h.
a
Reaction conditions: 1.0 equiv of pyrrolidin-2-one (0.2 M [1]), 1.0
equiv of aryl halide, 1.2 equiv of K3PO4, 1.5 mol % of Pd(dba)2, 1.5
mol % of xantphos, 1.5 mol % of M(OTf)3, 5.5 mL of anhydrous
toluene, 110 °C. Values given are averages and standard deviations of
three trials.
a significant increase in rate for arylations using PhI (Figure
3C). The addition of 1.5 mol % of Yb(OTf)3 increased the
initial rate slightly to (16 3.0) × 10−4 M min−1 (from (9.5
2.4) × 10−4 M min−1); however, the addition of 1.5 mol % of
In(OTf)3 produced a decrease in initial rate (ν = (6.2 2.6) ×
10−4 M min−1). In general, further increases in the Lewis
acid:Pd ratios resulted in a decrease in rate (Figure 3D).
The observed differences in Lewis acid acceleration effects
(PhBr vs PhI) lead us to believe that the transmetalation
mechanism may vary as a function of halide in the LPdPhX
intermediate. X− to NR2 ligand substitution may begin by
−
In cases where strongly deactivated heteroaryl halides are
used, reductive elimination may be turnover-limiting, as
amide association to LPd(Ph)(X), forming the five-membered
5866
DOI: 10.1021/acscatal.7b01317
ACS Catal. 2017, 7, 5862−5870

ACS Catalysis
Research Article
intermediate LPd(Ph)(X)(HNR2), or by a dissociative ligand
exchange via halide abstraction from LPd(Ph)(X), forming a
cationic LPd(Ph)+ species.6 (Figure 4) Assuming trans-
irreproducible, but when [1] was reduced, a slight decrease in
rate was observed (ν = (8.7
1:PhI)). Additional experiments involving exogenous salts
(NBu4I or NBu4OTf) resulted in biphasic reaction mixtures,
which precluded a quantitative kinetic analysis.
2.5) × 10−4 M min−1 (0.5:1
These observations support a mechanism in which trans-
metalation is turnover-limiting, given (1) the rate of catalysis is
dependent on the identity of aryl halide, (2) Lewis acid
acceleration is pronounced for PhBr reactions, but not PhI
reactions, (3) catalysis is inhibited by exogenous halide ions,
and (4) reactions using PhBr are zero order in [PhBr] and [1],
whereas reactions using PhI are more complex. In principle, the
sequence of amide coordination and dehydrohalogenation steps
is terminated by irreversible deprotonation of Pd-bound 1.
Therefore, the efficiency at which this step occurs would be
dependent on a number of pre-equilibria, which include
dissociation of X− from (xant)Pd(Ph)(X), coordination of 1,
dissociation of base, deprotonation of Pd-bound 1, and
precipitation of a HX:base complex. (Figure 5).
Figure 4. Two possible mechanisms for transmetalation: rate-limiting
amide coordination and rate-limiting halide dissociation.
metalation is turnover-limiting, the mechanisms could be
distinguished by determining the catalytic rate dependence
upon concentration of amide and of exogenous halide. An
associative mechanism would be promoted by increased
[amide], while a mechanism initiated by reversible halide
dissociation would be inhibited by increased [X−].
The initial rate of catalysis was then determined under
various conditions to establish reaction orders (Table 2). In the
Table 2. Observed Initial Rates of Catalysis for Pd-Catalyzed
Coupling of Aryl Halides and 1a
aryl
halide
modifications to std
initial rate
a
entry
conditions
(10−4 M min−1
)
1
2
3
4
5
6
7
8
9
PhBr
PhBr
PhBr
PhBr
PhBr
PhI
14 1.3
13 5.0
13 2.0
7.5 2.5
13 3.9
9.5 2.4
12 3.6
16 5.0
8.7 2.5
2 equiv PhBr
2 cquiv 1
Figure 5. Pre-equilibria potentially relevant to turnover-limiting
transmetalation.
1 equiv NBu4Br
1 equiv KOTf
Inhibition of PhBr catalysis by exogenous NBu4Br is
consistent with dissociation of Br− from (xant)Pd(Ph)(Br) as
a key step and may precede coordination of 1. For the case of
PhI the observed rate dependence on [1] is complicated by the
influence of [1] on the polarity of the reaction medium, a
known factor in amide N-arylation couplings.42 Nonetheless, it
is evident from our investigation that the halide identity
influences the various transmetalation equilibria and it is
possible that reactions employing PhI are more sensitive to
solution polarity than those employing PhBr (Table 2, entries 7
and 8). Additionally, the yields of reactions employing PhI are
also more variable in a screen of bases in comparison to those
of PhBr.43 We note that it is possible a Lewis acid adduct of 1
would be activated toward transmetalation (such as a metal
amidate complex), but we do not presently have experimental
evidence to support this claim, and such an effect should be
halide independent. Furthermore, increasing the Lewis acid
loading beyond 1:1 LA:Pd to 3:1 and 6:1 decreases the rate of
catalysis, where LA activation of 1 should further improve with
increased [LA].
Interaction of Oxidative Addition Complexes with
Metal Triflates. To better understand the interactions of metal
triflates with Pd intermediates prior to transmetalation, we
prepared (xant)Pd(Ph)(X) (22, X = Br; 23, X = I)22,44−49 and
studied their solution behavior with Lewis acids. Complexes of
the general structure (xant)Pd(Ar)(X) are known to be
fluxional,22,44,45 exhibiting rapid cis−trans isomerization at
room temperature (Figure 6), and the cis, trans, and ionized
forms of (xant)Pd(Ph)(X) represented in Figure 6 may have
different reactivities toward transmetalation in a catalytic
setting. If Lewis acids affect the kinetics and/or thermody-
namics of the isomerization pathway, the concentration of the
ionized intermediate and rates of formation and consumption
PhI
2 equiv PhI
2 equiv 1
PhI
PhI
0.5 equiv 1
a
Standard reaction conditions: 1.0−2.0 equiv of 1 (0.2−0.4 M [1]),
1.0−2.0 equiv of aryl halide, 1.2 equiv of K3PO4, 0−1 equiv of NBu4Br.
1.5 mol % of Pd(dba)2, 1.5 mol % of xantphos, 5.5 mL of anhydrous
toluene, 110 °C. Values given are averages and standard deviations of
three trials.
Lewis acid free coupling of amide 1 and PhBr with either 2:1
1:PhBr or 1:2 1:PhBr ratio, the rate of catalysis was unaffected
(ν = (13 5.0) × 10−4 M min−1 (2:1 1:PhBr), (13 2.0) ×
10−4 M min−1 (1:2 1:PhBr)), indicating that under our reaction
conditions N-arylations using PhBr are zero order in [1] and in
[PhBr]. Indeed, exogenous halide inhibits catalysis; the rate of
coupling of 1 and PhBr with tetrabutylammonium bromide
(NBu4Br, 1 equiv relative to amide) was reduced to (7.5 2.5)
× 10−4 M min−1. The rate of coupling of 1 and PhBr with
potassium triflate (KOTf, 1 equiv to amide) is roughly the same
as that without any additive (ν = (13.0
3.9) × 10−4
M
min−1). The latter experiment indicates that triflate anion itself
does not independently increase the rate. Further, rate
enhancements are not solely the consequence of an increase
of the ionic strength of the reaction medium.
The coupling of amide 1 and PhI exhibited different kinetic
behavior, accelerating when the concentration of PhI was
doubled (ν = (12 3.6) × 10−4 M min−1 (1:2 1:PhI), (9.5
2.4) × 10−4 M min−1 (1:1 1:PhI)) and when [1] was doubled
(ν = (16 5.0) × 10−4 M min−1). The nonlinear kinetic profile
of the PhI catalytic reaction suggests complex reaction orders.
Kinetic experiments with larger concentrations of [1] proved
5867
DOI: 10.1021/acscatal.7b01317
ACS Catal. 2017, 7, 5862−5870

ACS Catalysis
Research Article
κ3-P,O,P pincer-like bonding mode of xantphos44,45 can also
promote halide exchange. The 31P resonance of 22 and 23
shifts downfield in the presence of increasing amounts of
[(xant)Pd(Ph)][BArF ] (25), suggesting that bromide ligands
4
exchange between Pd centers on the NMR time scale. At large
mole fractions of 25 (X25 > 0.35), a single broad peak in the 31
P
spectrum (13.4 ppm, 292 K) is observed (Figure 8); this shift is
Figure 6. Proposed pathway of cis−trans isomerization in (xant)Pd-
(Ph)(X) complexes.
could changethereby influencing the reaction rate of our
catalytic reaction of interest. Variable-temperature NMR
experiments were therefore conducted to probe the interaction
of 22 and 23 with Lewis acids, with the understanding that our
experiments would be at temperatures and conditions different
from those of the catalytic reaction.
The 31P NMR spectrum of 22 at 273 K in CD2Cl2 features a
single resonance (singlet, 9.6 ppm) that broadens upon cooling
to 260 K and at 195 K decoalesces to two signals (10.2 ppm,
22-trans; 8.6 ppm, 22-cis; 13:1 trans:cis ratio) (Figure 7A). In
Figure 8. 31P NMR spectra (0.02M, CDCl3) at 292 K of mixtures of
22 and 25. X corresponds to the mole fraction of 25.
identical to the 31P NMR resonance of 25 in the absence of 22.
Similar behavior was observed for mixtures of 23 and 25. The
31P resonances of 22 and 23 therefore depend upon the
concentration of the dissociated (xant)Pd(Ph) fragment in
solution.
Figure 7. 31P NMR spectra (0.02M, CD2Cl2) at 273 and 195 K: (A)
22; (B) 22 and 1 equiv of Yb(OTf)3; (C) 23; (D) 23 and 1 equiv of
Yb(OTf)3.
To confirm the competency of Yb(OTf)3 to generate a
(xant)Pd(Ph) cation, 22 was combined with 5 equiv of
Yb(OTf)3 and 100 equiv of 1 in CDCl3. The 31P spectrum
revealed a single sharp resonance at 13.4 ppm, identical with
that of the 31P resonance of 25 (Figure 9). In a solvent mixture
akin to our catalytic conditions (0.4 M 1 in C7D8), the addition
of Yb(OTf)3 resulted in a broadening and downfield shift of the
31P resonance of the respective oxidative addition complex,
suggesting that Yb(OTf)3 is competent to reversibly abstract
halide ions from 22 and 23 in the catalytic reaction medium.
The combination of 22, 5 equiv of Yb(OTf)3, and 100 equiv of
1 in C7D8 results in a very broad 31P resonance centered at 8.82
ppm; while this is indicative of fluxional behavior, decoales-
cence was not observed under cryogenic conditions (220 K).
the presence of 1 equiv of Yb(OTf)3 (the most soluble metal
triflate explored in our study) in CD2Cl2 at 273 K, 22 resonates
further downfield (11.5 ppm) and at 195 K is broadened and
unresolved, without reaching a decoalescence point (Figure
7B)consistent with a change in cis−trans isomerization
behavior upon addition of Lewis acid.
Like 22, the 31P NMR spectrum of 23 is a singlet (10.6 ppm,
273 K, CD2Cl2) that broadens upon cooling. At 195 K, the
peak resolves into two signals resonating at 11.9 (23-trans) and
8.7 (23-cis) ppm (∼15:1 trans:cis ratio) (Figure 7C). Two
additional sharp singlets at 17 (24a) and 4.9 (24b) ppm are
evident in the spectrum at 260 K, corresponding to the cis and
trans isomers of Pd(xant)(I)2, formed as a decomposition
product.48,50 In the presence of 1 equiv of Yb(OTf)3 at 195 K,
only a single sharp resonance corresponding to 23-trans is
observed (Figure 7D). Figure 7 demonstrates that the
interaction of Yb(OTf)3 with 22 and 23 results in different
effects on the fluxional behavior of either complex, showing that
the effect of Lewis acid additive depends on the identity of
halide X in the putative catalytic intermediate (xant)Pd(Ar)(X).
One plausible role of metal triflate additives M(OTf)3 in
catalysis is the facilitation of halide exchange in the first step of
transmetalation, perhaps forming [(xant)Pd(Ph)][OTf]. Like
M(OTf)3, cationic xantphos-ligated Pd complexes featuring a
CONCLUSION
■
Co-catalytic metal triflate Lewis acids accelerate Pd/xantphos-
catalyzed amide N-arylation using PhBr by as much as a factor
of 3, while the acceleration of N-arylations using PhI is less
significant. The effect is general for several classes of amide
nucleophiles, and in several cases, improved selectivity for a
single N-arylamide product is obtained. Reduced rates are
observed at Lewis acid loadings beyond 1:1 LA:Pd. Various
observations suggest that different mechanisms may occur
depending on the aryl halide employed. For reactions with
PhBr, pseudo-zero-order kinetics are observed in both [PhBr]
and [1], as well as inhibition of catalysis by NBu4Br. For
5868
DOI: 10.1021/acscatal.7b01317
ACS Catal. 2017, 7, 5862−5870

ACS Catalysis
Research Article
ACKNOWLEDGMENTS
■
This work was supported by the National Science Foundation
(CHE-1565721). Acknowledgment is made to the Donors of
the American Chemical Society Petroleum Research Fund for
partial support of this research. Support for the NMR facility at
Temple University by a CURE grant from the Pennsylvania
Department of Health is gratefully acknowledged. The authors
thank Boulder Scientific for a gift of B(C6F5)3, Dr. Charles W.
DeBrosse for assistance with variable-temperature NMR, and
Prof. Nilay Hazari for helpful discussions.
REFERENCES
■
(1) Hartwig, J. F. Acc. Chem. Res. 1998, 31, 852−860.
(2) Hartwig, J. F. Nature 2008, 455, 314−322.
(3) Dugger, R. W.; Ragan, J. A.; Ripin, D. H. B. Org. Process Res. Dev.
2005, 9, 253−258.
(4) Schlummer, B.; Scholz, U. Adv. Synth. Catal. 2004, 346, 1599−
1626.
(5) Ruiz-Castillo, P.; Buchwald, S. L. Chem. Rev. 2016, 116, 12564−
12649.
(6) Hicks, J. D.; Hyde, A. M.; Cuezva, A. M.; Buchwald, S. L. J. Am.
Chem. Soc. 2009, 131, 16720−16734.
(7) Constable, D. J. C.; Dunn, P. J.; Hayler, J. D.; Humphrey, G. R.;
Leazer, J. L., Jr.; Linderman, R. J.; Lorenz, K.; Manley, J.; Pearlman, B.
A.; Wells, A.; Zaks, A.; Zhang, T. Y. Green Chem. 2007, 9, 411−420.
(8) Birkholz, M.-N.; Freixa, Z.; van Leeuwen, P. W. N. M. Chem. Soc.
Rev. 2009, 38, 1099−1118.
Figure 9. 31P NMR spectra (0.6 mM, C7D8) at 292 K: (A) mixture of
22, 100 equiv of 1, and x equiv of Yb(OTf)3; (B) mixture of 23, 100
equiv of 1, and x equiv of Yb(OTf)3; (C) 22, 100 equiv of 1, and
varying equivalents (x = 5, 1, 0.5, and 0) of Yb(OTf)3 (the spectrum of
pure 25 (bottom; CDCl3, 292 K) is shown for comparison).
(9) Beletskaya, I. P.; Cheprakov, A. V. Organometallics 2012, 31,
7753−7808.
(10) Falk, F. C.; Frohlich, R.; Paradies, J. Chem. Commun. 2011, 47,
̈
11095−11097.
reactions with PhI, the rate of catalysis is lower and somewhat
dependent on [PhI] and [1]. The rate of reaction of PhBr and
1 is not affected by KOTf, suggesting that acceleration effects
are not necessarily due to ionic strength or the triflate anion.
Transmetalation requires, at the very least, dissociation of X−
from (xant)Pd(Ph)(X), coordination of 1, dissociation of base,
deprotonation of Pd-bound 1, and precipitation of a HX:base
complex. One or more of these steps are evidently dependent
on whether X = Br or X = I. Complexes 22 and 23catalyst
resting state complexesundergo halide abstraction with
addition of Yb(OTf)3, forming [(xant)Pd(Ph)]+ as evidenced
by stoichiometric 31P NMR studies. We propose that metal
triflate additives facilitate halide abstraction in turnover-limiting
transmetalation, thereby accelerating the overall catalytic
process. Other productive influences of M(OTf)3 additives
are possible, and further study in our laboratory aims to better
understand additive effects across Pd catalysis.
(11) Falk, F. C.; Oechsle, P.; Thiel, W. R.; Daniliuc, C.-G.; Paradies,
J. Eur. J. Org. Chem. 2014, 21, 3637−3645.
(12) Surry, D. S.; Buchwald, S. L. Chem. Sci. 2011, 2, 27−50.
(13) Ikawa, T.; Barder, T. E.; Biscoe, M. R.; Buchwald, S. L. J. Am.
Chem. Soc. 2007, 129, 13001−13007.
(14) Shen, Q.; Shekhar, S.; Stambuli, J. P.; Hartwig, J. F. Angew.
Chem., Int. Ed. 2005, 44, 1371−1375.
(15) Crawford, S. M.; Lavery, C. B.; Stradiotto, M. Chem. - Eur. J.
2013, 19, 16760−16771.
(16) Sergeev, A. G.; Artamkina, G. A.; Beletskaya, I. P. Tetrahedron
Lett. 2003, 44, 4719−4723.
(17) Hamann, B. C.; Hartwig, J. F. J. Am. Chem. Soc. 1998, 120,
3694−3703.
(18) Hartwig, J. F.; Richards, S.; Baranano, D.; Paul, F. J. Am. Chem.
̃
Soc. 1996, 118, 3626−3633.
(19) Shekhar, S.; Hartwig, J. F. Organometallics 2007, 26, 340−351.
(20) Barder, T. E.; Biscoe, M. R.; Buchwald, S. L. Organometallics
2007, 26, 2183−2192.
(21) Barder, T. E.; Buchwald, S. L. J. Am. Chem. Soc. 2007, 129,
12003−12010.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.7b01317.
■
(22) Fujita, K.; Yamashita, M.; Puschmann, F.; Alvarez-Falcon, M.
M.; Incarvito, C. D.; Hartwig, J. F. J. Am. Chem. Soc. 2006, 128, 9044−
9045.
S
(23) Yin, J. J.; Buchwald, S. L. Org. Lett. 2000, 2, 1101−1104.
(24) Yin, J. J.; Buchwald, S. L. J. Am. Chem. Soc. 2002, 124, 6043−
6048.
Experimental procedures, characterization, spectral data,
and kinetic data (PDF)
(25) Chen, D.; Gau, M. R.; Dobereiner, G. E. Organometallics 2015,
34, 4069−4075.
AUTHOR INFORMATION
Corresponding Author
*E-mail for G.E.D.: dob@temple.edu.
ORCID
Graham E. Dobereiner: 0000-0001-6885-2021
Notes
The authors declare no competing financial interest.
(26) Huang, J.; Haar, C. M.; Nolan, S. P.; Marcone, J. E.; Moloy, K.
G. Organometallics 1999, 18, 297−299.
■
(27) Shen, Q.; Hartwig, J. F. J. Am. Chem. Soc. 2007, 129, 7734−
7735.
(28) Liberman-Martin, A.; Bergman, R. G.; Tilley, T. D. J. Am. Chem.
Soc. 2013, 135, 9612−9615.
(29) Liberman-Martin, A.; Levine, D. S.; Liu, W.; Bergman, R. G.;
Tilley, T. D. Organometallics 2016, 35, 1064−1069.
5869
DOI: 10.1021/acscatal.7b01317
ACS Catal. 2017, 7, 5862−5870

ACS Catalysis
Research Article
(30) Kobayashi, S.; Manabe, K. Acc. Chem. Res. 2002, 35, 209−217.
(31) Various additional Lewis acids have been evaluated. See Table
S2 in the Supporting Information.
(32) Cacchi, S.; Fabrizi, G.; Goggiamani, A.; Zappia, G. Org. Lett.
2001, 3, 2539−2541.
(33) Madar, D. J.; Kopecka, H.; Pireh, D.; Pease, J.; Pliushchev, M.;
Sciotti, R. J.; Wiedemann, P. E.; Djuric, S. W. Tetrahedron Lett. 2001,
42, 3681−3684.
(34) Dooleweerdt, K.; Fors, B. P.; Buchwald, S. L. Org. Lett. 2010, 12,
2350−2353.
(35) Su, M. J.; Buchwald, S. L. Angew. Chem., Int. Ed. 2012, 51,
4710−4713.
(36) Fox, J. M.; Huang, X.; Chieffi, A.; Buchwald, S. L. J. Am. Chem.
Soc. 2000, 122, 1360−1370.
(37) Manley, P. J.; Bilodeau, M. T. Org. Lett. 2004, 6, 2433−2435.
(38) The α-arylation products were identified by two-dimensional
NMR. See Figures S1−S3 in the Supporting Information.
(39) Kitagawa, O.; Takahashi, M.; Yoshikawa, M.; Taguchi, T. J. Am.
Chem. Soc. 2005, 127, 3676−3677.
(40) Kitagawa, O.; Yoshikawa, M.; Tanabe, H.; Morita, T.;
Takahashi, M.; Dobashi, Y.; Taguchi, T. J. Am. Chem. Soc. 2006,
128, 12923−12931.
(41) Due to the large error observed in the initial rate of reaction of 1
and PhBr with cocatalytic Al(OTf)3 and Sc(OTf)3, these Lewis acids
were not evaluated in kinetic experiments of the reaction of 1 and PhI.
(42) Dallas, A. S.; Gothelf, K. V. J. Org. Chem. 2005, 70, 3321−3323.
(43) See Table S1 in the Supporting Information.
(44) Guari, Y.; Van Strijdonck, G. P. F.; Boele, M. D. K.; Reek, J. N.
H.; Kamer, P. C. J.; Van Leeuwen, P. W. N. M. Chem. - Eur. J. 2001, 7,
475−482.
(45) Zuideveld, M. A.; Swennenhuis, B. H.; Maarten, D. K.; Guari, Y.;
Van Strijdonck, G. P. F.; Reek, J. N. H.; Kamer, P. C. J.; Goubitz, K.;
Fraanje, J.; Lutz, M.; Spek, A. L.; Van Leeuwen, P. W. N. M. J. Chem.
Soc., Dalton Trans. 2002, 11, 2308−2317.
(46) Grushin, V. V.; Marshall, W. J. J. Am. Chem. Soc. 2006, 128,
12644−12645.
(47) Miloserdov, F. M.; Grushin, V. V. Angew. Chem., Int. Ed. 2012,
51, 3668−3672.
(48) Miloserdov, F. M.; McMullin, C. L.; Belmonte, M. M.; Benet-
Buchholz, J.; Bakhmutov, V. I.; MacGregor, S. A.; Grushin, V. V.
Organometallics 2014, 33, 736−752.
(49) Andersen, T. L.; Friis, S. D.; Audrain, H.; Nordeman, P.; Antoni,
G.; Skrydstrup, T. J. Am. Chem. Soc. 2015, 137, 1548−1555.
(50) See Figures S6 and S7 in the Supporting Information.
5870
DOI: 10.1021/acscatal.7b01317
ACS Catal. 2017, 7, 5862−5870