Stabilized naked Sub-nanometric Cu clusters within a polymeric film catalyze C-N, C-C, C-O, C-S, and C-P bond-forming reactions

Article abstract of DOI:10.1021/jacs.5b00222

Sub-nanometric Cu clusters formed by endogenous reduction of Cu salts and Cu nanoparticles are active and selective catalysts for C-N, C-C, C-O, C-S, and C-P bond-forming reactions. Sub-nanometric Cu clusters have also been generated within a polymeric film and stored with full stability for months. In this way, they are ready to be used on demand and maintain high activity (TONs up to 10⁴) and selectivity for the above reactions. A potential mechanism for the formation of the sub-nanometric clusters and their electronic nature is presented.

Full text of DOI:10.1021/jacs.5b00222

Article
pubs.acs.org/JACS
Stabilized Naked Sub-nanometric Cu Clusters within a Polymeric
Film Catalyze C−N, C−C, C−O, C−S, and C−P Bond-Forming
Reactions
Judit Oliver-Messeguer,† Lichen Liu,† Saray García-García,† Claudia Canos-Gimenez,† Irene Domínguez,‡
̀
́
́
Rafael Gavara,‡ Antonio Domenech-Carbo,§ Patricia Concepcion,† Antonio Leyva-Perez,*,†
́
́
́
́
and Avelino Corma*,†
†Instituto de Tecnología Química, Universidad Politec
Naranjos s/n, 46022 Valencia, Spain
́
nica de Valencia-Consejo Superior de Investigaciones Científicas, Avda. de los
‡Packaging Lab, Instituto de Agroquímica y Tecnología de Alimentos, IATA-CSIC, Av. Agustín Escardino 7, 46980 Paterna, Valencia,
Spain
§Departament de Química Analítica, Universitat de Valencia, Dr. Moliner, 50, 46100 Burjassot, Valencia, Spain
S
* Supporting Information
ABSTRACT: Sub-nanometric Cu clusters formed by endog-
enous reduction of Cu salts and Cu nanoparticles are active
and selective catalysts for C−N, C−C, C−O, C−S, and C−P
bond-forming reactions. Sub-nanometric Cu clusters have also
been generated within a polymeric film and stored with full
stability for months. In this way, they are ready to be used on
demand and maintain high activity (TONs up to 104) and
selectivity for the above reactions. A potential mechanism for
the formation of the sub-nanometric clusters and their
electronic nature is presented.
clusters are formed in the reaction media.10−12 It appears then
that metal clusters formed by a small number of atoms, that
expose the specific frontier orbitals of the metal for reactivity,
may fill the gap between the single metal atom with organic
1. INTRODUCTION
Catalysis by supported metallic nanoparticles has been
performed for many years in academia and industry, and a
large number of commercial catalysts base their activity on
metal nanoparticles of less than 5 nm.1 However, the advances
ligands and the metal nanoparticles. If this is the case, it could
be expected that small clusters of non-noble metals, such as Cu,
made in materials science allow us today to prepare well-
could also present enhanced catalytic properties.
defined metal nanoparticles of nanometer or even sub-
nanometer size, opening new opportunities in heterogeneous
Copper shows unique catalytic activity13−21 for some cross-
catalysis on metals and oxides.2,3 It is interesting to notice that
coupling reactions, provided that suitable ligands, such as
diamines, are added to the reaction medium. These novel Cu-
catalysis by homogeneous metal catalysts either as transition
catalyzed cross-coupling reactions have rapidly found imple-
metal complexes or as metal salts in solution, and
mentation in organic synthesis,22−24 though it would be
heterogeneous catalysis by supported or unsupported metal
nanoparticles, have been advancing separately, with little
interaction and cross-fertilization. However, we believe that
they are just two different sides of the same coin, and new
desirable to decrease the amount of Cu (1−5 mol%) and
diamine (5−20 mol%) typically employed for these reactions.25
Cu cluster complexes have proved to be competent reaction
intermediates for these reactions,16,26 and, despite the fact that
unifying concepts need to be developed to bridging the gap
between homogeneous and heterogeneous catalysis.4
different ligand-stabilized Cu clusters and naked Cu clusters
have been synthesized,27−30 the use of well-defined Cu clusters
It has been shown5−9 for a number of reactions catalyzed by
gold and palladium organic complexes and salts that neither the
metal complex nor the salt was the active catalytic species, but
they decompose to form sub-nanometric clusters with 3−10
atoms which catalyze the reactions with very high turnover
frequencies (TOFs) and turnover numbers (TONs). For those
same reactions, an induction period is also observed when
nanometer-size gold or palladium nanoparticles are used as
catalysts, and the reaction starts when the sub-nanometric metal
in catalysis is very rare.31
In the present work, we will show that Cu clusters with up to
7 Cu atoms, formed in situ by endogeneous reduction, are
active and selective catalysts for C−N, C−C, C−O, C−S, and
C−P bond-forming reactions. Alternatively, Cu clusters will be
Received: January 12, 2015
Published: March 8, 2015
© 2015 American Chemical Society
3894
DOI: 10.1021/jacs.5b00222
J. Am. Chem. Soc. 2015, 137, 3894−3900

Journal of the American Chemical Society
Article
formed within a polymeric film to be stored for months,
keeping high activity and selectivity for the above-named
reactions, with TONs up to 104. The possible mechanism of
formation of the Cu clusters and their chemical nature will be
also discussed.
coupling starts and proceeds. Electrospray ionization mass
spectrometry (ESI-MS) confirmed the presence of Cu clusters
of 2−7 atoms in solution only when the coupling starts, but not
during the induction time. These results suggest that Cu
clusters of low atomicity are formed during a period that
corresponds to the induction time for the Goldberg reaction
between iodobenzene 1 and amide 2.
2. RESULTS AND DISCUSSION
The amount of Cu clusters in solution could be evaluated by
the method of Ohno et al.32 using a sample of well-defined Cu
clusters with an average size of 5 atoms (Cu5) as a reference
(SI Figure S3).30 The analytical measurements revealed that the
efficiency of formation of Cu5 from CuI is ∼3% for any
concentration of starting CuI. This means that the amount of
Cu clusters in solution for the Goldberg reaction in Figure 1 is
0.003 mol%, the TON being >104.
C−N, C−C, C−O, C−S, and C−P Cross-Coupling
Reactions with in Situ-Formed Sub-nanometric Cu
Clusters. Figure 1 shows the kinetic results for the cross-
At this point, the possible scope of the Cu clusters was tested
for other C−N, C−C, C−O, C−S, and C−P cross-coupling
reactions. The results given in Figure 2 show that a wide array
of nucleophiles can be coupled with different aryl iodides under
heating conditions in amide solvents, with 0.02−1 mol% Cu
clusters in solution. The reactivity of nucleophiles follows the
order thiols > phenols > amides ≈ alkynes > imidazoles >
phosphines ≫ amines, allowing selective functionalizations.
Blank experiments without any Cu gave up to 30% of products
in particular cases, but kinetic experiments showed that the Cu-
catalyzed reactions ran with at least 2 orders of magnitude
higher rates than the thermal reaction and with higher
selectivity. Notice that the Cu cluster system can couple
thiourea twice (product 9) and also give directly the one-pot
coupling−cyclization reaction to form indoles and benzofurans
(products 17 and 18).
The results obtained for each reaction under typical
conditions with diamine ligands, either reported in the
literature or obtained here by us, are presented as a reference
in Figure 2, and kinetic studies for the Goldberg coupling
between 1 and 2 are also included. From the results, it is
possible to say that, though the Cu cluster gives a higher TON
per atom of active Cu and higher selectivity toward different
nucleophiles than the Cu−diamine catalyst, the latter gives
higher yields than the cluster and is able to couple the more
difficult to react bromo and chloro derivatives. In situ UV−vis
measurements during the Cu−diamine-catalyzed reactions did
not show any evidence of the presence of Cu clusters, and, as it
can be seen in the kinetic curve in Figure 2, no induction time
was found. If diamines are added to the Cu cluster-catalyzed
reaction before or after the induction period, when the Cu
clusters are formed, the outcome of the reaction is not affected.
With all the above data in hand, we can say that the Cu
clusters and the Cu−diamine complexes present intrinsic
catalytic differences. While the Cu clusters are more active
for aryl iodides and have a better reactivity with different
nucleophiles under optimized conditions, the Cu−diamine
catalyst operates under relatively milder conditions, has a wider
functional group tolerance, and is successful for more-
demanding reactions. A Hammett plot for different aryl iodides
during the Goldberg reaction catalyzed by the Cu cluster (SI
Figure S4) shows ρ = +0.84, which indicates that the activation
of the carbon−halide bond by Cu plays a significant role during
the coupling as it occurs for the Cu−diamine catalyst.13,14,16
Mechanism of Formation of the Cu Clusters.
Endogenous Reduction of Cu Salts in Amide Solvents. The
cross-coupling reactions observed when using the amide as a
solvent do not occur with solvents such as toluene or dioxane,
Figure 1. Kinetic results for the cross-coupling reaction between
iodobenzene 1 and amide 2 catalyzed by commercially available Cu
compounds at 0.5 mol%: Cu(OAc)2 (blue diamonds), CuO
nanoparticles of 50 nm average size (red squares), Cu(acac)2 (orange
circles), and CuI (green triangles). The inset shows a magnification of
the initial points of the curves, where an ∼1 h induction time can be
observed for all the Cu catalysts tested. The reaction was followed by
gas chromatography using dodecane as an external standard after
dilution in acetonitrile. Each point is an average of three runs. Error
bars are also given.
coupling reaction between iodobenzene 1 and amide 2
(Goldberg reaction) in the presence of Cu(OAc)2, Cu(acac)2,
CuI, or CuO nanoparticles (∼50 nm) at 0.5 mol%, under basic
conditions in N-methylpyrrolidone (NMP) as a solvent. A ∼1 h
reaction induction time can be observed in all four cases, and,
after that, the formation of product 3 starts with a similar
reaction rate for all the Cu catalysts employed. The final yield
and selectivity to the coupled product was typically >90%, and
other bases (K2CO3, Cs2CO3) and amide solvents (N-
dimethylformamide, DMF) proved suitable for the reaction
under the present experimental conditions.
Other Cu salts (CuCl, CuBr, CuCN, CuSO4, CuBr2) gave
similar sigmoidal kinetic curves with high yields of product 3
after 24 h. It was possible to achieve high final yields of product
3 with just 0.1 mol% of CuI (see Supporting Information (SI)
Figure S1), though the induction period gets longer as the
amount of Cu salt decreases. Following the reaction by
absorption and emission ultraviolet−visible spectroscopy (UV−
vis) with the non-chromophoric 1-iodooctene as a reaction
substrate (see SI Figure S2) unveiled that the transition bands
corresponding to the Cu salts disappear during the induction
time, and new bands corresponding to Cu clusters of 2−7
atoms appear just after the induction time, when the Goldberg
3895
DOI: 10.1021/jacs.5b00222
J. Am. Chem. Soc. 2015, 137, 3894−3900

Journal of the American Chemical Society
Article
of starting Cu salt. In the case of DMF, gaseous CO2 and
dimethylamine were found as decomposition products after
heating the Cu salts.39
Examination of the DMF solution by UV−vis showed the
expected Cu clusters, and high-resolution transmission electron
microscopy (HR-TEM, see SI Figure S5) also showed Cu
nanoparticles of ∼5 nm in solution. This result indicates that,
apparently, both Cu clusters and Cu nanoparticles are formed
in the amide solvent when Cu salts are heated. The above
observations suggest that the amide solvent can act as a
stoichiometric reducing agent for Cu through the hydrogen
atoms in the α-carbon to the nitrogen. In accordance, no
Goldberg coupling reaction at all occurs when NMS is used as a
solvent.
Figure 3 shows two possible mechanisms for the reduction of
Cu(acac)2 in NMP under heating. In pathway A, the reduction
Figure 3. Proposed mechanism for the endogenous reduction of
Cu(acac)2 in NMP, and isotopic experiments.
of Cu(acac)2 in NMP occurs after coordination of the Lewis
acid Cu(II) to the amide oxygen and subsequent H-atom
abstraction by the anion.40 Reductive elimination gives Cu(0)
atoms that agglomerate into nanoparticles. In pathway B, the
reduction of Cu(II) occurs after spontaneous aerobic oxidation
of NMP23 to the α-amidoalcohol that then coordinates to
Cu(II) and triggers a reductive elimination to form Cu(0)
atoms and NMS.41,42 To check that the H-atom transfer from
the amide readily occurs, 5-d2-NMP was synthesized43,44 and
heated in the presence of Cu(II) acetylacetonate. Gas
chromatography−mass spectrometry (GC-MS) measurements
showed the formation of 3-d-pentane-2,4-dione, which was
confirmed by 1H and 13C NMR spectroscopy. Since that
product is common to both reaction pathways, a new isotopic
experiment was performed with an excess of H218O in the
reaction medium in order to differentiate between the two
pathways in Figure 3. Indeed, the presence or absence of the
18O atom from labeled water in the final oxidation product
NMS can differentiate pathway A from B. GC-MS measure-
ments showed that only 15% of the NMS formed contains 18O.
Figure 2. Scope of the Cu cluster-catalyzed cross-coupling reactions.
Isolated yields are reported, and the new bonds are shown in bold. The
Cu clusters are generated in situ from CuI or added within the EVOH
polymer (see ahead). For comparison, the values with Cu−diamine
catalysts are also given, reaction conditions: substrates (1 mmol), CuI
(5−20 mol%), N,N-dimethylethylenediamine (10−40 mol%), K3PO4
(2 mmol), anhydrous toluene or dioxane (0.25 M), nitrogen
atmosphere, 110 °C, 24 h. The inset shows the kinetics for the
Goldberg coupling between 1 and 2 under typical reaction conditions
for diamine-assisted coupling and the conditions reported here.
even at 135 °C, unless diamine is added. Furthermore,
formation of Cu clusters was not detected in toluene or
dioxane by UV−vis spectroscopy. These results indicate that
the amide solvent plays an active role in the formation of the
clusters, and we have attempted to reveal that role. Thus, when
NMP was used as a solvent, N-methylsuccinimide (NMS) was
found as a product after heating different Cu salts at 135 °C,
and the formation of NMS increased linearly with the amount
3896
DOI: 10.1021/jacs.5b00222
J. Am. Chem. Soc. 2015, 137, 3894−3900

Journal of the American Chemical Society
Article
Thus, the absence of significant amounts of 18O isotopic NMS
discards, in principle, the hydrolysis of the postulated iminium
carbon in pathway A, and it suggests pathway B as the main
route for the endogenous reduction of Cu salts to Cu clusters
and nanoparticles by the amide solvent under the present
heating conditions. Nevertheless, the question remains whether
the Cu clusters and nanoparticles form at the expense of each
other or they form concomitantly.
Figure 4. Left: Initial points of the kinetic curves for the cross-
coupling reaction between iodobenzene 1 and amide 2 in the presence
of diverse Cu clusters at 0.05 mol%. Right: Linear correlation between
the initial reaction rate for the Goldberg reaction and the amount of
Cu clusters containing EVOH polymer (Cu@EVOH) used as a
catalyst, no induction time was found in any case. The reactions were
followed by gas chromatography using dodecane as an external
standard after dilution with diethyl ether. Each point is an average of
three runs. Error bars are also given.
Dissolution of Cu Nanoparticles to Cu Clusters. When
freshly prepared Cu nanoparticles of 5 nm diameter in DMF,
free of clusters, were used as catalysts for the Goldberg reaction,
a short induction period (∼15 min) was still observed, after
which small Cu clusters were detected (see Figure S6). This
result confirms that the Cu clusters and not the nanoparticles
are the true active catalysts for the cross-coupling reaction.
Taking into account that any Cu salt gives an induction period
of ∼1 h, which is much longer than the induction period
observed when starting the reaction with the nanoparticles, and
while it was enough time to form Cu nanoparticles from the
different salts, it seems plausible than the Cu clusters are
formed at expenses of the nanoparticles. Small Cu nanoparticles
may very well melt at 135 °C (reaction conditions) to release
clusters in an auto-accelerated process, since as the nanoparticle
becomes smaller its melting point decreases, providing a
thermodynamically plausible explanation for the formation of
the clusters.45−48 In order to assess that the Cu nanoparticles,
and no other possible species in solution, are the catalytic
precursors of Cu clusters, monodisperse Cu nanoparticles free
of Cu salts and clusters were independently prepared by
reduction of Cu(I) iodide with NaBH4 and stabilized on
polyvinylpyrrolidone (PVP) with different particle average sizes
(1.5, 3.5, and 5.5 nm, see SI Figure S7 for characterization).
When the freshly prepared PVP-stabilized 1.5 nm Cu
nanoparticles were used as catalysts for the coupling reaction
between iodooctene and either amide 2 or imidazole (see SI
Figures S8 and S9), a very short induction time (∼15 min) was
found, after which the coupling reaction started with initial
rates and final yields comparable to those found for the Cu
clusters of 2−7 atoms. The induction time was very similar for
the larger nanoparticles (15 and 20 min for 3.5 and 5.5 nm
average particle size, respectively), with an overall catalytic
activity slightly lower than that of the smaller nanoparticles. In
situ UV−vis spectroscopy showed that the nanoparticles are
partially dissolved in <5 min under the heating conditions to
form Cu clusters that catalyze the Goldberg reaction after an
additional 15 min induction time. Thus, we can say that Cu
clusters are rapidly formed at the expense of the different Cu
nanoparticles but that these Cu clusters are not yet the true
catalysts of the reaction, since an additional ∼10 min induction
time is still found.
size of 13 atoms (Cu13) were also prepared electrochemically
(see SI Figure S12), and the kinetic results show an induction
time of 1 h. Notice that the induction time found with Cu salts
at this catalytic amount (0.05 mol%) is ∼4 h, much longer than
for the Cu5 and Cu13 clusters prepared by the electrochemical
method. The results indicate that these clusters are similar but
not exactly the true catalytic species of the Goldberg reaction,
although the atomicity of Cu5 fits very well with those clusters
formed during the reaction and which were associated with the
catalytically active species.
UV−vis measurements showed that the electrochemically
prepared Cu5 and Cu13 clusters rearrange during the coupling
reaction conditions to give new absorption and emission bands
after the induction time, corresponding exclusively to 2−7 atom
Cu clusters as it occurs with Cu salts, and then the coupling
reaction starts (see SI Figure S13).
Further characterization of Cu5 and Cu13 clusters formed
electrochemically was performed in order to gain more insight
into the nature of these clusters and their evolution during
reaction. Cyclic voltammetry of the Cu5 and Cu13 clusters
showed some oxidation and reduction peaks corresponding to
Cu(I); zeta-potential measurements of the Cu5 and Cu13
solutions indicate positively charged species (see SI Figure
S14A,B); ESI-MS measurements of freshly prepared Cu5 and
Cu13 clusters showed the presence of oxygen atoms in the
cluster; and X-ray photoelectron spectroscopy (XPS) showed
Cu(0) species in the clusters (see SI Figure S14C,D).49,50
These results indicate that the electrochemically prepared Cu
clusters in solution are not only the Cu(0) observed by XPS
but also some oxidized form of Cu(I) that, in any case, are not
the active catalyst for the coupling since they must rearrange
during the induction period.
Nature of the Catalytically Active Copper Clusters. Well-
defined Cu clusters with an average size of 5 atoms (Cu5) were
prepared according to a reported electrochemical method.30
We expected that these naked Cu5 clusters should catalyze the
Goldberg reaction, since the UV−vis and ESI-MS character-
ization of these Cu5 clusters correlates well with the clusters
formed during reaction (Figure S10). Disappointingly, Figure 4
shows that the electrochemically formed Cu5 clusters gave a
short but clear induction time of 30 min when used as a catalyst
for the Goldberg reaction at 0.05 mol%, which indicates that
the electrochemically prepared Cu5 clusters are not the true
catalysts of the reaction (for the whole kinetics, see SI Figure
S11). For the sake of comparison, Cu clusters with an average
Concerning Cu nanoparticles, XPS measurements showed
that the Cu nanoparticles on PVP contain 30% of Cu(I) oxide
already present in the freshly prepared samples. Remarkably,
the induction time (15 min) and TOF0 (45 h−1 at 0.2 mol%)
found for electrochemically prepared oxidized Cu clusters
correlate well with thosse of dissolved Cu nanoparticles (15−20
min induction time, TOF0 = 65 h−1).51 The dissolution of
Cu(I) oxide from Cu nanoparticles was observed some time
ago for the Ullmann coupling between iodobenzenes and
diarylamines; in fact, it was already proposed by Paine that
some common form of Cu(I) ions was formed and could
catalyze the reaction, regardless of the starting Cu source for
this reaction.52 Although a direct correlation with the work of
3897
DOI: 10.1021/jacs.5b00222
J. Am. Chem. Soc. 2015, 137, 3894−3900

Journal of the American Chemical Society
Article
Paine is difficult, the similarities between his conclusions and
our results here are striking.
After storage at room temperature without atmospheric
protection, the PVP-stabilized Cu nanoparticles showed further
oxidization to Cu(I) and then to Cu(II), according to XPS
measurements, and the smaller the nanoparticle, the easier the
oxidation. While the 5.5 nm (average size) Cu nanoparticles
still keep significant amounts of Cu(0) and Cu(I) after 1 day,
the 1.5 nm (average size) Cu nanoparticles are nearly fully
oxidized. Kinetic results showed a significantly lower catalytic
activity for the further oxidized nanoparticles containing Cu(II)
(see SI Figures S7−S9), and, in contrast with freshly prepared
samples, the catalytic activity of the 1.5 nm PVP-stabilized Cu
nanoparticles after 1 day was lower than that of the bigger (3.5
and 5.5 nm) Cu particles. In other words, the higher the
amount of Cu(I), the higher the catalytic activity.
Figure 5. Left: Photograph of the transparent Cu@EVOH film. The
circle marks the area in which the film is present. Right: Evolution of
oxygen flow through the developed film under isostatic conditions at
23 °C. From this curve, an oxygen permeability value of (4.1 0.2) ×
10−21 m3·m/(m2·s·Pa) was estimated. By applying the solution to
Fick’s law with the boundary conditions of a permeation test, the
diffusion coefficient was calculated to be 2.5 × 10−14 m2/s. This value
is in agreement with that reported for an excellent barrier to oxygen.
With these data in hand, we propose that the catalytically
active clusters are not Cu(0) or Cu(II) but some form of
deoxygenated Cu(I).
species.54 Diffuse reflectance UV−vis spectroscopy of Cu@
EVOH showed well-defined bands that correspond to Cu
clusters of ∼3−5 atoms (absorption at 275−300 nm, emission
at 300 nm) and ∼10−13 Cu atoms (absorption at 380 nm,
emission at 400 nm). No plasmonic Cu bands were found, and,
in agreement with this, HR-TEM measurements of the film did
not show any visible Cu nanoparticle ≥1 nm. Quantification by
the spectrophotometric method32 gives 4% of Cu5 clusters out
of the total Cu of the polymer. These results indicate that (1)
Cu(II) is reduced within the EVOH polymer to Cu(I) to some
extent, (2) Cu(I) clusters are formed and remain stable for
months, and (3) much of the starting Cu(II) still remains in the
outer part of the polymer. The last point can be due to an
incomplete reduction of the starting Cu(II) salt or to a
progressive re-oxidation of Cu(I) to Cu(II) in the outer layers
of the film that are more exposed to air.
The Cu@EVOH was introduced in DMF at the reaction
temperature, and the amount of Cu in solution was followed by
ICP-AES over time, while the size of the species in solution was
addressed by absorption and emission UV−vis and ESI-MS
(see SI Figure S20). We found that the velocity of diffusion of
Cu from the polymer to hot DMF is 2 orders of magnitude
higher than the reaction coupling, and that the Cu species
found in solution kept the same size as within the polymer, i.e.,
3−10 atoms. These results confirm that Cu@EVOH rapidly
leaches into solution the embedded Cu(I) clusters under
reaction conditions. One may then expect that these suitably
sized Cu(I) clusters could catalyze the Goldberg reaction
without a detectable induction time if they are the catalytically
active species. Indeed, Figure 4 shows that the induction time
for the coupling reaction between 1 and 2 completely
disappears with Cu@EVOH at 0.05 mol% and that a linear
correlation between the amount of Cu@EVOH and the initial
reaction rate exists. Additionally, a Hammett plot for different
iodides with the Cu@EVOH catalyst gives the same value as for
the Cu clusters generated in situ from Cu salts (see SI Figure
S4). Thus, we can conclude that de-oxygenated Cu(I) clusters
of 2−7 atoms are the catalytically active species during the
ligand-free Cu-catalyzed Goldberg reaction shown in Figure 1.
Synthesis of Deoxygenated Cu(I) Clusters by One-Pot
Reduction−Stabilization within an Oxygen-Protective Poly-
mer: The Catalytic Species. If we consider deoxygenated Cu(I)
as the plausible species to catalyze the cross-coupling, an
explanation for the induction time found in Figure 4 could be
the need to remove the oxygen atoms from the Cu(I) cluster.
To check this, we attempted the synthesis of non-oxidized
naked Cu clusters of low atomicity outside of the reaction. De-
oxygenation of the electrochemically prepared oxide Cu cluster
by reducing agents such as citric acid, hydrogen, and NaBH4
failed, and the direct synthesis of deoxygenated Cu clusters
from Cu salts with hydride agents also failed. Thus, we turned
our attention to prepare Cu clusters within an oxygen- and
water-free environment, in such a way that the Cu(I) clusters
could be liberated into the reaction without interacting with
oxygen. The strategy involves the following steps: (1)
incorporation of a Cu(II) salt within an oxygen-refractive
solid material, (2) in situ reduction of the embedded Cu, (3)
isolation of the Cu cluster-containing material, and (4) rapid
liberation of the Cu clusters into the reaction medium by
dissolution of the Cu-containing material to perform the
coupling reaction. For that purpose and after much
experimentation, we found that ethylene−vinyl alcohol
copolymer (EVOH), a well-known polymer that protects
against atmospheric oxygen and humidity,53 was able to carry
out the one-pot reduction−polymerization of Cu(NO3)2 using
carvacrol as an in situ mild reduction agent. A detailed
preparation of the Cu clusters−EVOH film (Cu@EVOH) can
be found in the SI (with characterization in Figures S15−S19),
and Figure 5 shows a photograph of the obtained film.
The transparent film of Cu@EVOH incorporates Cu
quantitatively, and experiments of adsorption of molecular
oxygen (see Figure 5) confirm that the new Cu-containing
material does not allow oxygen to diffuse easily. The
crystallinity of the polymer was not affected by Cu introduction
according to infrared measurements, and scanning electron
microscopy gives an estimated film thickness of 0.013 mm, in
concordance with the results obtained with a digital micro-
meter. The water adsorption of the material is very low
according to the water gain and the water diffusion coefficient
value of the EVOH polymer. The film does not change after
several months stored in plastic foil. Cyclic voltammetry
measurements showed that the inner layers of the polymer
contain Cu(I) species, while the external layers contain Cu(II)
3. CONCLUSIONS
De-oxygenated Cu(I) clusters of 2−7 atoms catalyze the cross-
coupling of iodoaromatics and iodoalkenes with nitrogen,
carbon, oxygen, sulfur, and phosphorus nucleophiles with high
3898
DOI: 10.1021/jacs.5b00222
J. Am. Chem. Soc. 2015, 137, 3894−3900

Journal of the American Chemical Society
Article
(5) Hagen, C. M.; Widegren, J. A.; Maitlis, P. M.; Finke, R. G. J. Am.
Chem. Soc. 2005, 127, 4423.
(6) Bayram, E.; Linehan, J. C.; Fulton, J. L.; Roberts, J. A. S.;
Szymczak, N. K.; Smurthwaite, T. D.; Ozkar, S.; Balasubramanian, M.;
Finke, R. G. J. Am. Chem. Soc. 2011, 133, 18889.
efficiency under ligand-free conditions. The clusters are
generated either by endogenous reduction of Cu salts in
amide solvents to form Cu nanoparticles that then dissolve into
Cu clusters, or by dislodging the clusters from PVP-stabilized
Cu nanoparticles, or by in situ reduction of Cu salts within an
oxygen-protective polymer.
(7) Oliver-Meseguer, J.; Cabrero-Antonino, J. R.; Domínguez, I.;
́
Leyva-Perez, A.; Corma, A. Science 2012, 338, 1452.
(8) Corma, A.; Concepcion, P.; Boronat, M.; Sabater, M. J.; Navas, J.;
Yacaman, M. J.; Larios, E.; Posadas, A.; Lopez-Quintela, M. A.; Buceta,
D.; Mendoza, E.; Guilera, G.; Mayoral, A. Nat. Chem. 2013, 5, 775.
4. EXPERIMENTAL SECTION
Typical Reaction Procedure. Tripotassium phosphate (424 mg, 2
mmol) was placed in a 1.5 mL vial equipped with a magnetic stir bar,
and anhydrous N-methylpyrrolidone or N-dimethylformamide (NMP
or DMF, 1 mL) was added. The desired amount of the copper iodide,
the corresponding halide (1.5 mmol), and the coupling counterpart (1
mmol) were then added. The vial was sealed, and the resulting mixture
was placed in a pre-heated oil bath at 135 °C and magnetically stirred.
After 24 h, slow addition of dimethyl ether or dichloromethane (5
mL) and water (6 mL), separation of the organic layer, and two
extractions of the aqueous layer with dimethyl ether or dichloro-
methane were carried out before the combined organic layers were
dried over MgSO4 and then filtered. The solvent was removed under
vacuum at room temperature, and the residue was purified by flash
column chromatography on silica gel to afford the corresponding
product. For kinetics, aliquots of 0.05 mL were periodically taken and
diluted in acetonitrile or diethyl ether (1 mL) for GC analysis, using
dodecane as an external standard. The same aliquots without dodecane
were used for UV−vis measurements. Further details can be found in
the SI.
(9) Leyva-Per
A. Angew. Chem., Int. Ed. 2013, 52, 11554.
(10) Oliver-Meseguer, J.; Leyva-Per
A. Chem. Commun. 2013, 49, 7782.
(11) Oliver-Meseguer, J.; Leyva-Per
2013, 5, 3509.
(12) Boronat, M.; Leyva-Per
47, 834.
(13) Klapars, A.; Huang, X.; Buchwald, S. L. J. Am. Chem. Soc. 2002,
124, 7421.
(14) Strieter, E. R.; Blackmond, D. G.; Buchwald, S. L. J. Am. Chem.
Soc. 2005, 127, 4120.
(15) Zhang, S.-L.; Liu, L.; Fu, Y.; Guo, Q.-X. Organometallics 2007,
26, 4546.
(16) Strieter, E. R.; Bhayana, B.; Buchwald, S. L. J. Am. Chem. Soc.
2009, 131, 78.
(17) Larsson, P.-F.; Correa, A.; Carril, M.; Norrby, P.-O.; Bolm, C.
Angew. Chem., Int. Ed. 2009, 48, 5691.
́ ́
ez, A.; Oliver-Meseguer, J.; Rubio-Marques, P.; Corma,
́
ez, A.; Al-Resayes, S. I.; Corma,
́
ez, A.; Corma, A. ChemCatChem
́
ez, A.; Corma, A. Acc. Chem. Res. 2014,
Preparation of Cu@EVOH. Pellets of EVOH polymer (with ∼4%
w/w carvacrol) were dissolved in a 1:1 (v:v) mixture of 1-propanol
and distilled water at 75 °C. Once dissolved, the mixture was left to
cool to room temperature, and Cu(II) nitrate was added in order to
obtain a metal loading of 0.031 mmol Cu/g dry polymer. Further
details can be found in the SI.
(18) Surry, D. S.; Buchwald, S. L. Chem. Sci. 2010, 1, 13.
(19) Sperotto, E.; van Klink, G. P. M.; van Koten, G.; de Vries, J. G.
Dalton Trans. 2010, 39, 10338.
(20) Zuidema, E.; Bolm, C. Chem.Eur. J. 2010, 16, 4181.
(21) Jones, G. O.; Liu, P.; Houk, K. N.; Buchwald, S. L. J. Am. Chem.
Soc. 2010, 132, 6205.
(22) Ley, S. V.; Thomas, A. W. Angew. Chem., Int. Ed. 2003, 42, 5400.
(23) Beletskaya, I. P.; Cheprakov, A. V. Coord. Chem. Rev. 2004, 248,
2337.
(24) Evano, G.; Theunissen, C.; Pradal, A. Nat. Prod. Rep. 2013, 30,
1467.
(25) Evano, G.; Blanchard, N. Copper-mediated cross-coupling
reactions; Wiley: Weinheim, 2014.
(26) Beletskaya, I. P.; Cheprakov, A. V. Organometallics 2012, 31,
ASSOCIATED CONTENT
■
S
* Supporting Information
Additional Figures S1−S20, experimental procedures, and
compound characterization, including NMR spectra. This
material is available free of charge via the Internet at http://
pubs.acs.org.
7753.
AUTHOR INFORMATION
(27) Nakamura, E.; Mori, S. Angew. Chem., Int. Ed. 2000, 39, 3750.
(28) Manbeck, G. F.; Lipman, A. J.; Stockland, R. A. J.; Freidl, A. L.;
Hasler, A. F.; Stone, J. J.; Guzei, I. A. J. Org. Chem. 2005, 70, 244.
(29) Haldo
Organometallics 2009, 28, 3815.
■
Corresponding Authors
*anleyva@itq.upv.es
*acorma@itq.upv.es
́
́ ́
n, E.; Alvarez, E.; Nicasio, M. C.; Perez, P. J.
(30) Vilar-Vidal, N.; Blanco, M. C.; Lopez-Quintela, M. A.; Rivas, J.;
Serra, C. J. Phys. Chem. C 2010, 114, 15924.
(31) Vilar-Vidal, N.; Rivas, J.; Lopez-Quintela, M. A. ACS Catal.
2012, 2, 1693.
(32) Ohno, S.; Teshima, N.; Zhang, H.; Sakai, T. Talanta 2003, 60,
1177.
(33) Swapna, K.; Murthy, S. N.; Nageswar, Y. V. D. Eur. J. Org. Chem.
2010, 6678.
(34) Feng, Y.; Wang, H.; Sun, F.; Li, Y.; Fu, X.; Jin, K. Tetrahedron
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial support by the Severo-Ochoa program, Consolider-
Ingenio 2010 (proyecto MULTICAT), and PLE2009 project
from MCIINN is acknowledged. J.O.-M., L.-L., and A.L.-P.
thank ITQ for the concession of a contract. I.D. and R.G. thank
the Spanish Ministry of Science and Innovation for financial
support through project AGL-2012-39920-C03-01. I.D. thanks
the CSIC for the postdoctoral contract concession (JAE-DOC).
2009, 65, 9737.
(35) Gelman, D.; Jiang, L.; Buchwald, S. L. Org. Lett. 2003, 5, 2315.
(36) Yanga, D.; Lia, B.; Yanga, H.; Fu, H.; Hu, L. Synlett 2011, 5, 702.
(37) Zou, L.-H.; Johansson, A. D.; Zuidema, E.; Bolm, C. Chem.
Eur. J. 2013, 19, 8144.
(38) Jaseer, E. A.; Prasad, D. J. C.; Sekar, G. Tetrahedron 2010, 66,
2077.
REFERENCES
■
(1) Srinivasan, R.; Davis, B. H. Platinum Metals Rev. 1992, 36, 151.
(2) Astruc, D.; Lu, F.; Aranzaes, J. R. Angew. Chem., Int. Ed. 2005, 44,
7852.
(39) Ai, K.; Liu, Y.; Lu, L.; Cheng, X.; Huo, L. J. Mater. Chem. 2011,
21, 3365.
(40) Jin, Z.; Xu, B.; Hammond, G. B. Tetrahedron Lett. 2011, 52,
1956.
(3) Kulkarni, A.; Lobo-Lapidus, R. J.; Gates, B. C. Chem. Commun.
2010, 46, 5997.
(4) Corma, A. Catal. Rev.Sci. Eng. 2004, 46, 369.
(41) Drago, R. S.; Riley, R. J. Am. Chem. Soc. 1990, 112, 215.
3899
DOI: 10.1021/jacs.5b00222
J. Am. Chem. Soc. 2015, 137, 3894−3900

Journal of the American Chemical Society
Article
(42) Jeon, S.-H.; Xu, P.; Mack, N. H.; Chiang, L. Y.; Brown, L.;
Wang, H.-L. J. Phys. Chem. C 2010, 114, 36.
(43) Duffield, A. M.; Budzikiewicza, H.; Carl Djerassi, N. D. J. Am.
Chem. Soc. 1964, 86, 5536.
(44) Pathak, T.; Thomas, N. F.; Akhtar, M.; Gani, D. Tetrahedron
1990, 48, 1737.
(45) Yeshchenko, O. A.; Dmitruk, I. M.; Alexeenko, A. A.; Dmytruk,
A. M. Phys. Rev. B 2007, 75, 085434.
(46) Park, S.-H.; Chung, W.-H.; Kim, H.-S. J. Mater. Proc. Technol.
2014, 214, 2730.
(47) Kaptay, G.; Janczak-Rusch, J.; Pigozzi, G.; Jeurgens, L. P. H. J.
Mater. Eng. Perform. 2014, 23, 1600.
(48) Huo, K.-T.; Chen, X.-M. Mod. Phys. Lett. B 2014, 28, 1450157.
(49) Vaughan, O. P. H.; Kyriakou, G.; Macleod, N.; Tikhov, M.;
Lambert, R. M. J. Catal. 2005, 236, 401.
(50) Espinos
́
, J. P.; Morales, J.; Barranco, A.; Caballero, A.; Holgado,
J. P.; Gonzalez-Elipe, A. R. J. Phys. Chem. B 2002, 106, 6921.
́
(51) Zuo, J. M.; Kim, M.; O’Keeffe, M.; Spence, J. C. H. Nature 1999,
401, 49.
(52) Paine, A. J. J. Am. Chem. Soc. 1987, 109, 1496.
(53) Hernandez-Munoz, P.; Gavara, R.; Hernandez, R. J. J. Membr.
Sci. 1999, 154, 195.
(54) Domínguez, I.; Domen
́ ́ ́
ech-Carbo, A.; Cerisuelo, J. P.; Lopez,
G.; Hernandez, P.; Gavara, R. J. Solid State Electrochem. 2014, 18, 2099.
́
3900
DOI: 10.1021/jacs.5b00222
J. Am. Chem. Soc. 2015, 137, 3894−3900