Formation of CC-type palladacycles with assistance from an apparently innocent NH(CO) functional group

Article abstract of DOI:10.1039/c2cc31299a

A novel cyclometalation pathway to form CC-type palladacycles is reported. Unlike common donor-assisted cyclometalation, the NH(CO) auxiliary group undergoes a deprotonation step to form a palladalactam intermediate. The coordinating nitrogen atom functions as an intramolecular base promoting selective C-H bond cleavage. Without the NH proton, the ortho-N-phenyl C-H is activated instead.

Full text of DOI:10.1039/c2cc31299a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 5632–5634
www.rsc.org/chemcomm
COMMUNICATION
Formation of CC-type palladacycles with assistance from an apparently
innocent NH(CO) functional groupw
Jhen-Yi Lee, Yao-Huei Huang, Shu-Ya Liu, Shun-Chu Cheng, Yang-Ming Jhou, Jenn-Huei Lii
and Hon Man Lee*
Received 21st February 2012, Accepted 5th April 2012
DOI: 10.1039/c2cc31299a
A novel cyclometalation pathway to form CC-type palladacycles
is reported. Unlike common donor-assisted cyclometalation, the
NH(CO) auxiliary group undergoes a deprotonation step to
form a palladalactam intermediate. The coordinating nitrogen
atom functions as an intramolecular base promoting selective
C–H bond cleavage. Without the NH proton, the ortho-N-phenyl
C–H is activated instead.
Scheme 1 Formation of palladacycle via sp3 C–H activation.
Palladacycles have recently received considerable attention
because they are key intermediates in palladium-catalyzed
ligand-directed C–H functionalization.1–14 The most popular
way to construct palladacycles is cyclometalation involving
donor-assisted C–H bond activation.15,16 Typical donor
groups contain N, O, P, S, Se, or As. Carbon-assisted bond
activation is also known to form CC-type palladacycles, but it is
much rarer. Other synthetic routes, sometimes complementary
to the donor-assisted approach,17 have also been developed.
Understanding the fundamental aspects of cyclometalation
processes can provide useful insights into the mechanistic features
of palladium-catalyzed C–H functionalization reactions.18–20
Herein, we reveal a unique cyclometalation pathway that leads to
the formation of CC-type palladacycles. Unlike common donor-
assisted approaches, the NH(CO) functionality on these ligands
serves as more than just a donor for a palladium atom to provide
chelation-assisted C–H bond activation. An intermediate palla-
dalactam actually involves a coordinating nitrogen atom, which
then functions as an intramolecular base, promoting selective
C–H bond cleavage. This mechanism could provide new path-
ways to create novel stoichiometric and catalytic transformations.
Preparation began by reaction of ligand L1 with PdCl2 in
pyridine, employing K2CO3 as a base. Heating at 80 1C
overnight afforded a five-membered palladacycle (1) in 66%
yield (Scheme 1). Deprotonation of L1 by pyridine led to
double C–H bond activation of the methyl and methylene
protons. Two groups of methylene protons were present in L1,
and the more acidic protons next to the carbonyl group were
selectively activated. Complex 1 was identified by 1H NMR
spectroscopy, exhibiting a characteristic singlet methine signal
at 5.32 ppm. The singlet for the methyl group was replaced by
the diastereotopic signals of the metal-bound methylene group
at 1.97 and 2.44 ppm. The NH proton (9.89 ppm) appears to
be intact after the reaction. The pyridine ligand in 1 can easily
be substituted with PPh3 to afford 2 in 84% yield. In the
13C{1H} NMR spectrum of 2, the methylene carbon cis to the
PPh3 was observed as a doublet at 7.6 ppm, with a 2JPC = 7.4 Hz.
The large coupling constant (2JPC = 97.6 Hz) for the doublet
at 60.2 ppm was consistent with a trans relationship between
the methine carbon and phosphorous atoms. X-ray diffraction
analyses confirmed the formation of 1 and 2 (Fig. 1 and ESIw).
The cis relationship between the methine carbon and chlorine
atoms in 2 allowed the NHÁ Á ÁCl intra-molecular hydrogen
bonds to stabilize the structures.
The NH proton on L1 was expected to be easier to
deprotonate than the methyl and methylene protons. We were,
however, rather surprised not to obtain a palladacycle con-
taining an amidate moiety. To determine if any plausible
palladium amidate complex could be isolated, we carried out the
same reaction at ambient temperature for 5 h; palladalactam (3),
1
an isomer of 1, was afforded in 51% yield. H NMR spectro-
scopy confirmed the formation of 3, showing a methyl signal
at 2.80 ppm and a methine singlet at 4.56 ppm. A downfield
signal attributable to the NH proton was absent; however,
signals from a coordinating pyridine were observed. The
pyridine ligand was substituted with PPh3, to afford 4. The
1H NMR spectrum of 4 also showed the characteristic methine
and methyl signals. An X-ray structure determination revealed
a highly distorted square planar coordination environment
around the palladium atom due to the very small bite of
Department of Chemistry, National Changhua University of Education,
Changhua 50058, Taiwan, R.O.C. E-mail: leehm@cc.ncue.edu.tw;
Fax: +886-4-7232105ext.3523; Tel: +886-4-7211190
w Electronic supplementary information (ESI) available: Details of
experiments, computations, and crystallography. CCDC 850737–850742.
For ESI and crystallographic data in CIF or other electronic format
see DOI: 10.1039/c2cc31299a
c
5632 Chem. Commun., 2012, 48, 5632–5634
This journal is The Royal Society of Chemistry 2012

View Article Online
Although the formation of 1 through intermediate 3
involves deprotonation of the NH proton, in its subsequent
conversion to 1, the NH proton was replaced. This NH proton
could have come from either the solvent or a proton transfer
from the methyl C–H bond. To gain mechanistic insight into
the proton source, we labeled the NH proton in L1 with
deuterium by employing the readily available PhND2 as a
starting material. First, we verified the inclusion of deuterium
into the amide group of L1. The 1H NMR spectrum of
deuterated L1 did not show a downfield signal corresponding
to the NH proton between 9 and 10 ppm. The 2H NMR
spectrum duly shows the deuterium signal at 10.21 ppm, con-
firming the successful formation of the deuterated compound.
Next, deuterated L1 was carried forward to prepare 1. The
synthetic procedure was slightly modified to avoid aqueous
washing, thus eliminating the possibility of deuterium exchange
with water solvent. After workup, the 1H NMR spectrum of the
product clearly showed the downfield NH proton integrating
with one proton at 9.89 ppm. The possibility of deuterium
exchange with pyridine solvent was ruled out by carrying out
the reaction between non-deuterated L1 and PdCl2 in pyridine-d5.
The results indicated that there was no deuterium incorporation
into the amide group. Thus, we confirmed that the amide proton
in 1 comes from the methyl proton exclusively. A similar
deuterium labeling experiment also confirmed that a similar
hydrogen migration occurred in the conversion of 6 to 5.
To further illustrate the importance of the NH proton in
the selective cyclometalation that formed 1, we prepared the
N-methyl ligand precursor L3 and investigated its cyclo-
metalation with PdCl2. Since the NH proton was replaced
by a methyl group, the formation of four-membered pallada-
cycles such as 3 and 6 was prohibited. The reaction thus took a
different course, affording a six-membered palladacycle (8) in
which the ortho-C–H bond of the N-phenyl ring was activated
(Scheme 3). Such C–H bond activation is commonly observed
in palladium-catalyzed C–H functionalization reactions involving
Fig. 1 Molecular structures of 1 (left) and 2 (right) with 50% probability
ellipsoids for non-H atoms.
Fig. 2 Molecular structures of 4 with 50% probability ellipsoids.
Only one of the two independent molecules is shown.
the organic ligand (Fig. 2 and ESIw). The N–Pd–C angle of 67.0(2)1
was comparable to that of a similar palladalactam compound
reported previously.21 To test whether 3 was a kinetic product,
pyridine or DMF solution of 3 was allowed to stir at 80 1C
overnight. Pure 1 was cleanly obtained in almost quantitative yield
after workup. Similarly, 4 could be isomerized to 2 in 87% yield.
Although, in general, sp3 C–H bonds are much less reactive
than sp2 C–H bonds,22 the methyl group in 1 was activated
owing to its direct connection to the positively charged
imidazolium ring.23 To test whether similar reactivity occurs
with an sp2 C–H bond distant from the imidazolium ring, we
synthesized L2 in which a phenyl ring was attached to the C2
position of the imidazolium salt. For ease of synthesis, L2
contained an N-1-naphthylmethyl instead of a 4-fluorobenzyl
group (Scheme 2). The reactivity pattern was indeed similar to
that of L1 and PdCl2, affording a four-membered palladacycle
(6) and a six-membered palladacycle (5) as kinetic and thermo-
dynamic products, respectively. Complex 7, derived from 6 by
substituting a pyridine with a PPh3 ligand, was also obtained.
The structures of 5 and 7 were confirmed by X-ray crystallo-
graphic analysis (Fig. 3 and ESIw).
1
amides as directing groups.3,8,24 The H NMR spectrum of 8
shows the characteristic peaks of the methyl and methine
signals at 3.38 and 5.61 ppm, respectively. Further heating a
THF solution of 8 for 3 h produces dimeric compound 9
(Fig. S1, ESIw). The pyridine ligand in 8 is more labile than
those in 1 and 5. Our results indicate that the amide a-methylene
on L3 was initially deprotonated and coordinated to the palladium
centre. Because no amide proton was available for activation,
chelation-assisted sp2 C–H cleavage was preferred over the sp3
C–H bond, forming 8 as the product. The possibility of
palladalactam was thus blocked and a five-membered pallada-
cycle such as 1 could not form. A preliminary study revealed a
similar reactivity pattern for the reaction between PdCl2 and a
Scheme 2 Formation of palladacycle via sp2 C–H activation.
Fig. 3 Molecular structures of 5 (left) and 7 (right) with 50% probability
ellipsoids for non-H atoms.
Scheme 3 Formation of palladacycle via o-N-phenyl C–H activation.
Chem. Commun., 2012, 48, 5632–5634 5633
c
This journal is The Royal Society of Chemistry 2012

View Article Online
barrier in the gas phase was calculated to be 38.1 kcal molÀ1
When solvent correction (pyridine as solvent) was applied, the
barrier to the transformation was reduced to 35.3 kcal molÀ1
.
.
We have demonstrated a novel cyclometalation process for
the formation of CC-type palladacycles via a unique NH(CO)-
assisted selective C–H bond activation pathway. Our results
contrast literature examples of cyclometalation involving donor-
assisted C–H bond activation. In the present case, the auxiliary
functional group is non-coordinating and appears to be intact in
the final product; however, a detailed investigation revealed a
mechanistic pathway involving the formation of a palladalactam.
After forming this kinetic product, the possibility of ortho-N-
phenyl C–H activation was blocked. Subsequent facile sp3 or sp2
C–H bond activation assisted by the coordinated nitrogen as a
base led to the formation of 1 and 5 while the amide moiety was
regenerated. By blocking this pathway with an N-methyl group,
we prevented access to the five-membered palladacycles. Instead
an ortho-N-phenyl C–H activation pathway was favored,
affording CC-type palladacycle 8. Work to apply this mecha-
nism to new modes of catalysis is in progress.
Chart 1 Five possible Pd isomers from L1. Relative energies
(in kcal molÀ1) for the three most stable isomers are shown (relative
free energies are in parentheses; possible coordination sites of ligand
are marked by asterisks).
derivative of L3 containing an N-methyl group and a phenyl
ring on the C2 position of the imidazolium ring.
L1 offers five coordination sites for metal complexation: the
ortho-N-phenyl carbon, the nitrogen atom, the methylene carbon,
the methyl carbon, and the C4 site of the imidazolium ring (to form
an abnormal NHC complex25). To understand the thermodynamic
stabilities of different isomers of palladium complexes of L1, we
performed density functional theory (DFT) calculations at the
B3LYP/LANL2DZ level of theory. The three most stable struc-
tural isomers are shown in Chart 1. The structural parameters of
S2 were in good agreement with the experimental structural data
for 1. The studies showed that S2 is more stable than S1 by
5.6 kcal molÀ1. Although there was a slight discrepancy between the
calculations and experimental observations in the stereochemistry
of S1 and S3 (chlorine and pyridine swapped coordination sites),
the relative energies of S2 and S1 agreed with the experimental
result that the five-membered palladacycle 1 is thermodynamically
preferred over the four-membered palladacyclic product 3. Also in
accord with the experimental results, the seven-membered
palladacycle (S3) with an amidate and alkyl donor was less
stable, and it was not observed experimentally.
Notes and references
1 M. Catellani, E. Motti and N. Della Ca’, Acc. Chem. Res., 2008, 41, 1512.
2 T. W. Lyons and M. S. Sanford, Chem. Rev., 2010, 110, 1147.
3 X. Chen, K. M. Engle, D.-H. Wang and J.-Q. Yu, Angew. Chem.,
Int. Ed., 2009, 48, 5094.
4 C. Huang, B. Chattopadhyay and V. Gevorgyan, J. Am. Chem.
Soc., 2011, 133, 12406.
5 B. Sezen, R. Franz and D. Sames, J. Am. Chem. Soc., 2002, 124, 13372.
6 W. Oloo, P. Y. Zavalij, J. Zhang, E. Khaskin and A. N.
Vedernikov, J. Am. Chem. Soc., 2010, 132, 14400.
7 T. Nishikata, A. R. Abela, S. Huang and B. H. Lipshutz, J. Am.
Chem. Soc., 2010, 132, 4978.
8 G.-W. Wang, T.-T. Yuan and D.-D. Li, Angew. Chem., Int. Ed.,
2011, 50, 1380.
9 B. Xiao, Y. Fu, J. Xu, T.-J. Gong, J.-J. Dai, J. Yi and L. Liu,
J. Am. Chem. Soc., 2009, 132, 468.
10 R. Samanta and A. P. Antonchick, Angew. Chem., Int. Ed., 2011,
50, 5217.
11 C. Wang, S. Rakshit and F. Glorius, J. Am. Chem. Soc., 2010,
132, 14006.
12 T. Kesharwani, A. K. Verma, D. Emrich, J. A. Ward and
R. C. Larock, Org. Lett., 2009, 11, 2591.
13 Y. Feng, Y. Wang, B. Landgraf, S. Liu and G. Chen, Org. Lett.,
2010, 12, 3414.
14 B. Qian, S. Guo, J. Shao, Q. Zhu, L. Yang, C. Xia and H. Huang,
J. Am. Chem. Soc., 2010, 132, 3650.
15 M. Albrecht, Chem. Rev., 2010, 110, 576.
16 J. Dupont, C. S. Consorti and J. Spencer, Chem. Rev., 2005, 105, 2527.
17 H. P. Dijkstra, M. Albrecht and G. van Koten, Chem. Commun.,
2002, 126.
18 D. L. Davies, S. M. A. Donald and S. A. Macgregor, J. Am. Chem.
Soc., 2005, 127, 13754.
19 D. Garcıa-Cuadrado, A. A. C. Braga, F. Maseras and A. M.
´
Echavarren, J. Am. Chem. Soc., 2006, 128, 1066.
20 A. J. Mota, A. Dedieu, C. Bour and J. Suffert, J. Am. Chem. Soc.,
2005, 127, 7171.
In the transformation of 3 to 1, a concerted, direct hydrogen
migration mechanism was postulated. The possibility of C–H
oxidative addition through an octahedral palladium(IV) species is
unlikely owing to the electron deficiency of 3. The preliminary
computational results for this transformation are shown in Fig. 4,
where TS1 has an imaginary frequency of À110 cmÀ1. The transi-
tion state exhibited a markedly negligible methyl C–H bond agostic
interaction with a PdÁ Á ÁC distance of 3.345 A (the sum of the
van der Waals radii of Pd and C equals 3.33 A)26 and a long
PdÁ Á ÁH contact of 2.997 A. On the other hand, the methyl C–H
bond was in close contact with the nitrogen atom at a very short
distance of 2.169 A (the sum of the van der Waals radii of N and H
equals 2.75 A).26 Agostic interactions have been commonly invol-
ved in many transition metal-mediated C–H bond activations.18,27
However, since the methyl C–H bond was acidic enough owing
to its attachment to the imidazolium ring,23 no significant pre-
activation via agostic interactions was required. The energy
21 W. Henderson, A. G. Oliver, C. E. F. Rickard and L. J. Baker,
Inorg. Chim. Acta, 1999, 292, 260.
22 M. Broring and C. Kleeberg, Chem. Commun., 2008, 2777.
23 G. Song, X. Li, Z. Song, J. Zhao and H. Zhang, Chem.–Eur. J.,
2009, 15, 5535.
24 D. Shabashov, J. s. R. M. Maldonado and O. Daugulis, J. Org.
Chem., 2008, 73, 7818.
25 S. Grundemann, A. Kovacevic, M. Albrecht, J. W. Faller and
¨
R. H. Crabtree, J. Am. Chem. Soc., 2002, 124, 10473.
26 A. Bondi, J. Phys. Chem., 1964, 68, 441.
27 M. Lafrance, S. I. Gorelsky and K. Fagnou, J. Am. Chem. Soc.,
2007, 129, 14570.
Fig. 4 TS1 for the conversion of 3 to 1. Selected H atoms have been
removed for clarity. Distances are in A.
c
5634 Chem. Commun., 2012, 48, 5632–5634
This journal is The Royal Society of Chemistry 2012