Synthesis of benzocyclobutenes by palladium-catalyzed C-H activation of methyl groups: Method and mechanistic study

Article abstract of DOI:10.1021/ja805598s

An efficient catalytic system has been developed for the synthesis of benzocyclobutenes by C-H activation of methyl groups. The optimal conditions employed a combination of Pd(OAc)₂ and P^tBu₃ as catalyst, K₂CO₃ as the base, and DMF as solvent. A variety of substituted BCB were obtained under these conditions with yields in the 44-92% range, including molecules that are hardly accessible by other methods. The reaction was found limited to substrates bearing a quaternary benzylic carbon, but benzocyclobutenes bearing a tertiary benzylic carbon could be obtained indirectly from diesters by decarboxylation. Reaction substrates bearing a small substituent para to bromine gave an unexpected regioisomer that likely arose from a 1,4-palladium migration process. The formation of this "abnormal" regioisomer could be suppressed by introducing a larger subsituent para to bromine. DFT(B3PW91) calculations on the reaction of 2-bromo-tert-butylbenzene with Pd(P^tBu₃) with different bases (acetate, bicarbonate, carbonate) showed the critical influence of the coordination mode of the base to induce both an easy C-H activation and to allow for a pathway for 1,4-palladium migration. Carbonate is shown to be more efficient than the two other bases because it can abstract the proton easily and at the same time maintain κ¹-coordination without extensive electronic reorganization.

Full text of DOI:10.1021/ja805598s

Published on Web 10/18/2008
Synthesis of Benzocyclobutenes by Palladium-Catalyzed C-H
Activation of Methyl Groups: Method and Mechanistic Study
Manon Chaumontet,† Riccardo Piccardi,‡ Nicolas Audic,† Julien Hitce,†
Jean-Louis Peglion,§ Eric Clot,*,| and Olivier Baudoin*,‡
Institut Charles Gerhardt, UMR5253, CNRS-UM2-UM1-ENSCM, case courrier 1501, Place
Euge`ne Bataillon, 34000 Montpellier, France, and UniVersite´ Lyon 1, CNRS UMR5246, Institut
de Chimie et Biochimie Mole´culaires et Supramole´culaires, 43 BouleVard du 11 NoVembre
1918, 69622 Villeurbanne, France
Received July 18, 2008; E-mail: clot@univ-montp2.fr; olivier.baudoin@univ-lyon1.fr
Abstract: An efficient catalytic system has been developed for the synthesis of benzocyclobutenes by
C-H activation of methyl groups. The optimal conditions employed a combination of Pd(OAc)2 and PtBu3
as catalyst, K2CO3 as the base, and DMF as solvent. A variety of substituted BCB were obtained under
these conditions with yields in the 44-92% range, including molecules that are hardly accessible by other
methods. The reaction was found limited to substrates bearing a quaternary benzylic carbon, but
benzocyclobutenes bearing a tertiary benzylic carbon could be obtained indirectly from diesters by
decarboxylation. Reaction substrates bearing a small substituent para to bromine gave an unexpected
regioisomer that likely arose from a 1,4-palladium migration process. The formation of this “abnormal”
regioisomer could be suppressed by introducing a larger subsituent para to bromine. DFT(B3PW91)
calculations on the reaction of 2-bromo-tert-butylbenzene with Pd(PtBu3) with different bases (acetate,
bicarbonate, carbonate) showed the critical influence of the coordination mode of the base to induce both
an easy C-H activation and to allow for a pathway for 1,4-palladium migration. Carbonate is shown to be
more efficient than the two other bases because it can abstract the proton easily and at the same time
maintain κ1-coordination without extensive electronic reorganization.
triggered by coordination of the metal to a directing group4,5
Introduction
or by oxidative addition,6,7 or intermolecular, nondirected C-H
In recent years, transition-metal-catalyzed C-H activation
has emerged as a powerful tool to transform otherwise unre-
active C-H bonds into carbon-carbon or carbon-heteroatom
bonds.1,2 This new area of organic chemistry is gradually
changing the way chemists functionalize molecules, providing
atom- and step-economical alternatives to more traditional
methods and facilitating the access to valuable and original
compounds. In regard to the wealth of methods recently
developed for the functionalization of arene and heteroarene
C(sp2)-H bonds,3 relatively little work has focused on the
functionalization of unreactive alkyl C(sp3)-H bonds by catalytic
C-H activation. Solutions to this rather difficult case have been
proposed using either intramolecular activation, which can be
activation.8 We recently reported on the palladium-catalyzed
C-H activation of benzylic alkyl groups borne by bromoben-
zenes 1 (eq 1).7 In this case, the oxidative addition of the C-Br
bond to palladium(0) triggers an intramolecular C-H activation
to give a five or six-membered palladacyclic intermediate,
(3) (a) Hassan, J.; Se´vignon, M.; Gozzi, C.; Schulz, E.; Lemaire, M. Chem.
ReV. 2002, 102, 1359. (b) Goj, L. A.; Gunnoe, T. B. Curr. Org. Chem.
2005, 9, 671. (c) Campeau, L.-C.; Fagnou, K. Chem. Commun. 2006,
1253. (d) Alberico, D.; Scott, M. E.; Lautens, M. Chem. ReV. 2007,
107, 174. (e) Seregin, I. V.; Gevorgyan, V. Chem. Soc. ReV. 2007,
36, 1173.
(4) (a) Chatani, N.; Asaumi, T.; Ikeda, T.; Yorimitsu, S.; Ishii, Y.;
Kakiuchi, F.; Murai, S. J. Am. Chem. Soc. 2000, 122, 12882. (b) Zhang,
X.; Fried, A.; Knapp, S.; Goldman, A. S. Chem. Commun. 2003, 2060.
(c) Murahashi, S.-I.; Komiya, N.; Terai, H.; Nakae, T. J. Am. Chem.
Soc. 2003, 125, 15312. (d) DeBoef, B.; Pastine, S. J.; Sames, D. J. Am.
Chem. Soc. 2004, 126, 6556. (e) Desai, L. V.; Hull, K. L.; Sanford,
M. S. J. Am. Chem. Soc. 2004, 126, 9542. (f) Kakiuchi, F.; Tsuchiya,
K.; Matsumoto, M.; Mizushima, E.; Chatani, N. J. Am. Chem. Soc.
2004, 126, 12792. (g) Giri, R.; Chen, X.; Yu, J.-Q. Angew. Chem.,
Int. Ed. 2005, 44, 2112. (h) Giri, R.; Liang, J.; Lei, J.-G.; Li, J.-J.;
Wang, D.-H.; Chen, X.; Naggar, I. C.; Guo, C.; Foxman, B. M.; Yu,
J.-Q. Angew. Chem., Int. Ed. 2005, 44, 7420. (i) Zaitsev, V. G.;
Shabashov, D.; Daugulis, O. J. Am. Chem. Soc. 2005, 127, 13154. (j)
Shabashov, D.; Daugulis, O. Org. Lett. 2005, 7, 3657. (k) Chen, X.;
Goodhue, C. E.; Yu, J.-Q. J. Am. Chem. Soc. 2006, 128, 12634. (l)
Giri, R.; Maugel, N.; Li, J.-J.; Wang, D.-H.; Breazzano, S. P.; Saunders,
L. B.; Yu, J.-Q. J. Am. Chem. Soc. 2007, 129, 3510. (m) Campeau,
L.-C.; Schipper, D. J.; Fagnou, K. J. Am. Chem. Soc. 2008, 130, 3266.
