Influence of the formation of the halogen bond ArX-N on the mechanism of diketonate ligated copper-catalyzed amination of aromatic halides

Article abstract of DOI:10.1021/om200952v

The mechanism of cross-coupling reactions of aryl halides PhX (X = I, Br, Cl) with cyclohexylamine catalyzed by CuI associated with a diketonate ligand (ket = 2-acetylcyclohexanoate) has been investigated via DFT. Phenyl halides which can be involved in h

Full text of DOI:10.1021/om200952v

Article
pubs.acs.org/Organometallics
Influence of the Formation of the Halogen Bond ArX- - -N on the
Mechanism of Diketonate Ligated Copper-Catalyzed Amination of
Aromatic Halides
Guillaume Lefevre,† Gregory Franc,† Carlo Adamo,‡ Anny Jutand,*,† and Ilaria Ciofini*,‡
̀
́
†Dep
France
́ ́
artement de Chimie, Ecole Normale Superieure, UMR CNRS-ENS-UPMC 8640, 24 rue Lhomond, F-75231 Paris Cedex 5,
‡Chimie ParisTech, ENSCP, UMR 7575, 11 rue P. et M. Curie, F-75231 Paris Cedex 5, France
S
* Supporting Information
ABSTRACT: The mechanism of cross-coupling reactions of
aryl halides PhX (X = I, Br, Cl) with cyclohexylamine catalyzed
by CuI associated with a diketonate ligand (ket = 2-acetyl-
cyclohexanoate) has been investigated via DFT. Phenyl halides
which can be involved in halogen-bond formation (X = I, Br)
with the anionic nitrogen of the intermediate [(ket)Cu-NHCy]− can undergo intramolecular oxidative addition, which is in
competition with the classical oxidative addition three-centered pathway, whereas those which cannot (X = Cl) undergo a
classical oxidative addition, leading to the common complex (ket)CuIIIPh(NHCy). The latter is involved in a faster reductive
elimination, leading to PhNHCy. C6F5I, which involves the strongest C6F5I- - -N halogen bond, generates the anion C6F5− and
consequently not the expected C−N cross-coupling product by substitution at the C−I bond, as evidenced experimentally.
Scheme 1. Mechanism of the Cu-Catalyzed Cross-Coupling
Reaction between PhI and NH2Cy in the Presence of
Cs2CO3 as the Base in DMF (Entry 1 in Table 1)6
INTRODUCTION
Copper complexes catalyze cross-coupling reactions of aryl halides
with amines in the presence of a base and a ligand L (eq 1).1
■
Cucat
Ar−X+H N−R⎯⎯⎯⎯⎯⎯→⎯ Ar−NHR
2
(1)
L, base
Reported mechanistic studies on dative N,N ligands have estab-
lished that the active catalyst is a (N,N)CuI−Nu species (Nu =
amidate, imidate) that reacts with PhI in a mechanism which is
still under debate, involving either an oxidative addition or an
inner- or outer-sphere electron transfer.2 Commercially available
preligands such as 1,3-diketones were found to be highly efficient
for C−N couplings from aryl iodide under mild conditions and
more severe conditions with aryl bromides.3 A few fully theoretical
papers dealing with 1,3-diketonate (ket) ligands recently appeared,
proposing either an outer-sphere single electron transfer
mechanism (SET)4 from an anionic [(ket)CuI−NHMe]− complex
toward PhI or an oxidative addition of a neutral (ket)-
CuI(NH2(CH2)5OH) complex to PhI via a prior PhI/amine
exchange.5
In contrast, in a previous study involving CyNH2 as primary
amine,6 we demonstrated, by a combined theoretical and
experimental approach, that the catalytically active species is the
anionic [(ket)CuI−NHCy]− (A), evidenced by cyclic voltam-
(C) which delivered the expected cross-coupling product after a
reductive elimination which is faster than the oxidative addition
(Scheme 1).6 The present work proposes a detailed analysis
of the mechanism of the reaction involving the active anionic
1
metry (CV) and H NMR, in which the negative charge is
mainly located on the nitrogen, as revealed by DFT cal-
culations. Consequently, the preliminary interaction of this
complex with PhI is a PhI- - -N interaction related to a halogen
bond (Scheme 1, Figure 1).6,7 An intramolecular oxidative addi-
tion then took place affording an aryl−copper(III) intermediary
Received: October 10, 2011
Published: February 1, 2012
© 2012 American Chemical Society
914
dx.doi.org/10.1021/om200952v | Organometallics 2012, 31, 914−920

Organometallics
Article
complex [(ket)Cu−NHCy]− (A) and aryl halides (ArX; X =
I, Br, Cl, F). Namely, the role played by the halide X is
studied in the context of possible formation of ArX- - -N
halogen bonds. It is worthwhile to note that a SET between
A and PhI was ruled out in our previous paper because of the
difference between the reduction potential of PhI and the
oxidation potential of complex A was too high (ΔE = ca. 3 V).6
Since PhBr and PhCl are reduced at more negative potentials
than is PhI, ΔE will be higher (ΔE = > 3 V) and the SET even
less favored.
unexpected products (Table 1, entry 5): the amino product 1,
formed by substitution of the fluorine in a position para to the
iodine, the product 2, with similar substitution but with loss of
the iodine atom, and pentafluorobenzene. This led us to
conduct a reaction in the absence of CuI and ligand (entry 6,
Table 1), which afforded product 1 as a result of a nucleophilic
substitution at the fluorine in a position para to the iodide, in
agreement with a classical SNAr mechanism8 (independently
tested in entry 7 on C6F6). Such reactions showed that, as
expected, the C−I bond could not be activated in a classical
SNAr mechanism, in contrast to C−F bonds. However, in the
presence of CuI and ligand, the formation of 2 and C6F5H,
which both had lost their iodide atoms, suggested an activation
of the C−I bond of C6F5I but without forming the desired C−
N product at the C−I bond. Such results will be rationalized
later on (vide infra).
RESULTS AND DISCUSSION
■
Experimental Copper-Catalyzed Cross-Coupling Re-
actions. From an experimental point of view, aryl halides
(ArX; X = I, Br, Cl, F) are known to show extremely different
reactivity in Cu-catalyzed C−N couplings, from the absence of
reaction observed in the case of aryl fluoride derivatives to good
yields observed for aryl iodides.3,4 The cross-coupling reactions
between ArI (Ar = Ph, p-F-C6H4−) and cyclohexylamine in the
presence of Cs2CO3 and a catalytic amount of CuI associated to
the diketone preligand ketH (2-acetylcyclohexanone) reported
in our previous work6 have been extended to PhX (X = Br, Cl)
(Table 1) and delivered the reactivity order PhI > PhBr ≫
Cross-Coupling Reactions via DFT Calculations. The
key point allowing an explanation of the difference of reactivity
might be that the oxidative addition appears to be promoted by
the formation of a stable complex (type BI, in Figure 1c)
a
Table 1. Amination of Aryl Halides
Figure 1. (a) Schematic structure and labeling scheme of the anionic
complex [(ket)Cu-NHCy]− (A). (b) Computed electrostatic surface
potential for A: the red color indicates negative charges. (c) Formation
of complex BI by halogen bonding.
exhibiting a sizable halogen bond between the halogen atom of
ArX and the negative N of [(ket)Cu−NHCy]− (complex A in
Figure 1a,b), as reported by us for PhI (Figure 1c).6
a
DMF 1 mL. Legend: (i) Product yield; (ii) Recovered ArX. Yields
determined after workup by 1H or 19F NMR of the crude mixture (see
the Supporting Information).
Density functional theory (DFT) was applied to determine the
structural and energetic features of the catalytic cycle involving
PhI, PhBr, PhCl, and C6F5I. This method, in conjunction with
hybrid exchange-correlation functionals, has become one of the
major theoretical tools of quantum chemistry, notably thanks to
its reliability, its low computational cost, and its range of applica-
tions. For these reasons, such an approach has been successfully
applied to the study of a large number of catalytic reactions for
more than a decade.9 More specifically, we made use of the
PhCl (entries 1, 3, and 4). The latter has also been established
with related ligands.2,3
In addition, similar catalytic tests were conducted with C6F5I,
expected to be more reactive than PhI due to the electron-
acceptor fluorines. Surprisingly, the desired cross-coupling pro-
duct was never obtained by substitution at the C−I bond.
Instead, slow and limited conversion of C6F5I led to three
915
dx.doi.org/10.1021/om200952v | Organometallics 2012, 31, 914−920

Organometallics
Article
Table 2. Computed NPA Charges for the Different Complexes and TS Involved in the Cross-Coupling Reaction
BX
TS1
A
PhI
PhBr
C6F5I
PhI
PhBr
PhCl
C6F5I
Cu
N
0.51
−1.15
−0.72
−0.74
0.55
−1.09
−0.77
−0.71
−0.28
C
0.53
−1.11
−0.76
−0.71
−0.17
0.57
−1.05
−0.77
−0.71
−0.40
0.79
−0.97
−0.72
−0.70
−0.18
0.79
−0.96
−0.72
−0.69
−0.14
0.66
−1.06
−0.72
−0.70
−0.12
0.77
−0.91
−0.72
−0.69
−0.37
O1
O2
C1
TS2
D
Ph
C6F5
Ph
C6F5
Ph
C6F5
Cu
0.87
−0.77
−0.68
−0.67
−0.15
0.87
−0.73
−0.65
−0.63
−0.30
0.86
−0.76
−0.71
−0.69
−0.11
0.86
−0.70
−0.70
−0.68
−0.24
0.65
0.65
−0.73
−0.77
−0.73
0.04
N
−0.74
−0.77
−0.73
0.14
O1
O2
C1
hybrid exchange correlation functional PBE0,10 which has been
proven to yield accurate and reliable thermochemistry data even
in the case of reactions involving transition-metal ions.11 This
hybrid functional mixes 25% of Hartree−Fock exchange to the
gradient corrected PBE exchange and correlation functional.12
All calculations were carried out using Gaussian09.13 A 6-
31+G(d) basis14 was used for all atoms (C, H, O, N, F, Cl, Br,
and I) except for the metal (Cu), which was treated using a
Stuttgart pseudopotential (SDD)15 and associated basis set.16
This pseudopotential has been proven to slightly overestimate
interaction energies, but general trends discussed here should be
conserved. All structures corresponding to stationary points were
obtained by fully optimizing in the absence of geometrical
constraints, and they were characterized by subsequent frequency
calculations as minima or first-order transition states (i.e., only
one imaginary frequency). Solvent effects (here DMF) were in-
cluded during both structural optimizations and frequency
calculations using an implicit solvation model based on the polar-
izable continuum model (PCM) of Tomasi and co-workers17 as
implemented in the Gaussian code.18 This method has been
proven to provide an accurate description of bulk solvent effects
on energy profiles, and it is actually one of the best compromises
between efficiency and accuracy to simulate solvent in
conjunction with a quantum chemical description of the reac-
tion.19 Charge analysis was performed using the NBO20 partition
scheme. Reaction energies were corrected for basis set super-
The structural and electronic features of the anionic complex
A, in the absence of the substrate, were first analyzed (Table 2).
As expected for a Cu(I) complex, the copper atom shows an in-
plane coordination of the three donor atoms (N, O1, and O2)
of the ligands with a small deviation from perfect planarity
(d(OOCuN) = 13.6°; Table S1 in the Supporting Informa-
tion). Of note, a bite angle of 83.8° was computed for the
diketonate ligand. This bite angle is comparable to what was
found in the case of the corresponding bidentate phenanthro-
line ligands (ca. 82° in CuI(dimethylphen)2-type complexes).24
Consequently, the copper atom is quite accessible to a substrate
(e.g., ArX) so that an oxidative addition to the metal atom is
fully compatible with its steric hindrance.
In order to define possible sites for the interaction between
aryl halides and complex A, both the electrostatic potential
surface of A (Figure 1b) and the atomic charges computed
using the NPA (Table 2) partition were analyzed. For the active
complex A, analyses pointed out that the most electron-rich
site was the nitrogen atom, whose NPA charge is computed to
be −1.15e, while the copper atom displayed a net positive
charge of (+0.51e) (Table 2). On the other hand, the halogen
atom of the aromatic halides showed a more marked
dependence of the halide atom. Indeed, while fluoride and
chloride derivatives respectively showed a negative (−0.347e)
and practically null (+0.006e) charge on the halide atom, both
bromide and iodide derivatives showed a net positive charge on
it, of +0.059e and +0.141e, respectively (Table 2). On the basis
of this charge analysis, two different scenarios could be
envisaged for the reaction of A with ArX to take place. Either
complex A gave rise to a standard oxidative addition, via a
three-center TS, or complex A could act as a nucleophile,
activating the ArX via the formation of a halogen−amino bond,
reminiscent of what occurs in metal−iodine exchange
reactions.7 Clearly depending on the substrate ArX, one or
the other reaction channel would be preferred. The possibility
of the latter reaction path to take place might be related to the
stability of the halogen−amido complex formed. The stabi-
lization energy computed for the different adducts is reported
in Table 3, together with the computed electron flux from
complex A to ArX. This latter value was computed as the
difference in total NPA charges between ArX in complex BX
and that of the free ArX, a negative value implying an electron
flux from complex A to ArX (nucleophilic behavior).
21a
position error (BSSE) using the counterpoise method.
Selected reaction paths were followed by integrating the intrinsic
reaction coordinate (IRC calculations).21b
Activation Step. On the basis of our previous inves-
6
1
tigations by CV and H NMR, it seems clear that, due to the
presence of a basic medium, the active CuI complex is the
anionic complex [(ket)CuI−NHCy]− (A) generated by
deprotonation by the base Cs2CO3 of the ligated NH2Cy in
(ket)CuI(NH2Cy) (Scheme 1).22 It is difficult to investigate the
equilibrium between the anionic complex A and its acidic form
(ket)CuI(NH2Cy) by DFT-type calculations because, under
experimental conditions, the base Cs2CO3 is used in high
concentration and it is present as a solid residue due to its
rather low solubility. Thus, all the homogeneous equilibria in
which this basis is involved are modified. In particular, it is
likely that the deprotonation process occurs at the interface
between the solid and the solution, making its modeling rather
complex.23
First of all, one notes that, in the case of the fluoride and
chloride compounds, no stable complex BF or BCl was found,
916
dx.doi.org/10.1021/om200952v | Organometallics 2012, 31, 914−920

Organometallics
Article
Table 3. Computed Stabilization Energies (in kcal mol−1)
Due to the Formation of Complex BX and Computed
a
Electron Flux
ArX
ΔE
flux (NBO)
PhF
PhCl
PhBr
PhI
0.72
−0.06
−0.15
−0.28
−4.92
−16.65
C6F5I
a
In e; see text for definition.
the interaction between complex A and PhX (X = F, Cl) giving
rise to an activated complex being fully repulsive. Thus, for these
two systems only a classical oxidative addition mechanism had to
be envisaged. On the other hand, the opposite behavior was
found for PhI, for which the stable adduct BI (−4.92 kcal mol−1)
was indeed computed. Less clear is the case of PhBr, for which
the adduct BBr formation was only very slightly endergonic
(+0.72 kcal mol−1), thus being fully possible under experimental
conditions, although not as stabilizing as in the case of PhI.
Introducing electron-withdrawing substituents (i.e., fluorine
atoms) on the PhI ring, when considering the C6F5I molecule,
further stabilized the formation of the adduct B′I, since it mag-
nified the electrophilic character of the iodine atom (computed
NPA charge for the iodine atom in C6F5I: +0.22e versus +0.17e
for PhI).
These theoretical outcomes suggested that PhI and PhBr
could undergo an intramolecular oxidative addition preceded
by the formation of a stable adduct of type BX activating the
Ar−X bond (X = I, Br). In contrast, a classical oxidative
addition for PhCl would most probablywith more difficulty
in comparison with the intramolecular casetake place due to
the absence of C−Cl activation by the nucleophilic N atom of
[(ket)CuI−NHCy]−.
Cross-Coupling Reaction As a Function of the Halide
of ArX. As reported in our previous work, the interaction of
complex A with PhI passed through the formation of a stable
iodine−amino interaction, giving rise to a stable intermediate
(BI), as discussed in the previous section (Figure 1). This latter
species easily evolved via a four-center TS (TS1; Figure 2a) to
give the “distorted” square-planar Cu(III) complex C (Figure 2a),
as in a classical oxidative addition.
In this context it is worth to note that at the TS state both
the N−I and C−I distances are significantly elongated with
respect to the stable adduct (BI). From this TS, the stable
complex C is formed by expulsion of the iodide ion. Once the
Cu(III) intermediate C is formed, it rapidly evolves toward the
product by reductive elimination, the second barrier (ca. 2 kcal
mol−1) being much smaller than the first one (Figure 2a). The
rate-determining step is thus the formation of the first TS (TS1,
four centers) and no accumulation of the Cu(III) intermediate
(C) is thus expected, in agreement with our previous experi-
mental study.6
This mechanism of the oxidative addition has then been
compared to the more classical three-center oxidative addition
when directly performed from A. The latter has been computed
at the same level of theory (Figure 2b). By comparing the two
reaction mechanisms, one notes that both pathways can be
envisaged: the classical three-center path is slightly thermody-
namically favored but begins with a strongly endothermic
π complexation of the copper onto the CC bond bearing the
Figure 2. Reaction paths computed for PhI: (a) intramolecular
oxidative addition from complex BI via halogen bond formation; (b)
classical three-center oxidative addition from A.
iodine atom, whereas the four-center path takes place from a
stabilized halogen-bonded precomplex (BI).25a
In contrast to the evolution of TS1 (from BI, Figure 2a), the
transition state TS′1 (from B″I, Figure 2b) evolved toward an
intermediate complex FI, in which the iodide is still coordinated
to the copper(III) complex, which is thus anionic. After expul-
sion of the I− anion, the Cu(III) complex (C) was formed exo-
thermically (ca. 12 kcal mol−1) (Figure 2b). This latter species
rearranged through a reductive elimination to yield the stable
product of the coupling reaction (D), as in Figure 2a.
As expected, if one substitutes the phenyl ring of PhI with
electron-withdrawing groups (such as by five fluorine atoms),
the stability of the halogen-bonded adduct B′I is enhanced
(−4.92 kcal mol−1 for PhI versus −16.65 kcal mol−1 for C6F5I)
(Figure 3). However, experimentally the cross-coupling did not
work; no amination at the C−I bond was observed (entry 5,
Table 1). To understand the origin of this apparent discrepancy
between the computed and experimental data, the full reaction
pathway for oxidative addition was studied in the case of C6F5I
(Figure 3 and Table S2 (Supporting Information)).
Overall, the theoretical reactivities of PhI and C6F5I are very
similar, and after the formation of the Cu(III) intermediate C
and C′, respectively, the energy barriers for the reductive elimi-
nation are comparable (2.21 kcal mol−1 for PhI versus 4.33 kcal
mol−1 for C6F5I) (compare Figure 2a and 3). Nonetheless, one
important difference can be noted: the formation of the very
stable adduct B′I in the case of C6F5I involves a larger barrier
917
dx.doi.org/10.1021/om200952v | Organometallics 2012, 31, 914−920

Organometallics
Article
C6F5I, favored the formation of the aryl anion and did not lead
to a copper-catalyzed C−N cross-coupling reaction. The three-
center oxidative addition from A (Figure S1) appears quite
unlikely, due to the high stability of the B′I complex.
As far as PhBr is concerned, the formation of the adduct BBr
was barrierless and slightly thermodynamically unfavorable
(+0.72 kcal mol−1) (Figure 5). The overall reaction followed a
Figure 3. Reaction path computed for C6F5I based on an oxidative
addition from complex B′I.
(25.64 kcal mol−1) for the oxidative addition leading to complex
C′. As a consequence, the overall reaction rate and yield
(dominated by this rate-determining step) is expected to
decrease on going from PhI to C6F5I, as experimentally
observed, since no C−N cross-coupling product was formed by
substitution at the C−I bond of C6F5I (Table 1, entry 5). In
other words, the adduct B′I formed with C6F5I seems to be too
stable so that the reaction was blocked at the formation of B′I.
Nevertheless, one observes the formation of C6F5H and com-
pound 2 (Table 1, entry 5). The C−I bond had been cleaved in
both products without formation of the C−N coupling at the
C−I bond, which rules out any oxidative addition. This suggests
the formation of C6F5− with further protonation leading to
C6F5H. The latter species will then undergo a SNAr reaction
with the amine in the presence of the base, ending with the
formation of product 2. The formation of C6F5− is similar to
what would happen in a classical halogen−metal exchange: for
example, in the reaction ArI + BuLi → ArLi + BuI.7 The
evolution of B′I according to this pathway was investigated by
DFT calculations, which did reveal the formation of C6F5−
together with the non-arylated Cu(III) complex E (Figure 4
Figure 5. Reaction path computed for PhBr.
pathway identical with what was observed for PhI, with reaction
barriers associated with the formation of both TS1 and TS2
very close to the values computed for PhI (24 and 2 kcal mol−1
for PhBr with respect to 22 and 2 kcal mol−1 for PhI in
Figure 2a). Due to the similar activation barriers computed for
the three-center and four-center pathways (22.5 kcal mol−1 vs
23.69 kcal mol−1; Figure S2 (Supporting Information) and
Figure 5, respectively) both channels must be envisaged.25b
In the case of PhCl, a complex of type BCl was not formed
(no halogen bond formation). A standard intermolecular
oxidative addition mechanism was thus computed, as depicted
in Figure 6. After a π complexation of the copper center onto
Figure 4. Reaction path computed for C6F5I involving the classical
evolution of a strong halogen bond formation leading to halogen−
metal exchange.
Figure 6. Reaction path computed for PhCl.
and Table S3 (Supporting Information)).26 Such activation
could not of course deliver the desired C−N cross-coupling
product by substitution at the C−I bond of C6F5−I, as
experimentally observed (Table 1, entry 5).
From this study, one concludes that the N- - -I halogen bond
formation in complex B′I, strongly exothermic in the case of
the CC bond bearing the Cl atom (thermodynamically un-
favored by 14.81 kcal mol−1) a three-center (Ph−Cu−Cl) tran-
sition state (TS′1, Figure 6) was computed, which is 26 kcal mol−1
higher in energy with respect to the reactants. This transition
state provided the intermediate anionic complex FCl in which the
halogen is directly coordinated to the copper(III) complex. After
918
dx.doi.org/10.1021/om200952v | Organometallics 2012, 31, 914−920

Organometallics
Article
expulsion of the Cl− anion, the Cu(III) complex C was formed
exothermically (ca. 9 kcal mol−1). This latter species could easily
rearrange through a reductive elimination to yield the stable
product of the coupling reaction D, as also established in the case
of PhI and PhBr (common intermediate complex C).
catalytic cycles, reaction paths computed for C6F5I and PhBr
following the classical oxidative addition path, and computed
imaginary frequencies associated with all TS. This material is
available free of charge via the Internet at http://pubs.acs.org.
The formation of C is highly exothermic, −16 kcal mol−1 for
PhI compared to −9 kcal mol−1 for PhCl, but the overall
reaction barrier computed for PhI is smaller (22.9 kcal mol−1
via complex BI (Figure 2a), 16.7 kcal mol−1 from complex A
(Figure 2b)) than for PhCl (26.04 kcal mol−1, Figure 6). The
larger activation energy of the oxidative addition step computed
for PhCl explains the experimentally observed lower reactivity
of PhCl (Table 1, entry 4).
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: ilaria-ciofini@chimie-paristech.fr (I.C.).
ACKNOWLEDGMENTS
■
The CNRS, ENS, UPMC, ENSCP, and ANR (Grant 07-CP2D-
08, H2OFerCat) are thanked for financial support.
Finally, it should be mentioned that no stable adduct nor
reasonably low in energy TS for oxidative addition was found
for PhF. This finding is in agreement with the results in entries
5−7 in Table 1, which showed that the arylated amine 1 formed
in the presence of copper by substitution of one F in C6F5I
(entry 5) was in fact not catalyzed by copper but was formed
via a SNAr reaction, in agreement with the experiments performed
in the absence of copper (entries 6 and 7). In other words, free
copper SNAr reactions on ArF are faster than copper-catalyzed
cross-coupling at a ArF5−I bond.
REFERENCES
■
(1) For reviews, see: (a) Cristau, H.-J.; Cellier, P. P.; Spindler, J.-F.;
Taillefer, M. Eur. J. Org. Chem. 2004, 695−709. (b) Beletskaya, I. P.;
Cheprakov, A. V. Coord. Chem. Rev. 2004, 248, 2337−2364. (c) Evano,
G.; Blanchard, N.; Toumi, T. Chem. Rev. 2008, 108, 3054−3131.
(d) Monnier, F.; Taillefer, M. Angew. Chem., Int. Ed. 2009, 48, 6954−
6971. (e) Sperotto, E.; van Klink, G. P. M; van Koten, G.; de Vries,
J. G. Dalton Trans. 2010, 39, 10338−10351.
(2) (a) Strieter, E. R.; Blackmond, D. G.; Buchwald, S. L. J. Am.
Chem. Soc. 2005, 127, 4120−4121. (b) Tye, Z.; Weng, J. W.; Johns,
A. M.; Incarvito, A. C. D.; Hartwig, J. F. J. Am. Chem. Soc. 2008, 130,
9971−9983. (c) Srieter, E. R.; Bhayana, B.; Buchwald, S. L. J. Am.
Chem. Soc. 2009, 131, 78−88. (d) Zhang, S. L.; Liu, L.; Fu, Y.; Guo,
Q.-X. Organometallics 2007, 26, 4546−4554.
CONCLUSIONS
■
We have reported herein the results of experimental and theo-
retical studies of diketonate-ligated Cu(I)-catalyzed C−N cross-
coupling between aryl halides and cyclohexylamine, leading to
C−N bond formation. It is established that a halogen-bonded
complex BX (X = I, Br) is formed by complexation of the
halogen atom of PhI and PhBr to the nucleophilic nitrogen of
[(ket)Cu(NHCy)]− (A). These complexes BX can undergo an
intramolecular oxidative addition (four-center pathway), which
is computed to be in competition with the classical three-center
oxidative addition pathway, slightly thermodynamically favored
for PhI but with a similar energy barrier for PhBr. Both path-
ways ultimately lead to the same (ket)Cu(Ph)(NHCy) copper-
(III) complex, which undergoes a faster reductive elimination,
leading to PhNHCy. In contrast, in the case of PhCl only the
classical oxidative addition takes place, due to the lack of halogen
bond formation. The DFT calculations reflect the observed
overall reactivity of aryl halides in the investigated C−N cross-
couplings (ArI > ArBr ≫ ArCl).
(3) (a) Shafir, A.; Buchwald, S. L. J. Am. Chem. Soc. 2006, 128, 8742−
8743. (b) Shafir, A.; Lichtor, P. A.; Buchwald, S. L. J. Am. Chem. Soc.
2007, 129, 3490−3491. (c) Altman, R. A.; Koval, E. D.; Buchwald,
S. L. J. Org. Chem. 2007, 72, 6190−6199. (d) Altman, R. A.; Hyde,
A. M.; Huang, X.; Buchwald, S. L. J. Am. Chem. Soc. 2008, 130, 9613−
9620. (e) Monnier, F.; Taillefer, M. Angew. Chem., Int. Ed. 2008, 47,
3096−3099. (f) Xia, N.; Taillefer, M. Angew. Chem., Int. Ed. 2009, 48,
337−339. (g) Larsson, P.-F.; Correa, A.; Carril, M.; Norrby, P.-O.;
Bolm, C. Angew. Chem., Int. Ed. 2009, 48, 5691−5693.
(4) Jones, G. O.; Liu, P.; Houk, K. N.; Buchwald, S. T. J. Am. Chem.
Soc. 2010, 132, 6205−6213.
(5) Yu, H.; Jiang, Y.; Fu, Y.; Liu, L. J. Am. Chem. Soc. 2010, 132,
18078−18091.
̀
(6) Franc, G.; Cacciuttolo, Q.; Lefevre, G.; Adamo, C.; Ciofini, I.;
Jutand, A. ChemCatChem 2011, 3, 305−309.
(7) For halogen bonds involving aryl iodides see: (a) Reich, H. J.;
Green, P.; Phillips, N. H. J. Am. Chem. Soc. 1989, 111, 3444−3445.
(b) Reich, H. J.; Whipple, W. L. Can. J. Chem. 2005, 83, 1577−1587.
(c) Metrangolo, P.; Neukirch, H.; Pilati, T.; Resnati, G. Acc. Chem. Res.
2005, 38, 386−395.
The lack of reactivity in C−N coupling at the C−I bond of
electron-deficient aryl iodides such as C6F5I has been explained.
Due to a very strong halogen bond in complex B′I, this inter-
mediate complex does not undergo intramolecular oxidative
addition as BI does but instead evolves via the classical formation
of the anion C6F5− and after protonation to SNAr-type products
by substitution at a C−F bond. No C−N product by substitution
at the C−I bond of C6F5I could thus be formed.
Therefore, the formation of stabilizing halogen-bonded
complexes between [(ket)Cu(NHCy)]− and aryl halides can
lead to a four-center oxidative addition pathway which is in
competition with the classical three-center path. Nevertheless, a
strong halogen-bonding complexation can disable both oxida-
tive addition pathways and lead to a reduction process of the
aryl halide.
(8) Battioni, P.; Brigaud, O.; Desvaux, H.; Mansuy, D.; Traylor, T. G.
Tetrahedron Lett. 1991, 32, 2893.
(9) van Santen, R. A., Sautet, P., Eds. Front Matter in Computational
Methods in Catalysis and Materials Science: An Introduction for Scientists
and Engineers; Wiley-VCH: Weinheim, Germany, 2009.
(10) Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110, 6158−6170.
(11) Coquet, R.; Tada, M.; Iwasawa, Y. Phys. Chem. Chem. Phys.
2007, 9, 6040−6046.
(12) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77,
3865−3868.
(13) Frisch, M. J.; et al. Gaussian 09, revision A.02; Gaussian, Inc.,
Wallingford, CT, 2009.
(14) (a) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem.
Phys. 1980, 72, 650−654. (b) Clark, T.; Chandrasekhar, J.; Spitznagel,
G. W.; Schleyer, P. v. R. J. Comput. Chem. 1983, 4, 294−301.
(15) Szentpaly, L. v.; Fuentealba, P.; Preuss, H.; Stoll, H. Chem. Phys.
Lett. 1982, 93, 555.
ASSOCIATED CONTENT
* Supporting Information
Text, figures, and tables giving experimental details, computed
structural parameters of the different species involved in the
■
S
(16) Kuchle, W.; Dolg, M.; Stoll, H.; Preuss, H. J. Chem. Phys. 1994,
100, 7535.
919
dx.doi.org/10.1021/om200952v | Organometallics 2012, 31, 914−920

Organometallics
Article
(17) Miertus, S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117−
129.
(18) Scalmani, G.; Frisch, M. J.; Mennucci, B.; Tomasi, J.; Cammi, R.;
Barone, V. J. Chem. Phys. 2006, 124, 094107.
(19) Mennicci, B.; Cammi, R. Continuum Solvation Models in
Chemical Physics: From Theory to Applications; Wiley: Chichester,
U.K., 2007.
(20) Foster, J. P.; Weinhold, F. J. Am. Chem. Soc. 1980, 102, 7211−
7218.
(21) (a) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 19, 553.
(b) Hratchian, H. P.; Schlegel, H. B. J. Chem. Phys. 2004, 120,
9918−24.
(22) For deprotonation of ligated amine, see: (a) Louie, J.; Hartwig,
J. F. Angew. Chem., Int. Ed. 1996, 35, 2359−2361. (b) Barder, T. E.;
Buchwald, S. T. J. Am. Chem. Soc. 2007, 129, 12003−12010. (c) Franc,
G.; Jutand, A. Dalton Trans. 2010, 39, 7873−7875.
(23) For related problems in Pd-catalyzed reactions see: Meyers, C.;
Maes, B. U. W.; Loones, K. T. J.; Bal, G.; Lemier
R. A. J. Org. Chem. 2004, 69, 6010−6017.
̀
e, G. L. F.; Dommisse,
(24) Lavie-Cambot, A.; Cantunel, M.; Leydet, Y.; Jonusauskas, G.;
Bassani, D. M.; McClenaghan, N. D. Coord. Chem. Rev. 2008, 252,
2572−2584.
(25) (a) The existence of a possible direct equilibrium between the
halogen-bonded complexes BI and B″I was also investigated. No
transition state was found. Indeed, in complex BI, the geometric
constraint is high between both reactants and the angle θ(NIC) is
close to 180° (Table S1, Supporting Information). The rotation of the
phenyl iodide from BI to its position in complex B″I cannot take place,
due to the repulsion of the lone pairs of the nitrogen atom and those
of the iodide atom. As soon as the NIC angle was set inferior to 150°,
the separated PhI and [(ket)Cu-NHCy]− were formed. Consequently,
there is no crossing between the two depicted mechanisms for
the oxidative addition. (b) Similar computations25a for BBr and B″Br
revealed that there is no crossing between the two mechanisms also in
this case.
(26) The Cu(I) has been oxidized to Cu(III), similar to what
happens in a classical halogen−metal exchange ArI + BuLi → ArLi +
BuI, in which the anionic Bu in BuLi has been oxidized to form BuI
while ArI has been reduced to Ar− in ArLi, except that in the present
case the halogen bond was not a C- - -I bond but a N- - -I bond.
920
dx.doi.org/10.1021/om200952v | Organometallics 2012, 31, 914−920