Mechanistic study of copper-catalyzed intramolecular ortho-C-H activation/carbon-nitrogen and carbon-oxygen cyclizations

Article abstract of DOI:10.1007/s11426-012-4795-3

Intramolecular ortho-C-H activation and C-N/C-O cyclizations of phenyl amidines and amides have recently been achieved under Cu catalysis. These reactions provide important examples of Cu-catalyzed functionalization of inert C-H bonds, but their mechanism

Full text of DOI:10.1007/s11426-012-4795-3

SCIENCE CHINA
Chemistry
May 2013 Vol.56 No.5: 619–632
doi: 10.1007/s11426-012-4795-3
• ARTICLES •
Mechanistic study of copper-catalyzed intramolecular ortho-C–H
activation/carbon-nitrogen and carbon-oxygen cyclizations
TANG ShiYa, GONG TianJun & FU Yao*
Department of Chemistry, University of Science and Technology of China, Hefei 230026, China
Received August 4, 2012; accepted August 27, 2012; published online November 28, 2012
Intramolecular ortho-C–H activation and C–N/C–O cyclizations of phenyl amidines and amides have recently been achieved
under Cu catalysis. These reactions provide important examples of Cu-catalyzed functionalization of inert C–H bonds, but
their mechanisms remain poorly understood. In the present study the several possible mechanisms including electrophilic aro-
matic substitution, concerted metalation-deprotonation (CMD), Friedel-Crafts mechanism, radical mechanism, and proton-
coupled electron transfer have been theoretically examined. Cu(II)-assisted CMD mechanism is found to be the most feasible
for both C–O and C–N cyclizations. This mechanism includes three steps, i.e. CMD with Cu(II), oxidation of the Cu(II) inter-
mediate, and reductive elimination from Cu(III). Our calculations show that Cu(II) mediates the C–H activation through an
six-membered ring CMD transition state similar to that proposed for many Pd-catalyzed C–H activation reactions. It is also in-
teresting to find that the rate-limiting steps are different for C–N and C–O cyclizations: for the former it is concerted
metalation-deprotonation with Cu(II), whereas for the latter it is reductive elimination from Cu(III). The above conclusions are
consistent with the experimental kinetic isotope effects (1.0 and 2.1 for C–O and C–N cyclizations, respectively), substituent
effects, and the reactions under O2-free conditions.
mechanism, DFT, copper, C–H activation, concerted metalation-deprotonation
Cu(OTf)2-catalyzed ortho-C–H activation/C–O cyclization
of amides [9].
1 Introduction
Despite the great contemporary interest in Cu-catalyzed
C–H activation, it remains unclear how these reactions take
place. For Cu(OAc)2-catalyzed pyridine-directed C–H acti-
Transition metal-catalyzed C–H activation has emerged as a
powerful method in synthetic organic chemistry [1]. Such
reactions promoted by Rh- [2], Ru- [3], and Pd-catalysts [4]
have been extensively studied, but Cu-catalyzed C–H acti-
vation began to draw attentions as a complementary method
only recently [5]. In this area, much work has been directed
to the Cu-catalyzed functionalization of acidic C–H bonds
as triggered by base [6], whereas base-free activation of
inert C–H bonds by Cu catalysts has been less studied. In
2006 Yu et al. reported Cu(OAc)2-catalyzed intermolecular
C–H functionalization (Scheme 1) [7]. Later, Buchwald et
al. reported Cu(OAc)2-catalyzed ortho-C–H activation/C–N
cyclization of amidines [8], while Nagasawa et al. reported
Scheme 1 Representative examples for Cu-catalyzed activation of inert
C–H bonds.
*Corresponding author (email: fuyao@ustc.edu.cn)
© Science China Press and Springer-Verlag Berlin Heidelberg 2012
chem.scichina.com www.springerlink.com

620
Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
vation, Yu et al. proposed a single electron transfer process
involving a cation-radical intermediate [7, 10]. On the other
hand, both Buchwald and Nagasawa suggested an electro-
philic attack mechanism in their C–N and C–O cyclization
reactions [8, 9]. Even so, the electrophilic attack may occur
at either Cu or Cu-coordinated nitrogen/oxygen atoms [5, 8].
Furthermore, although Wang [11] and Stahl [12] identified
macrocyclic aryl–Cu(III) complexes in some model C–H
activation processes [11], it is unclear whether Cu(III) is
still involved for the reactions without macrocyclic ligands
[8, 9, 13]. Finally, very few theoretical studies have been
carried out on Cu-catalyzed C–H activation [14]. Wu et al.
proposed a Heck-like transition state for Cu-mediated meta-
C–H functionalization, whereas Lin et al. proposed a
four-membered concerted metalation–deprotonation transi-
tion state for the coupling of Ar-H with PhI (Scheme 2) [14].
These theoretical results may only be applied to intermolec-
ular reactions, because neither the Heck-like nor the
four-membered concerted metalation-deprotonation transi-
tion state appears possible for intramolecular reactions due
to the skeleton restraint.
In an attempt to elucidate the mechanism, we have car-
ried out a thorough theoretical study of Cu-catalyzed
ortho-C–H activation/C–N and C–O cyclization reactions.
Several possible pathways are examined (Scheme 3) [5b, 8,
9]. In mechanism A, the N- or O-bound Cu atom in the
Cu-coordinated complex 2 attacks the aromatic ring to gen-
erate a Wheland intermediate 3, which then undergoes aro-
matization and reductive elimination to give the product 9.
This mechanism involves stepwise metalation and deproto-
nation. Different from mechanism A, mechanism B in-
volves a concerted metalation and deprotonation to form an
aryl–Cu complex 4 instead of a Wheland intermediate. Such
a process is similar to that proposed for Pd-catalyzed C–H
activation [15]. In mechanism C, the Cu-bound N or O atom
in 2 attacks the aromatic ring to form 5 with concerted re-
lease of a reduced Cu species, followed by re-aromatization
to provide 9. Mechanism A and C were both mentioned for
the benzimidazole-forming process in Buchwald’s work [8].
Moreover, mechanisms A and C may also proceed in a rad-
ical manner (i.e. mechanisms D and E) [5]. In mechanisms
D and E, a single electron transfer from the aromatic ring to
the coordinated Cu(II) takes place to generate a radical cation
intermediate 6 [7]. The following steps occur in a similar
way as that in mechanisms A and C. None of the above
mechanisms can be excluded by the available experimental
data. In addition, two specific details of these mechanisms
remain unknown, namely (1) which step is rate-determining,
and (2) is the oxidation state of the key Cu intermediate +2
or +3?
Here mechanisms A, B and C with Cu in either +2 or +3
oxidation state are examined. Our results indicate that the
mechanism B (concerted metalation–deprotonation, CMD)
is energetically favored for Cu-catalyzed intramolecular
C–H activation/C–N and C–O cyclizations. It consists of
three steps, i.e. concerted metalation-deprotonation with
Cu(II), oxidation of the Cu(II) intermediate, and reductive
elimination from Cu(III). Importantly, it is Cu(II) that me-
diates the C–H activation through an acetate/triflurome-
thanesulfonate (triflate)(OAc/OTf)-assisted CMD process.
Moreover, while reductive elimination is the rate-limiting
step for C–O cyclization, CMD is the rate-limiting step for
C–N cyclization. The theoretical mechanism is consistent
with the experimental kinetic isotope effects and substituent
effects. We further investigated the radical pathways
(mechanisms D and E) and proton-coupled electron transfer.
According to the radical scavenger experiment and related
calculations, we found that neither the radical pathway nor
proton-coupled electron transfer was plausible in the C–O
and C–N cyclization reactions.
2 Calculations of different mechanisms
2.1 Mechanisms A, B and C for C–O cyclization
2.1.1 Problem with mechanism A
Mechanism A (i.e. electrophilic aromatic substitution, SEAr)
has previously been proposed for Cu-catalyzed C–H func-
tionalization by Buchwald and Nagasawa et al., because they
found that the related reaction with an electron-donating sub-
stituent (R = OMe) is faster than that with an electron-
withdrawing group (R = Cl or COOMe) [8, 9]. However,
with the help of pulse electron paramagnetic resonance
(EPR) methods in combination with denstity functional the-
ory (DFT) calculations, Ribas et al. precluded the Wheland
intermediate (the core of SEAr mechanism) in the
Cu-catalyzed C–H cleavage reaction [10]. In the present
study, we employed the electrophilic cationic Cu(II) species
to investigate the SEAr pathway which involved the im-
portant intermediate O-3 (shown in Schemes 1 and 4).
However, all attempts resulted in the structure of O-IntA.
To examine whether or not O-IntA is a Wheland interme-
diate, we tested the aromaticity of Ring1 in O-IntA by the
harmonic oscillator model of aromaticity (HOMA) index
[16] and the nucleus-independent chemical shift (NICS)
index [17]. The reason for this is that significant loss of
aromaticity is a critical character of Wheland intermediate.
As a structure-based measure of aromaticity, HOMA in-
dexes were 0.968 and 0.994 for O-IntA and the amide, re-
Scheme 2 Previously proposed transition states for Cu-catalyzed C–H
activation.

Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
621
Scheme 3 Possible mechanisms for Cu-catalyzed ortho-C–H activation/C–N and C–O cyclizations.
tion; (ii) concerted metalation–deprotonation; (iii) reductive
elimination and catalyst regeneration. Cu has two possible
oxidation states, i.e. Cu(II) and Cu(III). Therefore, we have
to examine both oxidation states for the CMD mechanism
as shown below.
(i) CMD of Cu(III) intermediate. Cu(III) intermediates
are commonly proposed for Cu-catalyzed Ullmann reactions
[13, 18]. Recently, Stahl et al. proposed that Cu(III) species
may be involved in some Cu-catalyzed intramolecular
ortho-C–H activation [5, 19]. For the present reaction, we
propose that Cu(OTf)2 first coordinates to the O atom of the
reactant 1 to give a Cu(II) complex O-Int1 [20] (exergonic,
13.8 kcal/mol, Figures 1 and 2). Subsequently, O-Int1 is
oxidized by another molecule of Cu(OTf)2 to give a Cu(III)
intermediate O/CMD-Int1-III. Note that a similar dispro-
portionation reaction of Cu(II) salts was previously de-
scribed by Ribas and Stahl et al. [21]. With the help of tri-
flate, the Cu(III) atom of O/CMD-Int1-III reacts with the
N-phenyl ring through a six-membered-ring transition state
O/CMD-TS12-III. The activation energy for this transition
state is +42.1 kcal/mol as calculated from O-Int1 to
O/CMD-TS12-III. Because O/CMD-TS12-III corresponds
to the breaking of the C–H bond and concerted formation of
the HO and CuC bonds, this process can be considered as
a CMD process [15]. An aryl–Cu(III) [5, 12, 22] complex is
Scheme 4 Related intermediate in mechanism A.
spectively. As a magnetic measure of aromaticity, the
NICS(1) indexes were 8.69 and 10.16 ppm for O-IntA
and the amide, respectively. These data both indicated that
the aromaticity of Ring1 in O-IntA does not decrease sig-
nificantly as compared to the substrate amide [10]. There-
fore, O-IntA is not a Wheland intermediate. Without the
existence of the key Wheland intermediate (O-3), the SEAr
mechanism does not appear possible in Cu-catalyzed intra-
molecular cyclization.
2.1.2 Concerted metalation–deprotonation mechanism
Mechanism B (i.e. the CMD mechanism) is often proposed
for transition metal-catalyzed C–H activation [1]. This
mechanism consists of three main steps: (i) metal coordina-

622
Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
Figure 1 Cu(III)-assisted concerted metalation-deprotonation mechanism for C–O formation.
generated in the CMD process, which can easily undergo
reductive elimination to give the target product with a low
activation energy of +13.8 kcal/mol. After reductive elimi-
nation, the product and reduced Cu(I) species are formed.
With the aid of O2, the Cu(I) species is easily re-oxidized to
Cu(II) to finish the catalyst cycle.
the catalytic cycle (Figure 2).
Pathway 1: direct reductive elimination (unfavorable).
The Cu(II) intermediate O/CMD-Int2-II may directly un-
dergo reductive elimination to afford the product via a
three-center transition state O/CMD-TS23-II. Meanwhile, a
Cu(0) species is generated and will be converted to the ini-
The rate-determining step of the above pathway is the
concerted metalation–deprotonation process with an overall
barrier of +42.1 kcal/mol (Figure 1). This barrier is too high
for the reaction to proceed at the reported temperature (i.e.
140 °C). Therefore, the Cu(III)-assisted CMD mechanism
does not appear to be favored.
(ii) CMD of Cu(II) intermediate. This pathway also starts
from the Cu(II) complex O-Int1. However, instead of being
oxidized to Cu(III), the Cu(II) metal center attacks the
N-phenyl ring directly and at the same time the triflate ani-
on coordinated with Cu(II) abstracts the proton from the
aryl ring (Figures 3 and 4). This process proceeds via a cy-
clic six-membered ring transition state O/CMD-TS12-II
involving breaking of the C–H bond (1.42 Å) and concerted
formation of the H–O (1.20 Å) and Cu–C (2.02 Å) bonds.
Evidently this process is also a CMD process, but its activa-
tion energy is calculated to be as low as +20.9 kcal/mol (from
O-Int1 to O/CMD-TS12-II). A new four-coordinated Cu(II)
intermediate O/CMD-Int2-II is generated through the
CMD process. This intermediate has two pathways to finish
. Unfortunately, this
tial Cu(II) catalyst with the aid of O2
reductive elimination is calculated to be a very difficult
process with a high activation energy (+50.1 kcal/mol).
Pathway 2: oxidation and reductive elimination (favorable).
Alternatively O/CMD-Int2-II may be oxidized by another
molecule of Cu(OTf)2 to produce a four-coordinated
Cu(III) intermediate O/CMD-Int3-III (Note: O/CMD-
Int3-III is also involved in the Cu(III)-assisted CMD pro-
cess). Our calculation shows that this oxidation is only
slightly endergonic by +2.8 kcal/mol. Subsequently, O/CMD-
Int2-III can undergo reductive elimination easily to give
the product O-9 with a lower activation energy (+31.6
kcal/mol).
By comparing the above two pathways (Figure 2), we
conclude that the Cu(II)-assisted CMD mechanism should
involve three steps: concerted metalation-deprotonation
with Cu(II), oxidation of Cu(II) intermediate, and reductive
elimination from Cu(III). Reductive elimination is the
rate-limiting step and the overall barrier of +31.6 kcal/mol.
Evidently Cu(II)-assisted CMD mechanism is more favora-

Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
623
Figure 2 Cu(II)-assisted concerted metalation–deprotonation pathway for C–O formation.
Figure 3 Important structures in Cu(II)-assisted concerted metalation-deprotonation pathway for C–O formation.
ble than its Cu(III) counterpart.
kcal/mol (Figure 4). This electrophilic attack leads to the
intermediate O/FC-Int2-II and a reduced Cu(0) species.
After removal of the reduced Cu(0) from O/FC-Int2-II,
rearomatization takes place via the triflate-assisted depro-
tonation transition state O/FC-TS23-II. Then HOTf is re-
leased to generate benzoxazole as the final product. Mean-
while Cu(0) is re-oxidized to Cu(II). Note that the
rate-limiting step of the above process is deprotonation with
an overall activation energy of +49.8 kcal/mol. This barrier
is too high for the reaction to proceed at the reported tem-
perature (140 °C).
2.1.3 Friedel-Crafts mechanism
An alternative pathway to provide the C–H functionaliza-
tion products is Lewis acid-catalyzed Friedel-Crafts mecha-
nism (i.e. mechanism C) [23, 24]. With this mechanism the
reaction is proposed to include three steps: (i) metal coordi-
nation; (ii) electrophilic substitution of Cu-bound O atom;
(iii) deprotonation and catalyst regeneration. Different from
the CMD mechanism, electrophilic attack and deprotonation
are separated steps in mechanism C.
(i) Cu(II)-assisted Friedel-Crafts pathway. For this
pathway the reaction starts with the coordination of Cu(II)
with the substrate. After coordination, the Cu(II) complex
O-Int1 isomerizes to an active precursor O/FC-Int1-II
which can cause an electrophilic attack through the transi-
tion state O/FC-TS12-II with an activation energy of +48.8
(ii) Cu(III)-assisted Friedel–Crafts pathway. As dis-
cussed above, the active Cu(III) complex O/CMD-Int1-III
can also be generated in the system through oxidation. After
oxidation, the Cu(III)-bound O atom in O/CMD-Int1-III
can react with the aromatic ring through electrophilic attack
(Figure 5). A Cu(I) intermediate, i.e. O/FC-Int2-III-1, is

624
Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
Figure 4 Cu(II)-assisted Friedel-Crafts pathway for C–O formation.
Figure 5 Cu(III)-assisted Friedel-Crafts pathway in C–O formation.

Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
625
then generated, which can undergo deprotonation readily
through the transition state O/FC-TS23-III. Finally, HOTf
is released to generate benzoxazole as the final product.
Meanwhile Cu(I) is re-oxidized to Cu(II). In the above pro-
cess, the rate-determining step is electrophilic attack and the
overall barrier is +34.3 kcal/mol. This barrier is lower than
that for Cu(II)-assisted Friedel-Crafts pathway, but is still
higher than that for Cu(II)-assisted CMD mechanism.
its Cu(III) counterpart and Cu(II)/Cu(III)-assisted Friedel-
Crafts pathways [26] (Figure 7).
The Cu(II)-assisted CMD pathway also consists of three
steps in the C–N cyclization, i.e. concerted metalation-
deprotonation with Cu(II), oxidation of the Cu(II) interme-
diate, and reductive elimination from Cu(III). As Figure 8
shows, the imine moiety in the substrates can coordinate
with the catalyst Cu(OAc)2 to provide a stable Cu(II) com-
plex N-Int1 (exergonic, 4.8 kcal/mol). The Cu(II) center
of N-Int1 then attacks the N-phenyl ring via the
six-membered-ring transition state N/CMD-TS12-II. This
process is a CMD process involving the breaking of the
CPh–H bond and concerted formation of C–Cu bond. After
this CMD process, a new four-coordinated Cu(II) complex
N/CMD-Int2-II-1 is obtained which then may be oxidized
to a Cu(III) intermediate N/CMD-Int2-III-2. Subsequently,
N/CMD-Int2-III-2 undergoes reductive elimination to give
the product. This process shows a barrier of +13.2 kcal/mol
(from N/CMD-Int2-II-1 to N/CMD-TS23-III).
2.1.4 Comparing different pathways in C–O formation
To summarize the above calculations, the energy profiles
for different mechanisms of Cu-catalyzed ortho-C–H acti-
vation/C–O formation are shown in Figure 6. According to
these profiles, Cu(II)-assisted CMD pathway is favored as
compared to its Cu(III) counterpart. It is also more favora-
ble than either Cu(II)- or Cu(III)-assisted Friedel–Crafts
pathway. This Cu(II)-assisted CMD pathway consists of
three steps, i.e. concerted metalation- deprotonation with
Cu(II), oxidation of Cu(II) intermediate, and reductive
elimination from Cu(III). The overall barrier is +31.6
kcal/mol and the rate-limiting step is reductive elimination
from Cu(III) to generate the C–O bond.
Different from C–O cyclization reaction, the concerted
metalation–deprotonation step is the rate-limiting step for
the C–N cyclization. The overall activation energy for the
C–N cyclization is +22.0 kcal/mol.
Note that the overall energy barrier of the C–N cycliza-
tion (+22.0 kcal/mol) is lower than that of the C–O cycliza-
tion (+31.6 kcal/mol). This result is consistent with the ex-
periments that the C–N cyclization can be carried out under
milder conditions than the C–O cyclization (100 °C for the
former versus 140 °C for the latter). Nonetheless, we be-
lieve that the energy barrier for the C–N cyclization (+22.0
2.2 Mechanism for C–N cyclization
Just as for the C–O cyclization of amides, Cu-catalyzed
ortho-C–H activation/C–N cyclization of amidines may
proceed via either B or C mechanism with +2 or +3 oxida-
tion states of Cu [8, 25]. Based on our calculations, we also
find that the Cu(II)-assisted CMD pathway is favored over
Figure 6 Comparison of the energy profiles of mechanisms of Cu-catalyzed ortho-C–H activation/C–O cyclization reactions.

626
Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
Figure 7 Comparison of the energy profiles of mechanisms of Cu-catalyzed ortho-C–H activation/C–N cyclization reactions.
Figure 8 Cu(II)-assisted concerted metalation–deprotonation pathway in C–N formation. Details of other disfavored pathways (Cu(II)/Cu(III)-assisted
Friedel–Craft pathways and Cu(III)-assisted CMD pathway) are shown in the Supporting Information.

Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
627
kcal/mol) is underestimated to some extent. The reason is
that we assume that the reactants of the C–N cyclization are
compound N-1 and free Cu(OAc)2. Due to the strong bind-
ing ability of acetate anion, Cu(OAc)2 may exist as clusters
(i.e. Cun(OAc)2n) so that the effective concentration of
Cu(OAc)2 may be much lower than theoretically assumed.
This effect is difficult to estimate by current theoretical
methods, but it means that the energy barrier is underesti-
mated by our calculations.
the presence of 1, 4-cyclohexadiene (which is another effec-
tive radical trapper used by Sanford et al.) [29] afforded the
products in 60 [28] and 44% yields, respectively (Table 1).
The results indicate that the reactions are not inhibited by
the radical scavengers.
In addition, we tried to calculate the important interme-
diate O-6, an arene radical species in a single electron
transfer route. In theory O-6 and O-Int1 could not be com-
pared in energy because they have the same geometry as
O-2-II shown in Scheme 5. The difference between O-Int1
and O-6 is the aromaticity of ring1 and thus we calculated
the spin density of O-2-II to distinguish two species. The
total spin density observed over the six C atoms of ring1 is
0.007 electron (1 electron for the molecule O-2-II). More-
over, the HOMA index of ring1 is 0.999 (0.994 for the am-
ide O-1). These data both suggest that the ring1 must pos-
sess aromaticity, which is consistent with the Cu(II) com-
plex O-Int1 rather than the arene radical species O-6.
Without the existence of the arene radical species, the single
electron transfer is not plausible.
We also calculate the free energy of the proposed radical
intermediates (O-8 or N-8 [30]) (Scheme 6). It is found that
the formation of O-8 and N-8 from their Cu-bound precur-
sors is highly endergonic (+55.1 and +32.5 kcal/mol, re-
spectively). Consequently, the activation energies for gen-
eration of O-8 and N-8 must be higher than +55.1 and +32.5
kcal/mol, again indicating that the radical process is not
feasible under the reaction conditions.
2.3 Problems with the radical mechanisms
The radical mechanism via single electron transfer was pre-
viously proposed by Yu et al. to explain Cu(OAc)2-
catalyzed pyridine-directed ortho-C–H functionalization
[7a, 27]. Moreover, Stahl et al. suggested that the radical
pathway should be considered for a related reaction [5b].
Despite these proposals, there has been no experimental test
of the involvement of radical pathways for the C–N and
C–O cyclizations shown in Scheme 1.
To solve this problem, we performed the reactions (Table
1) with radical scavengers. Considering Cu-mediated in-
termolecular C–X (X = N, O) coupling as side reactions, the
commonly used radical scavengers containing N or O atom
(i.e. 2, 2, 6, 6-tetramethylpiperidine-1-oxyl (TEMPO) or
galvinoxyl) were not used. We noticed that Thomas et al.
previously ruled out the radical mechanism for the
Cu-mediated Ullmann arylation by showing that the reactions
was not quenched by 1, 1-diphenylethylene [18b]. Here, for
the Cu-catalyzed intramolecular C–H activation, we exam-
ined the effect of adding 1, 1-diphenylethylene (Table 1)
[18b]. The yields of benzoxazole and benzimidazole were
54 [28] and 41%, respectively. Similarly, the reactions in
2.4 Problems with proton-coupled electron transfer
Ribas et al. reported in 2010 that proton-coupled electron
transfer (PCET) could occur in Cu-mediated C–H cleavage
reactions in the presence of special macrocyclic ligands [10].
They revealed that the C–H···Cu(II) interaction is a vital fea-
Table 1 Effects of radical scavengers on the reaction efficiency
Entry Radical scavenger Yield a) of O-5 (%) Yield a) of N-5 (%)
1
81 (89 b))
45 c)

2
54
41
3
60
44
Scheme 5 The structure of O-2-II is best described as O-Int1, rather than
O-6.
a) Isolated yield; b) yield in the literature [9a]; c) the substrate is not
involved in Buchwald’s reaction [8].

628
Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
way. Therefore, the PCET with a bimolecular Cu-catalyst is
not plausible, either.
3 Comparisons with the experiments
The above theoretical calculations show that Cu(II)-assisted
CMD pathway is energetically favored in both the ortho-
C–H activation/C–O and C–N cyclization reactions. Such
pathway proceeds through three steps: concerted metala-
tion-deprotonation with Cu(II), oxidation of Cu(II) interme-
diate, and reductive elimination from Cu(III) (Scheme 7).
However, in the case of C–O cyclization reductive elimina-
tion is the rate-limiting step, whereas in C–N cyclization
concerted metalation–deprotonation is the rate-limiting step.
Below we use some experimental results to test these theo-
retical proposals.
Scheme 6 Free energies for the formation of proposed radical intermedi-
ates.
ture of the PCET pathway. However, in the present study,
with the aid of OAc/OTf, the C–H···Cu(II) distance is 2.90
Å, much longer than 2.14 Å in the reaction reported by
Ribas. Thus, the proton-coupled electron transfer with uni-
molecular Cu-catalyst does not appear possible in the
Buchwald’s and Nagasawa’s reactions.
Moreover, we also considered the PCET pathway in-
volving bimolecular Cu-catalysts. In such pathway, another
Cu(OAc)2 molecule abstracts the H from the substrate and
at the same time the extra Cu(II) center is reduced by one
electron. In other words, the CMD and oxidation take place
simultaneously with the help of another moleculer of Cu
catalyst. We only calculated the transition state (N/PCET-
TS12) of this pathway in Buchwald’s C–N coupling reac-
tion, because in Buchwald’s reaction the cleavage of C–H
bond is the rate-determining step. The free energy of
N/PCET-TS12 is +48.9 kcal/mol, which is much higher than
that of N/CMD-TS12-II (+17.2 kcal/mol) in the CMD path-
3.1 Kinetic isotope effects
Because concerted metalation-deprotonation (i.e. C–H acti-
vation) is the rate-limiting step in C–N but not in C–O cy-
clization, we expect to see different kinetic isotope effects
(KIE) in the two cyclizations reactions. In the case of C–O
cyclization, Nagasawa et al. previously determined the KIE
value to be 1.0 [9a], which is in good agreement with our
theory.
No KIE study has been reported for C–N cyclization re-
action. Thus we measure the intramolecular kinetic isotope
effect using the ortho-deuterium labeled substrate (Scheme
8). The ratio of the ortho-proton product versus the ortho-
deuterium product is 2.1:1.0. Thus the C–N cyclization
reaction exhibits a KIE that is different from the C–O
cyclization. This result verifies our theoretical prediction
Scheme 7 Proposed mechanism of Cu-catalyzed ortho-C–H activation/C–N and C–O cyclizations.

Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
629
Scheme 8 Intramolecular kinetic isotope effect.
that the C–H bond activation is the rate-limiting step for
C–N coupling.
reactivity of substituted amidines decreases in the order:
OMe > H > COOMe. This calculated trend is also con-
sistent with experiment, thus supporting the mechanism
proposed in Scheme 7.
3.2 Substituent effects
Nagasawa and Buchwald previously examined the reactions
with different substituents (Schemes 9 and 10). They found
that the reaction with an electron-donating substituent (R =
OMe) is faster than that with an electron-withdrawing group
(R = Cl or COOMe). These observations were used to sup-
port the hypothesis that the C–N/C–O cyclization occurs via
an electrophilic attack mechanism [8, 9].
Here we calculate the energy barriers for the reactions
with electron-donating OMe and electron-withdrawing Cl or
COOMe groups on the basis of the proposed mechanism (i.e.
the Cu(II)-assisted CMD pathway). It is found that for C–O
cyclization, the energy barriers are +31.4, +31.6 and +33.0
kcal/mol for OMe, H and Cl substituted substrates. Thus the
reactivity of substituted amides decreases in the order:
OMe > H > Cl. This calculated trend is consistent with ex-
periment. As to C–N cyclization, the energy barriers are
calculated to be +21.1, +22.0 and +22.7 kcal/mol for OMe,
H and COOMe substituted substrates, respectively. Thus the
3.3 O2-free reactions
According to the mechanism in Scheme 7, the generation of
the active Cu(III) species does not require the involvement
of O2. Instead, O2 is needed for the catalyst regeneration
step. If this proposal is correct, we must be able to observe
that the reaction can proceed to some extent without addi-
tion of O2. Indeed, when we carried out the same C–O and
C–N cyclization reactions in under Ar protection using
Schlenk techniques (Schemes 11 and 12), moderate isolated
yields of 28% and 18% were obtained for the C–O and C–N
cyclization products (theoretically, yields of the reaction
should be 50% in the presence of 1 equivalent of Cu salts
and the decrease in yield can be attributed to unknown
side-reactions suppressed by O2). These results confirm that
the active Cu(III) species can be generated via dispropor-
tionation in the absence of O2.
4 Conclusions
Cu-catalyzed C–H functionalization has recently attracted
considerable interest. In the present study, we report a com-
prehensive theoretical study (together with some experi-
mental tests) of the mechanism of the Cu-catalyzed intra-
molecular ortho-C–H activation/C–O and C–N cyclization
Scheme 9 Substituent effects on Cu-catalyzed C–H activation/C–O cy-
clization.
Scheme 11 O2-free C–H activation/C–O cyclization of amides.
Scheme 10 Substituent effects on Cu-catalyzed C–H activation/C–N
cyclization.
Scheme 12 O2-free C–H activation/C–N cyclization of amidines.

630
Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
reactions. By comparing the CMD and Friedel–Crafts
mechanisms with different oxidation states of Cu (+2, +3),
we find that the Cu(II)-assisted concerted metalation-
deprotonation (CMD) pathway is the most favorable for
both the C–O and C–N cyclization reactions. The Cu(II)-
assisted CMD mechanism consists of three steps, i.e. con-
certed metalation-deprotonation with Cu(II), oxidation of
the Cu(II) intermediate, and reductive elimination from
Cu(III). In particular, it is the Cu(II) species that mediates
the C–H activation process through an OAc/OTf-assisted
six-membered-ring CMD transition state similar to that
proposed for many Pd-catalyzed C–H activation reactions.
Also interestingly, while the rate-determining step in the
C–O cyclization reaction is reductive elimination, the
rate-determining step in the C–N cyclization reaction is the
concerted metalation–deprotonation step. This finding is
consistent with the experimental kinetic isotope effects (1.0
and 2.1 for C–O and C–N cyclization reactions, respective-
ly). We also show that the proposed mechanism is con-
sistent with the experimental substituent effects. Finally, we
carry out reactions under O2-free conditions and verify that
the Cu(III)-formation step can proceed without O2.
5.2 General procedure for Cu-catalyzed ortho-C–H
activation/C–O cyclization
To a Schlenk tube was added O-1 (0.25 mmol) and
Cu(OTf)2 (0.05 mmol). The tube was evacuated and back-
filled with O2 (or without O2 for O2-free reactions). Then
o-xylene (0.5 mL) was added via syringe. Note that the rad-
ical scavenger (0.25 mmol) was added for the radical scav-
enger experiment. The reaction mixture was put into a pre-
heated oil bath at 140 °C and stirred for 28 h. Subsequently
a small amount of ethyl acetate was added and the residue
was purified by silica gel chromatography. NMR spectra are
shown in the Supporting Information.
5.3 General procedure for Cu-catalyzed ortho-C–H
activation/C–N cyclization
To a Schlenk tube was added N-1 (0.25 mmol) and
Cu(OAc)2 (0.0375 mmol). The tube was evacuated and
backfilled with O2 (or without O2 for O2-free reaction).
Then HOAc (2.5 mmol, 5.00 equiv.), and DMSO (0.5 mL)
was added by syringe. Note that the radical scavenger (0.25
mmol) was added for the radical scavenger experiments.
The reaction mixture was put into a preheated oil bath at
100 °C and stirred for 18 h. Subsequently a small amount of
ethyl acetate was added and the residue was purified by
silica gel chromatography. NMR spectra are shown in the
Supporting Information.
5
Computational methods and experimental
details
5.1 Computational methods
Geometry optimization of all compounds in the gas phase
without any constraint was conducted by the density func-
tional theory method B3P86 [31] with the basis set
6-31+G(d) [32]. This method has been successfully used in
many recent studies to study the mechanism of transition
metal-catalyzed reactions [10, 33, 34]. Frequency calcula-
tion were conducted at the same level of theory as geometry
optimization to confirm whether the stationary points were
minima or saddle points. For saddle points, intrinsic reac-
tion coordinate (IRC) analysis [35] was performed to verify
that they connect the right reactants and products on the
potential energy surface. Single-point energy calculations
were performed by using the M06 method [36] with a more
flexible basis set 6-311++G(d, p). Note that this method was
recently used by Yates et al. to study Cu-mediated carbox-
ylation reaction [37]. The solvent effect was calculated with
a self-consistent reaction field (SCRF) method using the
SMD model [31a, 38]. o-Xylene ( = 2.545) and dimethyl
sulfoxide (DMSO) ( = 46.826) used in the experiments
were directly calculated. All the calculations were per-
formed with the Gaussian09 suite [39] of programs. Single-
point energies in the solution corrected by the Gibbs free
energy correction from frequency calculations were con-
ducted to describe the reaction energetics throughout the
study. All the solution-phase free energies presented in the
work correspond to the reference state of 1 mol/L, 298K.
We are grateful for the financial support from the National Basic Research
Program of China (973 program, 2012CB215306), the National Natural
Science Foundation of China (NSFC, 20832004, 20972148) and CAS
(KJCX2-EW-J02). The numerical calculations in this paper were carried
out on the supercomputing system in the Supercomputer Center of Univer-
sity of Science and Technology of China.
1
2
Balcells D, Clot E, Eisenstein O. C–H bond activation in transition
metal species from a computational perspective. Chem Rev, 2010,
110: 749–823
Colby DA, Bergman RG, Ellman JA. Rhodium-catalyzed C-C bond
formation via heteroatom-directed C–H bond activation. Chem Rev,
2010, 110: 624–655
3
4
Collet F, Lescot C, Dauban P. Catalytic C–H amination: The stere-
oselectivity issue. Chem Soc Rev, 2011, 40: 1926–1936
a) Lyons TW, Sanford MS. Palladium-catalyzed ligand-directed C–H
functionalization reactions. Chem Rev, 2010, 110: 1147–1169; b) Xu
LM, Li BJ, Yang Z, Shi ZJ. Organopalladium(IV) chemistry. Chem
Soc Rev, 2010, 39: 712–733; For recent examples, see: c) Wang DH,
Engle KM, Shi BF, Yu JQ. Ligand-enabled reactivity and selectivity
in a synthetically versatile aryl C–H olefination. Science, 2010, 327:
315–319; d) Mandal D, Yamaguchi AD, Yamaguchi J, Itami K. Syn-
thesis of dragmacidin D via direct C–H couplings. J Am Chem Soc,
2011, 133: 19660–19663; e) Xiao B, Gong TJ, Liu ZJ, Liu JH, Luo
DF, Xu J, Liu L. Synthesis of dibenzofurans via palladium-catalyzed
phenol-directed C–H activation/C–O cyclization. J Am Chem Soc,
2011, 133: 9250–9253; f) Xiao B, Gong TJ, Xu J, Liu ZJ, Liu L. Pal-
ladium-catalyzed intermolecular directed C–H amidation of aromatic
ketones. J Am Chem Soc, 2011, 133: 1466–1474; g) Ueda K,
Yanagisawa S, Yamaguchi J, Itami K. A general catalyst for the
-selective C–H bond arylation of thiophenes with iodoarenes.

Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
631
Angew Chem Int Ed, 2010, 49: 8946–8949; h) Dai HX, Yu JQ.
Pd-catalyzed oxidative ortho-C–H borylation of arenes. J Am Chem
Soc, 2012, 134: 134–137; i) Wasa M, Engle KM, Lin DW, Yoo EJ,
Yu JQ. Pd(II)-catalyzed enantioselective C–H activation of cyclo-
propanes. J Am Chem Soc, 2011, 133: 19598–19601
the harmonic oscillator model. Tetrahedron Lett, 1972, 13: 3839–
3842; b) Krygowski TM, Cyrañski MK. Structural aspects of aroma-
ticity. Chem Rev, 2001, 101: 1385–1419
17 NICS(1) estimated at 1A above the center of the ring. For details see:
Schleyer PvR, Maerker C, Dransfeld A, Jiao H, van Eikema Hommes
NJR. Nucleus-independent chemical shifts: A simple and efficient
aromaticity probe. J Am Chem Soc, 1996, 118: 6317–6318
18 a) Yu HZ, Jiang YY, Fu Y, Liu L. Alternative mechanistic explana-
tion for ligand-dependent selectivities in copper-catalyzed N- and
O-arylation reactions. J Am Chem Soc, 2010, 132: 18078–18091; b)
Zhang SL, Liu L, Fu Y, Guo QX. Theoretical study on copper(I)-
catalyzed cross-coupling between aryl halides and amides.
Organometallics, 2007, 26: 4546–4554
19 Casitas A, King AE, Parella T, Costas M, Stahl SS, Ribas X. Direct
observation of CuI/CuIII redox steps relevant to Ullmann-type cou-
pling reactions. Chem Sci, 2010, 1: 326–330
20 Note that O-Int1 is found to be the most stable. See Supporting
Information for details.
21 a) Ribas X, Jackson DA, Donnadieu B, Mahia J, Parella T, Xifra R,
Hedman B, Hodgson KO, Llobet A, Stack TDP. Aryl C–H activation
by CuII to form an organometallic aryl–CuIII species: A novel twist
on copper disproportionation. Angew Chem In Ed, 2002, 41: 2991–
2994; b) King AE, Brunold TC, Stahl SS. Mechanistic study of cop-
per-catalyzed aerobic oxidative coupling of arylboronic esters and
methanol: Insights into an organometallic oxidase reaction. J Am
Chem Soc, 2009, 131: 5044–5045
22 Huffman LM, Casitas A, Font M, Canta M, Costas M, Ribas X, Stahl
SS. Observation and mechanistic study of facile C–O bond formation
between a well-defined aryl-copper(III) complex and oxygen nucle-
ophiles. Chem Eur J, 2011, 17: 10642–10649
23 Sartori G, Maggi R. Use of solid catalysts in Friedel–Crafts acylation
reactions. Chem Rev, 2006, 106: 1077–1104
24 Kuang GC, Guha PM, Brotherton WS, Simmons JT, Stankee LA,
Nguyen BT, Clark RJ, Zhu L. Experimental investigation on the mech-
anism of chelation-assisted, copper(II) acetate-accelerated azide-alkyne
cycloaddition. J Am Chem Soc, 2011, 133: 13984–14001
25 An alternative intermediate (i.e. Cu-nitrene may also be involved as
proposed by Buchwald et al. for the C–N cyclization. However, our
calculation shows that the formation of Cu-nitrene is highly
endergonic (+35.2 kcal/mol) and therefore unfavorable (Figure 8).
26 For details about the disfavored routes, see Supporting Information.
27 Jones GO, Liu P, Houk KN, Buchwald SL. Computational explorations
of mechanisms and ligand-directed selectivities of copper-catalyzed
Ullmann-type reactions. J Am Chem Soc, 2010, 132: 6205–6213
28 In the case of C–O formation reaction with radical scavenger, the
small decrease of yield is probably attributed to radical species or
other alkene-promoted side-reactions. To further clarify the radical
pathways, we calculated the important radical species O-6 and O-8.
The data suggested that the radical pathway might be not plausible.
29 Lanci MP, Remy MS, Kaminsky W, Mayer JM, Sanford MS. Oxida-
tively induced reductive elimination from ((t)Bu2bpy)Pd(Me)2: Pal-
ladium(IV) intermediates in a one-electron oxidation reaction. J Am
Chem Soc, 2009, 131: 15618–15620
5
6
a) Wendlandt AE, Suess AM, Stahl SS. Copper-catalyzed aerobic
oxidative C–H functionalizations: Trends and mechanistic insights.
Angew Chem Int Ed, 2011, 50: 11062–11087; b) Yang L, Lu Z, Stahl
SS. Regioselective copper-catalyzed chlorination and bromination of
arenes with O2 as the oxidant. Chem Commun, 2009, 6460–6462
a) Li Z, Bohle DS, Li CJ. Cu-catalyzed cross-dehydrogenative cou-
pling: A versatile strategy for C-C bond formations via the oxidative
activation of sp(3) C–H bonds. Proc Natl Acad Sci U S A, 2006, 103:
8928–8933; b) Do HQ, Daugulis O. Copper-catalyzed arylation of
heterocycle C–H bonds. J Am Chem Soc, 2007, 129: 12404–12405; c)
Do HQ, Daugulis O. Copper-catalyzed arylation and alkenylation of
polyfluoroarene C–H bonds. J Am Chem Soc, 2008, 130: 1128–1129
Chen X, Hao XS, Goodhue CE, Yu JQ. Cu(II)-catalyzed functionali-
zations of aryl C–H bonds using O2 as an oxidant. J Am Chem Soc,
2006, 128: 6790–6791
Brasche G, Buchwald SL. C–H functionalization/C–N bond for-
mation: Copper-catalyzed synthesis of benzimidazoles from amidines.
Angew Chem Int Ed, 2008, 47: 1932–1934
a) Ueda S, Nagasawa H. Synthesis of 2-arylbenzoxazoles by copper-
catalyzed intramolecular oxidative C–O coupling of benzanilides.
Angew Chem Int Ed, 2008, 47: 6411–6413; b) Ueda S, Nagasawa H.
Copper-catalyzed synthesis of benzoxazoles via a regioselective C–H
functionalization/C–O bond formation under an air atmosphere. J
Org Chem 2009, 74: 4272–4277
7
8
9
10 Ribas X, Calle C, Poater A, Casitas A, Gomez L, Xifra R, Parella T
Benet-Buchholz J, Schweiger A, Mitrikas G, Sola M, Llobet A, Stack
TD. Facile C–H bond cleavage via a proton-coupled electron transfer
involving a C–H...Cu(II) interaction. J Am Chem Soc, 2010, 132:
12299–12306
11 a) Yao B, Wang DX, Huang ZT, Wang MX. Room-temperature aer-
obic formation of a stable aryl–Cu(III) complex and its reactions with
nucleophiles: Highly efficient and diverse arene C–H functionaliza-
tions of azacalix[1]arene[3]pyridine. Chem Commun, 2009, 2899–
2901; b) Wang ZL, Zhan L, Wang MX. Regiospecific functionaliza-
tion of azacalixaromatics through copper-mediated aryl C–H activa-
tion and C–O bond formation. Org Lett, 2011, 13: 6560–6563
12 a) Huffman LM, Casitas A, Font M, Canta M, Costas M, Ribas X,
Stahl SS. Observation and mechanistic study of facile C–O bond
formation between a well-defined aryl-copper(III) complex and oxy-
gen nucleophiles. Chem Eur J, 2011, 17: 10643–10650; b) Huffman
LM, Stahl SS. Carbon-nitrogen bond formation involving well-
defined aryl-copper(III) complexes. J Am Chem Soc, 2008, 130:
9196–9197
13 Garcia-Lopez J, Yanez-Rodriguez V, Roces L, Garcia-Granda S,
Martinez A, Guevara-Garcia A, Castro GR, Jimenez-Villacorta F,
Iglesias MJ, Lopez Ortiz F. Synthesis and characterization of a cou-
pled binuclear Cu(I)/Cu(III) complex. J Am Chem Soc, 2010, 132:
10665–10667
14 a) Chen B, Hou XL, Li YX, Wu YD. Mechanistic understanding of
the unexpected meta selectivity in copper-catalyzed anilide C–H
bond arylation. J Am Chem Soc, 2011, 133: 7668–7671; b) Wang M,
Fan T, Lin Z. DFT Studies on copper-catalyzed arylation of aromatic
C–H Bonds. Organometallics, 2012, 31: 560–569; c) Santoro S, Liao
RZ, Himo F. Theoretical study of mechanism and selectivity of copper-
catalyzed C–H bond amidation of indoles. J Org Chem, 2011, 76:
9246–9252
15 a) Gorelsky SI, Lapointe D, Fagnou K. Analysis of the concerted
metalation–deprotonation mechanism in palladium-catalyzed direct
arylation across a broad range of aromatic substrates. J Am Chem Soc,
2008, 130: 10848–10849; b) García-Cuadrado D, Mendoza P De;
Braga AAC, Maseras F, Echavarren AM. Proton-abstraction mecha-
nism in the palladium-catalyzed intramolecular arylation: Substituent
effects. J Am Chem Soc, 2007, 129: 6880–6886
30 The wavefunction is stable under the perturbations considered.
31 a) Becke AD. Density-functional thermochemistry. III. The role of
exact exchange. J Chem Phys, 1993, 98: 5648–5652; b) Perdew JP.
Density-functional approximation for the correlation energy of the
inhomogeneous electron gas. J Phys Rev B, 1986, 33: 8822–8824
32 Rassolov VA, Pople JA, Ratner MA, Windus TL. 6-31G* basis set
for atoms K through Zn. J Chem Phy, 1998, 109: 1223–1229
33 Pavelka M, Šimánek M, Šponer J, Burda JV. Copper cation interac-
tions with biologically essential types of ligands: A computational
DFT study. J Phys Chem A, 2006, 110: 4795–4809
34 a) Li Z, Fu Y, Zhang SL, Guo QX, Liu L. Heck-type reactions of
imine derivatives: A DFT study. Chem Asian J, 2010, 5: 1475–1486;
b) Shang R, Yang ZW, Wang Y, Zhang SL, Liu L. Palladium-
catalyzed decarboxylative couplings of 2-(2-azaaryl)acetates with
aryl halides and triflates. J Am Chem Soc, 2010, 132: 14391–14393;
c) Zhang SL, Fu Y, Shang R, Guo QX, Liu L. Theoretical analysis of
16 a) Kruszewski J, Krygowski TM. Definition of aromaticity Basing on

632
Tang SY, et al. Sci China Chem May (2013) Vol.56 No.5
factors controlling Pd-catalyzed decarboxylative coupling of carbox-
actions, excited states, and transition elements: two new functionals
and systematic testing of four M06-class functionals and 12 other
functionals. Theor Chem Acc, 2008, 120: 215–241
ylic acids with olefins. J Am Chem Soc, 2010, 132: 638–646; d)
Shang R, Fu Y, Wang Y, Xu Q, Yu HZ, Liu L. Copper-catalyzed
decarboxylative cross-coupling of potassium polyfluorobenzoates
with aryl iodides and bromides. Angew Chem Int Ed, 2009, 48:
9350–9354; e) Li Z, Zhang SL, Fu Y, Guo QX, Liu L. Mechanism of
Ni-catalyzed selective C–O bond activation in cross-coupling of aryl
esters. J Am Chem Soc, 2009, 131:8815–8823
37 Ariafard A, Zarkoob F, Batebi H, Stranger R, Yates BF. DFT Studies
on the carboxylation of the C–H bond of heteroarenes by copper(I)
complexes. Organometallics, 2011, 30: 6218–6224
38 Cramer CJ, Marenich AV, Truhlar DG. Universal solvation model
based on solute electron density and on a continuum model of the
solvent defined by the bulk dielectric constant and atomic surface
tensions. J Phys Chem B, 2009, 113: 6378–6396
35 Gonzalez C, Schlegel HB. An Improved Algorithm for Reaction Path
Following. J Chem Phys, 1989, 90: 2154–2161
36 Truhlar DG, Zhao Y. The M06 suite of density functionals for main
group thermochemistry, thermochemical kinetics, noncovalent inter-
39 Gaussian 09, Revision B.01, Frisch MJ, Trucks GW, Schlegel HB, et
al. Wallingford CT: Gaussian Inc. 2010