(n) Mousseau, J. J.; Larive´e, A.; Charette, A. B. Org. Lett. 2008, 10,
1641.
† Institut de Chimie des Substances Naturelles, CNRS UPR2301, avenue
de la Terrasse, 91198 Gif-sur-Yvette, France.
‡ Institut de Chimie et Biochimie Mole´culaires et Supramole´culaires.
§ Institut de Recherche Servier, 11 rue des Moulineaux, 92150 Suresnes,
France.
| Institut Charles Gerhardt.
(1) Selected reviews: (a) Dyker, G. Angew. Chem., Int. Ed. 1999, 38, 1698.
(b) Kakiuchi, F.; Chatani, N. AdV. Synth. Catal. 2003, 345, 1077. (c)
Dick, A. R.; Sanford, M. S. Tetrahedron 2006, 62, 2439. (d) Godula,
K.; Sames, D. Science 2006, 312, 67.
(2) In this article, C-H activation refers to transition-metal-catalyzed
reactions that involve the cleavage of a C-H bond by an inner-sphere
mechanism (see ref 1c). This definition excludes reactions that involve
an outer-sphere mechanism, such as C-H insertions: (a) Modern
Rhodium-Catalyzed Organic Reactions; Evans, P. A., Ed.; Wiley-VCH:
Weinheim, 2005. (b) Davies, H. M. L.; Manning, J. R. Nature 2008,
451, 417. (c) Christmann, M. Angew. Chem., Int. Ed. 2008, 47, 2740.
9
10.1021/ja805598s CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 15157–15166 15157

A R T I C L E S
Chaumontet et al.
Table 1. Optimization of C-H Activation Reaction Parameters
depending on the nature of the activated alkyl. This palladacycle
subsequently evolves by intramolecular C-C coupling to give
carbocycles 2-3 or ꢀ-H elimination to form olefins 4. In
particular, we were able to observe the direct formation of
benzocyclobutenes 2 by C(sp3)-H activation of benzylic methyl
groups.7a,6c,9
conversion GC yield
(%)b
entry solvent
base
Pd source
liganda
(%)b
1
2
3
4
5
6
7
8
9
DMF
DMA
NMP
K2CO3 Pd(OAc)2
K2CO3 Pd(OAc)2
K2CO3 Pd(OAc)2
P(o-tol)3
P(o-tol)3
P(o-tol)3
P(o-tol)3
P(o-tol)3
P(o-tol)3
P(o-tol)3
P(o-tol)3
P(o-tol)3
P(o-tol)3
P(o-tol)3
P(o-tol)3
100
72
100
35
80
40
30
8
54
7
67
38
47
22
51
9
10
1
14
1
xylenes K2CO3 Pd(OAc)2
DMF
DMF
DMF
DMF
DMF
KHCO3 Pd(OAc)2
Na2CO3 Pd(OAc)2
Cs2CO3 Pd(OAc)2
KF
Pd(OAc)2
K3PO4 Pd(OAc)2
KOAc Pd(OAc)2
iPrNEt2 Pd(OAc)2
K2CO3 Pd2dba3
K2CO3 PdCl2(MeCN)2 P(o-tol)3
K2CO3 PdCl2(MeCN)2 F-TOTP
K2CO3 PdCl2(MeCN)2 MeO-TOTP
10 DMF
11 DMF
12 DMF
13 DMF
14 DMF
15 DMF
16 DMF
17 DMF
18 DMF
19 DMF
20 DMF
21 DMF
22 DMF
23 DMF
60
1
100
100
88
100
23
19
5
70
73
52
71
2
1
2
13
89
93 (73)
90
8
Benzocyclobutenes (BCB) are both important structural
elements in active pharmaceutical ingredients and valuable
intermediates for organic synthesis.10 In particular, they undergo
thermal electrocyclic ring-opening to give ortho-xylylenes that
can subsequently react in pericyclic reactions such as Diels-Alder
cycloadditions.11 This property was largely exploited in the past
decades in the total synthesis of polycyclic natural products.10
Despite their high synthetic value, very few general and
chemoselective methods exist for the preparation of function-
alized BCB, which seriously limits their availability and their
use as synthetic intermediates. In this article, we describe a
general method for the preparation of substituted BCB by
K2CO3 PdCl2(MeCN)2 PPh3
K2CO3 PdCl2(MeCN)2
d
dppp
K2CO3 PdCl2(MeCN)2 (Cy3PH)BF4
K2CO3 PdCl2(MeCN)2 DavePhos
K2CO3 PdCl2(MeCN)2 JohnPhos
K2CO3 PdCl2(MeCN)2 (tBu3PH)BF4
K2CO3 PdCl2(MeCN)2 Q-Phos
K2CO3 PdCl2(MeCN)2 IPr
43
100
100
100
26
c
a 10 or
5
mol% ligand for PdII or Pd0 source, respectively.
b Calculated using tetradecane as internal standard. Other observed
products included the starting bromobenzene 1a and the corresponding
proto-debrominated product. c Yield of the isolated product. d 5 mol%.
(5) Recent work in allylic C-H activation: (a) Chen, M. S.; Prabagaran,
N.; Labenz, N. A.; White, M. C. J. Am. Chem. Soc. 2005, 127, 6970.
(b) Delcamp, J. H.; White, M. C. J. Am. Chem. Soc. 2006, 128, 15076.
(c) Olsson, V. J.; Szabo´, K. J. Angew. Chem., Int. Ed. 2007, 46, 6891.
(d) Fraunhoffer, K. J.; White, M. C. J. Am. Chem. Soc. 2007, 129,
7274. (e) Reed, S. A.; White, M. C. J. Am. Chem. Soc. 2008, 130,
3316. (f) Covell, D. J.; White, M. C. Angew. Chem., Int. Ed. 2008,
47, 6448.
palladium-catalyzed C-H activation. We show that this reaction
is compatible with a number of functional groups, which should
stimulate new developments in the chemistry of BCB. In
addition, we report experimental observations and DFT calcula-
tions on the key C-H activation reaction that show for the first
time the dual role of the carbonate base in the C-H bond
cleavage as well as in palladium 1,4-migration.
(6) (a) Dyker, G. Angew. Chem., Int. Ed. Engl. 1992, 31, 1023. (b) Dyker,
G. J. Org. Chem. 1993, 58, 6426. (c) Dyker, G. Angew. Chem., Int.
Ed. Engl. 1994, 33, 103. (d) Toivola, R. J.; Savilampi, S. K.; Koskinen,
A. M. P. Tetrahedron Lett. 2000, 41, 6207. (e) Catellani, M.; Motti,
E.; Ghelli, S. Chem. Commun. 2000, 2003. (f) Sole´, D.; Vallverdu´,
L.; Solans, X.; Font-Bardia, M. Chem. Commun. 2005, 2738. (g)
Barder, T. E.; Walker, S. D.; Martinelli, J. R.; Buchwald, S. L. J. Am.
Chem. Soc. 2005, 127, 4685. (h) Dong, C.-G.; Hu, Q.-S. Angew.
Chem., Int. Ed. 2006, 45, 2289. (i) Ren, H.; Knochel, P. Angew. Chem.,
Int. Ed. 2006, 45, 3462. (j) Ren, H.; Zi, L.; Knochel, P. Chem. Asian
J. 2007, 2, 416. (k) Liron, F.; Knochel, P. Tetrahedron Lett. 2007,
48, 4943. (l) Lafrance, M.; Gorelsky, S. I.; Fagnou, K. J. Am. Chem.
Soc. 2007, 129, 14570. (m) Dong, C.-G.; Hu, Q.-S. Tetrahedron 2008,
64, 2537. (n) Motti, E.; Catellani, M. AdV. Synth. Catal. 2008, 350,
565. (o) Watanabe, T.; Oishi, S.; Fujii, N.; Ohno, H. Org. Lett. 2008,
10, 1759. (p) Salcedo, A.; Neuville, L.; Zhu, J. J. Org. Chem. 2008,
73, 3600.
Results and Discussion
1. Conditions Optimization. We quickly discovered that the
initial set of conditions that were used in our first report was
not optimal for BCB synthesis (Table 1, entry 1).7a Using
bromobenzene 1a as model substrate for the C-H activation
giving BCB 2a, all reaction parameters were reinvestigated.
Representative examples of modifications of the solvent, base,
palladium source and ligand are listed in Table 1 with fixed
temperature (150 °C) and reaction time (1.5 h). First, it appears
that the nature of both the solvent (entries 1-4) and the base
(entries 1, 5-11) is crucial to achieve high yields, DMF and
potassium carbonate providing the optimal combination. It is
noteworthy that the carbonate countercation had also a major
impact on the reaction efficiency (entries 1, 6-7). In contrast,
the nature of the palladium source was relatively indifferent
(entries 1, 12-13). Next, the palladium ligand was varied using
dichlorobis(acetonitrile)palladium(II) as precatalyst (entries
13-23 and Figure 1). The nature of the ligand had again a
profound influence on the reaction yield, with bulky, electron-
rich monophosphines of the type RPtBu2 giving the highest yield
in 2a (entries 20-22). Among these,12 tri-tert-butylphosphine,
(7) (a) Baudoin, O.; Herrbach, A.; Gue´ritte, F. Angew. Chem., Int. Ed.
2003, 42, 5736. (b) Hitce, J.; Retailleau, P.; Baudoin, O. Chem.-
Eur. J. 2007, 13, 792. (c) Hitce, J.; Baudoin, O. AdV. Synth. Catal.
2007, 349, 2054.
(8) (a) Baudry, D.; Ephritikhine, M.; Felkin, H.; Holmes-Smith, R.
J. Chem. Soc., Chem. Commun. 1983, 788. (b) Burk, M. J.; Crabtree,
R. H. J. Am. Chem. Soc. 1987, 109, 8025. (c) Chen, H.; Schlecht, S.;
Semple, T. C.; Hartwig, J. F. Science 2000, 287, 1995. (d) Lawrence,
J. D.; Takahashi, M.; Bae, C.; Hartwig, J. F. J. Am. Chem. Soc. 2004,
126, 15334. (e) Murphy, J. M.; Lawrence, J. D.; Kawamura, K.;
Incarvito, C.; Hartwig, J. F. J. Am. Chem. Soc. 2006, 128, 13684.
(9) Catellani, M.; Ferioli, L. Synthesis 1996, 769.
(10) (a) Mehta, G.; Kotha, S. Tetrahedron 2001, 57, 625. (b) Sadana, A. K.;
Saini, R. K.; Billups, W. E. Chem. ReV. 2003, 103, 1539.
(11) (a) Oppolzer, W. A. Synthesis 1978, 793. (b) Charlton, J. L.; Alauddin,
M. M. Tetrahedron 1987, 43, 2873. (c) Nemoto, H.; Fukumoto, K.
Tetrahedron 1998, 54, 5425.
9
15158 J. AM. CHEM. SOC. VOL. 130, NO. 45, 2008

Benzocyclobutenes by C-H Activation
A R T I C L E S
observed. This limitation was overcome in an indirect manner
using diester-type substrates and decarboxylation, as will be
shown later.
We next examined the reaction of bromobenzenes bearing
different substituents on the benzylic carbon (entries 3-13).
First, the presence of a functional group was not absolutely
required, since the reaction worked on 2-bromo-tert-butylben-
zene 1c (entry 3). The isolated yield was only 46% but this
was due to the volatility of BCB 2c that was the only observed
product. This result stands in sharp contrast to the production
of a dimeric BCB in the seminal work of Dyker starting from
2-iodo-tert-butylbenzene.6c This dimer was never observed
under our conditions. This probably reflects the influence of
the phosphine ligand on the outcome of the reaction (see
mechanistic considerations). The avoidance of the production
of oligomers renders this process particularly predictable and
synthetically useful for the synthesis of BCB. A tert-butyl ester
(entry 4) and other heteroatom-containing functional groups
(entries 5-8) were tolerated on the benzylic position. However
alcohols (entries 5-6) and amines (entry 7) had to be protected
to avoid side reactions such as intramolecular C-N or C-O
coupling. Importantly, the reaction was not restricted to gem-
dimethyl substrates (entries 6, 9-12). In particular, the C-H
activation proved remarkably regioselective with substrates
bearing one methyl group and another alkyl group (entries
9-11). No sign of C-H activation on the other alkyl group,
which would have given rise to the corresponding olefin by ꢀ-H
elimination,7b was observed. This shows the strong steric
preference for C-H activation at benzylic methyl groups.6l In
addition, no intramolecular Mizoroki-Heck coupling was ob-
served when a terminal olefin was present on the alkyl chain
(entry 10), provided there was a sufficiently long spacer (four
methylenes) between the olefin and the benzylic carbon. With
shorter spacers (three methylenes or less), intramolecular
carbopalladation became predominant. Starting from diesters
1l-m (entries 12-13), the corresponding BCB 2l-m were
again produced in good yield. This provides an indirect entry
into BCB containing a tertiary benzylic carbon, that could not
be obtained directly by C-H activation (entry 2). Indeed BCB
2l-m could be both decarboxylated (Scheme 2), either from
2l using a Krapcho-type procedure to yield carboxylic acid 4,15
or from 2m by TFA-mediated cleavage of the t-butyl ester and
thermal decarboxylation of the corresponding carboxylic acid
to yield methyl ester 5.16
Figure 1. Ligands screened for C-H activation.
Scheme 1. Optimized Conditions for the Synthesis of BCB 2a
which was employed for convenience as its phosphonium salt,13
proved the optimal ligand (entry 21). The higher activity of this
ligand compared to the initial P(o-tol)3 allowed us to decrease
the reaction temperature down to 120 °C for the synthesis of
2a (Scheme 1). The reaction, when performed on a 10 g scale
(39 mmol of 1a) using Pd2dba3 (5 mol%)/(tBu3PH)BF4 (10
mol%) as catalyst precursors and as low as 1.3 equiv potassium
carbonate, gave BCB 2a in a reproducible 81% yield.14 This is
the minimal temperature required for an efficient conversion
of the substrate to the desired BCB product with this catalyst.
2. Reaction Scope. The formation of other substituted BCB
by C(sp3)-H activation was next examined using the Pd/PtBu3
catalytic system (Table 2). All starting bromoarenes were
obtained in 2-5 steps in good overall yields from commercially
available 2-substituted bromoarenes using standard chemistry
(see Supporting Information). The most generally applicable
conditions were found with Pd(OAc)2 (10 mol%) as palladium
source in combination to (tBu3PH)BF4 (20 mol%) in DMF at
140 °C. On a multigram scale, the amounts of palladium and
phosphonium were routinely decreased to 6 mol% and 12 mol%,
respectively, without significant impact on the yield. As shown
in Table 2, the C-H activation reaction showed a good tolerance
to a variety of substituents on the aromatic or 4-membered ring.
