DOI: 10.1002/chem.

201406109 Full Paper

& Palladium Catalysis

Is PdII-Promoted s-Bond Metathesis Mechanism Operative for the

PdPEPPSI Complex-Catalyzed Amination of Chlorobenzene with
Aniline? Experiment and Theory
Feiqun Wang,[a] Lei Zhu,[a] Yunfei Zhou,[a] Xiaoguang Bao,*[a] and Henry F. Schaefer, III[b]

Abstract: Reduction of the PdPEPPSI precatalyst to a Pd0 bond metathesis in pathways leading to the diphenylamine
species is generally thought to be essential to drive product have relatively low barriers. Such pathways are
Buchwald–Hartwig amination reactions through the well- more favorable energetically than the corresponding
documented Pd0/PdII catalytic cycle and little attention has reductive elimination reactions resulting in Pd0 species and
been paid to other possible mechanisms. Considered here is other putative routes, such as the PdII/PdIV mechanism,
the PdPEPPSI-catalyzed aryl amination of chlorobenzene single electron transfer mechanism, and halide atom transfer
with aniline. A neat reaction system was used in new ex- mechanism. In some special cases, if reactants/additives are
periments, from which the potentially reductive roles of the inadequate to reduce a PdII precatalyst, a PdII-involved
solvent and labile ligand of the PEPPSI complex in leading s-bond metathesis mechanism might be feasible to drive
to Pd0 species are ruled out. Computational results demon- the Buchwald–Hartwig amination reactions.
strate that anilido-containing PdII intermediates involving s-

Introduction The first well-defined NHCPdII complexes for catalyzing the

BH amination reaction was reported by Nolan and co-workers
Palladium-catalyzed Buchwald–Hartwig (BH) amination in 2002 (1a, Figure 1).[5] Subsequently, Nolan and co-workers
reactions are among the most powerful synthetic methods for synthesized many other NHC-containing PdII complexes, such
the construction of CN bonds and have been widely em- as [NHCPd-cycle] (1b),[6] [NHCPd(acac)Cl] (1c),[7]
[8] [9]
ployed in both academic research and modern chemical indus- [NHCPd(allyl)Cl] (1d), and [NHCPd(cinnamyl)Cl] (1e), all of
try.[1] Phosphine-containing Pd catalysts have been extensively which are capable of driving BH amination reactions. In 2006,
studied and optimized to mediate BH amination reactions.[2] Organ and co-workers developed the PdPEPPSI catalyst
However, phosphine-based ligands are toxic, expensive, and (PEPPSI is the abbreviation for pyridine-enhanced precatalyst
air-sensitive. Therefore, development of phosphine-free Pd preparation stabilization and initiation).[10] The PdPEPPSI cata-
catalysts would be highly desirable. Over the last two decades, lyst 1f, for example, is also able to catalyze BH amination reac-
N-heterocyclic carbene (NHC) ligands have attracted tremend- tions.[11a] Subsequently, many other PdPEPPSI-like complexes
ous interest because they are more than mimics of phosphine have been reported.[11] Both modification of the NHC moiety
ligands.[3] The unique electronic properties and versatile steric and replacement of 3-chloropyridine by other ligands have
bulk of effective NHC ligands render improved performance in attracted great interest in order to obtain better catalytic
many Pd-catalyzed cross-coupling reactions.[4] In addition, performance.[11]
NHCPdII complexes are usually stable with respect to air and The commonly proposed reaction mechanism for the
moisture and relatively easy to synthesize.[4b,d] More attractively, NHCPdII catalyzed amination reaction[4c, e] is through a Pd0/PdII
structures of NHCPdII complexes are well-defined,[4b–e] a catalytic cycle (Scheme 1). In order to enter the catalytic cycle,
characteristic particularly appealing for mechanistic studies. a PdII complex, which is denoted a “precatalyst”, is supposed
to undergo reduction and ligand dissociation steps to afford
the Ln-Pd0 species first. Subsequently, the aryl halide undergoes
an oxidative addition (OA) reaction at the activated Pd0 site to
[a] F. Wang, L. Zhu, Y. Zhou, Prof. X. Bao
College of Chemistry, Chemical Engineering and Materials Science yield a Pd-aryl intermediate (structure I in Scheme 1). Next, an
Soochow University amine-Pd-aryl complex (II) can be generated via coordination
199 Ren-Ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123 (China) of the amine. Then, deprotonation of the coordinated amine is
E-mail: xgbao@[Link]
followed with the assistance of base to yield an aryl-Pd-amido
[b] Prof. H. F. Schaefer, III
intermediate (III). Finally, a reductive elimination (RE) reaction
Center for Computational Quantum Chemistry
University of Georgia, Athens, Georgia 30602 (United States) of III occurs to afford the cross-coupling product, and the Pd0
Supporting information for this article is available on the WWW under catalyst is regenerated. Computational studies have been
[Link]

Chem. Eur. J. 2015, 21, 4153 – 4161 4153  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

[allylPdalkoxide] complex fol-

lowed by a RE step.[8, 9] It is con-
ceivable that the presence of
active transmetalating agents,
such as organomagnesium, -zinc,
-tin, or -boron reagents, might
reduce PdII species to Pd0 effec-
tively.[10, 13] In addition, the pres-
ence of b-H in substrates or ad-
ditives may convert PdII to Pd0
via a b-H elimination followed
by RE.[14]
In the 1f-promoted amination
of chlorobenzene with aniline,
using KOtBu and 1,4-dioxane as
base and solvent, respectively,
II
Figure 1. Representative NHCPd complexes serving to catalyze Buchwald–Hartwig (BH) amination reactions.
the product, diphenylamine, can
be obtained in excellent yield.
However, it seems that the re-
duction pathways involving the
PdII precatalyst, mentioned in the literature, cannot apply here.
The mechanism of the reduction of 1f to the Pd0 species,
which is essential to start BH amination reactions in the pro-
posed Pd0/PdII mechanism, remains elusive. However, with the
exception of the Pd0/PdII mechanism, no other possible reac-
tion mechanism for the PdPEPPSI catalyzed amination reac-
tion has, to our knowledge, been examined, such as a s-bond
metathesis mechanism, PdII/PdIV mechanism, single electron
transfer (SET) mechanism, or halide atom transfer (HAT) mecha-
nism.
Whether the BH amination reaction could be carried out via
another mechanism, in some cases, instead of the well-estab-
lished Pd0/PdII catalytic cycle is an interesting and potentially
important question. Herein we propose that a PdII-promoted
s-bond metathesis mechanism might be feasible in the
PdPEPPSI-catalyzed amination of chlorobenzene with aniline,
in which the PdII complex has little chance to be reduced to
the Pd0 species.

Results and Discussion

0 II
Scheme 1. Generally proposed Pd /Pd mechanism for Buchwald–Hartwig Excluding the potentially reductive role of the solvent and
(BH) amination catalyzed by NHCPdII complexes. labile ligand of the PEPPSI complex: Experimental
investigation
carried out to provide an energy profile for the Pd0/PdII We have explored whether an alternative reaction mechanism
mechanism.[12] for the 1f-catalyzed amination of chlorobenzene with aniline
Some modes for the activation of PdII complexes to Pd0 takes place. Instead of the commonly proposed Pd0/PdII
species have been suggested in the literature. Nolan and co- catalytic cycle, we first ruled out the potential reducing agents,
workers proposed that the precatalyst 1b might be converted solvents, and labile ligands of the PEPPSI complex, which
to Pd0 by an attack of alkoxide on Pd to form an [arylPd might be able to reduce PdII to Pd0. Our experiments in this
alkoxide] complex followed by a RE step.[6a] This pathway can context are summarized in Table 1.
also apply to the activation of 1c after a rearrangement of the In the 1f-catalyzed amination of chlorobenzene with aniline,
acetylacetonate (acac) moiety from the O,O-bound to the C- using 1,4-dioxane and KOtBu as solvent and base, respectively,
bound PdII tautomer.[7] The reduction of 1d and 1e to Pd0 diphenylamine can be obtained in excellent yield (Table 1,
might be achieved either through a direct nucleophilic attack entry 1). It might be possible that coordination of a 1,4-diox-
at the allyl moiety by alkoxide, or through the formation of an ane to PdII followed by a b-H elimination could afford a PdH

Chem. Eur. J. 2015, 21, 4153 – 4161 [Link] 4154  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

PEPPSI analog 1g (Figure 2), in which the 3-chloropyridine

Table 1. Experimental yields when PdPEPPSI and its analogs serve
effectively to catalyze the amination of chlorobenzene with aniline. ligand was replaced by 1-methylimidazole, can also effectively
promote the same reaction.[11e] Moreover, we synthesized
another PEPPSI analog, 1h (Figure 2), in which the labile ligand
is replaced by aniline (see the Supporting Information for
Entry[a] Catalyst Solvent Yield [%][b]
detailed experimental procedures and characterization). The
crystal structure of the new compound 1h shows that, similar
1 1f 1,4-dioxane 96
to the complexes of PEPPSI family, the NHC ligand and the
2 1f toluene 95
3 1f aniline 96 labile aniline ligand of 1h are trans to each other. Catalytic
4[c] 1g toluene 79 studies for 1h showed that this NHCPdII complex can also
5 1h toluene 89 catalyze the amination reaction remarkably well, using either
6 1h aniline 95
toluene or aniline as solvent (Table 1, entries 5 and 6). The
7 PdCl2 1,4-dioxane trace
excellent performance of 1h can rule out a crucial reductive
[a] Reaction conditions: Chlorobenzene (1.0 mmol), aniline (1.1 mmol for
role of a “throwaway” ligand in forming the Pd0 species.
solvent ¼
6 aniline), KOtBu (1.2 mmol), solvent (2 mL), 90 8C, 4 h; [b] yield
isolated after two runs; [c] reference [11e]. PEPPSI complexes are usually bench-stable, so dissociation
of the NHC ligand is very unlikely to occur. This is also support-
ed by computational evidence that DG for the ligand ex-
change is significantly endothermic, by 30 kcal mol1 [Eq. (1);
complex and an oxonium ion intermediate. Next, reduction of 1 kcal = 4.184 kJ]. The experimental fact that PdCl2 is unable to
the PdH complex might lead to Pd0. To rule out the potential- catalyze the amination reaction also demonstrates the crucial
ly critical role of the 1,4-dioxane solvent in reducing the PdII role of the NHC ligand (Table 1, entry 7).
complex, amination reactions in which 1,4-dioxane was re-
placed by toluene or aniline (Table 1, entries 2 and 3, respec-
tively),[15] were carried out and proceeded with excellent yields
of the desired product. Therefore, the solvent, 1,4-dioxane, is
not indispensable in the amination reaction.
One might suspect that the “throwaway” ligand of the
PEPPSI catalyst could somehow play a role in reducing the PdII
complex. In fact, Shao and co-workers demonstrated that the
Computational investigation of anilido-containing PdII
intermediates and possible subsequent reactions.
One may suggest that aniline is most likely to function as a
potential reducing agent in converting PdII to the Pd0 spe-
cies.[16] Thus, intermediates involving aniline and any possible
subsequent reactions were computationally explored.

Anilido-containing PdII intermediates

For the well-defined complex 1f, Organ and co-workers pro-
posed that the 3-chloropyridine moiety may be regarded as
a “throwaway” ligand.[11a] In the presence of aniline, 1f can be
converted to 1h via a ligand-exchange step (Figure 3). The
present computational study suggests that this step is slightly
exothermic in free energy, by 1.2 kcal mol1. Subsequently, the
PdII–aniline complex formed may undergo a deprotonation
step facilitated by the base, KOtBu. This results in the elimina-
tion of KCl and tBuOH and affords a [(NHC)Pd(Cl)(NHPh)] inter-
mediate (3). The latter step is computed to be endothermic by
3.4 kcal mol1.[17] In addition, the resulting intermediate 3 may
undergo an anionic ligand-exchange step with base to gener-
ate an intermediate 4. A very slight exothermicity (0.4 kcal
mol1) is found for this step. Coordination of aniline with 3
Figure 2. Top) Molecular structures of 1g and 1h; bottom) crystal structure and subsequent elimination of KCl and tBuOH assisted by
of 1h. Hydrogen atoms are omitted for clarity. Selected bond lengths []
KOtBu can afford a [(NHC)Pd(NHPh)2] intermediate (6). This
and angles [o]: Pd1C1 1.968(2), Pd1N3 2.1064(17), Pd1Cl2 2.2970(5),
Pd1Cl1 2.2993(5); C1-Pd1-N3 175.52(8), C1-Pd1-Cl2 90.44(6), N3-Pd1-Cl2 step is predicted to be substantially exothermic, by 13.2 kcal
90.52(5), C1-Pd1-Cl1 91.81(6), N3-Pd1-Cl1 87.53(5). mol1 (Figure 3).

Chem. Eur. J. 2015, 21, 4153 – 4161 [Link] 4155  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

lead to the formation of a PdIII

radical cation and a radical anion
of the aryl halide (Scheme 2,
path d). This radical pair could
form the amination product or
generate a PdIV intermediate
after a subsequent SET step. An
alternative PdIII-mediated mecha-
nism is hypothesized that pro-
ceeds via a halide atom transfer
(HAT) step, yielding a phenyl
radical and a PdIII radical
(Scheme 2, path e). The resulting
radical pair could recombine to
afford the desired product. Both
the SET and HAT mechanisms
have been proposed for CuI-cat-
alyzed Ullmann-type reactions.[18]
A PdIII-based mechanism was
proposed by Ritter and co-work-
Figure 3. Free-energy profile for the conversion of 1f to the three anilido-containing PdII intermediates.
ers for the oxidation of PdII
dimer with suitable oxidizing
Any or all of the three anilido-containing PdII intermediates, agents[19] and subsequently computationally studied by Schoe-
3, 4, and 6, might undergo a RE reaction to yield a Pd0 species nebeck and co-workers.[20] To our knowledge, the question of
(Scheme 2, path a). Accordingly,
the BH reaction could be driven
via the well-documented Pd0/PdII
catalytic cycle. Apart from the RE
reaction, these three intermedi-
ates might utilize a four-cen-
tered s-bond metathesis mecha-
nism in the presence of an aryl
halide to produce the desired
cross-coupling product in a con-
certed manner (Scheme 2,
path b). The breaking of the C
Cl bond and the formation of
the CN bond occur simultane-
ously. The oxidation state of Pd
is unchanged throughout the
catalytic cycle. A mechanism in-
volving a PdII/PdIV catalytic cycle
is also proposed (Scheme 2,
path c). The aryl halide might
undergo oxidative addition to an
anilido-containing PdII intermedi-
ate, affording a PdIV intermedi-
ate. A subsequent RE step would
generate the amination product.
In addition, a PdIII-mediated
mechanism[19] has been postulat-
ed for the BH amination reac-
tion, regardless of the limited ex-
perimental evidence supporting
this hypothesis. Single electron
transfer (SET) from a PdII inter-
mediate to an aryl halide can Scheme 2. Proposed reaction pathways for anilido-containing PdII intermediates.

Chem. Eur. J. 2015, 21, 4153 – 4161 [Link] 4156  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

whether the PdII/PdIII catalytic cycle could be operative in the

BH reaction has not been evaluated.

Possible reaction routes for the [(NHC)Pd(Cl)(NHPh)]

intermediate
Five plausible reaction pathways (Scheme 2) associated with
the [(NHC)Pd(Cl)(NHPh)] intermediate 3 were computationally
studied (Figure 4). The first is RE of 3 and the others are differ-
ent reaction routes between 3 and PhCl, leading to the cross-
coupling product. The RE of 3 proceeds via a three-membered
ring transition state (TS), where the NCl bond (2.088 ) is
being formed and the PdCl bond is breaking (3.023 )
(Figure 5, TS(3–7)). The activation barrier for this step is com-
puted to be 42 kcal mol1 and the RE reaction is endothermic
by 36 kcal mol1 (Figure 4, path a). The computational results
suggest that the RE of 3 is unlikely.
It should be noted that the anilido ligand of 3 (Figure 4) has
nucleophilic character. In the presence of PhCl, it is possible to Figure 5. Optimized TS structures pertaining to the reaction routes for RE
generate the amination product via a s-bond metathesis step (TS(3-7)), s-bond metathesis (TS(8-9)), and PdII/PdIV (TS(8-10)) for 3. Bond
lengths are shown in . Hydrogen atoms are omitted for clarity.

Figure 4. Free-energy profiles of the five reaction pathways examined for 3.

Chem. Eur. J. 2015, 21, 4153 – 4161 [Link] 4157  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

in a concerted manner. In the structure of TS(8–9), the CCl shown in Figure S2 in the Supporting Information). The RE
and PdN bond lengths are lengthened to 1.989 and 2.032 , reaction of 4 is predicted to have DGRE° = 36 kcal mol1 and to
respectively, whereas the Pd-Cl and CN bond lengths are lead to an endergonic Pd0 product with DGRE = 8 kcal mol1
shortened to 2.409 and 2.063 , respectively (Figure 5). A typi- (see the Supporting Information, Figure S1, path a).
cal four-center s-metathesis feature is seen at the TS structure. As for the s-bond metathesis mechanism, a four-centered TS
The predicted free energy barrier for the s-bond metathesis structure is located (see the Supporting Information, Figure S2,
route is approximately 27 kcal mol1 relative to the separated 3 TS(14–15)). Theory shows that the free energy barrier associ-
plus PhCl, which is approximately 14 kcal mol1 lower than that ated with this pathway is 30 kcal mol1 relative to separated 4
for the reductive elimination of 3. Formation of the amination and PhCl. The resulting product complex is significantly exo-
product complex is computed to be very exothermic, by thermic, by 31 kcal mol1 (see the Supporting Information, Fig-
41 kcal mol1 (Figure 4, path b). The complex 1h can be regen- ure S1, path b). Therefore, the RE pathway of 4 is unfavorable
erated after diphenylamine is replaced by aniline. Interestingly, from both kinetic and thermodynamic perspectives compared
the s-bond metathesis pathway leading to the Ph2NH product with the corresponding s-bond metathesis mechanism. In
directly is much more favorable than the generation of Pd0 via addition, similar to 3, the PdII/PdIV, SET, and HAT routes of 4 are
the RE pathway, both kinetically and thermodynamically. energetically unfavorable in comparison with the corres-
The oxidative addition of PhCl to 3, forming a PdIV inter- ponding s-bond metathesis mechanism (see the Supporting
mediate, is computed to have a free energy barrier of 49 kcal Information for details).
mol1. This step results in an endergonic PdIV intermediate
with DG = 32 kcal mol1 (Figure 4, path c). The computational
Possible reaction routes of the [(NHC)Pd(NHPh)2] intermediate
results show that the oxidative addition of PhCl to 3 is highly
unfavorable. For [(NHC)Pd(NHPh)2] intermediate 6, five possible routes were
Intermolecular SET from 3 to PhCl can yield a PdIII radical also investigated. The RE of 6 was computed to have a barrier
cation (11) and an aryl chloride radical anion. Subsequently, height of 33 kcal mol1 (the TS structure of this step, TS(6–19),
dissociation of the CCl bond of the aryl chloride radical anion is shown in Figure S4 in the Supporting Information). The gen-
yields a free aryl radical and a chloride anion. Finally, the free erated Pd0 product complex 19 is endothermic by 3 kcal mol1
aryl radical reacts with the PdIII radical cation to produce the (see the Supporting Information, Figure S3, path a). Compared
desired product. However, the theoretical results demonstrate with the RE of intermediates 3 and 4, the RE of 6 has the
that the free energy barrier of the SET pathway is prohibitively lowest energy barrier and the smallest endothermicity.
high, suggesting that the SET mechanism is very unlikely to The s-bond metathesis pathway for 6 has a computed DG°
occur (Figure 4, path d). of 24 kcal mol1 relative to separated 6 and PhCl. The
The chlorine atom transfer pathway involves homolytic four-centered TS results in a product complex,
cleavage of the CCl bond of PhCl and simultaneous formation [(NHC)Pd(Cl)(NHPh)(NHPh2)] (21), the generation of which is ex-
of the PdCl bond to afford a PdIII radical (12) and an aryl radi- ergonic by 44 kcal mol1 (see the Supporting Information, Fig-
cal. The generated aryl radical favors reaction with the anilido ure S3, path b). Similar to 3 and 4, intermediate 6 also favors
ligand of 12 to afford the coupling product, rather than com- the s-bond metathesis pathway in the presence of PhCl, in-
bining with the PdIII radical to form a PdIV intermediate. This is stead of the corresponding RE route leading to Pd0. Computa-
because the formation of the coupling product is dramatically tional results also suggest that the s-bond metathesis pathway
exothermic, whereas the formation of the PdIV intermediate is proceeding through 6 is more favorable than the analogous
endothermic (Figure 4, path e). Similar to the CuI-catalyzed Ull- route via 3 and 4. The other three routes applied to 6, PdII/PdIV,
mann reactions via the HAT mechanism, the PdCl bond for- SET, and HAT, are less competitive than the corresponding s-
mation can compensate the energy needed to break the CCl bond metathesis mechanism (see the Supporting Information
bond of PhCl.[18] Nevertheless, the predicted energy resulting for details).
in formation of 12 and an aryl radical is approximately 27 kcal Overall, the present computational results indicate that the
mol1 higher than the energy of separated 3 and PhCl. Al- three plausible deprotonated aniline-containing PdII species
though the TS of the HAT pathway has not been located, the may not undergo RE reactions to form Pd0 easily in the 1f-cat-
energy barrier height of HAT mechanism must be higher than alyzed amination of chlorobenzene with aniline. Instead, the s-
27 kcal mol1 and, therefore, higher than that of the s-bond bond metathesis mechanism is more facile for all three inter-
metathesis mechanism, so the HAT pathway is less feasible mediates than the corresponding RE reactions, both kinetically
than the s-bond metathesis route.[21] and thermodynamically. Other proposed reaction pathways,
such as the PdII/PdIV mechanism, the SET mechanism, and the
HAT mechanism, cannot compete with the s-bond metathesis
Possible reaction routes of the [(NHC)Pd(OtBu)(NHPh)]
mechanism either.
intermediate
Similarly, the five reaction pathways in Scheme 2 can also
Frontier orbital analysis of the s-bond metathesis mechanism
apply to the [(NHC)Pd(OtBu)(NHPh)] intermediate 4. Of primary
interest is the possibility of the RE of 4 to afford the Pd0 A frontier orbital analysis provides more insight into the
species (the optimized TS structure of this step, TS(4–13), is amination reaction via the s-bond metathesis mechanism. The

Chem. Eur. J. 2015, 21, 4153 – 4161 [Link] 4158  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

Figure 6. Frontier orbitals for the transition states of the three s-bond metathesis reactions.

highest occupied molecular orbitals (HOMO) of the transition solvent, the studied BH reactions are not very sensitive to the
states for all three s-bond metathesis reactions qualitatively counterion.
show a lone pair located on N of the deprotonated aniline One may envision that the base employed, KOtBu, might
ligand, and this lone pair is not employed in bonding to Pd play a reductive role in converting the PdII precatalyst to the
(Figure 6). Thus, the deprotonated aniline moiety can partici- Pd0 species, since KOtBu has been shown to act as a one-elec-
pate in the subsequent nucleophilic substitution reaction. The tron reducing agent.[23] Therefore, we carried out the amination
corresponding lowest unoccupied molecular orbitals (LUMO) reaction in the laboratory, replacing KOtBu with KOH. About
display the out-of-phase combinations of the Pd ds atomic or- 20 % yield of the diphenylamine product was obtained for
bital (AO) and the Cl ps AOs (Figure 6). Both the existence of both 1f and 1h (Table 2, entries 1 and 2). The low efficiency of
an empty ds orbital on PdII and the presence of a lone pair on KOH compared with KOtBu in mediating the amination reac-
the anilido ligand are important for the amination reaction to tion may be due to the following two factors. One is that KOH
proceed via s-bond metathesis mechanism. Such a PdII-mediat- is less basic than KOtBu, and hence less effective in assisting
ed s-bond metathesis mechanism is consistent with the reac- the deprotonation of aniline. The other factor may be that
tion mechanism of water/alcohol with palladium hydride com- KOH is less soluble in aniline than KOtBu. One may envisage
plexes proposed by Dedieu and co-workers.[22] In addition, it that the production of water by the action of KOH on the
should be noted that the decomposition of all three anilido- acidic anilinium proton might have any effect on the catalytic
containing PdII intermediates via the RE route is difficult. Thus, performance. The addition of molecular sieves to sequester
the stability of the three intermediates, which cannot undergo water leads to a decrease of the yield of Ph2NH (Table 2, en-
RE reactions easily, is also responsible for the viability of
subsequent s-bond metathesis reaction.

Table 2. Experimental yields for the 1f and 1h catalyzed amination of

chlorobenzene with aniline using KOH as a base.
Replacement of KOtBu with NaOtBu and KOH in the
amination reaction: More experiments
In order to explore whether the counterion of the base used
Entry[a] Catalyst Additive Yield [%][b]
has any effect on the efficiency of the catalytic reaction, the
reactions catalyzed by 1f and 1h described in Table 1 were 1 1f – 22
2 1h – 22
performed using NaOtBu in addition to those already 3 1f 4  MS[c] 12
described using KOtBu. The yield of the desired product 4 1h 4  MS[c] 11
decreased slightly with 1,4-dioxane and aniline as the solvents 5 1f water[d] 23
(see the Supporting Information, Table S2, entries 1, 3, and 5), 6 1h water[d] 31

whereas the yield of diphenylamine diminished significantly [a] Reaction conditions: chlorobenzene (1.0 mmol), aniline (2 mL), KOH
when using toluene as the solvent (Table S2, entries 2 and 4). (1.2 mmol), 90 8C, 4 h; [b] yield isolated after two runs; [c] 4  MS
(200 mg); [d] water (1.0 mmol).
Nevertheless, in the neat reaction system with aniline as the

Chem. Eur. J. 2015, 21, 4153 – 4161 [Link] 4159  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

tries 3 and 4). The addition of water in stoichiometric amount Gibbs free energies for all species were obtained at 298.15 K and
can slightly increase the yield of Ph2NH (Table 2, entries 5 and 1 atm at their respective optimized structures.
6). The yielded/added water may increase the solubility of To consider solvation effects, single-point energy computations
KOH, hence resulting in slightly higher yield of the product. Re- using the SMD model[28] with toluene as solvent were performed
gardless of the low efficiency of using KOH as a base to drive based on the optimized gas-phase geometries of all species.
Larger basis sets (LANL2DZ for Pd and 6-311 + + G(d,p) for other
the amination reaction, KOtBu is not indispensable. If reactants
atoms) were utilized for single-point energy calculations on
and additives have little chance to reduce the PdII precatalyst, stationary points. The solution-phase Gibbs free energy was
a PdII-mediated s-bond metathesis mechanism might be determined by adding the solvation single-point energy and the
operative to drive the BH amination reaction. gas-phase thermal correction to the Gibbs free energy obtained
from the vibrational frequencies. Unless otherwise specified, the
solution-phase Gibbs free energy was used in the present discus-
sions. The Gaussian 09 suite of programs[29] was used throughout.
Conclusions
In summary, it is commonly suggested that the Buchwald–
Hartwig (BH) amination reactions promoted by the PdPEPPSI Acknowledgements
catalyst proceed through the well-documented Pd0/PdII
catalytic cycle. Therefore, the initial reduction of PdII species to XB was supported by the Natural Science Foundation of China
Pd0 is essential to drive the amination reactions. However, in (NSFC 21302133), the startup fund from the Soochow Universi-
the special reaction system of 1f-catalyzed aryl amination of ty, and the Priority Academic Program Development of Jiangsu
chlorobenzene with aniline, there is no explicit reducing agent. Higher Education Institutions (PAPD). H.F.S. was supported by
Moreover, a neat reaction system was experimentally created, the U.S. National Science Foundation, (Grant CHE-1361178).
from which the potentially reductive roles of the solvent and
labile ligand of the PEPPSI complex in leading to Pd0 species
Keywords: amination · density functional calculations ·
were ruled out. Nevertheless, whether the reduction of PdII
palladium · reaction mechanisms · s-bond metathesis
species could be realized via plausible anilido-containing PdII
intermediates were computationally evaluated. Computational
results suggest that all reductive elimination reactions of the [1] Reviews: a) J. F. Hartwig in Modern Arene Chemistry (Ed.: D. Astruc),
Wiley-VCH, Weinheim, 2002, pp. 107 – 168; b) L. Jiang, S. L. Buchwald in
anilido-containing PdII intermediates have relatively high
Metal-Catalyzed Cross-Coupling Reactions, 2nd ed. (Eds.: A. de Meijere, F.
energy barriers and are thermodynamically disfavored. In con- Diederich), Wiley-VCH, Weinheim, 2004, pp. 699 – 760.
trast, the corresponding PdII-mediated s-bond metathesis [2] Reviews: a) J. P. Wolfe, S. Wagaw, J.-F. Marcoux, S. L. Buchwald, Acc.
pathway has a lower energy barrier leading to the CN cross- Chem. Res. 1998, 31, 805 – 818; b) J. F. Hartwig, Acc. Chem. Res. 1998, 31,
852 – 860; c) J. F. Hartwig, Nature 2008, 455, 314 – 322; d) J. F. Hartwig,
coupling product and is much more feasible than the RE route
Acc. Chem. Res. 2008, 41, 1534 – 1544; e) D. S. Surry, S. L. Buchwald,
both kinetically and thermodynamically. Other proposed reac- Angew. Chem. Int. Ed. 2008, 47, 6338 – 6361; Angew. Chem. 2008, 120,
tion pathways, such as PdII/PdIV mechanism, single electron 6438 – 6461; f) D. S. Surry, S. L. Buchwald, Chem. Sci. 2011, 2, 27 – 50;
transfer (SET) mechanism, and halide atom transfer (HAT) g) D. Maiti, B. P. Fors, J. L. Henderson, Y. Nakamura, S. L. Buchwald,
mechanism, cannot compete with the s-bond metathesis Chem. Sci. 2011, 2, 57 – 68.
[3] a) W. A. Herrmann, Angew. Chem. Int. Ed. 2002, 41, 1290 – 1309; Angew.
mechanism either. Therefore, in some special cases, if reactants Chem. 2002, 114, 1342 – 1363; b) T. M. Trnka, R. H. Grubbs, Acc. Chem.
and additives are insufficient to reduce a PdII precatalyst, a PdII- Res. 2001, 34, 18 – 29.
mediated s-bond metathesis mechanism might be operative [4] Reviews: a) M. S. Viciu, S. P. Nolan, Top. Organomet. Chem. 2005, 14,
241 – 278; b) N. Marion, S. P. Nolan, Acc. Chem. Res. 2008, 41, 1440 –
to drive the BH amination reaction. The unprecedented
1449; c) G. C. Fortman, S. P. Nolan, Chem. Soc. Rev. 2011, 40, 5151 –
mechanism proposed herein is expected to stimulate further 5169; d) E. A. Kantchev, C. J. O’Brien, M. G. Organ, Angew. Chem. Int. Ed.
mechanistic studies of BH amination reactions. 2007, 46, 2768 – 2813; Angew. Chem. 2007, 119, 2824 – 2870; e) C. Val-
ente, S. C˛alimsiz, K. H. Hoi, D. Mallik, M. Sayah, M. G. Organ, Angew.
Chem. Int. Ed. 2012, 51, 3314 – 3332; Angew. Chem. 2012, 124, 3370 –
3388.
Computational Methods [5] a) M. S. Viciu, R. M. Kissling, E. D. Stevens, S. P. Nolan, Org. Lett. 2002, 4,
2229 – 2231; For derivatives of 1a, see b) O. Esposito, P. M. P. Gois,
The wB97XD[24] density functional method, which contains both A. K. de K. Lewis, S. Caddick, F. G. N. Cloke, P. B. Hitchcock, Organometal-
long-range exchange and empirical dispersion corrections, was em- lics 2008, 27, 6411 – 6418.
ployed in this work to carry out the computations. This functional [6] a) M. S. Viciu, R. A. Kelly, E. D. Stevens, F. Naud, M. Studer, S. P. Nolan,
has shown good performance in describing processes involving Org. Lett. 2003, 5, 1479 – 1482; For derivatives of 1b, see b) J. Broggi, H.
weak interactions and bond breaking/making.[25] The LANL2DZ Clavier, S. P. Nolan, Organometallics 2008, 27, 5525 – 5531.
basis set, in conjunction with the LANL2DZ pseudopotential,[26] was [7] a) O. Navarro, N. Marion, N. M. Scott, J. Gonzlez, D. Amoroso, A. Bell,
used for the Pd atom, whereas the 6-31G(d) basis set[27] was used S. P. Nolan, Tetrahedron 2005, 61, 9716 – 9722; For derivatives of 1c, see
b) N. Marion, E. C. Ecarnot, O. Navarro, D. Amoroso, A. Bell, S. P. Nolan, J.
for other atoms in the geometry optimizations. Vibrational
Org. Chem. 2006, 71, 3816 – 3821; c) N. Marion, P. Frmont, I. M. Puijk,
frequency analyses at the same level of theory were performed on E. C. Ecarnot, D. Amoroso, A. Bell, S. P. Nolan, Adv. Synth. Catal. 2007,
all optimized structures to characterize stationary points as local 349, 2380 – 2384; d) O. H. Winkelmann, A. Riekstins, S. P. Nolan, O. Navar-
minima or transition states. Further, intrinsic reaction coordinate ro, Organometallics 2009, 28, 5809 – 5813; e) N. Marion, O. Navarro, E. D.
(IRC) computations were carried out to confirm that transition Stevens, E. C. Ecarnot, A. Bell, D. Amoroso, S. P. Nolan, Chem. Asian J.
states connect appropriate reactants and products. The gas-phase 2010, 5, 841 – 846; f) S. Meiries, A. Chartoire, A. M. Z. Slawin, S. P. Nolan,

Chem. Eur. J. 2015, 21, 4153 – 4161 [Link] 4160  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

Organometallics 2012, 31, 3402 – 3409; g) S. Meiries, K. Speck, D. B. [16] H. Nakao, H. Shiigi, Y. Yamamoto, S. Tokonami, T. Nagaoka, S. Sugiyama,
Cordes, A. Z. Slawin, S. P. Nolan, Organometallics 2013, 32, 330 – 339. T. Ohtani, Nano Lett. 2003, 3, 1391 – 1394.
[8] a) M. S. Viciu, R. F. Germaneau, O. Navarro-Fernandez, E. D. Stevens, S. P. [17] Only a thermodynamic evaluation of the deprotonation step was per-
Nolan, Organometallics 2002, 21, 5470 – 5472; b) M. S. Viciu, O. Navarro, formed, because the deprotonation step has been suggested not to be
R. F. Germaneau, R. A. Kelly, W. Sommer, N. Marion, E. D. Stevens, L. Cav- the rate-limiting step when a stronger base, such as KOtBu, is employe-
allo, S. P. Nolan, Organometallics 2004, 23, 1629 – 1635; c) M. J. Cawley, d.[12a]
F. G. N. Cloke, R. J. Fitzmaurice, S. E. Pearson, J. S. Scott, S. Caddick, Org. [18] a) G. O. Jones, P. Liu, K. N. Houk, S. L. Buchwald, J. Am. Chem. Soc. 2010,
Biomol. Chem. 2008, 6, 2820 – 2825. 132, 6205 – 6213; b) S. Zhang, Z. Zhu, Y. Ding, Dalton Trans. 2012, 41,
[9] a) O. Navarro, N. Marion, J. Mei, S. P. Nolan, Chem. Eur. J. 2006, 12, 13832 – 13840.
5142 – 5148; b) N. Marion, O. Navarro, J. Mei, E. D. Stevens, N. M. Scott, [19] a) D. C. Powers, T. Ritter, Nat. Chem. 2009, 1, 302 – 309; b) D. C. Powers,
S. P. Nolan, J. Am. Chem. Soc. 2006, 128, 4101 – 4111; c) A. Chartoire, X. M. A. L. Geibel, J. E. M. N. Klein, T. Ritter, J. Am. Chem. Soc. 2009, 131,
Frogneux, S. P. Nolan, Adv. Synth. Catal. 2012, 354, 1897 – 1901; d) A. 17050 – 17051; c) D. C. Powers, D. Benitez, E. Tkatchouk, W. A. Goddard,
Chartoire, A. Boreux, A. R. Martin, S. P. Nolan, RSC Adv. 2013, 3, 3840 – T. Ritter, J. Am. Chem. Soc. 2010, 132, 14092 – 14103; d) D. C. Powers,
3843; e) P. D. Hruszkewycz, D. Balcells, L. M. Guard, N. Hazari, M. Tilset, J. D. Y. Xiao, M. A. L. Geibel, T. Ritter, J. Am. Chem. Soc. 2010, 132, 14530 –
Am. Chem. Soc. 2014, 136, 7300 – 7316. 14536; e) D. C. Powers, T. Ritter, Acc. Chem. Res. 2012, 45, 840 – 850.
[10] a) C. J. O’Brien, E. A. B. Kantchev, C. Valente, N. Hadei, G. A. Chass, A. [20] a) M. C. Nielsen, E. Lyngvi, F. Schoenebeck, J. Am. Chem. Soc. 2013, 135,
Lough, A. C. Hopkinson, M. G. Organ, Chem. Eur. J. 2006, 12, 4743 – 1978 – 1985; b) K. J. Bonney, F. Schoenebeck, Chem. Soc. Rev. 2014, 43,
4748; b) M. G. Organ, S. Avola, I. Dubovyk, N. Hadei, E. A. B. Kantchev, 6609 – 6638.
C. J. O’Brien, C. Valente, Chem. Eur. J. 2006, 12, 4749 – 4755. [21] We evaluated the effect of the addition of TEMPO (1.0 mmol) to experi-
[11] a) M. G. Organ, M. Abdel-Hadi, S. Avola, I. Dubovyk, N. Hadei, E. A. B. mentally examine whether radical pathways are involved in mediating
Kantchev, C. J. O’Brien, M. Sayah, C. Valente, Chem. Eur. J. 2008, 14, the BH reactions. The BH amination reactions were not completely sup-
2443 – 2452; For derivatives of 1f, see b) A. Chartoire, X. Frogneux, A. pressed, although the addition of TEMPO played a subtle role in de-
Boreux, A. M. Z. Slawin, S. P. Nolan, Organometallics 2012, 31, 6947 – creasing the yield of Ph2NH (see the Supporting Information, Table S1).
6951; c) K. H. Hoi, J. A. Coggan, M. G. Organ, Chem. Eur. J. 2013, 19, [22] A. Milet, A. Dedieu, G. Kapteijn, G. van Koten, Inorg. Chem. 1997, 36,
843 – 845; d) L. Zhu, T.-T. Gao, L.-X. Shao, Tetrahedron 2011, 67, 5150 – 3223 – 3231.
5155; e) L. Zhu, Y.-M. Ye, L.-X. Shao, Tetrahedron 2012, 68, 2414 – 2420; [23] E. C. Ashby, A. B. Goel, J. N. Argyropoulo, Tetrahedron Lett. 1982, 23,
f) W.-X. Chen, L.-X. Shao, J. Org. Chem. 2012, 77, 9236 – 9239; g) Z. Jin, L.- 2273 – 2276.
L. Qiu, Y.-Q. Li, H.-B. Song, J.-X. Fang, Organometallics 2010, 29, 6578 – [24] J.-D. Chai, M. Head-Gordon, Phys. Chem. Chem. Phys. 2008, 10, 6615 –
6586; h) Y.-J. Li, J.-L. Zhang, X.-J. Li, Y. Geng, X.-H. Xu, Z. J. Jin, Organo- 6620.
met. Chem. 2013, 737, 12 – 20; i) Z. Jin, S.-X. Guo, X.-P. Gu, L.-L. Qiu, H.-B. [25] a) L. Dang, M. F. Shibl, X. Yang, D. J. Harrison, A. Alak, A. J. Lough, U.
Song, J.-X. Fang, Adv. Synth. Catal. 2009, 351, 1575 – 1585; j) T. Tu, W. Fekl, E. N. Brothers, M. B. Hall, Inorg. Chem. 2013, 52, 3711 – 3723; b) H.
Fang, J. Jiang, Chem. Commun. 2011, 47, 12358 – 12360; k) W. Fang, J. Li, M. B. Hall, J. Am. Chem. Soc. 2014, 136, 383 – 395.
Jiang, Y. Xu, J. Zhou, T. Tu, Tetrahedron 2013, 69, 673 – 679; l) Y. Takeda, [26] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299 – 310.
Y. Ikeda, A. Kuroda, S. Tanaka, S. Minakata, J. Am. Chem. Soc. 2014, 136, [27] P. C. Hariharan, J. A. Pople, Theor. Chim. Acta. 1973, 28, 213 – 222.
8544 – 8547. [28] A. V. Marenich, C. J. Cramer, D. G. Truhlar, J. Phys. Chem. B 2009, 113,
[12] a) K. H. Hoi, S. C˛alimsiz, R. D. J. Froese, A. C. Hopkinson, M. G. Organ, 6378 – 6396.
Chem. Eur. J. 2011, 17, 3086 – 3090; b) K. H. Hoi, S. C˛alimsiz, R. D. J. [29] Gaussian 09, Revision C.01, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Froese, A. C. Hopkinson, M. G. Organ, Chem. Eur. J. 2012, 18, 145 – 151. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
[13] a) M. G. Organ, G. A. Chass, D. C. Fang, A. C. Hopkinson, C. Valente, Syn- nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
thesis 2008, 2776 – 2797; b) C. Valente, S. Baglione, D. Candito, C. J. A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,
O’Brien, M. G. Organ, Chem. Commun. 2008, 735 – 737; c) J. Nasielski, N. K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Hadei, G. Achonduh, E. A. B. Kantchev, C. J. O’Brien, A. Lough, M. G. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, J. E. Peralta, F. Ogliaro, M.
Organ, Chem. Eur. J. 2010, 16, 10844 – 10853. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobaya-
[14] a) R. O. Navarro, R. A. Kelly, S. P. Nolan, J. Am. Chem. Soc. 2003, 125, shi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J.
16194 – 16195; b) O. Navarro, N. Marion, Y. Oonishi, R. A. Kelly, S. P. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B. Cross, V.
Nolan, J. Org. Chem. 2006, 71, 685 – 692. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
[15] To rule out the potential dimerization or oligomerization of the aniline A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Moro-
under the conditions of excess aniline, a stoichiometric reaction of 1f kuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dap-
(0.2 mmol) in the presence of 1.2 equivalents of KOtBu in aniline (3 mL) prich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski,
was performed as a control experiment. The crude reaction mixture D. J. Fox, Gaussian, Inc., Wallingford CT, 2010.
was subsequently analyzed by LC-MS. The molecular ion peaks of dime-
rization or other oligomerizations of the aniline were not detected in
the LC-MS analysis. Nevertheless, the molecular ion peak ([M+H] + =
171) was found, which is assigned to the cross-coupling product of ani- Received: November 16, 2014
line with 3-chloropyridine, the labile ligand of 1f. Published online on January 29, 2015

Chem. Eur. J. 2015, 21, 4153 – 4161 [Link] 4161  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim