Copper Nanoparticles Stabilized in a Porous Chitosan Aerogel as a Heterogeneous Catalyst for C-S Cross-coupling

Article abstract of DOI:10.1002/cctc.201500565

Cu nanoparticles (NPs) embedded into chitosan fibrils act as a heterogeneous catalyst for the C-S coupling of aryl halides and thiophenol in toluene. The catalyst can be reused under optimal conditions with some activity decrease from the first to the second use. Characterization of the reused catalyst shows that Cu NPs do not undergo agglomeration. Cu-chitosan is more active for aryl iodides than aryl bromides and chlorides. The presence of chitosan in a large excess is detrimental for the catalytic activity, probably because of the increase of solvent viscosity. It was found that halides formed as byproducts in the reaction act as a catalyst poison and may influence the extent of Cu leaching.

Full text of DOI:10.1002/cctc.201500565

DOI: 10.1002/cctc.201500565
Full Papers
Copper Nanoparticles Stabilized in a Porous Chitosan
Aerogel as a Heterogeneous Catalyst for CÀS Cross-
coupling
Sana Frindy,[a, b] Abdelkrim el Kadib,[c] Mohamed Lahcini,[b] Ana Primo,*[a] and
Hermenegildo García*[a]
Cu nanoparticles (NPs) embedded into chitosan fibrils act as
a heterogeneous catalyst for the CÀS coupling of aryl halides
and thiophenol in toluene. The catalyst can be reused under
optimal conditions with some activity decrease from the first
to the second use. Characterization of the reused catalyst
shows that Cu NPs do not undergo agglomeration. Cu-chito-
san is more active for aryl iodides than aryl bromides and
chlorides. The presence of chitosan in a large excess is detri-
mental for the catalytic activity, probably because of the in-
crease of solvent viscosity. It was found that halides formed as
byproducts in the reaction act as a catalyst poison and may in-
fluence the extent of Cu leaching.
Introduction
Heterogeneous catalysis using metal nanoparticles (MNPs) has
become a very important tool to promote many organic reac-
tions.[1,2] The stabilization of metastable metallic entities
against aggregation, with the minimal alteration of the intrinsic
catalytic activity, is one of the key points to design highly dura-
ble reactive MNPs. Generally, this fundamental prerequisite can
be attained by confining MNPs in porous solid supports,
wherein the limited available space in the cavities impedes
MNP growth and possible interactions with the support may
prevent metal leaching from the solid to the liquid phase.[3,4]
Besides inorganic metal oxides, organic polymers have been
used frequently as insoluble MNP supports,[5,6] and synthetic
polymers are used preferentially in comparison to natural bio-
polymers. However, the last two decades have witnessed an in-
creasing interest in the exploitation of the properties of natural
biopolymers in catalysis, particularly with regard to sustainabili-
ty and environmental issues.[7,8] Among natural biopolymers,
chitosan, a polysaccharide of glucosamine obtained from the
partial deacetylation of chitin, meets many required criteria to
be a valuable solid support in the field of catalysis. For in-
stance, although amorphous resins suffer from the burying of
catalytic sites within the polymer backbone, which thereby
limits the access of reactants, chitosan can be shaped as micro-
spherical beads in which the nanofibrils of the polymer are en-
tangled to form an open 3D network with remarkably high
macroporosity and surface area compared to synthetic poly-
mers.[9] In addition, chitin, from which chitosan is extracted, is
the most abundant waste residue in the fishery industry. Chito-
san has proven its efficiency for metal ion removal from pollut-
ed media and as an insoluble support for transition-metal com-
plexes and active noble MNPs.[10–15]
The increasing interest for the valorization of biomass
wastes as well as the lack of toxicity in these renewable resour-
ces prompted us to explore the potential use of chitosan as
a support for MNPs. We report the use of chitosan as support
of Cu nanoparticles (NPs) and the catalytic activity of the re-
sulting material for the CÀS coupling of thiophenol with aryl
halides. Coupling reactions catalyzed by transition metals are
currently among the most versatile reactions in organic synthe-
sis[16] and they can be applied not only for cross-coupling CÀC
reactions but also for the coupling of carbon with heteroa-
toms, which include O, N, P, and, in particular, S.[14,17] However,
one drawback of the current processes is that many of these
coupling reactions are still based on the use of Pd and other
noble metals as catalysts. Hence, for economic reasons, there
is an urgent need to find an alternative to Pd catalysts based
on first-row transition metals.[18,19] In the present case, the CÀS
coupling has been performed using Cu as the active metal. If
we consider that Cu interacts strongly with amines and N-ter-
minated ligands,[13] we speculate that the presence of sequen-
tial glucosamine units within the chitosan backbone will make
this natural biopolymer an adequate support for Cu NPs to
form a leaching-resistant catalyst. Precedents in the literature
have already shown the use of chitosan as a support of Cu
[a] S. Frindy, Dr. A. Primo, Prof. H. García
Instituto Universitario de Tecnología Química (CSIC-UPV)
Av. de los Naranjos s/n, 46022 Valencia (Spain)
E-mail: aprimoar@itq.upv.es
hgarcia@qim.upv.es
[b] S. Frindy, Prof. M. Lahcini
Laboratory of Organometallic and Macromolecular
Chemistry-Composites Materials
Faculty of Sciences and Technologies
Cadi Ayyad University
Avenue Abdelkrim Elkhattabi, B. P. 549,40000 Marrakech (Morocco)
[c] Prof. A. el Kadib
Euromed Research Institute, Engineering Division
Euro-Mediterranean University of Fes (UEMF), F›s-Shore
Route de Sidi Hrazem, 30070 F›s (Morocco)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201500565.
ChemCatChem 2015, 7, 3307 – 3315
3307
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
complexes or NPs as heterogeneous catalysts for redox and cy-
cloaddition reactions.[20–23] In a related study, CuII and CuI salts
were embedded onto chitosan and used for the coupling of
aryl halides and sodium sulfinates to form aryl sulfones.[24]
high specific surface area and porosity to be obtained. In the
present case, the BET surface area measured by N2 adsorption
isotherm was ꢀ187 m2 gÀ1, which is significantly lower than
the BET surface area of native chitosan, for which a similar
drying process allows values around 300 m2 gÀ1 to be reached.
This decrease in the surface area is probably a consequence of
the presence of Cu NPs that block some of the pores that
would be present in pristine chitosan microspheres.
Results and Discussion
In the first part of this report, we describe catalyst preparation
and provide analytical and spectroscopic evidence for the for-
mation of Cu NPs embedded into the fibrils of chitosan micro-
spheres. In the second part, we describe our investigation of
the catalytic activity of these materials for CÀS cross-coupling.
SEM images of the Cu-chitosan microspheres show the po-
rosity of the bead (Figure 1). Higher magnification shows that
Catalyst preparation and characterization
The preparation of Cu-chitosan microspheres was performed
in a stepwise manner (Scheme 1). This multistep preparation
method consists of the solvothermal formation of Cu NPs start-
ing from Cu(NO3)2 dissolved in ethylene glycol and reduction
in situ at 1508C. The solvothermal reduction of Cu2+ by heat-
ing in ethylene glycol, generally known as the polyol reduction
method, is well documented in the literature and is known to
render Cu NPs of a small average particle size between 1 and
5 nm.[25,26] These Cu NPs were supported on chitosan by
adding an acidic aqueous solution of this biopolymer to the
ethylene glycol solution in which Cu NPs were formed previ-
ously, and the mixture was stirred to obtain a homogeneous
hydrogel. The formation of highly porous Cu-chitosan micro-
spheres was finally achieved by dropping soluble Cu NPs-chito-
san hydrogel into a basic aqueous solution, which causes the
instantaneous precipitation of the microspheres. These wet
Cu-chitosan microspheres were dried by the gradual replace-
ment of water by ethanol, followed by supercritical CO2 proc-
essing. Previous studies have shown that before supercritical
drying, water has to be exchanged completely by ethanol.[14,27]
Otherwise, the direct removal of water affords shrunken xero-
gels that feature a very low porosity and low surface area be-
cause of the operation of strong capillary forces and hydrogen
bonding interactions during sudden water removal, which lead
to the collapse of porosity in chitosan.[27]
Figure 1. SEM image of Cu-chitosan microspheres that shows the intensity
along the bead of the Ka1 line characteristic of Cu in EDX superimposed in
blue. The intensity of the Ka1 line is proportional to Cu concentration.
the microspheres are constituted by interwoven fibrils of chito-
san polymer that leave an empty space between them. The
small size of the Cu NPs prepared by the solvothermal method
does not allow the detection of these MNPs at the highest
possible magnification used in the SEM study. However, ele-
mental mapping using energy dispersive X-ray (EDX) analysis
detects the presence of Cu distributed quasi-homogeneously
throughout the microspheres, although there is some higher
concentration on the external surface of the bead (Figure 1). In
other studies, it was found that MNPs are located preferentially
near the external surface, and a gradual decrease of the con-
centration of the metal is observed on going deeper into the
sub-millimeter particle.[14] In the present case, the fact that the
beads have been prepared by the intimate mixing of Cu NPs
in dissolved chitosan results apparently in a better dispersion
of Cu NPs in the whole polymer bead that are located not only
near the external surface but also in similar proportions at the
center of the microspheres.
Three samples were prepared in which the Cu loading was
varied from 0.5 to 2.3 wt% by controlling the amount of
Cu(NO3)2 reduced by ethylene glycol. The above experimental
preparation procedure, particularly the supercritical CO2 drying
of alcogels is similar to that used previously for the preparation
of chitosan-based solid catalysts.[13–15,27,28] These reports have
shown that supercritical drying allows chitosan particles with
The XRD pattern of Cu-chitosan does not show any peaks
characteristic of chitosan because of the amorphous nature of
this natural biopolymer. In con-
trast, despite the low loading of
Cu in these materials, its pres-
ence is detectable by XRD, and
weak peaks at 2q=43.25, 50.34,
and 74.048 were recorded and
assigned unambiguously to the
crystalline [111], [200], and
Scheme 1. Procedure for the preparation of Cu-chitosan; i) addition of chitosan solution to preformed Cu NPs,
ii) addition of chitosan that contains Cu NPs into a NaOH solution, iii) gelification of the chitosan, and iv) super-
critical CO2 drying.
[220] facets of metallic Cu NPs
(Figure S1). The average particle
ChemCatChem 2015, 7, 3307 – 3315
www.chemcatchem.org
3308
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
size for the three Cu-chitosan samples under study with Cu
loadings of 0.5, 1, and 2.3 wt% could be determined by apply-
ing the Scherrer equation using the most intense [111] peak of
the diffractogram, and the value was similar for the three sam-
ples (ꢀ0.5 nm). Thus, no influence of Cu loading on the aver-
age particle size at the loading range in this study was ob-
served. Notably, although XRD does not rule out the presence
of some Cu atoms in higher oxidation states, it proves the
presence of Cu0 clearly as expected based on previous reports
of the solvothermal reduction of Cu in ethylene glycol.[29–31]
The Cu-chitosan beads were also characterized by high-reso-
lution TEM. This study reveals the presence of homogeneous,
ultra-small Cu NPs of 1–2 nm, located throughout the entire
bead. Representative TEM images of Cu-chitosan are shown in
Figure 2, in which the homogeneous particle size of Cu NPs
Figure 3. XPS peaks that correspond to Cu-chitosan and the best deconvolu-
tion to individual components. Left: fresh Cu-chitosan; right: Cu-chitosan
after four consecutive uses as a catalyst in the CÀS cross-coupling of thio-
phenol and iodobenzene.
Figure 2. TEM image of a Cu-chitosan microsphere that shows the presence
of uniform, small Cu NPs. The inset shows the particle size distribution esti-
mated by counting a statistically relevant number of Cu NPs.
tion of the N atom to the metal centers (NH2!M, M=Ti, Zr, Al,
V, W, Mo) has been proposed.[15,33,34] The spectrum of the fresh
Cu-chitosan catalyst shows a weak Cu2p peak that can be de-
convoluted into two individual components with maxima at
BE=931.5 (0.07%) and 934.3 eV (0.1%) that correspond to the
contributions of CuI and CuII, respectively. The assignment of
CuI was confirmed by observation of the corresponding Auger
peak with a kinetic energy of 916.5 eV. Notably, there is an ap-
parent discrepancy between the detection of CuI and CuII by
XPS and Cu0 NPs by XRD and TEM. This discrepancy has been
observed frequently in metals and is because XPS is a surface
technique that only reports on the shallow thin layer of the
material of a few nm, whereas the other techniques report on
the bulk of the material. Thus, most probably exposure to the
air has formed a thin layer of Cu oxides on the external surface
of Cu-chitosan. For this reason and to ensure the reproducibili-
ty of the catalytic results, Cu-chitosan materials were submit-
ted to a preactivation step immediately before their use as cat-
alyst that consists of reduction with H2 at 2008C.
After four consecutive uses as a catalyst for the CÀS cou-
pling of thiophenol and iodobenzene, significant changes
were observed by XPS both in Cu NPs and the chitosan sup-
port. The most remarkable changes are related to the signifi-
cant decrease in the percentage of external N atoms, the
atomic proportion of which decreases from 5 to 0.1%, and at
the same time the XPS peak splits and exhibits a new compo-
nent at a significantly lower BE of 396.4 eV, which corresponds
to N atoms not coordinated to Cu. This decrease in the N con-
and their distribution in the biopolymer matrix can be ob-
served. A comparison of the Cu NP size distribution of fresh
Cu-chitosan and after its use as a catalyst does not show signif-
icant changes. This indicates that there is no evident agglom-
eration of Cu NPs under the reaction conditions. There are
precedents in the literature that show that Pd NPs embedded
on chitosan also exhibit an average particle size of ꢀ1 nm.[32]
Notably, the average particle size determined by XRD and TEM
are within the same range, ꢀ1 nm, which is close to the reso-
lution limit of TEM.
In the case of the Cu-chitosan catalyst, an insightful under-
standing of the Cu NPs–chitosan interaction in the fresh
sample and the changes undergone under the reaction condi-
tions was gained by analysis of the X-ray photoelectron spec-
tra (XPS; Figure 3). For the fresh Cu-chitosan sample, this tech-
nique shows the expected C1s, N1s, and O1s peaks, which
correspond to chitosan, together with a Cu2p peak. The XPS
C1s peak can be deconvoluted into three components of a rel-
ative proportion of 23, 30, and 6%, the positions of which are
attributed to C bonded to C (binding energy (BE)=284.5 eV), C
bonded to N (BE=286.7 eV), and C bonded to O (BE=
288.5 eV). Furthermore, the spectrum of the fresh catalyst
shows the N1s peak at BE=400.0 eV. This shift to a higher
binding energy (compared to BE=399 eV for NH2 in native chi-
tosan) suggests a dative coordination of NH2 to Cu NPs (NH2!
Cu). A similar shift for the N1s BE has been observed for a set
of chitosan-metal oxide hybrid materials for which the dona-
ChemCatChem 2015, 7, 3307 – 3315
www.chemcatchem.org
3309
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
tent is also reflected in the C1s peak, which becomes simpler,
and the contribution of C atoms bonded to N observed previ-
ously is almost absent. Besides the change in chitosan, the XPS
Cu2p peak also undergoes remarkable changes (Figure 3). Al-
though the total concentration of Cu atoms at the surface
does not vary and remains constant, the contribution of the
CuII component at BE=934.76 eV almost disappears and only
the peak for CuI at BE=931.5 eV remains. This change could
be the result of the activation of the catalyst by H2 at 2008C
before the reaction and/or to the reduction by IÀ ions formed
during the reaction that leads to some I2 and CuI. In any case,
XPS evidences clearly that as consequence of the reaction, the
external surface of chitosan loses N atoms and that the active
sites of the reaction seem to be Cu in low oxidation states,
probably Cu0 and CuI.
Figure 4. Time–conversion plots for the CÀS coupling of thiophenol and io-
dobenzene catalyzed by Cu-chitosan as a function of the Cu loading on chi-
~
&
*
tosan. conversion and yield for Cu-chitosan with 2.3 wt% of Cu; con-
version and + yield for Cu-chitosan with 1 wt% of Cu; * conversion and ”
&
yield for Cu-chitosan with 0.5 wt% of Cu. Inset: diphenyl sulfide yield from
*
2 mol% Cu-chitosan as catalyst; yield of diphenyl sulfide using 2 mol%
Further insight on the interplay between the Cu NPs and
chitosan support has been gained by FTIR spectroscopy.
Indeed, besides the typical signals of the carbohydrate scaffold
recognized in Cu-chitosan, the band assigned to ÀNH2 disap-
peared, and we observed a new shoulder at n˜ =1557 cmÀ1,
which indicates a weak interaction of the NH2 fragment with
Cu NPs. This shift to lower values is typical for hybrid materials
that exhibit an interaction between chitosan and its metal or
metal oxide partner. As an illustration, a shift to n˜ =1520 cmÀ1
has been observed for chitosan-vanadium oxide[15] and for chi-
tosan-stabilized Au NPs,[14] in which NH2!V and NH2!Au take
place, respectively. Similar values were also observed for chito-
san-sepiolite[35] and chitosan-montmorillonite hybrid materi-
als.[36]
~
Cu-chitosan with 50 wt% chitosan; yield of diphenyl sulfide in the pres-
ence of 2 mol% Cu-chitosan with 100 wt% chitosan. Reaction conditions:
1 mmol of iodobenzene, 1.2 mmol of thiophenol, 3 mL of toluene, 1308C,
3 mmol of NEt3, 2 mol% of catalyst.
ieved for low loadings. In this case, it could be that the
amount of chitosan plays a negative role in the catalytic reac-
tion. To maintain the Cu/substrate molar ratio while we vary
the Cu loading on the support, the amount of chitosan sup-
port has to be over four times higher at 0.5 wt% loading than
that at 2.3 wt% in Cu-chitosan.
After we selected the most convenient Cu NP loading on
chitosan as 2.3 wt%, additional experiments to optimize the
Cu/substrate mol ratio were performed (Figure 5). Indeed, the
Catalytic study
Three samples of Cu-chitosan that have Cu loadings of 0.5, 1,
and 2.3 wt%, determined by inductively coupled plasma opti-
cal emission spectroscopy (ICP-OES), were evaluated as cata-
lysts for the coupling of thiophenol and iodobenzene to form
diphenyl sulfide in toluene as the solvent at 1308C (Scheme 2).
The time–conversion plots for these three catalysts are shown
Figure 5. Time–yield plots for the reaction of thiophenol and iodobenzene
*
catalyzed by Cu-chitosan as a function of Cu loading on the support
~
&
2 mol%; 4 mol%; 1 mol%. Reaction conditions: 1 mmol of iodobenzene,
1.2 mmol of thiophenol, 3 mL of toluene, 1308C, 3 mmol of NEt3.
Scheme 2. CÀS coupling of thiophenol and aryl halides to form diaryl
initial reaction rates at Cu/substrate molar ratios of 1 and 4%
are identical, and the optimal Cu/substrate molar ratio is 2%.
Again, this effect of the increase of Cu/substrate molar ratio
unusual as it is expected that the reaction rate should increase
along with the Cu/substrate molar ratio. The fact that there is
an optimal ratio of 2 mol% could be the consequence of two
opposite effects. On one hand, the reaction rate should in-
crease with the amount of Cu present, but on the other hand
the presence of higher amounts of chitosan should be detri-
mental as stated previously. Therefore, a balance between
these two opposite effects leads to a compromise, and the
sulfides.
in Figure 4. The initial reaction rate and performance of the
catalyst increase with the Cu loading; the best-performing Cu-
chitosan sample is the one with the highest Cu loading. This
pattern contrasts with the typical trend observed commonly
for supported MNPs in which the catalytic activity generally in-
creases as the loading of metal on the support decreases if the
metal/substrate molar ratio is maintained because of the small-
er particle size and higher metal dispersion that can be ach-
ChemCatChem 2015, 7, 3307 – 3315
www.chemcatchem.org
3310
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
highest performance is reached at a Cu/substrate molar ratio
of 2%.
Cu leaching in toluene as a solvent and NEt3 as base was sig-
nificantly lower (2% of the initial Cu content of the catalyst,
25 mg in 3 mL) and at the same time exhibits an adequate ini-
tial reaction rate. It could be that compared to Pd, Cu is more
prone to undergo leaching and becomes partly dissolved in or-
ganic solvents. There are precedents in the literature that
show the absence of metal leaching in the case of Pd NPs em-
bedded in chitosan in xylene at 1308C for ꢀ7 h.[37] Accordingly,
toluene was selected for subsequent studies on the catalytic
activity of Cu-chitosan.
To provide experimental evidence in support of the negative
role that an excess of chitosan can play on the catalytic activity
of the Cu NPs for the CÀS coupling, a series of experiments
using Cu-chitosan as the catalyst to which additional amounts
of unmodified chitosan was added were performed. The results
are presented inset in Figure 4. As illustrated above, Cu-chito-
san in the absence of any added chitosan exhibits the highest
initial reaction rate and diphenyl sulfide yield at the final time.
The addition of 50 or 100 wt% of chitosan decreases both
the initial reaction rate and the diphenyl sulfide yield. This neg-
ative effect of chitosan addition increases, although not linear-
ly, if the amount of chitosan added was doubled. Therefore, it
can be concluded that although the presence of chitosan as
the support is necessary to confine and stabilize the Cu NPs to
allow their recovery and separation from the reaction mixture,
an increase in the amount in chitosan to values higher than
the minimum necessary plays an undesirable role by decreas-
ing the activity of the Cu catalyst. One possible way in which
chitosan can have a negative effect on the reaction is by in-
creasing the viscosity of the reaction medium or by adsorbing
reagents and (by)products, which blocks the Cu active sites.
After the optimization of Cu loading on chitosan and the
Cu/substrate molar ratio, the influence of solvent on the reac-
tion was studied. Notably, the variation of the nature of the
solvent is accompanied unavoidably by the variation of the
base used in the reaction because of solubility limitations. Spe-
cifically, the coupling of thiophenol and iodobenzene with
KOH in acetonitrile, Na2PO4 in 1,4-dioxane, and NEt3 in toluene
were screened. The temporal profile of the CÀS coupling reac-
tion in each of these solvent/base pairs is shown in Figure 6.
Indeed, the lowest reaction rate was observed for 1,4-dioxane,
and the fastest transformation took place in acetonitrile. Un-
fortunately, chemical analysis of the liquid phase after the re-
moval of the solid catalyst showed that the acetonitrile con-
tained a significant amount of the total Cu present initially in
the catalyst, which indicates the occurrence of a high degree
of Cu leaching. The occurrence of leaching is favored by rela-
tively long reaction times and high temperatures. In contrast,
The next parameter that was optimized was the reaction
temperature. As a result of its structure, there is a limitation of
the maximum temperature that chitosan can stand without
becoming dissolved or undergoing a substantial change of the
morphology of its particles.[38] In the present case, the influ-
ence of the temperature on the catalytic activity of Cu-chitosan
was determined in the range of 120–1508C, and the results are
presented in Figure 7. The initial reaction rate increases with
Figure 7. Influence of the temperature on the time–yield plot for the reac-
&
*
tion of thiophenol and iodobenzene catalyzed by Cu-chitosan 140, 130,
~
120, * 1508C; Reaction conditions: 1 mmol of iodobenzene, 1.2 mmol of
thiophenol, 3 mL of toluene, 3 mmol of NEt3, 2 mol% of catalyst. Inset: Ar-
rhenius plot of the logarithm of the initial reaction rate versus the reciprocal
of the absolute temperature.
the temperature from 120 to 1408C but then diminishes at
1508C. This lower initial reaction rate observed at 1508C is re-
lated to variations of the morphology of chitosan spheres at
this temperature and, accordingly, it represents the tempera-
ture limit in this case. The influence of temperature on the ini-
tial reaction rate in the range of 120–1408C allows us to deter-
mine the activation energy of the process (Ea =95.5 KJmolÀ1
)
from the Arrhenius plot (Figure 7, inset). In view of the above
data, 1308C was selected as a convenient temperature to per-
form CÀS couplings using Cu-chitosan as the catalyst.
According to the previous characterization data that show
the presence of low-oxidation Cu0 and CuI states, the most rea-
sonable reaction mechanism for the CÀS coupling compatible
with the literature[39] is proposed in Scheme 3. We propose
that PhÀI splits on the Cu NP surface to form CuÀPh and CuÀI
species. Similarly, in the presence of base, PhÀSÀ anions should
also bind to the Cu NP surface. Thus, the surface of Cu NP
should have a population of adsorbed phenyl, phenylthio, and
iodo species on different sites. CÀS coupling will take place on
Figure 6. Time–yield plot for the reaction of thiophenol and iodobenzene
*
catalyzed by Cu-chitosan in three combinations of solvent and base: ace-
~
&
tonitrile/KOH; toluene/Et3N; 1,4-dioxane/Na2PO4. Reaction conditions:
1 mmol of iodobenzene, 1.2 mmol of thiophenol, 3 mL of solvent, 1308C,
3 mmol of base, 2 mol% of catalyst.
ChemCatChem 2015, 7, 3307 – 3315
www.chemcatchem.org
3311
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
the initial Cu amount present in the catalyst). The extent of Cu
leaching was almost the same in the subsequent reactions
(2.23 and 2.17 wt%). Based on these analytical data, it appears
that the leaching of Cu into toluene is not responsible for the
change from the fresh to the used catalyst and it is more likely
that changes in the porosity and structure of the chitosan
beads should be responsible for this decrease in activity upon
reuse. It is proposed that a reorganization of the porosity takes
place in the first run and, then, the resultant Cu-chitosan is
considerably more stable from a catalytic point of view and
maintains its structure relatively unaltered in subsequent runs.
Below, we will comment on other deactivation pathways that
also occur as consequence of the formation of iodide as by-
product during the course of the reaction.
One general feature in cross-coupling reactions is the much
higher reactivity of iodides with respect to bromides, and aryl
chlorides are notoriously less reactive. This general trend is
a consequence of the CÀX bond strength and its influence on
the kinetics of the reaction in which the rate-determining step
involves the cleavage of this bond. As expected, Cu-chitosan
was inactive to promote the CÀS coupling of thiophenol with
bromo- or chlorobenzene. However, if a nitro group, which ac-
tivates the CÀX bond, is present, then CÀS coupling between
2- and 4-nitrobromobenzene and 4-nitrochlorobenzene takes
place. The initial reaction rate and final conversion depend on
the substrate and follow the order 4-nitrochlorobenzene<4-ni-
trobromobenzeneꢀ2-nitrobromobenzene<iodobenzene (Fig-
ure 9). Although the initial reaction rate should be a function
of the nature and strength of the CÀX bond, it was observed
that the final yield also follows the order of the initial rate.
The conversion of both 4-nitrobromobenzene and 4-nitro-
chlorobenzene stops after ꢀ25 h (Figure 9). This could indicate
catalyst deactivation. To verify this possibility and to under-
stand the origin of the possible deactivation, the reaction of
thiophenol and iodobenzene was selected as a model and
0.1 equivalents of NaX (X: IÀ, BrÀ, and ClÀ) was added on pur-
pose as a possible poison. As the CÀS coupling progresses, an
increasing build-up of the XÀ concentration should develop as
byproduct of the coupling, and it was reasoned that this
halide could act as a poison to deactivate the catalyst because
Scheme 3. Simplified mechanistic proposal for CÀS coupling on Cu NPs.
i) PhÀI splitting on low-coordinate Cu atoms on the surface; ii) Adsorption of
PhSÀ on the surface; iii) PhSPh desorption; iv) IÀ desorption.
the Cu NP surface, and the subsequent desorption of PhÀSPh
and IÀ will complete the catalytic cycle.
An important issue in heterogeneous catalysis is to show
the stability of the catalyst under the reaction conditions. To
test the reusability of Cu-chitosan in toluene, a series of con-
secutive reactions between iodobenzene and thiophenol were
performed using the same catalyst sample. At the end of each
reaction, the solid catalyst was recovered by filtration, washed
with fresh toluene, and used in a consecutive reaction under
the same conditions. The temporal profile of diphenyl sulfide
formation in consecutive runs is presented in Figure 8. Al-
though the initial reaction rate of the fresh catalyst was signifi-
cantly higher than that in the subsequent reuses, the differen-
ces in the initial reaction rate and yield at the final time from
the second to the fifth reuses were relatively minor. To under-
stand this pattern, chemical analysis of the Cu present in the
liquid phase in each of the consecutive runs was performed.
As noted earlier, the chemical analysis of toluene in the first
run revealed the presence of a small amount of Cu (2.3% of
&
*
Figure 8. Time–yield plot for the reuse of Cu-chitosan; run 1, run 2,
Figure 9. Time–yield plots for the coupling of thiophenol with substituted
~
~
&
*
run 3, * run 4; + run 5. Reaction conditions: 1 mmol of iodobenzene,
aryl halides. iodobenzene; 4-nitrobromobenzene; 4-nitrochloroben-
zene. Reaction conditions: 1 mmol of aryl halide, 1.2 mmol of thiophenol,
3 mL of toluene, 1308C, 3 mmol of NEt3, 2 mol% of catalyst.
1.2 mmol of thiophenol, 3 mL of toluene, 1308C, 3 mmol of NEt3, 2 mol% of
Cu catalyst.
ChemCatChem 2015, 7, 3307 – 3315
www.chemcatchem.org
3312
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Table 1. Influence of NBu4Cl concentration on the percentage of Cu that
leaches into toluene.
ClÀ conc.
Initial Cu
amount [mg]
Cu amount
leached [mg]
Proportion of initial
Cu leached to toluene [%]
[ppm]
0
0.25
1.25
0.7
0.7
0.7
0
1.1
3.4
0
0.15
0.48
that can facilitate the dissolution and leaching of Cu. This
effect should be especially important in the case of ClÀ for
which the interactions with Cu are stronger.
Figure 10. Influence of the absence or presence of NaX on the time–yield
plot for the reaction of thiophenol and iodobenzene catalyzed by Cu-chito-
~
&
*
san: without base; NaI; NaBr; * NaCl. Reaction conditions: 1 mmol of
iodobenzene, 1.2 mmol of thiophenol, 3 mL of toluene, 1308C, 3 mmol of
NEt3, 2 mol% of catalyst.
To gain a deeper insight into the possible changes that
occur in Cu-chitosan during catalysis, the solid catalyst was re-
covered after the fourth run and characterized by TEM and
XRD. Selected TEM images of the Cu-chitosan used four times
and EDX analysis of Cu-chitosan after this consecutive reuse as
a catalyst are presented in Figure 11. As in the case of the
fresh Cu-chitosan material, EDX elemental mapping of the mi-
crosphere indicates that Cu is still present homogeneously and
highly dispersed in the form of nanometric particles through-
of a strong CuÀX interaction. The results of the influence of
NaX on the catalytic activity of Cu-chitosan are shown in
Figure 10. The initial reaction rate and product yield at 6 h de-
crease gradually in the order of XÀ >IÀ >BrÀ >ClÀ. These kinet-
ic data support the role of the halide as a catalyst poison, and
the poisoning effect increases with the strength of the CuÀX
interaction.
According to these data, aryl chlorides as substrates are not
only the least reactive because of the strength of the CÀCl
bond but also the chloride formed as a byproduct displays
a strong inhibitory effect attributed to its strong CuÀCl bond
compared to the other acyl halides. This poisoning effect is re-
sponsible for the deactivation of the catalyst as the reaction
progresses, which stops the reaction before it reaches the
equilibrium concentration.
Another possible effect of the interaction of halides with Cu
NPs is to increase the Cu leaching from the solid to the solu-
tion. In MNPs and particularly in catalysis by supported Au NPs
in which HAuCl4 is the most frequent precursor, it has been
found that the presence of residual amounts of ClÀ from the
precursor is highly detrimental from the point of view of the
stability of the catalyst.[40,41] The presence of ClÀ on the Au cat-
alyst increases the Au NPs particle size, an effect that has been
attributed to the ability of ClÀ to mobilize metal atoms.[42] We
were also interested to determine if besides deactivation, the
presence of ClÀ as a mobilizing agent increases Cu leaching
from the Cu-chitosan solid to toluene. To check this possibility,
a series of experiments in which Cu-chitosan was heated at
the reaction temperature in toluene that contained increasing
amounts of (Bu)4NCl was performed. This quaternary ammoni-
um chloride was selected to overcome the solubility problems
of NaCl and other inorganic chlorides in toluene. After 6 h re-
action, the solid catalyst was removed by filtration, and the
clear toluene solution was analyzed for its Cu content. The re-
sults are presented in Table 1 and they show clearly that al-
though no Cu was detected in the liquid phase in the absence
of ClÀ, the Cu content in toluene increases as the concentra-
tion of ClÀ increases. Accordingly, it can also be expected that
one of the main reasons for Cu leaching in the present CÀS
coupling reaction could be the complexation of Cu with XÀ
Figure 11. TEM and bright-field SEM images of four times reused Cu-chito-
san. The images show that Cu is distributed all over the chitosan beads.
out the beads. In XRD, fresh Cu-chitosan shows the pattern
that corresponds to metallic Cu, although the peaks are weak
(Figure S1). In contrast, after reuse this diffraction peak is
barely detectable or disappears (Figure S1). This could indicate
either a decrease on the average Cu NP size or the occurrence
of Cu leaching with a decrease of the total Cu loading. The
second possibility seems less likely if we consider that chemical
analysis of the solution showed that the accumulated amount
of Cu leaching in the fourth reuse was below 10% of the initial
Cu content (127 mg leached Cu out from 1.27 mg Cu-chitosan).
In any case, it can be concluded that the Cu NPs do not ag-
glomerate during the catalytic reaction.
Finally, the scope of the reaction was screened beyond the
parent Ph2S and nitro derivative by performing a series of CÀS
couplings using Cu-chitosan as the catalyst, and the results are
shown in Table 2. The CÀS coupling promoted by Cu-chitosan
seems to be general for aryl iodides, although the yield de-
pends on the substituent, and electron-withdrawing substitu-
ChemCatChem 2015, 7, 3307 – 3315
www.chemcatchem.org
3313
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
ized by GC–MS and H and 13C NMR spectroscopy of the reaction
mixtures.
1
Table 2. Results of the CÀS coupling catalyzed by Cu-chitosan. Reaction
conditions: 0.5 mmol of iodobenzene, 0.6 mmol of thiophenol, 3 mL of
toluene, 1308C, 3 mmol of NEt3, 2 mol% of catalyst.
Characterization
Entry
Substrate
X-p-C6H4-Y
CÀS coupling
product yield [%]
SEM images were acquired by using a JEOL JSM 6300 apparatus.
TEM images were recorded by using a Philips CM 300 FEG system
with an operating voltage of 100 kV. Diffuse reflectance infrared
Fourier transform spectroscopy (DRIFTS) was performed by using
a NICOLET iS10 spectrometer. XPS were recorded by using a SPECS
spectrometer equipped with a Phoibos 150 9MCD detector using
a non-monochromatic X-ray source (Al and Mg) operating at
200 W. The samples were evacuated in the prechamber of the
spectrometer at 1”10À9 mbar. Some of the samples were activated
in situ by N2 flow at 4508C for 3 h, followed by evacuation at
10À8 mbar. The measured intensity ratios of the components were
obtained from the area of the corresponding peaks after nonlinear
Shirley-type background subtraction and corrected by the trans-
mission function of the spectrometer. Quantitative ICP-OES meas-
urements were performed by using a 715-ES Varian apparatus. BET
analysis was performed by using a Micromeritics ASAP2420 instru-
ment. GC was performed by using a Bruker 430-GC instrument
equipped with a flame-ionization detector (FID) and a HP-5MS
column (30 m”0.25 mm”0.25 mm), the stationary phase of which
is constituted by cross-linked 5% phenylmethylsilicone, using n-do-
decane as standard. A known weight of n-dodecane was added to
an aliquot of the reaction diluted in CH2Cl2, and the resulting solu-
tion (10 mL) was injected into the GC (2508C injector temperature).
The product yields were determined by 1H NMR spectroscopy of
the reaction mixture after the addition of p-dimethoxybenzene as
the internal standard by using a 300 MHz Varian Geminis Instru-
ment and CDCl3 as the solvent.
1
2
3
4
5
6
7
8
X=I; Y=H
96;[a] 92[b]
X=Br; Y=H
X=I; Y=NO2
X=Br; Y=NO2
X=Cl; Y=NO2
X=I; Y=F
0
99;[a] 88[b]
84;[a] 77[b]
46;[a] 40[b] (isolated 38)
86[b]
87[b]
44[b]
X=I; Y=COCH3
X=I; Y=OCH3
[a] Determined by GC using dodecane as internal standard; [b] Deter-
mined by 1H NMR spectroscopy of the reaction mixture using p-dime-
thoxybenzene as internal standard.
ents afford higher yields. Unactivated phenyl bromide fails to
give the CÀS coupling product. In contrast, nitro-substituted
aryl halides render the p-nitro diphenyl sulfide in various
yields, which depends on the halide, as commented on earlier.
Conclusions
We have shown that highly dispersed Cu nanoparticles (NPs)
can be prepared and embedded within the chitosan fibril
matrix. The resulting microspheres exhibit catalytic activity for
CÀS coupling. The presence of a large amount of chitosan is
detrimental for the catalytic activity, probably because of solu-
tion viscosity changes or reagent adsorption and, accordingly,
there is an optimal value for the Cu loading on chitosan and
the Cu/substrate molar ratio. The catalyst can be reused under
optimal conditions at least four times, but a decrease in the ac-
tivity from the first to the second use is noticed accompanied
by minor Cu leaching. Cu-chitosan is more active for aryl io-
dides than aryl bromides and chlorides. Catalytic tests have
shown that halides released during the reaction may act as
a poison for Cu NPs and as mobilizers of Cu atoms, which in-
crease the leaching of Cu to the solution, particularly ClÀ. The
characterization of the catalyst after use reveals the absence of
Cu NP agglomeration, but a decrease in the average particle
size is observed based on the disappearance of the characteris-
tic Cu peaks in the XRD pattern. If we consider the wide availa-
bility of chitosan, which can be obtained in large amounts as
waste of the fishery industry, and that Cu is a widely abundant
transition metal, Cu-chitosan has many advantages from a sus-
tainability point of view and exhibits an adequate activity for
the CÀS coupling with aryl iodides.
Catalyst preparation
Preparation of Cu NPs
Cu NPs were synthesized by solvothermal reduction in ethylene
glycol. Briefly, Cu(NO3)2 (0.054 g) was dissolved in ethylene glycol
(0.5 mL). The mixed solution was heated for 24 h at 1508C under
reflux. After the solution was cooled, a chitosan solution (0.5 g of
chitosan in 28 mL of 1% v/v acetic acid solution) was added drop-
wise, and the mixture was stirred until the total dispersion of Cu.
Then, this solution was added dropwise into a NaOH solution (4m)
by using a syringe with a 0.8 mm needle. The Cu-chitosan micro-
spheres were aged in the alkaline solution for 24 h and then fil-
tered and washed until a neutral pH solution was obtained. The
hydrogel microspheres were first dehydrated by immersion for
30 min with stirring in a series of successive ethanol/water baths of
increasing alcohol concentration (10, 30, 50, 70, 90, and 100 v/v%).
Finally, Cu-chitosan aerogel microspheres were formed by super-
critical CO2 drying (73.8 bar, 31.58C) by using a Polaron 3100 appa-
ratus.
CÀS coupling reaction
Experimental Section
Typically, iodobenzene (1 mmol), thiophenol (1.2 mmol), and Et3N
(3 mmol) were dissolved in toluene (3 mL). Then, the catalyst
(2 mol% of Cu respect to iodobenzene) was added. The reaction
mixture pressurized with Ar (5 bar) was stirred and placed in a pre-
heated silicon bath and heated at 1308C for 60 h. The reaction
progress was monitored by periodic analysis of aliquots of the
liquid phase by GC using n-dodecane as the external standard.
General remarks
Commercially available reagents and solvents were purchased
from Across and Aldrich and used without further purification. Chi-
tosan of medium molecular weight and a deacetylation degree of
80% was purchased from Sigma–Aldrich. All the CÀS coupling
products have been reported in the literature and were character-
ChemCatChem 2015, 7, 3307 – 3315
www.chemcatchem.org
3314
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[17] G. Y. Li, Angew. Chem. Int. Ed. 2001, 40, 1513–1516; Angew. Chem. 2001,
113, 1561–1564.
Reusability of heterogeneous Cu-chitosan catalysts
After the completion of the first cycle as described above, the Cu-
chitosan microspheres were recovered from the reaction mixture
by filtration and washed with toluene three times. The recycled
Cu-chitosan catalyst was used again in the next coupling reaction
under the same reaction conditions.
[18] N. Miyaura, A. Suzuki, Chem. Rev. 1995, 95, 2457–2483.
[19] F. S. Han, Chem. Soc. Rev. 2013, 42, 5270–5298.
[20] R. B. N. Baig, R. S. Varma, Green Chem. 2013, 15, 1839–1843.
[21] Y. Chang, Y. Wang, F. Zha, R. Wang, Polym. Adv. Technol. 2004, 15, 284–
286.
[22] E. Guibal, Prog. Polym. Sci. 2005, 30, 71–109.
[23] A. V. Kucherov, N. V. Kramareva, E. D. Finashina, A. E. Koklin, L. M. Kustov,
J. Mol. Catal. A 2003, 198, 377–389.
[24] C. Shen, J. Xu, W. Yu, P. Zhang, Green Chem. 2014, 16, 3007–3012.
[25] K. J. Carroll, J. U. Reveles, M. D. Shultz, S. N. Khanna, E. E. Carpenter, J.
Phys. Chem. C 2011, 115, 2656–2664.
[26] B. K. Park, S. Jeong, D. Kim, J. Moon, S. Lim, J. S. Kim, J. Colloid Interface
Sci. 2007, 311, 417–424.
Acknowledgements
Financial support by the Spanish Ministry of Economy and Com-
petitiveness (Severo Ochoa and CTQ-2012-32315) is gratefully ac-
knowledged. S.F. thanks the EMAG for a fellowship to support her
stage at Valencia.
[27] R. Valentin, K. Molvinger, F. Quignard, D. Brunel, New J. Chem. 2003, 27,
1690–1692.
[28] A. Primo, M. Liebel, F. Quignard, Chem. Mater. 2009, 21, 621–627.
[29] C.-Y. Lu, M.-Y. Wey, Y.-H. Fu, Appl. Catal. A 2008, 344, 36–44.
[30] W. Yu, H. Xie, L. Chen, Y. Li, C. Zhang, J. Dispersion Sci. Technol. 2010, 31,
364–367.
Keywords: copper · cross-coupling · heterogeneous catalysis ·
nanoparticles · polymers
[31] H. Kawasaki, R. Arakawa, 32 pp., Chemical Indexing Equivalent to
161:758379 (JP), 2013.
[32] M. Adlim, M. Abu Bakar, K. Y. Liew, J. Ismail, J. Mol. Catal. A 2004, 212,
141–149.
[1] A. Corma, H. García, Chem. Soc. Rev. 2008, 37, 2096–2126.
[2] D. Astruc, F. Lu, J. Ruiz Aranzaes, Angew. Chem. Int. Ed. 2005, 44, 7852–
7872; Angew. Chem. 2005, 117, 8062–8083.
[3] A. Roucoux, J. Schulz, H. Patin, Chem. Rev. 2002, 102, 3757–3778.
[4] Y. Wang, Z. Xiao, L. Wu, Curr. Org. Chem. 2013, 17, 1325–1333.
[5] M. Krµlik, A. Biffis, J. Mol. Catal. A 2001, 177, 113–138.
[6] J. Lu, P. H. Toy, Chem. Rev. 2009, 109, 815–838.
[7] L. F. Xiao, F. W. Li, C. G. Xia, Appl. Catal. A 2005, 279, 125–129.
[8] A. Kadib, M. Bousmina, D. Brunel, J. Nanosci. Nanotechnol. 2014, 14,
308–331.
[9] A. El Kadib, M. Bousmina, Chem. Eur. J. 2012, 18, 8264–8277.
[10] M. G. Dekamin, M. Azimoshan, L. Ramezani, Green Chem. 2013, 15, 811 –
820.
[33] A. El Kadib, K. Molvinger, M. Bousmina, D. Brunel, J. Catal. 2010, 273,
147–155.
[34] A. Aqil, A. El Kadib, M. Aqil, M. Bousmina, A. Elidrissi, C. Detrembleur, C.
JØrôme, RSC Adv. 2014, 4, 33160–33163.
[35] M. Darder, M. López-Blanco, P. Aranda, A. J. Aznar, J. Bravo, E. Ruiz-
Hitzky, Chem. Mater. 2006, 18, 1602–1610.
[36] A. El Kadib, K. Molvinger, T. Cacciaguerra, M. Bousmina, D. Brunel, Micro-
porous Mesoporous Mater. 2011, 142, 301–307.
[37] B. C. E. Makhubela, A. Jardine, G. S. Smith, Appl. Catal. A 2011, 393,
231–241.
[11] P. K. Sahu, P. K. Sahu, S. K. Gupta, D. D. Agarwal, Ind. Eng. Chem. Res.
2014, 53, 2085–2091.
[12] D. A. Barskiy, K. V. Kovtunov, A. Primo, A. Corma, R. Kaptein, I. V. Kop-
tyug, ChemCatChem 2012, 4, 2031–2035.
[13] M. Chtchigrovsky, A. Primo, P. Gonzalez, K. Molvinger, M. Robitzer, F.
Quignard, F. Taran, Angew. Chem. Int. Ed. 2009, 48, 5916–5920; Angew.
Chem. 2009, 121, 6030–6034.
[14] A. Primo, F. Quignard, Chem. Commun. 2010, 46, 5593–5595.
[15] A. El Kadib, A. Primo, K. Molvinger, M. Bousmina, D. Brunel, Chem. Eur. J.
2011, 17, 7940–7946.
[38] A. El Kadib, ChemSusChem 2015, 8, 217–244.
[39] S. V. Ley, A. W. Thomas, Angew. Chem. Int. Ed. 2003, 42, 5400–5449;
Angew. Chem. 2003, 115, 5558–5607.
[40] A. S. K. Hashmi, G. J. Hutchings, Angew. Chem. Int. Ed. 2006, 45, 7896–
7936; Angew. Chem. 2006, 118, 8064–8105.
[41] G. J. Hutchings, Catal. Today 2005, 100, 55–61.
[42] B. Nkosi, N. J. Coville, G. J. Hutchings, M. D. Adams, J. Friedl, F. E.
Wagner, J. Catal. 1991, 128, 366–377.
Received: May 20, 2015
[16] G. Dyker, Angew. Chem. Int. Ed. 1999, 38, 1698–1712; Angew. Chem.
1999, 111, 1808–1822.
Published online on September 1, 2015
ChemCatChem 2015, 7, 3307 – 3315
www.chemcatchem.org
3315
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim