Discriminating role of bases in diketonate copper(i)-catalyzed C-O couplings: Phenol versus diarylether

Article abstract of DOI:10.1039/c3dt32948h

The mechanism of the formation of phenol from PhI and CsOH catalysed by copper(i) ligated to the 1,3-diketonate ket′^- generated from 2,2,6,6-tetramethyl-3,5-heptanedione (TMHD) has been investigated by DFT calculations associated with experimental techniques: cyclic voltammetry, ¹H NMR, and ESI-MS. Weak halogen bonding between the negatively charged O atom of [(ket′)Cu^I-OH]^- and PhI leads to an oxidative addition that gives (ket′)Cu^III(Ph)-OH. The latter undergoes a faster reductive elimination that delivers (ket′)Cu ^I(PhOH) from which PhOH is released. PhOPh is formed in the presence of an extra base Cs₂CO₃. The two catalytic cycles of formation of PhOH or PhOPh are branched at the level of (ket′)Cu ^I(PhOH) that can either afford PhOH in the presence of CsOH or be deprotonated by Cs₂CO₃ to generate [(ket′)Cu ^I-OPh]^-. The oxidative addition of [(ket′)Cu ^I-OPh]^- to PhI leads to (ket′)Cu^III(Ph)- OPh involved in a faster reductive elimination that delivers PhOPh and the Cu^I catalyst.

Full text of DOI:10.1039/c3dt32948h

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Discriminating role of bases in diketonate copper(I)-
catalyzed C–O couplings: phenol versus diarylether†
Cite this: Dalton Trans., 2013, 42, 5348
Guillaume Lefèvre,a Anis Tlili,b Marc Taillefer,*b Carlo Adamo,c Ilaria Ciofini*c and
Anny Jutand*a
The mechanism of the formation of phenol from PhI and CsOH catalysed by copper(I) ligated to the 1,3-
diketonate ket’− generated from 2,2,6,6-tetramethyl-3,5-heptanedione (TMHD) has been investigated by
DFT calculations associated with experimental techniques: cyclic voltammetry, 1H NMR, and ESI-MS. Weak
halogen bonding between the negatively charged O atom of [(ket’)CuI–OH]− and PhI leads to an oxi-
dative addition that gives (ket’)CuIII(Ph)–OH. The latter undergoes a faster reductive elimination that deli-
vers (ket’)CuI(PhOH) from which PhOH is released. PhOPh is formed in the presence of an extra base
Cs2CO3. The two catalytic cycles of formation of PhOH or PhOPh are branched at the level of (ket’)
CuI(PhOH) that can either afford PhOH in the presence of CsOH or be deprotonated by Cs2CO3 to gener-
ate [(ket’)CuI–OPh]−. The oxidative addition of [(ket’)CuI–OPh]− to PhI leads to (ket’)CuIII(Ph)–OPh
involved in a faster reductive elimination that delivers PhOPh and the CuI catalyst.
Received 8th December 2012,
Accepted 18th January 2013
DOI: 10.1039/c3dt32948h
www.rsc.org/dalton
Introduction
Ullmann condensations, arylation of nucleophiles by aryl
halides in the presence of copper, are key reactions leading to
the formation of C(aryl)–C and C(aryl)–heteroatom bonds,
allowing the synthesis of intermediates and synthetic targets
of interest in the field of life sciences and materials. Until
2000, the reactions were not used to their full potential,
because they suffered from limited synthetic scope due to
harsh conditions. A renaissance occurred in 2001, after break-
throughs by two groups who described very efficient copper–
ligand systems for C–C, C–N and C–O coupling reactions.1
These systems and those developed afterwards by other groups
enable the use of catalytic amounts of copper under mild
temperature conditions, thanks to oxygen- and/or nitrogen-
ligands.2 Neutral N,N3 or anionic 1,3-diketonate ligands4 have
shown their efficiency for C–N bond formation. We recently
established experimentally and theoretically that a key step of
the 1,3-diketonate (ket)/Cu-catalyzed C–N coupling is the for-
mation of an anionic complex [(ket)CuI–NHR]− which reacts
with PhI via halogen bonding with the negatively charged N
Scheme 1 Mechanism of the C–N coupling when the copper(I) catalyst is
ligated by the 1,3-diketonate ligand ket− (base = Cs2CO3).5
atom of [(ket)CuI–NHR]− (Scheme 1).5 This intermediate
complex is then involved in an intramolecular oxidative
addition followed by
(Scheme 1).5,6
a
faster reductive elimination
Anionic 1,3-diketonate ligands (such as ket− in Scheme 1)
also promote copper-catalyzed C–O bond formation from ArI
and ROH (R = alkyl, aryl) to generate Ar–OR.4i,7
aEcole Normale Supérieure, Département de Chimie, UMR CNRS-ENS-UPMC 8640,
24 rue Lhomond, F-75231 Paris Cedex 5, France. E-mail: Anny.Jutand@ens.fr
bInstitut Charles Gerhardt Montpellier, AM2N, ENSCM, 8 rue de l’Ecole Normale,
F-34296 Montpellier Cedex 5, France
CuI associated with the pre-ligand ket′H (2,2,6,6-tetra-
methyl-3,5-heptanedione, TMHD) selectively catalyzed the
cross-coupling of aryl iodides or bromides with CsOH leading
to phenols7a,b (eqn (1) in Scheme 2) or to diarylethers when
the reaction was performed in the presence of an additional
base Cs2CO3 (eqn (2), Scheme 2).8
cChimie ParisTech, ENSCP, UMR 7575, 11 Rue Pierre et Marie Curie, F-75005 Paris,
France
†Electronic supplementary information (ESI) available: 1H NMR, ESI/MS, DFT
calculations. See DOI: 10.1039/c3dt32948h
5348 | Dalton Trans., 2013, 42, 5348–5354
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Scheme 2 Formation of PhOH (eqn (1)) or PhOPh (eqn (2)) from PhI catalysed
by copper(I) ligated by the 1,3-diketonate ligand ket’− (TMHD).
The mechanism of formation of PhOPh from PhI and
PhOH catalyzed by CuI ligated to 1,10-phenanthroline (phen)
is reported to proceed via oxidative addition of PhI to (phen)
CuI–OPh.9 In contrast, a single electron transfer (SET) between
PhI and the model complex [(ket)CuI–OMe]− (see ket in
Scheme 1) is proposed from DFT calculations for the for-
mation of PhOMe from PhI and MeOH.4i To our knowledge,
the mechanism of the formation of PhOH and PhOPh from
PhI, hydroxide and a base catalyzed by CuI ligated to the di-
ketonate ligand ket′− generated from TMHD (Scheme 2) has
never been elucidated. Preliminary results by our group have
been included in a review on the mechanism of copper-cata-
lyzed reactions.10 We report herein full experimental and DFT
analyses on the mechanism of formation of PhOH and on the
specific role of the base Cs2CO3 that is required for the for-
mation of PhOPh.
Results and discussion
Mechanism of the CuI-catalysed formation of phenol from
PhI and CsOH (eqn (1), Scheme 2)
Since DFT calculations have been used to establish the exist-
ence of a halogen bond between PhI and the negatively
charged N atom of [(ket)CuI–NHCy]− as a template for oxi-
dative addition (Scheme 1), similar DFT calculations11 have
been performed on the model complex [(ket′)CuI–OH]− (A)
(Fig. 1a–c). The electrostatic potential surface of A and the
atomic charges were computed using the NPA partition
(Fig. 1b).11 The most electron rich site was the oxygen atom of
the ligated OH whose NPA charge was computed to be −1.26
|e−| (Fig. 1b) while the copper atom displayed a net positive
charge of +0.59 |e−| (Fig. 1b). On the other hand, phenyl
iodide showed a positive charge on the iodine atom of +0.14
|e−|.5b The computed interaction between A and PhI revealed
the formation of complex B (slightly stabilized by −1.37 kcal
mol−1) involving a halogen bond12 between the negatively
charged oxygen atom of A and the partly positively charged I
Fig. 1 (a) [(ket’)Cu–OH]− (A): labelling and computed molecule. (b) Computed
electrostatic surface potential of A with the red colour for negative charge. (c)
Second HOMO (−5.42 eV) of A. (d) LUMO (−0.92 eV) of PhI, exhibiting the mole-
cular plane as a symmetry element. (e) Ball-and-stick computed structure for
complex B. (f) Occupied molecular orbital of complex B exhibiting a bonding
density between the O and I atoms (−8.49 eV).
atom of PhI (Scheme 3, Fig. 1e). The second HOMO of [(ket′)- atoms (Fig. 1f). Consequently the C–I bond in B (2.14 Å) is
CuI–OH]− (−5.42 eV) and the LUMO of PhI (−0.92 eV) involved longer than that in PhI (2.12 Å). It is reminiscent of what was
in that halogen bond are displayed in Fig. 1c and d respect- computed for the complex [(ket)Cu–NHCy]− (Scheme 1) but
ively. The molecular orbital of B exhibiting the O–I bonding with a lower stability (−1.37 kcal mol−1 versus −4.92 kcal
density also bore an antibonding density between the C and I mol−1).5 Interestingly, the C–I bond ligated via the iodine atom
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 5348–5354 | 5349

View Article Online
Paper
Dalton Transactions
Scheme 3 Computed pathway for the CuI-catalyzed formation of PhOH.
was not in the local plane defined by the three atoms Cu, H
and O (of OH), but the bonding O–I interaction developed in
an orthogonal direction (Fig. 1e and f). It is consistent with
the fact that the molecular orbital of A involved in this halogen
bonding (Fig. 1c) is anti-symmetric with respect to this plane
and can only develop bonding interaction with a σ-type orbital
such as the LUMO of PhI (Fig. 1d) orthogonally to this plane.
At this stage, one can envisage either an oxidative addition
of PhI from complex B or a classical oxidative addition from
complex A. Both pathways have been studied computationally
on the singlet potential energy surface for all CuI and CuIII
complexes (calculated to be the more stable spin state).
Fig. 2 (a) HOMO (−4.88 eV) of A. (b) Ball-and-stick computed structure for
complex B’’. (c) Labelling scheme and calculated energy for B’’.
Release of the iodide ion from the anionic CuIII complex F
delivered complex C (ket′)CuIII(Ph)–OH and after reductive
The values of formation energies ΔE (including BSSE and elimination complex D (ket′)CuI(PhOH) in which (ket′)CuI was
thermal corrections) computed for the different copper com- ligated by the phenol (Scheme 3). The transition state of the
plexes are given in Scheme 3 and structural data in Table S1.† reductive elimination TS2 (4.37 kcal mol−1, Scheme 3) required
The two pathways are competitive and lead to the same inter- lower activation energy than that of the overall oxidative
mediate E on the way to F, the anionic Ph–CuIII complex first addition via TS1 (14.69 kcal mol−1), making the reductive elimi-
formed in the oxidative addition (Scheme 3, see the computed nation faster than the oxidative addition which became rate-
3-center TS1). Nevertheless, the first stabilizing interaction is determining.
the formation of the halogen-bonded stabilized complex B. In
In the course of the DFT calculations, another less ener-
was found
contrast to the stronger N–I bonding (−4.92 kcal mol−1) calcu- getic stabilizing interaction (0.2 kcal mol−1
)
lated for the interaction of [(ket)Cu–NHCy]− with PhI between the Cu centre of complex A and the iodide of PhI
(Scheme 1), which induced a 4-center transition state for an (Fig. 2) leading to complex B′′ with a C–I bond of 3.18 Å
intramolecular oxidative addition without cleavage of the N–I (Fig. 2b,c and Table S4†) in a square planar geometry (Fig. 2b).
bond,5 the weaker O–I bond calculated in complex
B
The molecular orbital of A involved in this halogen bond is
(−1.37 kcal mol−1) resulted in an easier O–I bond cleavage to displayed in Fig. 2a and that of PhI in Fig. 1d. The weakness of
give the intermediate complex E followed by a 3-center tran- the Cu–I bond is due to the fact that the HOMO of A (Fig. 2a)
sition state (Scheme 3). A low energy barrier for the oxidative is very diffuse and also interacts with the lone pairs of PhI
addition of PhI to [(ket′)Cu–OH]− has been calculated (Fig. S3†), leading to a global destabilization of the system.
(Scheme 3). In contrast, a higher barrier was obtained in Such Cu–I interaction was not found in the case of [(ket)CuI–
related DFT studies for the oxidative addition of PhI to [(ket)- NHCy]− due to the trigonal geometry around the Cu centre
Cu–OMe]− (see the formula of ket in Scheme 1), however calcu- that could not display a tetrahedral distortion to allow coordi-
lated for the less stable triplet spin state of the CuIII complex.4i nation by PhI, in contrast to the pseudo “T” shaped complex A
5350 | Dalton Trans., 2013, 42, 5348–5354
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Table 1 Reduction or oxidation peak potentials of ligand, copper complexes
and reagents in DMF
EpRed [E0]a (V)
−2.23 (R1)
EpOxa (V)
ket′H
ket′−
+0.51 (O1)
+0.41
+0.94 (O2)
+1.23 (O3)
+1.01 (O4)
CuI
[(ket′)CuI]−
[(ket′)Cu–OH]−
[(ket′)Cu–OPh]−
PhI
[−1.91]
PhOH
+1.52 (O5)
+0.30 (O6)
PhO−
a Peak potentials Ep (in Volts) are relative to the reference SCE. They
were determined at a glassy carbon disk electrode (d = 1 mm) at the
scan rate of 0.5 V s−1 at 22 °C. [E0]: standard potential.14a
Fig. 3 Cyclic voltammetry at a glassy carbon disk electrode (d = 1 mm) at the
scan rate of 0.5 V s−1 in DMF containing nBu4NBF4 (0.3 M) as the supporting
electrolyte, at 22 °C. (a) R1: reduction of 2,2,6,6-tetramethyl-3,5-heptanedione,
ket’H (40 mM). (b) O1: oxidation of ket’− formed after addition of CsOH·H2O
(100 mM) to ket’H (40 mM). (c) O2: oxidation of [(ket’)CuI]− formed by addition
of CuI (10 mM) to ket’− (40 mM). (d) Oxidation of [(ket’)Cu–OH]− at O3 formed
after addition of CsOH·H2O (415 mM) to [(ket’)CuI]− generated as in Fig. 3c. (e)
Oxidation of ket’− (O1) and [(ket’)CuI]− (O2) generated in the stoichiometric
mixture [ket’H] = [CsOH] = [CuI] = 20 mM. (f ) After addition of PhOH (150 mM)
to 3e. O5: oxidation of PhOH. (g) Oxidation of [(ket’)Cu–OPh]− at O4 after
addition of Cs2CO3 (150 mM) to 3f. O6: oxidation of PhO−.
Scheme 4 Formation of [(ket’)Cu–OH]−.
generated [(ket′)CuI]− was an equilibrium (Scheme 4, the oxi-
dation peak of free CuI at +0.41 V overlapped with O1). This
explains why an excess of 1,3-diketonate is always required in
C–O (or C–N) couplings4–8 in order to fully convert CuI into
[(ket′)CuI]− and to maintain a high catalyst loading.
(Fig. 1b). However, complex B′′ did not lead to any oxidative
After addition of CsOH·H2O (415 mM) to a solution con-
addition because the π-complex (E in Scheme 3) could not be taining [(ket′)CuI]− generated as in Fig. 3c, a new oxidation
generated.
peak O3 appeared at EpO3 = +1.23 V which was assigned to the
The mechanism established in Scheme 3 on the basis of anionic complex [(ket′)Cu–OH]− (Fig. 3d, vide infra)
DFT calculations was then investigated from an experimental (Scheme 4).13
point of view. The mechanistic studies were performed in DMF
Complexes [(ket′)CuI]− and [(ket′)Cu–OH]− were then
also used in the catalytic reactions8 at room temperature. characterised by 1H NMR performed in DMF-d7 under stoi-
Cyclic voltammetry (CV) was first used to characterise the key chiometric conditions. The preligand ket′H (0.11 M) was
complex [(ket′)Cu–OH]− proposed to be able to react with PhI accompanied by its minor enol form in the ratio 2.8 : 1
(Scheme 3).
(Fig. S1a†). After addition of CsOH·H2O (1.09 equiv.), ket′− was
The 1,3-diketone preligand ket′H (TMHD) (40 mM) was characterised by its vinylic proton at 5.41 ppm and a singlet
characterised by an irreversible reduction peak at EpR1 = −2.23 V for the two tBu groups at 1.07 ppm (Fig. S1b†). The addition of
vs. SCE (Fig. 3a, Table 1). This reduction peak partly dis- CuI (1 equiv.) led to [(ket′)CuI]− characterised by a new singlet
appeared after addition of CsOH·H2O (1 equiv., 40 mM) and for the two tBu groups at 0.994 ppm and a singlet for the
the oxidation peak of the deprotonated ligand ket′− appeared at vinylic proton at 5.56 ppm, together with free ket′− whose
EpO1 = +0.51 V (Table 1). The reaction performed under stoichio- signals were still detected as the minor ones (Fig. S1c†). Inter-
metric conditions was quite slow. The deprotonation was estingly, addition of CsOH·H2O (1.09 equiv.) afforded a new
made to completion in the presence of excess CsOH·H2O complex [(ket′)Cu–OH]− that did not exhibit a unique singlet
(100 mM) (Fig. 3b). After addition of CuI (10 mM) in the ratio for the protons of tBu but two close singlets at 1.22 and
TMHD/Cu = 4, the oxidation peak of ket′− at O1 partly dis- 1.18 ppm (Fig. S1d†), suggesting a dissymmetrisation of the
appeared in a slow reaction and a new oxidation peak O2 devel- ket′− ligand in [(ket′)Cu–OH]− with consequently a broader
oped at EpO2 = +0.94 V (Fig. 3c). Peak O2 was assigned to [(ket′)- singlet for the vinylic proton observed at 5.06 ppm. The com-
CuI]− (Scheme 4, vide infra). When the reaction was performed puted structure of [(ket′)Cu–OH]− revealed that indeed the
under stoichiometric conditions ([ket′H] = [CsOH] = [CuI] = ligand ket′− was no longer symmetrical due to the ligated
20 mM), the CV performed after 1 h 30 min revealed the pres- hydroxo (Fig. 1a). The Cu–O2 distance (2.306 Å) was indeed
ence of the oxidation peak of ket′− at O1 and [(ket′)CuI]− at O2 found to be longer than the Cu–O1 one (1.897 Å) (Table S1†).
(Fig. 3e), indicating that the reaction of ket′− with CuI that The complex [(ket′)Cu–OH]− was then generated in THF to
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 5348–5354 | 5351

View Article Online
Paper
Dalton Transactions
allow its characterization by MS. The ESI(−)/MS spectrum free phenol and [(ket′)Cu–OH]− is proposed for the last step e
(eluent: acetonitrile) exhibited two copper complexes with the of cycle I in Scheme 5 (vide infra).
right isotopic pattern for Cu centres: two main peaks separated
Mechanism of the CuI-catalysed formation of diphenyl ether
from PhI, CsOH and Cs2CO3 (eqn (2) in Scheme 2)
by two mass units (100 and 45% respectively): [(ket′)Cu–OH]−,
THF,MeCN and [(ket′)CuI]−,2MeCN together with the free ket′−
ligand (see ESI†). The detection of [(ket′)CuI]− and free ket′− in In experiments reported by some of us, PhOPh was the major
a reaction performed in the presence of a stoichiometric product formed from PhI (1 mmol) and CsOH (1.5 mmol)
amount of the ligand ket′− confirmed a posteriori the equili- when an extra base such as Cs2CO3 (0.5 mmol) was added,
brium in Scheme 4. Indeed, an excess of ket′− was required to which inhibited the formation of PhOH (Scheme 2).8a Two
push all equilibrium towards the formation of ket′-ligated bases CsOH or Cs2CO3 are formally in competition for the
copper complexes (vide supra, Scheme 4).
deprotonation of phenol in (ket′)Cu(PhOH) (D). Cyclic voltam-
Phenyl iodide (1.8 mmol, 15 equiv.) was then introduced metry was thus performed to discriminate those two pathways.
into the electrochemical cell containing [(ket′)Cu–OH]− PhOH (50 mM) was added to a solution of [(ket′)CuI]− in DMF
formed step by step as explained above from CuI (0.12 mmol, generated under stoichiometric conditions ([ket′H] = [CsOH] =
10 mM), ket′H (0.48 mmol, 40 mM) and CsOH·H2O (1.7 mmol, [CuI] = 20 mM, Fig. 3e). The CV was not significantly modified
140 mM). No significant change occurred. The oxidation peak (Fig. 3f) suggesting that D (whenever formed) was present at
of [(ket′)Cu–OH]− was still observed. The too high potential low concentration. After addition of Cs2CO3 (150 mM), the CV
gap (ΔE = 3.1 V) between the standard reduction potential of exhibited a new oxidation peak O4 at EpO4 = +1.01 V which was
PhI (Table 1) and the oxidation potential of [(ket′)Cu–OH]− at assigned to [(ket′)Cu–OPh]− (A′) (Fig. 3g). The same complex
O3 ruled out any fast outer-sphere electron transfer (SET).14b was indeed formed in the reaction of an authentic sample of
After one night at r.t., PhOH was formed at trace level, in con- PhO− with [(ket′)CuI]− and characterised in CV by peak O4
1
trast to the catalytic reaction which was however performed at (Table 1) and by H NMR (Fig. S2†). The well-resolved signals
130 °C. Nevertheless, the C–O coupling did take place but was of the aromatic protons of the ligated OPh in A′, 7.11 ppm (dd,
not efficient at r.t. suggesting a very slow reaction of PhI with J = 7.5 Hz, 2H, m-H), 6.83 ppm (d, J = 7.5 Hz, 2H, o-H),
[(ket′)Cu–OH]−. Under the same experimental conditions, 6% 6.65 ppm (dd, J = 7.5 Hz, 1H, p-H), were different from those of
of phenol was formed after 14 h at 60 °C and 21% after 24 h at the free PhO− with a higher shift of 0.44 ppm for the o-H
90 °C (yields relative to CsOH used under substoichiometric (Fig. S2b–d†). The two tBu groups in complex [(ket′)Cu–OPh]−
conditions). No PhOPh was detected in those reactions.
(A′) appeared as a broad signal, suggesting a flexibility of the
The DFT calculation in Scheme 3 revealed that the reductive Cu–OPh bond. The calculated Cu–O2 and Cu–O1 distances
elimination from (ket′)CuIII(Ph)–OH was much faster than the (2.18 Å and 1.92 Å respectively, Table S2†) were indeed found
oxidative addition; this is why (ket′)CuIII(Ph)–OH which did to be closer than those in [(ket′)Cu–OH]− (A) (vide supra). The
not accumulate could not be observed by cyclic voltammetry hydroxy group was more strongly ligated to the copper centre
when PhI was added to [(ket′)Cu–OH]−.
than the phenoxy group. The complex [(ket′)Cu–OPh]− was
A full catalytic cycle is now proposed (paths a–d in cycle I in also generated in THF and characterised by ESI(−)-MS together
Scheme 5) that is based on experiments (up to the formation with [(ket′)CuI]−, 2MeCN (see ESI†).
of A) and DFT calculations (from A to D). A substitution of the
The oxidation peak O4 of [(ket′)Cu–OPh]− did not disappear
ligated phenol in (ket′)Cu(PhOH) (D) by OH− to generate the in the presence of an excess of CsOH (50 mM). On the other
hand, when CsOH was added instead of Cs2CO3 to a solution
containing PhOH and [(ket′)CuI]− (as in Fig. 3f) the complex
[(ket′)Cu–OH]− (A) was exclusively formed (O3) and not [(ket′)-
Cu–OPh]−. This suggests that once the ligated PhOH in (ket′)-
Cu(PhOH) (D) was deprotonated by Cs2CO3, the resulting
complex [(ket′)Cu–OPh]− (A′) was stable in the presence of
CsOH and could thus enter into a new catalytic cycle on the
way to the formation of PhOPh.
DFT calculations were performed to characterise the
Ph–OPh coupling,11 starting from (ket′)Cu(PhOH) D formed in
the first catalytic cycle I (Scheme 6 and Table S2† for structural
data). Deprotonation of the ligated phenol by Cs2CO3 gave
complex [(ket′)Cu–OPh]− (A′). The negative charge of the O
atom in the ligated PhO was calculated to be −0.84 |e−|. Con-
sequently, a complex B′ was formed by O–I halogen bonding
with PhI which was slightly endothermic (0.21 kcal mol−1
,
Scheme 6). The oxidative addition from B′ delivered complex
C′ via complexes E′ and F′ (see computed TS′1 in Scheme 6) in
a sequence similar to that computed for the hydroxo complex.
Scheme 5 Mechanisms of the formation of Ph–OH (paths a–e in cycle I) and
PhOPh (paths a–d in cycle I, then e’–j’ in cycle II) when the copper(I) catalyst is
ligated by the 1,3-diketonate ligand ket’−.
5352 | Dalton Trans., 2013, 42, 5348–5354
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
reductive elimination from (ket′)CuIII(Ph)–OPh, delivers PhOPh
and the CuI catalyst.
Experimental
Cyclic voltammetry was performed under argon in a 3-elec-
trode cell connected to a Schlenk line with a Pt wire as the
counter-electrode, a saturated calomel electrode (SCE) as the
reference and a glassy carbon disk electrode (d = 1 mm) as the
working electrode, at the scan rate of 0.5 V s−1, at 22 °C. 12 mL
of anhydrous degassed DMF containing nBu4NBF4 (0.3 M)
were poured into the cell followed by 100 μL (0.48 mmol,
40 mM) of 2,2,6,6-tetramethyl-3,5-heptanedione (TMHD). A CV
was first performed towards reduction potentials (Fig. 3a,
Table 1). The same CV was performed after addition of 80 mg
(0.48 mmol, 40 mM) of CsOH·H2O but first towards oxidation
potentials (Fig. 3b). 23 mg (0.12 mmol, 10 mM) of CuI was
then added and CV performed just as above (Fig. 3c). 837 mg
(5 mmol, 425 mM) of CsOH·H2O was then added (Fig. 3d).
In other experiments, 50 μL (0.24 mmol, 20 mM) of TMHD
was added, followed by 40 mg (0.24 mmol, 20 mM) of
CsOH·H2O. After 1 h 30 min, the reduction peak of TMHD was
not detected. The oxidation peak of ket′− was observed at O1
attesting to full deprotonation. 46 mg (0.24 mmol, 20 mM) of
CuI was then added and CV performed towards oxidation
potentials (Fig. 3e). 56 mg (0.6 mmol, 50 mM) of PhOH was
added and CV performed just after (Fig. 3f). After addition of
587 mg (1.8 mmol, 150 mM) of Cs2CO3, the CV exhibited the
oxidation peak O4 (Fig. 3g).
200 μL (1.8 mmol) of PhI was added into the electrochemi-
cal cell containing [(ket′)Cu–OH]− formed step by step from
23 mg (0.12 mmol) of CuI, 100 μL (0.48 mmol) of ket′H and
285 mg (1.7 mmol) of CsOH·H2O. The reaction mixture was
heated at 60 °C during 24 h. In another similar experiment the
reaction mixture was heated at 90 °C during 24 h. Dichloro-
methane was added followed by aqueous HCl (37% in
volume). After extraction by dichloromethane and evaporation
of the solvent, the yield of PhOH was determined on the crude
mixture using CHCl2CHCl2 as an authentic standard. 6% of
phenol was formed at 60 °C and 21% at 90 °C (yields relative
to CsOH·H2O used under substoichiometric conditions).
No PhOPh was detected.
Scheme 6 Computed pathway for the formation of PhOPh from complex D.
Comparison of the energy required for the oxidative addition
from complex B (14.69 kcal mol−1, Scheme 3) with that from
complex A′ (11.31 kcal mol−1, Scheme 6) revealed that the
latter was slightly lower. This is in agreement with the weaker
halogen bonding in complex B′ than in complex B.
The anionic CuIII complex F′ formed in the oxidative
addition released its iodide to give (ket′)CuIII(Ph)–OPh (C′)
which underwent reductive elimination, leading to complex
(ket′)CuI(PhOPh) (D′) with an activation energy of 6 kcal mol−1
(see TS′2 in Scheme 6) making again the oxidative addition
rate-determining.
Therefore, the formation of PhOPh first requires the for-
mation of (ket′)Cu(PhOH) (D) via routes a–d in cycle I of
Scheme 5 and then follows route e′–j′ thanks to the extra base
Cs2CO3 (cycle II in Scheme 5). The bases CsOH vs. Cs2CO3
exert thus a discriminating role at the level of complex D
(Scheme 5).
Conclusion
Mechanisms are proposed for the diketonate Cu-catalyzed for-
mation of phenol from PhI and CsOH and that of PhOPh in
the presence of an additional base such as Cs2CO3, which
rationalizes the discriminating role of the bases. Reaction of
[(ket′)CuI–OH]− and PhI (via weak halogen bonding) leads to
an oxidative addition (with a 3-center TS) with a higher energy
barrier than the ensuing reductive elimination from (ket′)
CuIII(Ph)–OH that generates (ket′)CuI(PhOH). The two catalytic
cycles of formation of PhOH or PhOPh are branched at the
level of (ket′)CuI(PhOH), which can either release PhOH via
reaction of CsOH (path e in Scheme 5) or be intercepted by
Cs2CO3 to generate [(ket′)CuI–OPh)]− (path e′ in Scheme 5)
which, after oxidative addition with PhI followed by a faster
Acknowledgements
CNRS, ENS, ENSCP-ParisTech and ANR (grant ANR-07-
CP2D-08, H2OFERCAT) are thanked for financial support.
Notes and references
1 Two important breakthroughs in 2001 for C–C, C–N or C–O
coupling in Taillefer and Buchwald groups: (a) M. Taillefer,
H.-J. Cristau, P. P. Cellier and J.-F. Spindler, French Patent
16547, WO 03053885, 2001; US Patent 10/159,506, US
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 5348–5354 | 5353

View Article Online
Paper
030171593, 2002; US Patent 10/159,829, US 030236413,
Dalton Transactions
8732; (d) D. Yang and H. Fu, Chem.–Eur. J., 2010, 16, 2366–
2370; (e) A. Mahmood and N. N. Leadbeater, Catal.
Commun., 2010, 12, 64–66; (f) L. Jing, J. Wei, L. Zhou,
Z. Huang, Z. Li and X. Zhou, Chem. Commun., 2010, 46,
4767–4769; (g) S. Maurer, W. Liu, X. Zhang, Y. Jiang and
D. Ma, Synlett, 2010, 976–978; (h) R. Paul, M. Ali and
2002; (b) M. Taillefer, H.-J. Cristau, P. P. Cellier, A. Ouali
and J.-F. Spindler, French Patent 06717, WO 031101966,
2002; (c) S. L. Buchwald, A. Klapars, J. C. Antilla, G. E. Job,
M. Wolter, F. Y. Kwong, G. Nordmann and E. J. Hennessy,
US Patent 0286268, WO 02085838, 2001.
2 For reviews see: (a) I. P. Beletskaya and A. V. Cheprakov,
Coord. Chem. Rev., 2004, 248, 2337–2364; (b) G. Evano,
N. Blanchard and M. Toumi, Chem. Rev., 2008, 108, 3054–
3131; (c) F. Monnier and M. Taillefer, Angew. Chem., Int.
Ed., 2009, 48, 6954–6971; (d) E. Sperotto, G. P. M. van
Klink, G. van Koten and J. G. de Vries, Dalton Trans., 2010,
39, 10338–10351.
3 (a) H.-J. Cristau, P. P. Cellier, J.-F. Spindler and M. Taillefer,
Chem.–Eur. J., 2004, 10, 5607–5622; (b) H.-J. Cristau,
P. P. Cellier, J.-F. Spindler and M. Taillefer, Eur. J. Org.
Chem., 2004, 695; (c) E. R. Strieter, D. G. Blackmond and
S. L. Buchwald, J. Am. Chem. Soc., 2005, 127, 4120–4121;
(d) S. L. Zhang, L. Liu, Y. Fu and Q.-X. Guo, Organometallics,
2007, 26, 4546–4554; (e) Z. Tye, J. W. Weng, A. M. Johns,
A. C. D. Incarvito and J. F. Hartwig, J. Am. Chem. Soc., 2008,
T.
Punniyamurthy,
Synthesis,
2010,
4268–4272;
(i) K. G. Thakur and G. Sekar, Chem. Commun., 2011, 47,
4767–4769; ( j) K. Yang, Z. Li, Z. Wang, Z. Yao and S. Jiang,
Org. Lett., 2011, 13, 4340–4343; (k) For reactions without
any ligand, see: J. F. Marcoux, S. Doye and S. L. Buchwald,
J. Am. Chem. Soc., 1997, 119, 10539–10540.
8 (a) A. Tlili, M. Monnier and M. Taillefer, Chem.–Eur. J.,
2010, 16, 12299–12302; (b) M. Taillefer, F. Monnier and
A. Tlili, French Patent 57099, PCY 2011/05044, 2010.
9 (a) J. W. Tye, Z. Weng, R. Giri and J. F. Hartwig, Angew.
Chem., Int. Ed., 2010, 49, 2185–2189; (b) For a mechanistic
study of the C–O bond formation between a well-defined
aryl-copper(III) complex and phenol, see: L. M. Huffman,
A. Casitas, M. Font, M. Canta, M. Costas, X. Ribas and
S. Stahl, Chem.–Eur. J., 2011, 17, 10642–10649.
130, 9971–9983; (f) E. R. Srieter, B. Bhayana and 10 G. Lefevre, G. Franc, A. Tlili, C. Adamo, M. Taillefer,
S. L. Buchwald, J. Am. Chem. Soc., 2009, 131, 78–88;
(g) R. Giri and L. F. Hartwig, J. Am. Chem. Soc., 2010, 132,
15860–15863.
4 (a) A. Shafir and S. L. Buchwald, J. Am. Chem. Soc., 2006,
128, 8742–8743; (b) B. de Lange, M. H. Lambers-Verstap-
pen, L. Schmieder-van de Vondervoort, N. Sereinig, R. de
I. Ciofini and A. Jutand, Organometallics, 2012, 31, 7694–
7707.
11 All calculations were carried out using the hybrid PBE0
exchange correlation functional and modelling solvent
effects (here DMF) using a polarizable continuum model
(PCM). See ESI† for more details.
Rijk, A. H. M. de Vries and J. G. de Vries, Synlett, 2006, 18, 12 For halogen bonds involving ArI, see: (a) H. J. Reich,
3105–3109; (c) A. Shafir, P. A. Lichtor and S. L. Buchwald,
J. Am. Chem. Soc., 2007, 129, 3490–3491; (d) X. Li and
W. Bao, J. Org. Chem., 2007, 72, 3863–3867; (e) M. Taillefer
and N. Xia, French Patent 07 06827, 2007; (f) F. Monnier
and M. Taillefer, Angew. Chem., Int. Ed., 2008, 47, 3096–
P. Green and N. H. Phillips, J. Am. Chem. Soc., 1989, 111,
3444–3445; (b) H. J. Reich and W. L. Whipple,
Can. J. Chem., 2005, 83, 1577–1587; (c) P. Metrangolo,
H. Neukirch, T. Pilati and G. Resnati, Acc. Chem. Res., 2005,
38, 386–395.
3099; (g) Z. Xi, F. Liu, Y. Zhou and W. Chen, Tetrahedron, 13 (a) The addition of excess CsOH to [(ket′)CuI]− resulted in
2008, 64, 4254–4259; (h) N. Xia and M. Taillefer, Angew.
Chem., Int. Ed., 2009, 48, 337–339; (i) G. O. Jones, P. Liu,
K. N. Houk and S. T. Buchwald, J. Am. Chem. Soc., 2010,
132, 6205–6213; ( j) H.-Z. Yu, Y.-Y. Jiang, Y. Fu and L. Liu,
J. Am. Chem. Soc., 2010, 132, 18078–18091; (k) S. Zhang and
Y. Ding, Organometallics, 2011, 30, 633–641.
5 (a) G. Franc, Q. Cacciuttolo, G. Lefèvre, C. Adamo, I. Ciofini
and A. Jutand, ChemCatChem, 2011, 3, 305–309;
(b) G. Lefèvre, G. Franc, C. Adamo, A. Jutand and I. Ciofini,
Organometallics, 2012, 31, 914–920.
the full solubility (via reaction) of the initial CuI complex
as expressed by the sum of the oxidation peak currents
ipO2 + iOp3 (Fig. 3d) which was higher than iOp2 observed at
lower CsOH concentrations (Fig. 3c).13b Importantly, the
ratio ipO2/iOp3 did not reflect the ratio of the thermodynamic
concentrations of the two complexes but dynamic concen-
trations since [(ket′)CuI]− was oxidized first at O2. Its con-
sumption by oxidation in the diffusion layer resulted in the
shift of the equilibrium towards [(ket′)CuI]− (CE mechan-
ism);13c (b) Currents are proportional to the concentration
of electroactive species;13c (c) A. J. Bard and L. R. Faulkner,
in Electrochemical Methods. Fundamentals and Applications,
WILEY & Sons, New York, 1980, pp. 442–450.
6 For other proposed mechanisms based on DFT, see ref. 4i–k.
7 (a) A. Tlili, N. Xia, F. Monnier and M. Taillefer, Angew.
Chem., Int. Ed., 2009, 48, 8725–8728; (b) M. Taillefer,
F. Monnier, A. Tlili and N. Xia, French Patent 07767, 14 (a) R. J. Enemaerke, T. B. Christensen, H. Jensen and
WO2010/142913, 2009; (c) For reactions involving N,N, N,O
or O,O ligands, see: D. Zaho, N. Wu, S. Zhang, P. Xi, X. Su,
J. Lan and J. You, Angew. Chem., Int. Ed., 2009, 48, 8729–
K. Daasbjerg, J. Chem. Soc., Perkin Trans., 2001, 2, 1620–
1630; (b) The higher the potential gap, the slower the elec-
tron transfer. For potential gaps allowing SET, see ref. 14a.
5354 | Dalton Trans., 2013, 42, 5348–5354
This journal is © The Royal Society of Chemistry 2013