Palladium-catalyzed intramolecular C(sp3)-H functionalization: Catalyst development and synthetic applications

Article abstract of DOI:10.1002/chem.200600811

A novel catalytic system, based on a mixture of palladium acetate and tris(5-fluoro-2-methylphenyl)-phosphane (F-TOTP), has been designed for the intramolecular C-H functionalization of alkane segments. Among other analogues of tris(2-methylphenyl)phospha

Full text of DOI:10.1002/chem.200600811

DOI: 10.1002/chem.200600811
3
ꢀ
Palladium-Catalyzed Intramolecular C(sp ) H Functionalization: Catalyst
G
Development and Synthetic Applications
Julien Hitce,[a] Pascal Retailleau,[a] and Olivier Baudoin*[b]
Dedicated to K. C. Nicolaou on the occasion of his 60th birthday
Abstract:
A
novel catalytic system,
under milder reaction conditions that
allowed the regioselective production
of various olefins adjacent to a quater-
nary benzylic carbon atom, as well as
novel bi- and tricyclic molecules. A
general mechanism was proposed, with
the preferential formation of a six-
membered palladium(ii) palladacycle
after oxidative addition and cyclopalla-
based on a mixture of palladium ace-
tate and tris(5-fluoro-2-methylphenyl)-
phosphane (F-TOTP), has been de-
signed for the intramolecular C H
functionalization of alkane segments.
Among other analogues of tris(2-meth-
ꢀ
dation. The regioselective C H func-
ꢀ
tionalization of alkyl groups into ole-
fins was illustrated in the synthesis of
the antihypertensive drug verapamil
(14) and an analogue (19). A particu-
larly mild ruthenium-catalyzed direct
hydroamidation of the intermediate
olefin in this synthesis is also reported.
ylphenyl)phosphane (P(o-tol)3), F-
A
ꢀ
Keywords: C H
activation
·
·
TOTP was shown to have the optimal
metal-bonding properties for this reac-
tion. This catalytic system operated
catalysis
·
dehydrogenation
hydroamidation · palladium
Introduction
tions of metal carbenes and metal nitrenes have proven to
be particularly efficient, both in nonstereoselective and ster-
eoselective forms, and have been successfully applied to nat-
The activation and functionalization of unactivated alkane
and arene C H bonds directed toward organic synthesis is
ural product synthesis.[3] Other milestones in C H function-
ꢀ
ꢀ
currently of great interest.[1] A number of transition-metal-
alization have recently been accomplished in the absence[4]
or presence[5] of a metal-directing group. In the latter case,
the metal is precoordinated by at least one heteroatom of
the substrate molecule, which triggers the cleavage of a par-
ꢀ
mediated C H functionalization processes have been descri-
bed in the literature in the past decade; these processes pro-
vide chemists with atom- and step-economical alternatives
to standard methods, allow access to novel molecular scaf-
folds, and offer new bond disconnections for target-oriented
synthesis. Compared to those for arenes,[1,2] there are far
ꢀ
ticular C H bond through cyclometalation. This approach
was utilized by the Sames groupin the total syntheses of
rhazinilam and the core of teleocidin B4, with a stoichiomet-
ric amount of transition metal (PtII or PdII).[6]
ꢀ
fewer reports on the C H functionalization of alkane seg-
ꢀ
ꢀ
ments of organic molecules. Within this field, C H inser-
In 2003, we reported a palladium(0)-catalyzed C H func-
tionalization of benzylic alkyl groups that gives rise to ole-
[a] J. Hitce, P. Retailleau#
ICSN-CNRS, Gif-sur-Yvette (France)
[b] Prof. Dr. O. Baudoin
UniversitØ Claude Bernard–Lyon I, bât. CPE
43 bd du 11 Novembre 1918
69622 Villeurbanne (France)
Fax : (+33)472-432-963
E-mail: Olivier.baudoin@univ-lyon1.fr
[#] X-ray crystal structure analysis.
Supporting information for this article (full characterization of all
new compounds, detailed experimental procedures, and copies of
NMR spectra for the target compounds) is available on the WWW
under http://www.chemeurj.org/ or from the author.
ꢀ
Scheme 1. Palladium(0)-catalyzed C H functionalization of benzylic
alkyl groups. P
A
792
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 792 – 799

FULL PAPER
fins (alkane dehydrogenation) or benzocyclobutenes
(Scheme 1).[7] In contrast to other reports featuring direct
precoordination of PdII by heteroatoms of the substrate,[5]
the PdII intermediate in our case was generated by oxidative
addition of an aryl halide to Pd0. This initiated an intramo-
base, with both ingredients having a strong impact on the re-
action yield. We concentrated our reoptimization efforts on
the reaction of compound 1a to give olefin 1b at 1508C
(Scheme 2).
ꢀ
lecular C H bond cleavage, followed by b-H elimination
and reductive elimination to give an olefin or by direct re-
ductive elimination to give a benzocyclobutene. In contrast
to the seminal work of Dyker[8] and the recent report by
Buchwald and co-workers,[9] our process is purely intramo-
ꢀ
lecular and is not accompanied by intermolecular C C bond
ꢀ
ꢀ
formation, but instead by intramolecular C H or C C bond
formation.[10] Therefore this reaction can best be termed in-
3
ꢀ
tramolecular C
(sp ) H functionalization. In view of the rela-
tively harsh conditions that we first reported (1508C,
10 mol% Pd), we decided to reoptimize the catalytic system
in order to come upwith milder conditions, better suited to
applications in target-oriented synthesis. In this article, we
report the design of a more efficient catalyst, the extension
of the reaction scope (for the formation of olefins and bi- or
tricyclic molecules), and the application of the method to
the synthesis of the antihypertensive drug verapamil.
Results and Discussion
ꢀ
Design of a new catalytic system: Aside from C H inser-
ꢀ
Scheme 2. Effect of ligand basicity on the intramolecular C H functional-
tions,[3] we are not aware of previous specific ligand design
ization of 1a. Reaction conditions: 1a (0.4 mmol), Pd
A
3
ꢀ
in the context of C
(sp ) H functionalization. Our initial
10 mol%), ligand (2 mol% or 20 mol%), K2CO3 (2 equiv), DMF ([1a]=
0.2m), 1508C.
ligand screening revealed P
(o-tol)3 (L1) to be the most
active ligand among a variety of aryl and alkyl phosphanes
for the dehydrogenation of alkyl groups.[7] Optimal condi-
tions were found with DMF as the solvent and K2CO3 as the
First, different sources of palladium(ii) or palladium(0)
were screened, including PdCl2, PdBr2, Pd
(acac)2] (acac=acetylacetone), and [Pd2(dba)3] (10 mol%
Pd; dba=trans,trans-dibenzylideneacetone), in combination
with P
(o-tol)3 (2 equiv for PdII, 1 equiv for Pd0); among
these sources, Pd(OAc)2 gave the highest yield (68%) and
A
G
ACHTREUNG
AHCTREUNG
Abstract in French: Un nouveau syst›me catalytique, formØ à
partir de lꢀacØtate de palladium et de la tris(5-fluoro-2-me-
thylphenyl)phosphine (F-TOTP), a ØtØ conÅu pour la fonc-
AHCTREUNG
reaction rate (0.5 h, turnover frequency (TOF)=14 hꢀ1). In
this case, analysis of the reaction mixture run in [D7]DMF
by 31P NMR spectroscopy revealed the presence of the Herr-
mann–Beller palladacycle (d=34.7 ppm),[11] which had been
generated in situ, together with the free phosphane (d=
ꢀ30.4 ppm) and the phosphane oxide (d=35.6 ppm). When
run directly with the Herrmann–Beller palladacycle
(5 mol%), the reaction gave the same yield as the mixture
ꢀ
tionnalisation C H intramolØculaire de groupements alkyles.
Parmi d’autres analogues de P(o-tol)3, il a ØtØ montrØ que F-
A
TOTP poss›de les propriØtØs de coordination du mØtal opti-
males pour cette rØaction. Le nouveau syst›me catalytique
op›re dans des conditions rØactionnelles plus douces qui per-
mettent d’obtenir de faÅon rØgiosØlective diffØrentes olØfines
adjacentes à un carbone quaternaire benzylique ainsi que de
nouvelles molØcules bi- et tricycliques. Un mØcanisme gØnØral
est proposØ, mettant en jeu un palladacycle de palladium(ii) à
six chaînons qui est formØ intermØdiairement apr›s addition
of Pd
N
N
(even when free phosphane was added to the palladacycle).
This indicates that the active palladium(0) species is gener-
ꢀ
ated faster from the mixture of Pd
from the palladacycle.
We then decided to study the influence of the phosphane
basicity, and to this purpose, we synthesized P(o-tol)3 ana-
logues L2–L5 bearing electron-donating or -withdrawing
groups at the 4- or 5-positions (Scheme 2). They were ob-
tained in one stepand high yield from the corresponding
aryl bromides, through formation of the Grignard reagent
A
G
oxydante et cyclopalladation. La fonctionnalisation C H rØ-
giosØlective de groupements alkyles en olØfines a ØtØ illustrØe
par la synth›se du vØrapamil (14), mØdicament anti-hyperten-
seur, et d’un analogue (19). Au cours de cette synth›se, une
hydroamidation directe catalysØe par le ruthØnium, qui s’ef-
fectue dans des conditions particuli›rement douces, est Øgale-
ment dØcrite.
AHCTREUNG
Chem. Eur. J. 2007, 13, 792 – 799
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
793

O. Baudoin et al.
Table 1. Comparison of the basicity of P
A
complex I was generated thermally from a mixture of Pd-
AHCTREUNG
Ligand
L2
L1
L3
L4
L5
[a]
nCO [cmꢀ1
]
1965
689
1970
705
1974
716
1979
726
1983
740
ACHTREUNG
1J(P,Se) [Hz][b]
ACHTREUNG
giving approximately 10% conversion of 1a in 10 h at
1008C (with 2.5 mol% I). This shows that the active catalyt-
ic species was again generated faster from the Pd/ligand
mixture than from the isolated palladacycle.
[a] CO absorption band of the trans-[(R3P)2Rh(CO)Cl] complex, as
measured from the FTIR spectrum. [b] 31
77Se coupling constant meas-
ꢀ
P
31
ꢀ
ured from the P NMR spectrum (121.5 MHz, CDCl3) of the R3P Se
complex.
Extension of the reaction scope: Next, we studied the scope
of the new catalytic system at 1008C (Table 2). The reaction,
which was before limited to substrates bearing a bulky ben-
and reaction with PCl3.[12] Among them, the meta-fluoro
analogue
L4
(tris(5-fluoro-2-methylphenyl)phosphane,
F-TOTP), which has not been
reported previously, was ob-
tained in 82% yield from 2-
bromo-4-fluorotoluene.
Table 2. Synthesis of olefins.[a]
Entry
Starting material
Product(s)
Yield [%][b] (ratio)
The classification of phos-
phanes L1–L5 by decreasing
basicity, L2>L1>L3>L4>L5,
was established by two different
spectroscopic methods: 1) the
measurement of nCO in the IR
1
77
1a
2a
3a
1b
2b
3b
2
3
4
67[c]
82[c]
68
spectrum
of
the
trans-
(R3P)2Rh(CO)Cl complexes[13]
and 2) the measurement of the
1J
the
(P,Se) coupling constants of
phosphane
selenides
(Table 1).[14] The screening of
phosphanes L1–L5 in the reac-
tion of compound 1a with Pd-
A
(10 mol%)/ligand
(20 mol%; Scheme 2) revealed
F-TOTP (L4) to be the most
active ligand in terms of yield
(82%, turnover number (TON)
= 8.2) and rate (TOFꢁ16 hꢀ1).
F-TOTP thus seems to have the
optimal basicity for this reac-
tion. A decrease in the amount
of palladium to 1 mol% slowed
down the reaction, which differ-
entiated further the different li-
4a
5a
4b
5b
5
6
70
gands and showed again
a
68 (93:7)
higher activity for F-TOTP
(TON=71, TOF=71 hꢀ1). The
reaction yield decreased mark-
edly when less than 1 mol% Pd
was employed. With these re-
sults in hand, we attempted to
decrease the reaction tempera-
ture. Gratifyingly, the same re-
action could be run at 1008C
6a
7a
6b+6c
7
8
63
7b
85 (55:45)
with 5 mol% Pd(OAc)2 and
A
10 mol% F-TOTP, to give
olefin 1b in 77% yield in 1 h
(see Table 2, entry 1). The
8a
8b+8c
[a] Pd(OAc)2 (5 mol%), F-TOTP (L4, 10 mol%), K2CO3 (2 equiv), DMF, 1008C. [b] Yield after isolation by
A
Herrmann-type
palladacyclic flash chromatography. [c] With 20 mol% of F-TOTP.
794
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 792 – 799

3
ꢀ
Palladium-Catalyzed C(sp ) H Functionalization
A
FULL PAPER
Table 3. Synthesis of bi- and tricyclic molecules.[a]
Entry Starting ma- Product(s)
terial
Yield [%][b]
(ratio)
1
62 (4:1)
9a
9b (ꢂ)+9c
2
62
Figure 1. X-ray crystal structure of complex I (30% thermal ellipsoids
10a
10b
plot, hydrogen atoms were omitted for clarity).[15]
3
70 (7:3)
zylic substituent,[7] now works ideally with a benzylic nitrile.
This is a clear improvement as a greater variety of alkyl
groups (in particular, bulkier groups), undergoing the intra-
11a
11b+11c
[a] Pd(OAc)2 (5 mol%), F-TOTP (L4, 20 mol%), K2CO3 (2 equiv),
DMF, 1008C. [b] Yield after isolation by flash chromatography.
G
ꢀ
molecular C H functionalization, can be introduced a to
the nitrile groupby standard alkylation.
Nevertheless, groups other than nitrile, for instance, ester
groups, could be employed successfully (Table 2, entries 2
and 6). Dialkyl substitution was required as before, as the
monoethyl analogue of 1a did not furnish the corresponding
olefin. Electron-rich (Table 2, entry 3) or electron-poor
(Table 2, entry 4) substituents on the aromatic ring were
well tolerated, with a decreased reaction rate for the CF3
substituent. Di-n-propyl substrate 5a furnished olefin 5b
(Table 3, entry 3) in 70% combined yield. Compounds 11b
and 11c were isolated from this mixture in 50 and 20%
yield, respectively, as single diastereoisomers (as determined
1
at the precision of H NMR spectroscopy). The reduction of
the nitrile groupin
11b provided primary amine 13
(Scheme 3), which again revealed a cis ring fusion by
with complete regioselectivity and
E stereoselectivity
(Table 2, entry 5). It is noteworthy that the corresponding
ethyl ester derivative 6a gave a 93:7 mixture of the substi-
tuted and terminal olefins (Table 2, entry 6). The reaction of
cycloalkanes 7a and 8a (Table 2, entries 7 and 8) gave inter-
esting results: in the first case olefin 7b was produced regio-
selectively, whereas in the second case an approximately 1:1
mixture of inseparable regioisomers 8b and 8c was ob-
tained.
The reaction of substrates bearing a tertiary a-benzylic
carbon atom also gave interesting results (Table 3). First, di-
Scheme 3. Reduction of nitrile groups to primary amines to determine
the relative configuration of 10b.
A
indane 9b and olefin 9c (Table 3, entry 1). Compound 9b
was isolated as a single diastereoisomer in 50% yield, and
its relative stereochemistry was deduced from NOESY ex-
ꢀ
periments. The synthesis of 9b via intramolecular C C bond
NOESY experiments. The relative configuration of 11c was
also ascribed from NOESY experiments. The formation
mechanism of 11b and 11c will be commented on in the
next section. Compound 11b can be seen as a nor-abietane-
type diterpenoid analogue (benzene–cyclopentane–cyclohex-
formation was exploited further with the intramolecular
functionalization of (2R,5R)-dimethylcyclopentane 10a
(Table 3, entry 2), obtained in two steps from commercially
available (2S,5S)-hexanediol. Gratifyingly, the major product
was tricyclic molecule 10b, isolated as a single diaster-
eoisomer, together with a small amount of olefin. To deter-
mine the relative, and by extension absolute, configuration
of 10b, the nitrile groupwas reduced with LiAlH 4 to the pri-
mary amine 12 (Scheme 3), from which the cis ring fusion
was deduced. (2S,6S)-Dimethylcyclohexane 11a was ob-
tained in a similar fashion from commercially available
(2R,6R)-heptanediol. The intramolecular functionalization
of 11a gave a 7:3 mixture of tricycle 11b and olefin 11c
ane fused ring system),[16] therefore C H functionalization
ꢀ
could provide an original and stereoselective entry into this
series of natural products.
ꢀ
In conclusion, it appears that the C H functionalization is
a useful, versatile, and selective method for the synthesis of
olefins adjacent to quaternary benzylic carbon atoms
(Table 2).[17] In addition, when the a-benzylic carbon atom is
tertiary (Table 3), interesting functionalized polycyclic mole-
cules are obtained at the expense of olefin formation.
Chem. Eur. J. 2007, 13, 792 – 799
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
795

O. Baudoin et al.
mechanism, as the smaller ni-
trile groupleads to an increased
proportion of substituted olefin
(100% 5b) as compared to the
ethyl ester (6b/6c 93:7). By
contrast, the formation of ole-
fins 5–8b might be interpreted
either as the formation of palla-
dacycle C followed by endocy-
clic b-H elimination or as the
formation of palladacycle
D
with exocyclic b-H elimination.
However, we suggest that in
these cases six-membered palla-
dacycle C is also formed pref-
erably. This proposal is support-
ed by the deuterium-labeling
experiments that were reported
in our initial communication
and that were confirmed under
the new reaction conditions.[7]
Further experiments and theo-
ꢀ
Scheme 4. Proposed general mechanism for the intramolecular C H functionalization.
Mechanistic proposal: Based on the distribution of products
observed in this study (Tables 2 and 3), we propose the fol-
ꢀ
lowing general mechanism for the intramolecular C H func-
tionalization (Scheme 4).
Oxidative addition of substrate A to Pd0 would generate
PdII complex B, which would undergo cyclopalladation with
ꢀ
concomitant C H bond cleavage to furnish six-membered
palladacycle C or five-membered palladacycle D, complexes
ꢀ
corresponding to cleavage of a b- or a-benzylic C H bond,
respectively (Scheme 4). Recent theoretical studies support
a base-induced proton-abstraction mechanism for the cyclo-
palladation step, rather than a second oxidative addition
IV
ꢀ
ꢀ
leading to an energetically disfavored H Pd Br palladacy-
cle.[18] A b-H elimination stepfrom C or D (Scheme 4,
path a), R2 =H) would then produce palladium hydride E,
which by reductive elimination would give the olefin prod-
uct F. In complex D, the exocyclic position of the CH2R3 res-
idue would enable a classical syn-b-hydride elimination,
whereas in palladacycle C, b elimination could occur either
ꢀ
ꢀ
from a distorted syn Pd C/C H conformation or through
anti elimination.[19] Alternatively, a direct reductive elimina-
tion from palladacycle C (Scheme 4, path b), R2 ¼H) would
furnish indane-type product G. While the latter route is un-
ambiguous for the production of indane 9b and tricyclic
molecules 10b and 11b (Table 3), the formation of olefins
(Table 2) could arise from either palladacycle C or D (or a
mixture of both).
Scheme 5. Probable intermediate palladacycles C1–C3 in the reactions of
9–11a.
A closer examination of the structure of these olefins
(Table 2) suggests that six-membered palladacycle C is also
preferred in this case. Indeed the intervention of palladacy-
cle C, undergoing exocyclic b-H elimination, seems the most
probable explanation for the production of terminal olefin
6c and cyclic olefin 8c.[20] As shown with 5a and 6a, the size
of the benzylic R1 grouphas a significant influence on the
retical calculations are underway to determine whether con-
current mechanistic pathways may also operate.
In the proposed catalytic cycle, paths a) and b) seem to
coexist, as illustrated by the reaction of bromide 9a to give
the mixture of indane 9b and olefin 9c (Table 3, Scheme 5).
Intermediately, palladacycle C1, in which the CN and Me
796
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 792 – 799

3
ꢀ
Palladium-Catalyzed C(sp ) H Functionalization
A
FULL PAPER
groups lie cis to each other, is
probably produced in a diaster-
eoselective fashion and then
evolves by reductive elimina-
tion (path b)) to give indane 9b
and
by
b-H
elimination
(path a)) to give olefin 9c. An
alternative explanation would
be that olefin 9c is produced
from the trans diastereoisomer
of C1. Similarly, the formation
of palladacycle C2 explains the
formation of tricycle 10b from
bromide 10a (Scheme 5). Final- Scheme 6. Synthesis of verapamil (14) and analogue 19: a) LiHMDS (1.0 equiv), DMPU (1.0 equiv), EtI
(1.0 equiv), THF, 208C, 1 h; b) LiHMDS (5 equiv), DMPU (5 equiv), iPrI (5 equiv), THF, 208C, 1 h; c) Pd-
ly, the production of tricycle
A
11b and olefin 11c from bro-
mide 11a can be explained by
the formation of the same pal-
ladacyclic intermediate, C3.
Indeed, the CN and Me groups
Table 4, entry 8; f) NaHMDS (1.5 equiv), MeI (2.0 equiv), THF, 08C, 1 h; g) BH3·SMe2 (1.85 equiv), THF, 0!
208C, 18 h, then 1m HCl, 1008C, 4 h (37% for 19 and 46% for 14 for two steps). HMDS=1,1,1,3,3,3-hexame-
thyldisilazane, DMPU=1,3-dimethyl-3,4,5,6-tetrahydro-2-(1H)-pyrimidinone.
of 11c have the same cis relationshipas those of 11b, where-
as a trans stereochemistry would have been observed if 11c
was generated from the C3 diastereoisomer. It can be noted
that the relative stereochemistry of bicyclic compound 9b is
the opposite to that of tricyclic compounds 10b and 11b.
This probably originates from the fact that different fac-
tors—axial versus equatorial position of substituents in the
first case, ring strain in the second case—govern the forma-
tion of each type of molecule.
In conclusion, the analysis of product distribution shows
that six-membered palladacycle C is a very likely intermedi-
ate in this reaction, at least with certain substrates. At this
point, there seems to be no evidence of a five-membered
palladacycle intermediate D in the formation of olefins.
However, this intermediate is unambiguously involved in
the regioselective formation of benzocyclobutenes from sub-
strates bearing at least one benzylic methyl substituent.[7]
Hence the preference for a five- or six-membered palladacy-
cle in these reactions seems to be related to the substitution
of the a-benzylic carbon atom bound to the palladium: pal-
ladacycle D is formed with a primary a-benzylic carbon
atom (Me group), while palladacycle C is probably formed
with more substituted a-benzylic carbon atoms (Et, nPr, iPr,
etc.).
veratrylamine (2-(3,4-dimethoxyphenyl)ethylamine) under a
CO atmosphere and ruthenium catalysis (Table 4). In the
seminal work, primary amines were treated at 120–1808C
Table 4. Hydroamidation of olefins 1b and 17 with homoveratrylamine.[a]
Entry Olefin Equiv of amine
A
1
2
3
4
5
6
7
8
1b
1b
1b
1b
1b
17
2
4
4
4
4
4
4
4
33
33
5
10
20
33
10
20
Ar
Ar
47
67
34
67
69
36
32
50
CO
CO
CO
Ar
CO
CO
17
17
[a] Olefin (1 equiv), homoveratrylamine, [Ru3(CO)12], DMF, gas atmos-
phere (balloon), 1208C, 17 h. [b] Yield after isolation by chromatography.
with an excess of simple olefins under CO pressure (39 atm)
and with 0.2 mol% [Ru3(CO)12],[23,24] conditions which
would not be reasonable for the homologation of the more
elaborate olefin 17. We were pleased to discover that 1b un-
derwent mild hydroamidation with excess homoveratryla-
mine by using a stoichiometric amount of Ru under an
argon atmosphere (Table 4, entry 1). Optimal reaction con-
ditions were found with DMF as the solvent, at 1208C, and
with 4 equivalents of amine (Table 4, entry 2), to furnish
amide 18a in 67% yield. The optimal 1:4 ratio of olefin/
amine is reversed compared to literature precedents[23,24]
and is better adapted to the homologation of elaborate ole-
fins, such as 17, with commercially available amines. Gratify-
ingly, catalytic amounts of Ru could be employed by run-
ning the reaction under a CO atmosphere, with an optimal
quantity of 10 mol% [Ru3(CO)12] (entries 3–5). In all cases
the conversion was complete, and the identifiable byprod-
ucts included homoveratrylformamide and the reduced
olefin. Based on these byproducts and on literature propo-
sals,[24b] we believe that the active catalytic species in this re-
Synthesis of verapamil: To demonstrate further the utility of
the current method for the functionalization of quaternary
centers, we applied it to the synthesis of verapamil (14), a
well-known calcium-channel blocker (Scheme 6).[21,22] Com-
pound 16, obtained from commercially available bromide 15
by sequential alkylations with EtI and iPrI, underwent re-
ꢀ
gioselective intramolecular C H functionalization catalyzed
by Pd
A
on the more bulky iPr groupcould be observed. The one-
carbon homologation of the bulky olefin 17 proved trouble-
some, so olefin 1b was chosen for model studies. We decid-
ed to examine the direct hydroamidation of 1b with homo-
Chem. Eur. J. 2007, 13, 792 – 799
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
797

O. Baudoin et al.
turated solution of NH4Cl and extracted with diethyl ether. After evapo-
ration, a small volume of ethanol was added to the residue and the mix-
ture was filtered to give F-TOTP as a white powder (1.67 g, 82%). M.p.
1628C; 1H NMR (300 MHz, CDCl3): d=2.34 (s, 3H), 6.39 (ddd, J=9.0,
3.3, 3.3 Hz, 1H), 6.98 (ddd, J=8.5, 8.5, 3.0 Hz, 1H), 7.21 ppm (ddd, J=
8.4, 4.8, 4.8 Hz, 1H); 13C NMR (62.9 MHz, CDCl3): d=20.2 (d, J=
21.6 Hz), 116.1 (d, J=20.6 Hz), 119.1 (d, J=23.2 Hz), 131.7 (dd, J=5.8,
5.8 Hz), 135.7 (dd, J=13.2, 3.8 Hz), 138.2 (d, J=26.0 Hz), 161.4 ppm (d,
J=246.3 Hz); 31P NMR (121.5 MHz, CDCl3); d=ꢀ27.1 ppm; 19F NMR
(282.4 MHz, CDCl3): d=ꢀ116.5 ppm; elemental analysis: calcd for
C21H18F3P: C 70.39, H 5.06; found: C 70.22, H 5.06; HRMS (ESI,
MeOH): m/z: calcd for C21H19F3P: 359.1176 [M+H]+; found: 359.1197.
action is a triruthenium cluster hydride. Transposition of
these conditions to the more bulky olefin 17 gave a 32–36%
yield of amide 18b (Table 4, entries 6 and 7). Finally, an in-
crease in the amount of [Ru3(CO)12] to 20 mol% (60 mol%
Ru) under a CO atmosphere furnished an optimal 50%
yield of amide 18b (Table 4, entry 8). Further investigation
of this particularly mild hydroamidation process is under-
way. The synthesis was completed with methylation of
amides 18a and 18b, followed by chemoselective reduction
[25]
of the amide groupas previously described,
to give vera-
pamil (14) and analogue 19 in 46 and 37% yield, respective-
ly (overall yields: 17% for 14 and 18% for 19).
Palladacycle I: A mixture of palladium acetate (20 mg, 0.09 mmol) and
phosphane L4 (42 mg, 0.12 mmol) in toluene (2 mL) was heated to
1008C for 5 min. After evaporation of the toluene, a mixture of dichloro-
methane (1 mL) and heptanes (1 mL) was added to the residue and the
suspension was filtered through celite. The solution was evaporated and
the solid residue was crystallized from chloroform/heptanes, to give pale
yellow crystals (31 mg, 67%). A monocrystal was analyzed by X-ray dif-
fraction at low temperature;[15] 1H NMR (300 MHz, CDCl3): d=1.95 (s,
3H), 2.0–3.3 (brm, 8H), 6.4–7.3 ppm (brm, 9H); 13P NMR (121.5 MHz,
CDCl3): d=32.3 ppm; 19F NMR (282.4 MHz, CDCl3): d=ꢀ117.1, ꢀ116.2,
ꢀ115.1 ppm.
Representative procedure for the preparation of brominated substrates
(1a): Lithium bis(trimethylsilyl)amide (18.3 mL of a 1m solution in THF,
18.3 mmol) was added dropwise to a solution of 2-bromophenylacetoni-
trile (1.2 g, 6.1 mmol) and DMPU (2.2 mL, 18.3 mmol) in THF (40 mL)
under argon. After the reaction mixture had been stirred for 30 min at
room temperature, iodoethane (1.47 mL, 18.3 mmol) was added with a
syringe and the reaction mixture was stirred for 1 h. The reaction was
quenched with a saturated NH4Cl solution (40 mL) and the aqueous
layer was extracted with diethyl ether (40 mL”3). The combined organic
layers were washed with a 1m HCl solution (40 mL) and with a saturated
solution of NaHCO3 (40 mL). The extracts were then dried over magne-
sium sulfate and evaporated under reduced pressure. The residue was pu-
rified by flash chromatography (silica gel, heptanes/ethyl acetate 95:5) to
afford 1a as a colorless oil (1.42 g, 92%).[26] 1H NMR (250 MHz, CDCl3):
d=0.72 (t, J=7.5 Hz, 6H), 2.06 (m, 2H), 2.65 (m, 2H), 7.17 (dt, J=7.5,
1.3 Hz, 1H), 7.30 (dt, J=8.0, 1.8 Hz, 1H), 7.62 (dd, J=7.8, 1.8 Hz, 1H),
7.69 ppm (dd, J=8.0, 1.8 Hz, 1H); 13C NMR (75.5 MHz, CDCl3): d=
10.4, 28.6, 52.4, 118.7, 122.6, 127.7, 129.4, 131.2, 131.9, 136.1 ppm.
Conclusion
We have reported a ligand design specifically adapted to the
Pd0-catalyzed intramolecular C H functionalization of
ꢀ
alkane segments. Bulky electron-poor F-TOTP (L4) was re-
vealed to be the optimal palladium ligand. The novel cata-
lytic system allowed milder reaction conditions and proved
efficient on a range of interesting substrates, with the pro-
duction of olefins adjacent to a quaternary benzylic carbon
atom (dehydrogenation) or of bi- and tricyclic molecules of
potential synthetic use. A general mechanism was suggested
for this reaction based on the structure of products and liter-
ature data. Furthermore, the utility of the method was illus-
trated by the synthesis of the drug verapamil (14) and an an-
alogue, 19. During this synthesis, a particularly mild Ru-cat-
alyzed hydroamidation stepwas also reported. We believe
that this work might have implications in other catalytic
processes and will help extend the use of nontraditional
bond disconnections in multistepsynthesis.
ꢀ
Representative procedure for the intramolecular C H functionalization
Experimental Section
(Table 2, entry 1): A dry resealable Schlenk tube containing a magnetic
rod was charged with bromide 1a (150 mg, 0.59 mmol), palladium acetate
(6.7 mg, 0.030 mmol), F-TOTP (21.3 mg, 0.059 mmol), and potassium car-
bonate (164 mg, 1.19 mmol). The Schlenk tube was twice evacuated and
backfilled with argon, before it was capped with a rubber septum. Dry
DMF (3 mL) was injected under argon, then the septum was replaced by
a screwcapand the mixture was stirred at 100 8C (preheated oil bath) for
45 min. After cooling, the mixture was diluted with diethyl ether and fil-
tered through celite. The organic solution was washed with water (6 mL)
and the aqueous layer was extracted with diethyl ether (6 mL”3). The
combined organic layers were then dried over magnesium sulfate and
evaporated. The residue was purified by flash chromatography (silica gel,
heptanes/ethyl acetate 95:5) to afford 1b as an oil (78 mg, 77%).[7,17]
1H NMR (250 MHz, CDCl3); d=1.70 (t, J=7.5 Hz, 3H), 2.10 (m, 2H),
5.36 (d, J=10.5 Hz, 1H), 5.57 (d, J=17.4 Hz, 1H), 5.97 (dd, J=17.5,
10.5 Hz, 1H), 7.35–7.49 ppm (m, 5H); 13C NMR (75.5 MHz, CDCl3): d=
9.7, 33.1, 51.2, 116.5, 120.6, 126.2, 129.0, 130.1, 137.4, 138.6 ppm.
General: Reagents were commercially available and were used without
further purification unless otherwise stated. All solvents were distilled
from the appropriate drying agents immediately before use. Yields refer
to chromatographically and spectroscopically homogeneous materials.
Merck silica gel 60 (particle size 40–63 mm) was used for flash column
chromatography; 1 and 2 mm SDS silica-gel-coated glass plates (60F254)
were used for preparative TLC with UV light as the visualizing agent.
NMR spectra were recorded on Bruker AC-250, AMX-300, AMX-400,
or AMX-500 instruments at 295 K with tetramethylsilane or residual pro-
tiated solvent used as an internal reference for 1H and 13C spectra. 31P
and 19F NMR spectra were calibrated with H3PO4 and CCl3F as external
references. Attributions were made on the basis of 2D experiments.
Products that had been reported previously were isolated in greater than
95% purity, as determined by 1H NMR spectroscopy and capillary GC.
GC analyses were performed with a Shimadzu QP2010 GCMS apparatus,
with simultaneous double injection on a DB-5ms column lined with a
mass (EI) or an FID detection system.
Representative hydroamidation procedure (Table 4, entry 4):
A dry
Schlenk tube containing a magnetic rod was charged with olefin 1b
(50 mg, 0.29 mmol) and triruthenium dodecacarbonyl (18.7 mg,
0.096 mmol). The schlenk tube was twice evacuated and backfilled with
argon, before it was capped with a rubber septum. Dry DMF (0.6 mL)
and homoveratrylamine (197 mL, 1.17 mmol) were injected under argon.
The Schlenk tube was then purged with carbon monoxide and the mix-
ture was stirred under carbon monoxide at 1208C (preheated oil bath)
for 17 h. After cooling, the solvent was evaporated under reduced pres-
Tris(5-fluoro-2-methylphenyl)phosphane (F-TOTP, L4): 2-Bromo-4-fluo-
rotoluene (2.25 mL, 18.1 mmol) was added dropwise to a stirred suspen-
sion of crushed Mg turnings (0.48 g, 19.7 mmol) in THF (10 mL) under
argon at 208C (water bath). The mixture was heated to 758C for 2 h then
cooled to 258C. A solution of PCl3 (500 mL, 5.72 mmol) in THF (5 mL)
was added with a syringe pump over 30 min, then the mixture was re-
fluxed for 1 h. Once cooled to 08C, the reaction was quenched with a sa-
798
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 792 – 799

3
ꢀ
Palladium-Catalyzed C(sp ) H Functionalization
A
FULL PAPER
sure. The residue was purified by flash chromatography (silica gel, hep-
tanes/ethyl acetate 1:4) to afford 18a as a colorless oil (74 mg, 67%).
1H NMR (500 MHz, CDCl3): d=0.90 (t, J=7.4 Hz, 3H), 1.87–2.08 (m,
3H), 2.22–2.37 (m, 3H), 2.69 (t, J=7.4 Hz, 2H), 3.34–3.49 (m, 2H), 3.84
(s, 3H), 3,85 (s, 3H), 5.42 (brs, 1H), 6.67 (s, 1H), 6.68 (d, J=8.4 Hz,
1H), 6.79 (d, J=8.4 Hz, 1H), 7.26–7.29 (m, 1H), 7.30–7.32 ppm (m, 4H);
13C NMR (75.5 MHz, CDCl3): d=9.6, 32.4, 34.5, 35.7, 35.1, 40.7, 48.6,
55.8, 55.9, 111.4, 111.8, 120.6, 121.9, 126.0, 127.9, 129.0, 131.2, 137.3,
147.7, 149.1, 171.2 ppm; HRMS (ESI): m/z: calcd for C23H28N2O3Na:
403.1998 [M+Na+]; found: 403.2002; IR (film): n=3302, 2232, 1645,
[10] Two related intramolecular processes appeared recently: a) C.-G.
Dong, Q.-S. Hu, Angew. Chem. 2006, 118, 2347–2350; Angew.
Chem. Int. Ed. 2006, 45, 2289–2292; b) H. Ren, P. Knochel, Angew.
Chem. 2006, 118, 3541–3544; Angew. Chem. Int. Ed. 2006, 45, 3462–
3465.
[11] a) W. A. Herrmann, C. Brossmer, K. Öfele, C.-P. Reisinger, T. Prier-
meier, M. Beller, H. Fischer, Angew. Chem. 1995, 107, 1989–1992;
Angew. Chem. Int. Ed. Engl. 1995, 34, 1844–1848; b) W. A. Herr-
mann, V. P. W. Bçhm, C.-P. Reisinger, J. Organomet. Chem. 1999,
576, 23–41.
1513 cmꢀ1
.
[12] C. B. Ziegler, Jr., R. F. Heck, J. Org. Chem. 1978, 43, 2941–2946.
[13] K. G. Moloy, J. L. Petersen, J. Am. Chem. Soc. 1995, 117, 7696–
7710, and references therein.
[14] D. W. Allen, B. F. Taylor, J. Chem. Soc. Dalton Trans. 1982, 51–54.
[15] CCDC-291540 (complex I) contains the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif/
[16] Recent syntheses: a) T. Lomberget, E. Bentz, D. Bouyssi, G. Balme,
Org. Lett. 2003, 5, 2055–2057; b) M. Banerjee, R. Mukhopadhyay,
B. Achari, A. K. Banerjee, Org. Lett. 2003, 5, 3931–3933; c) E. Fil-
lion, D. Fischlock, J. Am. Chem. Soc. 2005, 127, 13144–13145;
d) M. Banerjee, R. Mukhopadhyay, B. Achari, A. K. Banerjee, J.
Org. Chem. 2006, 71, 2787–2796; e) L. Planas, M. Mogi, H. Takita,
T. Kajimoto, M. Nobe, J. Org. Chem. 2006, 71, 2896–2898.
[17] For the direct olefination of phenylacetonitriles with alkynes, to give
generally Z/E olefin mixtures, see: a) M. Makosza, Tetrahedron Lett.
7, 1966, 5489–5492; b) M. Makosza, J. Czyzewski, M. Jawdosiuk,
Org. Synth. 1976, 55, 99–102; c) C. Koradin, A. Rodriguez, P. Kno-
chel, Synlett 2000, 1452–1454.
[18] a) A. J. Mota, A. Dedieu, C. Bour, J. Suffert, J. Am. Chem. Soc.
2005, 127, 7171–7182; b) D. L. Davies, S. M. A. Donald, S. A. Mac-
gregor, J. Am. Chem. Soc. 2005, 127, 13754–13755; c) D. Garcꢁa-
Cuadrado, A. A. C. Braga, F. Maseras, A. M. Echavarren, J. Am.
Chem. Soc. 2006, 128, 1066–1067; d) I. Alonso, M. Alcamꢁ, P. Mau-
león, J. C. Carretero, Chem. Eur. J. 2006, 12, 4576–4583.
[19] Recent publications featuring anti b-H elimination: a) M. Ikeda,
S. A. A. El Bialy, T. Yakura, Heterocycles 1999, 51, 1957–1970; b) I.
Schwarz, M. Braun, Chem. Eur. J. 1999, 5, 2300–2305; c) K. Maeda,
E. J. Farrington, E. Galardon, B. D. John, J. M. Brown, Adv. Synth.
Catal. 2002, 344, 104–109; d) M. Lautens, Y.-Q. Fang, Org. Lett.
2003, 5, 3679–3682.
Acknowledgements
This work was financially supported by ICSN-CNRS. We thank Prof. M.
Gómez for stimulating discussions.
[1] Recent reviews: a) G. Dyker, Angew. Chem. 1999, 111, 1808–1822;
Angew. Chem. Int. Ed. 1999, 38, 1698–1712; b) Y. Guari, S. Sabo-
Etienne, B. Chaudret, Eur. J. Inorg. Chem. 1999, 1047–1055; c) C.
Jia, T. Kitamura, Y. Fujiwara, Acc. Chem. Res. 2001, 34, 633–639;
d) V. Ritleng, C. Sirlin, M. Pfeffer, Chem. Rev. 2002, 102, 1731–
1769; e) F. Kakiuchi, N. Chatani, Adv. Synth. Catal. 2003, 345, 1077–
1101; f) S. Ma, Z. Gu, Angew. Chem. 2005, 117, 7680–7685; Angew.
Chem. Int. Ed. 2005, 44, 7512–7517; g) A. R. Dick, M. S. Sanford,
Tetrahedron 2006, 62, 2439–2463; h) K. Godula, D. Sames, Science
2006, 312, 67–72.
[2] a) J. Hassan, M. SØvignon, C. Gozzi, E. Schultz, M. Lemaire, Chem.
Rev. 2002, 102, 1359–1469; b) F. Kakiuchi, S. Murai, Acc. Chem.
Res. 2002, 35, 826–834; c) L. A. Goj, T. B. Gunnoe, Curr. Org.
Chem. 2005, 9, 671–685; d) L.-C. Campeau, K. Fagnou, Chem.
Commun. 2006, 1253–1264.
[3] a) H. M. L. Davies, R. E. J. Beckwith, Chem. Rev. 2003, 103, 2861–
2903; b) P. Müller, C. Fruit, Chem. Rev. 2003, 103, 2905–2919;
c) H. M. L. Davies, Ø. Loe, Synthesis 2004, 2595–2608; d) Modern
Rhodium-Catalyzed Organic Reactions (Ed.: P. A. Evans), Wiley-
VCH, Weinheim, 2005.
[4] Recent examples: a) Z. Li, C.-J. Li, J. Am. Chem. Soc. 2004, 126,
11810–11811; b) J. D. Lawrence, M. Takahashi, C. Bae, J. F. Hart-
wig, J. Am. Chem. Soc. 2004, 126, 15334–15335; c) Z. Li, C.-J. Li, J.
Am. Chem. Soc. 2005, 127, 3672–3673; d) M. S. Chen, N. Prabagar-
an, N. A. Labenz, M. C. White, J. Am. Chem. Soc. 2005, 127, 6970–
6971.
[20] In these cases, the formation of a seven-membered palladacycle can
also be considered.
[21] a) L. M. Prisant, Heart Dis. 2001, 3, 55–62; b) R. N. Brogden, P.
Benfield, Drugs 1996, 51, 792–819.
[22] Recent syntheses of verapamil: a) A. H. Mermerian, G. C. Fu,
Angew. Chem. 2005, 117, 971–974; Angew. Chem. Int. Ed. 2005, 44,
949–952; b) L. Wu, J. F. Hartwig, J. Am. Chem. Soc. 2005, 127,
15824–15832.
[23] Y. Tsuji, T. Ohsumi, T. Kondo, Y. Watanabe, J. Organomet. Chem.
1986, 309, 333–344.
[5] Recent examples: a) L. V. Desai, K. L. Hull, M. S. Sanford, J. Am.
Chem. Soc. 2004, 126, 9542–9543; b) V. G. Zaitsev, D. Shabashov, O.
Daugulis, J. Am. Chem. Soc. 2005, 127, 13154–13155; c) R. Giri, X.
Chen, J.-Q. Yu, Angew. Chem. 2005, 117, 2150–2153; Angew. Chem.
Int. Ed. 2005, 44, 2112–2115.
[6] a) J. A. Johnson, D. Sames, J. Am. Chem. Soc. 2000, 122, 6321–6322;
b) J. A. Johnson, N. Li, D. Sames, J. Am. Chem. Soc. 2002, 124,
6900–6903; c) B. D. Dangel, K. Godula, S. W. Youn, B. Sezen, D.
Sames, J. Am. Chem. Soc. 2002, 124, 11856–11857.
[7] O. Baudoin, A. Herrbach, F. GuØritte, Angew. Chem. 2003, 115,
5914–5918; Angew. Chem. Int. Ed. 2003, 42, 5736–5740.
[8] a) G. Dyker, Angew. Chem. 1992, 104, 1079–1081; Angew. Chem.
Int. Ed. Engl. 1992, 31, 1023–1025; b) G. Dyker, J. Org. Chem. 1993,
58, 6426–6428; c) G. Dyker, Chem. Ber. 1994, 127, 739–742; d) G.
Dyker, Angew. Chem. 1994, 106, 117–119; Angew. Chem. Int. Ed.
Engl. 1994, 33, 103–105.
[24] For related Ru-catalyzed hydroamidation of olefins with forma-
mides, see: a) Y. Tsuji, S. Yoshii, T. Ohsumi, T. Kondo, Y. Watanabe,
J. Organomet. Chem. 1987, 331, 379–385; b) T. Kondo, T. Okada, T.
Mitsudo, Organometallics 1999, 18, 4123–4127; c) S. Ko, H. Han, S.
Chang, Org. Lett. 2003, 5, 2687–2690.
[25] R. M. Bannister, M. H. Brookes, G. R. Evans, R. B. Katz, N. D. Tyr-
rell, Org. Process Res. Dev. 2000, 4, 467–472.
[26] C. Pascal, J. Dubois, D. GuØnard, F. GuØritte, J. Org. Chem. 1998,
63, 6414–6420.
[9] T. E. Barder, S. D. Walker, J. R. Martinelli, S. L. Buchwald, J. Am.
Chem. Soc. 2005, 127, 4685–4696.
Received: June 9, 2006
Published online: October 17, 2006
Chem. Eur. J. 2007, 13, 792 – 799
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
799