As only major limitation, we found it was restricted to substrates
bearing a quaternary benzylic carbon, as the reaction of
bromobenzene 1b bearing a tertiary benzylic carbon (entry 2)
failed to produce the corresponding BCB. In this case a mixture
of products including the proto-debrominated product (major)
and the styrene arising from ꢀ-H elimination (minor)6n,7 were
The C-H activation of bromoarenes containing various aryl
substituents was next examined (Table 2, entries 14-28). First,
substrates bearing a substituent para to the quaternary benzylic
carbon were reacted under the usual conditions (entries 14-20),
to give the corresponding BCB in moderate to good yields
regardless of their electron-withdrawing or donating nature.
Remarkably, pyridine 1t also underwent C-H activation to give
pyridocyclobutene 2t in 62% yield (entry 20). This constitutes
a new entry to this synthetically useful motif that is otherwise
obtained by [2 + 2] cycloaddition of a pyridyne and a ketene
acetal.17 Next, the C-H activation of bromobenzenes bearing
a substituent para to the bromine atom was studied (entries
21-27). Surprisingly, the reaction of p-fluorobromobenzene 1u
gave an inseparable mixture of two BCB regioisomers 2u and
(12) (a) Nishiyama, M.; Yamamoto, T.; Koie, Y. Tetrahedron Lett. 1998,
39, 617. (b) Littke, A. F.; Fu, G. C. Angew. Chem., Int. Ed. 1998, 37,
3387. (c) Aranyos, A.; Old, D. W.; Kiyomori, A.; Wolfe, J. P.; Sadighi,
J. P.; Buchwald, S. L. J. Am. Chem. Soc. 1999, 121, 4369. (d) Kataoka,
N.; Shelby, Q.; Stambuli, J. P.; Hartwig, J. F. J. Org. Chem. 2002,
67, 5553. (e) Christmann, U.; Vilar, R. Angew. Chem., Int. Ed. 2005,
44, 366.
(15) Evans, P. A.; Kennedy, L. J. J. Am. Chem. Soc. 2001, 123, 1234.
(13) Netherton, M. R.; Fu, G. C. Org. Lett. 2001, 3, 4295.
(14) On this scale the Pd2dba3/(tBu3PH)BF4 combination is economically
advantageous and at this temperature using 10 mol% palladium(0)
was necessary to ensure a reproducible yield.
(16) Oppolzer, W. A.; Ba¨ttig, K.; Petrzilka, M. HelV. Chim. Acta 1978,
61, 1945.
(17) Mariet, N.; Ibrahim-Ouali, M.; Santelli, M. Tetrahedron Lett. 2002,
43, 5789.
9
J. AM. CHEM. SOC. VOL. 130, NO. 45, 2008 15159

A R T I C L E S
Chaumontet et al.
Table 2. Scope of the synthesis of BCB by C(sp3)-Ha
a Reaction conditions: Pd(OAc)2 (10 mol%), (tBu3PH)BF4 (20 mol%), K2CO3 (1.3 equiv), DMF, 140 °C. b The ratio of regioisomers (when
applicable) was measured by 1H NMR and GC. c Yields of isolated products. d Ts ) tosyl. e Tr ) triphenylmethyl. f Yield after recrystallization. g Ns )
4-nitrophenylsulfonyl.
3u (entry 21), that were clearly identified by NMR analysis.
This phenomenon, that gives invaluable information on the
reaction mechanism (vide infra), was never described before to
our knowledge.6 Increasing the steric bulk with the replacement
9
15160 J. AM. CHEM. SOC. VOL. 130, NO. 45, 2008

Benzocyclobutenes by C-H Activation
A R T I C L E S
Scheme 2. Reactions of BCB Obtained by C-H Activation
Figure 2. X-ray Crystal Structure of BCB 2z (30% probability ellipsoids
plot).
(Table 1). Moreover, reactions on substituted aromatic rings
evidenced the occurrence of an isomerization process during
the catalytic cycle (Table 2, compounds 3). To shed more light
on these observations, a computational study of the catalytic
cycle was carried out at the DFT level with the hybrid functional
B3PW91 (see Supporting Information for computational details).
There are many parameters to take into account but the focus
of the theoretical study is (i) to delineate the basic features of
the mechanism, (ii) to study the influence of the nature of the
base, and (iii) to explain the origin of the isomerization process
observed with the substituted aromatic rings.
The basic features of the mechanism were studied on the
model system 2-bromo-tert-butylbenzene 1c and Pd(PtBu3).
Bromoarene 1c was selected as it was shown to react experi-
mentally and lacks any influence of a functional group close to
the activated methyl. The phosphine considered was PtBu3 as
it was shown to be the most efficient and it allows to exclude
a mechanism involving two phosphines coordinated at Pd.6g,12c,19
The mechanism of base-assisted cyclometalation has been
studied computationally by several groups showing the critical
influence of coordination of the base to the metal to achieve
the C-H activation.6l,20 Experimentally, it was shown that
carbonate is the most efficient base, bicarbonate was shown to
be slightly less active and the activity with acetate was
significantly lower (Table 1, entries 1, 5, 10). Generally, even
though the experimental results are often better with carbonate,
the computational studies already published considered either
bicarbonate,20c,e-g or acetate6l,20h as a model for the base. In
the present work, the three different bases have been considered
to study their influence on the C-H activation. To take into
account the charged species involved in the mechanism, the
geometries were determined in the gas phase at the B3PW91
level but the energies were evaluated including a continuum to
model the solvent according to the PCM procedure. The solvent
of the fluorine atom by a trifluoromethyl group (entry 22)
suppressed the production of the “abnormal” regioisomer, and
BCB 2v was formed as sole product. The same trend was
observed with an oxygenated substituent: the reaction of
p-methoxybromobenzene 1w gave a ca. 10:1 mixture of
inseparable BCB regioisomers 2w and 3w (entry 23). Dimethoxy-
bromobenzene 1x (entry 24) gave rise to BCB regioisomers 2x
and 3x in a similar ratio. To suppress the formation of the
“abnormal” regioisomer in this series of oxygenated molecules,
substrate 1y containing a bulkier trityl ether was prepared (entry
25). As expected, the C-H activation reaction of 1y furnished
BCB 2y as the sole observable regioisomer, in 70% isolated
yield. The trityl group of 2y could be removed by treatment
with HCl in diethyl ether to give phenol 6 in 80% yield (Scheme
2). The reaction of diester 1z (entry 26) gave rise to the
corresponding BCB 2z in 44% yield after recrystallization to
remove undesired byproduct (including the abnormal regioiso-
mer and decarboxylation products). The three-dimensional
structure of 2z was deduced from X-ray diffraction analysis
(Figure 2), which shows the characteristic distortion of the
quaternary carbon-containing cyclobutene ring. The reaction of
nosyl-protected p-aminobromobenzene 1aa furnished BCB 2aa
in 72% yield, with no observable trace of regioisomer (entry
27). Again this result is noteworthy since it is notorious that
amino substituents are not compatible with most methods used
to form the 4-membered ring of BCB. The nosyl group of 2aa
was cleaved easily as described by Fukuyama et al.18 to give
the corresponding aniline 7 quantitatively (Scheme 2). Finally,
the reaction of chlorobromobenzene 1ab furnished BCB 2ab
in satisfying yield (entry 28), which shows the lack of reactivity
of the C-Cl bond at this position of the aromatic ring under
these reaction conditions. All synthesized BCB showed good
stability under the reaction conditions and upon storage. This
property can be imputed to the presence of the quaternary
benzylic carbon and the absence of a strong electron-donating
substituent directly on the cyclobutene ring, that both disfavor
electrocyclic ring-opening and thus the production of the reactive
ortho-xylylene form.10,11
(19) (a) Stambuli, J. P.; Bu¨hl, M.; Hartwig, J. F. J. Am. Chem. Soc. 2002,
124, 9346. (b) Stambuli, J. P.; Incarvito, C. D.; Bu¨hl, M.; Hartwig,
J. F. J. Am. Chem. Soc. 2004, 126, 1184. (c) Ikawa, T.; Barder, T. E.;
Biscoe, M. R.; Buchwald, S. L. J. Am. Chem. Soc. 2007, 129, 13001.
(20) (a) Biswas, B.; Sugimoto, M.; Sakaki, S. Organometallics 2000, 19,
3895. (b) Davies, D. L.; Donald, S. M. A.; Macgregor, S. A. J. Am.
Chem. Soc. 2005, 127, 13754. (c) Garc´ıa-Cuadrado, D.; Braga,
A. A. C.; Maseras, F.; Echavarren, A. M. J. Am. Chem. Soc. 2006,
128, 1066. (d) Alonso, I.; Alcam´ı, M.; Mauleo´n, P.; Carretero, J. C.
Chem.-Eur. J. 2006, 12, 4576. (e) Garc´ıa-Cuadrado, D.; de Mendoza,
P.; Braga, A. A. C.; Maseras, F.; Echavarren, A. M. J. Am. Chem.
3. Computational Studies. The synthesis of Pd-catalyzed
benzocyclobutene described in this work is highly dependent
on the nature of the base, the phosphine ligand and the solvent
¨
Soc. 2007, 129, 6880. (f) Ozdemir, I.; Demir, S.; C¸ etinkaya, B.;
Gourlaouen, C.; Maseras, F.; Bruneau, C.; Dixneuf, P. H. J. Am. Chem.
Soc. 2008, 130, 1156. (g) Lafrance, M.; Rowley, C. N.; Woo, T. K.;
Fagnou, K. J. Am. Chem. Soc. 2006, 128, 8754. (h) Gorelsky, S. I.;
Lapointe, D.; Fagnou, K. J. Am. Chem. Soc. 2008, 130, 10848.
(18) Fukuyama, T.; Jow, C.-K.; Cheung, M. Tetrahedron Lett. 1995, 36,
6373.
9
J. AM. CHEM. SOC. VOL. 130, NO. 45, 2008 15161

A R T I C L E S
Chaumontet et al.
Figure 3. Optimized geometries for the extrema located along the pathway
for C-Br oxidative addition of 1c to Pd(PtBu3). H atoms have been omitted
for clarity.
Figure 4. Optimized geometry for the extrema located along the pathway
for C-H activation by the coordinated carbonate after Br- substitution in
10. H atoms not involved in the process have been omitted for clarity.
giving the best results experimentally, DMF, was selected and
the parameters used are those of Stassen and co-workers.21 In
the remaining of the paper, only PCM energies on gas phase
optimized geometries are given.
Scheme 3. Reaction Energies (kJ mol-1) for the Substitution
Reaction of Br- by the Various Bases KO2CX in 10 (∆tEY) and 9
(∆cEY)
3.1. C-Br Oxidative Addition. The oxidative addition of aryl-
halide bond to Pd0 has been extensively studied computationally
as it is the first step in any Pd-catalyzed cross-coupling
reaction.22 C-Br activation is generally considered to proceed
through oxidative addition from an η2-CdC complex of
bromobenzene. In the case of iodobenzene, an alternative
mechanism with direct I · · ·Pd interaction has been proposed
22g
by Thiel et al.
However this mechanism is operating when
two phosphine ligands are present on Pd, which is not the case
with PtBu3. Therefore only the oxidative addition pathway has
been considered. Figure 3 shows the geometry of the various
extrema located along the pathway for C-Br bond cleavage.
The η2-CdC complex 8 is easily obtained with a formation
energy of -67.2 kJ mol-1. The activation barrier to reach TS-
8-9, 19.7 kJ mol-1, is low and very similar to that obtained
by Lin (22.6 kJ mol-1) for Ph-Br reacting with Pd(PMe3),22h
and lower than that obtained by Ahlquist and Norrby (36.7 kJ
mol-1) for Ph-Cl reacting with Pd(PtBu3).22k In the transition
state TS-8-9, the C-Br bond is longer than in 8 (1.962 Å, 8;
2.161 Å, TS-8-9) and the Pd-C bond is shorter (2.130 Å, 8;
2.019 Å, TS-8-9) in agreement with an oxidative addition
process.
electron-rich phosphine PtBu3. The C-Br oxidative addition is
thus an easy process with a low activation barrier and a
significant negative reaction energy. Two isomers are possible
for the resulting PdII ML3 complex, 9 and 10, with the isomer
featuring the vacant site trans to the aryl ligand, 10, being more
stable.
3.2. C-H Activation from 10. Several computational studies
have shown how proton abstraction from an aromatic C(sp2)-H
bond is made easier upon coordination of the base trans to the
aryl ring.20 Recently, C(sp3)-H activation has also been ad-
dressed computationally.6l In some instances this intramolecular
pathway has been compared with an intermolecular pathway
where the base is not coordinated to the metal.20e,23 For Pd-
complexes with a bidentate phosphine, it was shown experi-
mentally and computationally that an intermolecular pathway
is preferred.23 The C-H activation can even be performed by
the coordinated Br in the case of electron-deficient arenes as
shown by Fagnou.20g Nevertheless, in the majority of the cases
the preferred pathway is the intramolecular with coordination
of the base. The only geometry considered in these computa-
tional studies for the Pd(Ar)(PR3) complexes is with the
phosphine cis to the aromatic ring, leaving thus two cis-vacant
sites for the base to interact with Pd.6l,20c,e,g
The product of C-Br activation, 9, is different from the X-ray
structure obtained by Hartwig et al. on Pd(Ph)(Br)(PtBu3).19b
In the latter compound, the aryl is trans to the vacant site with
the phenyl ring perpendicular to the plane, whereas the Br atom
is trans to the phosphine. A structure similar to that observed
by Hartwig, 10 (Figure 3), was found to be 27.2 kJ mol-1 more
stable than 9. In both structures 9 and 10, contrary to the
complex of Hartwig, there was no clear indication of any agostic
t
interaction developing between a C-H bond on a Bu group
(phosphine or aryl) and the palladium center. The C-Br
oxidative addition reaction is exothermic with a reaction energy
of -34.5 kJ mol-1 to form 9 from 8. The reaction energy for
Ph-Br reacting with Pd(PMe3) was computed by Lin et al. to
Following these works, substitution of Br- in 10 by the
various bases studied yielded (κ2-O2CX) Pd complexes, 11Y (X
) Me, acetate, Y ) A; X ) OH, bicarbonate, Y ) B; X ) O-,
carbonate, Y ) C, see Figure 4 for geometry of 11C). The
complexes with acetate, 11A, and bicarbonate, 11B, are neutral
and the complex with carbonate, 11C, is anionic. The energy of
the reaction ∆tEY is computed to be negative for all cases
(Scheme 3). The reaction is more exothermic in the case of
carbonate because the dianionic nature of the base allows a more
efficient (κ2-O2CX) bonding interaction.
From these complexes, C-H activation is effective through
TS-11Y-12Y (see Figure 4 for geometry of TS-11C-12C). The
geometry of these TS is very similar to that obtained in other
computational studies. However, the activation barriers are
significantly higher than those reported in the literature (Table
be -14.2 kJ mol-1 22h
The larger exothermicity in our case
.
could be explained by a greater stabilization of PdII by the more
(21) Bo¨es, E. S.; Livotto, P. R.; Stassen, H. Chem. Phys. 2006, 331, 142.
(22) (a) Bickelhaupt, F. M.; Ziegler, T.; von Schleyer, P. R. Organometallics
1995, 14, 2288. (b) Diefenbach, A.; Bickelhaupt, F. M. J. Chem. Phys.
2001, 5, 4030. (c) Albert, K.; Gisdakis, P.; Ro¨sch, N. Organometallics
1998, 17, 1608. (d) Jakt, M.; Johannissen, L.; Rzepa, H. S.;
Widdowson, D. A.; Wilhelm, R. J. Chem. Soc., Perkin Trans. 2 2002,
576. (e) Sundermann, A.; Uzan, O.; Martin, J. M. L. Chem.-Eur. J.
2001, 7, 1703. (f) Senn, H. M.; Ziegler, T. Organometallics 2004,
23, 2980. (g) Gooꢀen, L. J.; Koley, D.; Hermann, H.; Thiel, W. Chem.
Commun. 2004, 2141. (h) Lam, K. C.; Marder, T. B.; Lin, Z.
Organometallics 2007, 26, 758. (i) Goossen, L. J.; Koley, D.; Hermann,
H. L.; Thiel, W. Organometallics 2005, 24, 2398. (j) Ahlquist, M.;
Fristrup, P.; Tanner, D.; Norrby, P.-O. Organometallics 2006, 25, 2066.
(k) Ahlquist, M.; Norrby, P.-O. Organometallics 2007, 26, 550.
(23) Pascual, S.; de Mendoza, P.; Braga, A. A. C.; Maseras, F.; Echavarren,
A. M. Tetrahedron 2008, 64, 6021.
9
15162 J. AM. CHEM. SOC. VOL. 130, NO. 45, 2008

Benzocyclobutenes by C-H Activation
A R T I C L E S
Table 3. Activation Energy, ∆E#, and Reaction Energy, ∆E (kJ
mol-1) for the C-H Activation Process by the Coordinated Base in
11Y and 13Y (Y ) A, Acetate; Y ) B, Bicarbonate; Y ) C,
Carbonate)
is only slightly favored, whereas coordination of carbonate is
likely to displace Br-.
Contrary to the bromo complex 9, complexes 13Y feature a
clear agostic interaction of one C-H bond of the tBu group on
the aromatic ring (see Figure 5 for 13C, Pd· · ·C ) 2.436 Å,
Pd· · ·H ) 1.878 Å, C-H ) 1.135 Å). The proton transfer from
C(sp3) to O is effective in a plane perpendicular to the P-Pd-Ph
axis as illustrated in the geometry of TS-13C-14C shown in
Figure 5. The agostic Pd · · ·C bond distance has been reduced
by ca. 0.1 Å (2.350 Å, TS-13C-14C; 2.436 Å, 13C), the C-H
bond is elongated to 1.4 Å and the H · · ·O distance has decreased
to 1.3 Å, while the Pd · · ·H bond is still long (2.024 Å), speaking
against any direct Pd · · ·H bonding in the TS. The reaction is
best described as a proton transfer to the basic uncoordinated
oxygen atom of the carbonate. Table 3 shows a comparison of
the activation barriers and the reaction energies for the
transformation for the three different bases. The activation
barriers are all lower than the corresponding value obtained for
the C-H activation from the (κ2-O2CX) complexes 11Y, only
slightly for acetate (Y ) A) and bicarbonate (Y ) B) but
significantly for carbonate (Y ) C). As a result the activation
barrier is computed to be the lowest for carbonate, in agreement
with the experimental results. The differences in magnitude
between the computed activation barriers qualitatively reproduce
the difference in activity (3:1 for carbonate vs bicarbonate and
7.3:1 for carbonate vs acetate). This is in qualitative agreement
with the conversions observed experimentally (Table 1, entries
1, 5, 10), albeit on a different system (1a).
Contrary to the case of 11Y, the C-H activation in 13Y is
computed to be exothermic (Table 3). As the phosphine ligand
is already trans to the strongest σ-donor (aryl), there is no
particular energetic penalty introduced on the Pd-PR3 interaction
in the C-H activation. The Pd...P bond distance is in fact
identical in the reactant 13C (2.532 Å) and in the product 14C
(2.542 Å) and even slightly shorter in TS-13C-14C (2.450 Å).
The forming Pd-C(sp3) bond is trans to the coordinated oxygen
of the base. In the cases of acetate and bicarbonate, the proton
transfer is accompanied by a switch of the nature of the
coordinated oxygen from alkoxy in 13Y (Pd-O ) 2.057 Å, Y
) A; 2.051 Å, Y ) B) to ketone in 14Y (Pd-O ) 2.365 Å, Y
) A; 2.399 Å, Y ) B). The loss of Pd-O interaction is
compensated by the energy gain of creating the O-H bond. In
the case of the carbonate, the Pd-O bond distance only
elongates by 0.15 Å upon going from 13C to 14C (1.99 vs 2.148
Å) as result of the “dianionic” nature of the carbonate. There is
no need for extensive reorganization of the electronic structure
of the base to achieve the proton transfer. As a consequence
the activation barrier is lower and the reaction is more
exothermic (Table 3).
11Y
13Y
∆E#
∆E
∆E#
∆E
Y ) A
Y ) B
Y ) C
138.0
143.5
187.9
85.0
100.1
78.8
122.3
118.3
115.3
-19.0
-6.8
-82.5
Figure 5. Optimized geometry for the extrema located along the pathway
for C-H activation by the coordinated carbonate after Br- substitution in
9. H atoms not involved in the process have been omitted for clarity.
3). For C(sp2)-H activation by HCO3-, Maseras, Echavarren et
al. reported an activation barrier of 98.4 kJ mol-1 20c,e,23
The
.
present activation barrier with bicarbonate is significantly higher
(143.5 kJ mol-1) reflecting both the lower acidity of C(sp3)-H
bonds and the increased steric bulk of the phosphine (PH3 was
the model phosphine in the study by Maseras). The importance
of the bulk of the phosphine is illustrated by a comparison with
a study by Fagnou.6l In this case, C(sp3)-H activation was
studied with AcO- as base and PMe3 as phosphine. The
computed activation barrier (113 kJ mol-1) is again lower than
the value obtained in the present work (138 kJ mol-1). Thus,
in the present system with PtBu3 cis to the aromatic ring and a
C(sp3)-H bond to cleave, the computed activation barrier are
significantly higher than in other computational studies. The
value obtained for CO32- reaches a high 187.9 kJ mol-1. This
is a consequence of the necessity to break the very stable (κ2-
O2CX) interaction in 11C.
The trends in the computed barrier (Table 3) are completely
opposite to the experimental results (Table 1). Carbonate is the
worst base and acetate is the best. Moreover the C-H activation
reaction is in all cases strongly endothermic with reaction
energies greater than 80 kJ mol-1 (Table 3). This is the result
of creating a Pd-C(sp3) bond trans to P and losing the (κ2-
O2CX) coordination of the base. As a matter of fact, in the case
of carbonate, the Pd-P bond distance elongates by 0.05 Å in
the TS (2.351 Å, 11C; 2.402 Å, TS-11C-12C), but the elongation
reaches a high 0.3 Å in the product (2.685 Å, 12C). There is
thus significant loss of Pd-PR3 interaction in the product of C-H
activation, leading to a strongly endothermic transformation.
The energetic pattern is thus very unfavorable and the C-H
For the C-H activation step in the presence of a bulky
phosphine there is thus a clear preference for a pathway with
κ1-coordination of the base cis both to the aromatic ring and to
the phosphine (Table 3). The substitution of Br- by the base is
computed to be an exothermic process with a larger thermo-
dynamic driving force for carbonate (Scheme 3). The activation
barrier computed with the various bases are similar in magnitude
but with a slightly lower value for carbonate as observed
experimentally. The lower activation barrier and the more
exothermic reaction energy (Table 3) for the proton transfer in
13C are due to the ease for carbonate to reorganize its electronic
structure both to achieve good interaction with Pd through one
oxygen atom and to develop good basic properties through the
other oxygen atom during the whole process.
t
activation of the Bu group is not likely to proceed from
coordination of the base to 10.
3.3. C-H Activation from 9. The actual product of C-Br
activation is not the more stable compound 10 but complex 9
featuring a bromide trans to the vacant site. Substitution in this
complex of Br- for the various bases yielded the three (κ1-
O2CX) complexes 13Y (X ) Me, acetate, Y ) A; X ) OH,
bicarbonate, Y ) B; X ) O-, carbonate, Y ) C, see Figure 5
for geometry of 13C). The substitution reaction ∆cEY is
exothermic (Scheme 3) but the thermodynamic preference for
the (κ1-O2CX) bonding in 13Y is less strong than for the (κ2-
O2CX) bonding in 11Y. Substitution by acetate or bicarbonate
9
J. AM. CHEM. SOC. VOL. 130, NO. 45, 2008 15163

A R T I C L E S
Chaumontet et al.
Scheme 4. Deuterium-Labeling Experimentsa
Figure 7. Optimized geometry for the extrema located along the pathway
for proton transfer to the aromatic ring from 14C. H atoms not involved in
the process have been omitted for clarity.
3.5. 1,4-Migration of Palladium. With bromoarenes bearing
a substituent para to Br, the regioisomeric product 3 resulting
formally to a 1,4-migration of Pd has been observed in some
instances together with the expected product 2 (Table 2, entries
21, 23, 24). The 1,4-palladium migration was also evidenced
experimentally using deuterated substrate 1ad (Scheme 4).
Partial incorporation of deuterium (10%) was indeed observed
on the ortho aromatic position by 2H NMR.24 The observation
of regioisomers 3 and the partial incorporation of D in product
2ad implies that a hydrogen, initially on the alkyl group, ends
up on the aromatic ring. From the product of C-H activation,
14C, a transition state, TS-14C-17C, for proton transfer to the
aromatic ring has been located on the potential energy surface
(Figure 7). The activation barrier is computed to be 81.9 kJ
mol-1 and the reaction is moderately exothermic (-27.8 kJ
mol-1) to form the κ2-carbonate complex 17C. Thus the proton
transfer to form 17C is competitive with the dissociation of
HCO3- from 14C to form 15 and then to form 2c through C-C
coupling. In fact the activation barrier for protonation of the
aromatic ring is lower than the one for C-C reductive
elimination (94 kJ mol-1). Thus formation of 17C en route to
forming 2c would explain the experimental observations. Indeed
rotation of the aromatic ring in 17C gives access to a proton
transfer from the other ortho position to reform 14C that would
ultimately give the benzocyclobutene resulting from 1,4-
palladium migration. This process is similar to the hydrogen
shifts studied computationally by Dedieu et al.25 However in
their case the migration was effective through interaction with
the Pd center. In the present case the coordinated base is
responsible for the proton migration. 1,4-Migrations of pal-
ladium were observed experimentally in many instances,6g,9,26,27
but their exact mechanism has remained elusive. The present
work suggests that Pd-bound carbonate plays a key role as
proton shuttle in these migration processes.
a Reactions conditions: see Table 2.
Figure 6. Optimized geometry for the extrema located along the pathway
for C-C coupling from 15. H atoms have been omitted for clarity.
3.4. C-C Reductive Coupling. The C-H activation step is
always considered to be the rate-limiting step as confirmed by
experimental values for kinetic isotope effect usually larger than
3. However, generally, the C-C coupling step does not involve
the creation of a constrained 4-membered ring. It was thus
important to check if the C-H activation step remains the rate-
determining step in the present case.
The reaction of partially deuterated substrate 1ac under the
reoptimized conditions (Table 2) allowed the determination of
a primary intramolecular isotope effect of 5.8 (Scheme 4), that
was similar to those reported in the literature for similar types
of reactions.6l,20c,e,23 This value is in agreement with C-H
activation being the rate-limiting step.
All attempts to locate a transition state for C-C coupling
with the protonated base still coordinated failed. Dissociation
of the protonated base from 14Y yielded the complex 15 (Figure
6) featuring a vacant site trans to the alkyl group. The
dissociation is endothermic but the binding energy of HO2CX
is low (20.5 kJ mol-1, X ) CH3; 7.4 kJ mol-1, X ) OH; 21.0
kJ mol-1, X ) O-), therefore entropy effects associated to the
dissociation will override the enthalpy contributions and will
drive the reaction toward 15.
The transition state for C-C coupling from 15 has been
located on the potential energy surface (Figure 6) and the
activation barrier amounts to 94 kJ mol-1. In the transition state,
the forming C · · ·C bond distance is 1.893 Å, while the Pd · · ·C
bond distances are 2.079 Å (aryl) and 2.174 Å (alkyl). The
product of the reaction, 16, is a π-complex of the aromatic ring
(Figure 6) and the transformation from 15 to 16 is exothermic
(-15.7 kJ mol-1). The coordination energy of the benzocy-
clobutene 2c to Pd(PtBu3) amounts to -79.1 kJ mol-1, a
quantity similar to the binding energy of 1c (-67.2 kJ mol-1).
Therefore the catalyst Pd(PtBu3) will be easily regenerated by
dissociation of 2c to enter a new catalytic cycle upon coordina-
tion of 1c. The C-C coupling step is thus easier than the C-H
activation step, even in the case of the formation of the
4-membered ring. The highest energy barrier along the catalytic
cycle is thus that corresponding to the C-H activation step.
For this step the nature of the base has an influence and it was
The above transformation not only explains the regioisomers
observed experimentally, but it is also a strong support for the
(24) The low level of incorporation, compared to the amount of “abnormal”
regioisomer observed with substituents para to the bromine and in
particular with fluorine (Table 2, entry 21), might reflect a tendency
of electron-withdrawing substituents to favour the 17C f 3 pathway:
see refs 20c, e, and 23. Alternatively, an intermolecular H/D exchange
with traces of water or another protic source could be operating: see
ref 26e.
(25) (a) Mota, A. J.; Dedieu, A.; Bour, C.; Suffert, J. J. Am. Chem. Soc.
2005, 127, 7171. (b) Mota, A. J.; Dedieu, A. J. Org. Chem. 2007, 72,
9669.
(26) (a) Ca´mpora, J.; Lo´pez, J. A.; Palma, P.; Valerga, P.; Spillner, E.;
Carmona, E. Angew. Chem., Int. Ed. 1999, 38, 147. (b) Catellani, M.;
Cugini, F.; Bocelli, G. J. Organomet. Chem. 1999, 584, 63. (c) Karig,
G.; Moon, M.-T.; Thasana, N.; Gallagher, T. Org. Lett. 2002, 4, 3115.
(d) Huang, Q.; Fazio, A.; Dai, G.; Campo, M. A.; Larock, R. C. J. Am.
Chem. Soc. 2004, 126, 7460. (e) Zhao, J.; Campo, M.; Larock, R. C.
Angew. Chem., Int. Ed. 2005, 44, 1873. (f) Kesharwani, T.; Larock,
R. C. Tetrahedron 2008, 64, 6090.
shown that the height of the barrier varies as CO32- < HCO3
-
< AcO- in qualitative agreement with the experimental results.
(27) Ma, S.; Gu, Z. Angew. Chem., Int. Ed. 2005, 44, 7512.
9
15164 J. AM. CHEM. SOC. VOL. 130, NO. 45, 2008

Benzocyclobutenes by C-H Activation
A R T I C L E S
Scheme 5. Overall Mechanism Based on Experimental and Computational Results (L ) PtBu3)
overall mechanism for formation of 2c from 1c (Scheme 5).
Only coordination of the base cis to the aromatic ring would
allow such a transformation. In the products of C-H activation
resulting from coordination trans to the aryl (12C, see Figure
4), the steric bulk introduced by PtBu3 would impede any
interaction between the formed O-H and the aromatic carbon
bonded to Pd. In the product from C-H activation with the
base coordinated cis (14C), the steric bulk of the phosphine is
not interfering with the proton transfer to the aromatic ring (see
Figure 7). Moreover the observation of the 1,4-palladium
migration and the proposal of a base-assisted mechanism for
such transformation is a strong support for the intramolecular
C-H activation by an external base and would be more difficult
to explain if an intermolecular deprotonation with a noncoor-
dinated base were to be operative.
In the κ1-adduct, the proton transfer from the sp3 carbon is
computed to be easier for carbonate compared to bicarbonate
and acetate, in agreement with the experimental results. The
lower activation barrier is the result of both a higher basicity
and a smaller electronic reorganization of the coordinated base
upon protonation. Dissociation of the protonated base generates
an unsaturated metallacycle from which C-C coupling to form
the benzocyclobutene is effective. The activation barrier for
C-C coupling is lower than the one for C-H cleavage, thus
setting the latter process as the rate-limiting step in agreement
with the experimental kinetic isotope effect. The κ1-coordination
of the base cis to the aromatic ring gives access to a competitive
pathway. The proton on the base can be transferred to the
aromatic ring, thus yielding an alkyl intermediate. The transfer
is reversible and, finally, the BCB is obtained, but rotation of
the phenyl ring in the alkyl intermediate gives access to the
other regioisomer. This product results formally from a 1,4-
migration of Pd and the mechanism presented therein (Scheme
5) could be more general and could allow to explain other 1,4-
migrations of palladium observed experimentally.
Conclusions
We reported on an efficient catalytic system for the synthesis
of benzocyclobutenes by C-H activation of methyl groups. The
most general conditions were found with a combination of
Pd(OAc)2 and PtBu3 as catalyst, K2CO3 as the base, and DMF
as solvent. A variety of substituted BCB were obtained using
this method with yields in the 44-92% range, including
molecules that are hardly accessible by other methods. The
reaction was found limited to substrates bearing a quaternary
benzylic carbon, but BCB bearing a tertiary benzylic carbon
could be obtained indirectly from diesters by decarboxylation.
Reaction substrates bearing a small substituent para to bromine
gave an unexpected regioisomer that likely arouse from a 1,4-
palladium migration process. The formation of this “abnormal”
regioisomer could be suppressed by introducing a larger
subsituent para to bromine. Given the synthetic interest of BCB
in pericyclic reactions via their o-xylylene form, this work
should open new perspectives in the chemistry of BCB and their
use for the construction of polycyclic molecules.
Experimental Section
General C-H Activation Procedure (Table 2). A dry reseal-
able Schlenk tube containing a magnetic rod was charged with the
aryl bromide (1 mmol), Pd(OAc)2 (0.1 equiv), P(t-Bu)3 ·HBF4 (0.2
equiv), and dry K2CO3 (1.3 equiv). The Schlenk tube was evacuated
and backfilled with argon twice, then capped with a rubber septum.
Dry DMF (4.0 mL) was injected under argon, then the septum was
replaced by a screwcap and the mixture was stirred at 140 °C
(preheated oil bath) until disappearance of the starting material in
GCMS analysis (generally 1 to 1.5 h). After cooling, the mixture
was diluted with Et2O and filtered through Celite. The organic
solution was washed with brine, dried over MgSO4, and the solvent
evaporated under reduced pressure. The residue was purified by
flash chromatography to afford the benzocyclobutene.
Computational Details. The calculations were performed with
the Gaussian03 package28 at the B3PW91 level.29 Palladium was
represented by the relativistic effective core potential (RECP) from
the Stuttgart group and the associated basis set,30 augmented by a
DFT calculations on the reaction mechanism highlighted
important features associated to this transformation. The bulk
of the phosphine is not only necessary to create a monoligated
Pd0 complex as the active species, it also allows for a
κ1-coordination of the base cis to the metallated aromatic ring.
This geometry prevents the formation of very stable κ2-adducts
of the base that would make the C-H activation step difficult.
(28) Pople, J. A. et al. Gaussian 03, revision C.02; 2003.
(29) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Perdew, J. P.;
Wang, W. Phys. ReV. B 1992, 45, 13244.
9
J. AM. CHEM. SOC. VOL. 130, NO. 45, 2008 15165

A R T I C L E S
Chaumontet et al.
f polarization function.31 Bromine and phosphorus were represented
by the relativistic effective core potential (RECP) from the Stuttgart
group and the associated basis set,32 augmented by a d polarization
function.33 A 6-31G(d,p) basis set was used for all the other atoms
(C, H, O).34 The geometry optimizations were performed without
any symmetry constraint followed by analytical frequency calcula-
tions to confirm that a minimum or a transition state had been
reached. The nature of the species connected by a given transition
state structure was checked by optimization as minima of slightly
altered TS geometries along both directions of the transition state
vector. The energies of all the systems studied at the B3PW91 level
in gas phase were computed with inclusion of solvent effects (DMF)
according to the PCM scheme as implemented in Gaussian. The
geometries obtained in gas phase were used without further
reoptimization within the PCM methodology35 and the united atom
topological model with UAKS radii was used. Optimizations in
solvent have been shown by Senn and Ziegler to yield qualitative
difference in C-X oxidative addition reactions,22f but in the present
study, attempts to converge geometry optimizations within PCM
failed.
Acknowledgment. This work was financially supported by
Institut de Recherche Servier (fellowships to R.P. and N.A.), ICSN
(fellowships to M.C. and J.H.), and CNRS (ATIP grant to O.B.).
Supporting Information Available: Full characterization of
all new compounds, detailed experimental procedures, copies
of NMR spectra for target benzocyclobutenes and X-ray crystal
structure data (CIF) for compound 2z. Complete ref 28 and
Cartesian coordinates of all optimized structures. This material
is available free of charge via the Internet at http://pubs.acs.org.
(30) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor.
Chim. Acta 1990, 77, 123.
(31) Ehlers, A. W.; Bohme, M.; Dapprich, S.; Gobbi, A.; Hollwarth, A.;
Jonas, V.; Kohler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
Chem. Phys. Lett. 1993, 208, 111.
(32) Bergner, A.; Dolg, M.; Kuchle, W.; Stoll, H.; Preuss, H. Mol. Phys.
1993, 80, 1431.
JA805598S
(33) Hollwarth, A.; Bohme, M.; Dapprich, S.; Ehlers, A. W.; Gobbi, A.;
Jonas, V.; Kohler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
Chem. Phys. Lett. 1993, 208, 237.
(34) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213.
(35) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. ReV. 2005, 105, 2999.
9
15166 J. AM. CHEM. SOC. VOL. 130, NO. 45, 2008