New Cu-based catalysts supported on TiO2 films for Ullmann SNAr-type C-O coupling reactions

Article abstract of DOI:10.1002/chem.201102151

New routes for the preparation of highly active TiO₂-supported Cu and CuZn catalysts have been developed for C-O coupling reactions. Slurries of a titania precursor were dip-coated onto glass beads to obtain either structured mesoporous or non-porous titania thin films. The Cu and CuZn nanoparticles, synthesized using a reduction by solvent method, were deposited onto calcined films to obtain a Cu loading of 2 wt %. The catalysts were characterized by inductively coupled plasma (ICP) spectroscopy, temperature-programmed oxidation/reduction (TPO/TPR) techniques, ⁶³Cu nuclear magnetic resonance (NMR) spectroscopy, X-ray diffraction (XRD), scanning and transmission electron microscopy (S/TEM-EDX) and X-ray photo-electron spectroscopy (XPS). The activity and stability of the catalysts obtained have been studied in the C-O Ullmann coupling of 4-chloropyridine and potassium phenolate. The titania-supported nanoparticles retained catalyst activity for up to 12 h. However, catalyst deactivation was observed for longer operation times due to oxidation of the Cu nanoparticles. The oxidation rate could be significantly reduced over the CuZn/TiO₂ catalytic films due to the presence of Zn. The 4-phenoxypyridine yield was 64 % on the Cu/nonporous TiO₂ at 120 °C. The highest product yield of 84 % was obtained on the Cu/mesoporous TiO₂ at 140 °C, corresponding to an initial reaction rate of 104 mmol g_cat^-1 s^-1. The activation energy on the Cu/mesoporous TiO₂ catalyst was found to be (144±5) kJ mol^-1, which is close to the value obtained for the reaction over unsupported CuZn nanoparticles (123±3 kJ mol^-1) and almost twice the value observed over the catalysts deposited onto the non-porous TiO₂ support (75±2 kJ mol^-1). Copyright

Full text of DOI:10.1002/chem.201102151

DOI: 10.1002/chem.201102151
New Cu-Based Catalysts Supported on TiO2 Films for Ullmann SNAr-Type
ꢀ
C O Coupling Reactions
Faysal Benaskar,[a] Volker Engels,[b] Evgeny V. Rebrov,[c] Narendra G. Patil,[a]
Jan Meuldijk,[a] Peter C. Thꢀne,[a] Pieter C. M. M. Magusin,[a] Brahim Mezari,[a]
Volker Hessel,[a] Lumbertus A. Hulshof,[a] Emiel J. M. Hensen,[a]
Andrew E. H. Wheatley,*[b] and Jaap C. Schouten*[a]
Abstract: New routes for the prepara-
tion of highly active TiO2-supported
Cu and CuZn catalysts have been de-
scanning and transmission electron mi-
croscopy (S/TEM-EDX) and X-ray
photo-electron spectroscopy (XPS).
The activity and stability of the cata-
lysts obtained have been studied in the
idation rate could be significantly re-
duced over the CuZn/TiO2 catalytic
films due to the presence of Zn. The 4-
phenoxypyridine yield was 64% on the
Cu/nonporous TiO2 at 1208C. The
highest product yield of 84% was ob-
tained on the Cu/mesoporous TiO2 at
1408C, corresponding to an initial reac-
tion rate of 104 mmolgcatꢀ1 sꢀ1. The ac-
tivation energy on the Cu/mesoporous
TiO2 catalyst was found to be (144ꢁ
5) kJmolꢀ1, which is close to the value
obtained for the reaction over unsup-
ported CuZn nanoparticles (123ꢁ
3 kJmolꢀ1) and almost twice the value
observed over the catalysts deposited
onto the non-porous TiO2 support
(75ꢁ2 kJmolꢀ1).
ꢀ
veloped for C O coupling reactions.
Slurries of a titania precursor were dip-
coated onto glass beads to obtain
either structured mesoporous or non-
porous titania thin films. The Cu and
CuZn nanoparticles, synthesized using
a reduction by solvent method, were
deposited onto calcined films to obtain
a Cu loading of 2 wt%. The catalysts
were characterized by inductively cou-
pled plasma (ICP) spectroscopy, tem-
perature-programmed oxidation/reduc-
tion (TPO/TPR) techniques, 63Cu nu-
clear magnetic resonance (NMR) spec-
troscopy, X-ray diffraction (XRD),
ꢀ
C O Ullmann coupling of 4-chloropyr-
idine and potassium phenolate. The ti-
tania-supported nanoparticles retained
catalyst activity for up to 12 h. Howev-
er, catalyst deactivation was observed
for longer operation times due to oxi-
dation of the Cu nanoparticles. The ox-
Keywords: copper · cross-coupling
reactions · mesoporous materials ·
nanoparticles · titania · Ullmann re-
action
Introduction
lysts for fine-chemical synthesis has been the subject of in-
creasing research efforts during the past decade.[1] The ultra-
high surface area available is a major contributor to in-
creased reactivity in the particle size range 5–10 nm. Howev-
er for nanocluster sizes of less than 50 nm this is accompa-
nied by a tunable electron-band structure.[2] The exploitation
of such properties promises the applicability of metals and
polymetallic nanoparticles in which particle lattice strain[3]
and surface energy[4] lead to an entirely changed metal-sub-
strate interaction for both mono- and poly-metallic clusters
in comparison to bulk systems.[5] In this vein, we have re-
cently published a novel liquid-type synthetic route by
which to obtain chemically stable Cu0 nanocatalysts.[6] We
have shown that both monometallic Cu and bimetallic CuSn
or CuZn catalysts demonstrate excellent yields of up to
ꢀ
ꢀ
Nanostructured metal catalysts in C O and C C coupling
reactions: The employment of nanostructured metal cata-
[a] F. Benaskar, N. G. Patil, Prof. Dr. J. Meuldijk, Dr. P. C. Thꢀne,
Dr. P. C. M. M. Magusin, B. Mezari, Prof. Dr. V. Hessel,
Prof. Dr. L. A. Hulshof, Prof. Dr. E. J. M. Hensen,
Prof. Dr. J. C. Schouten
Department of Chemical Engineering and Chemistry
Eindhoven University of Technology
P.O. Box 513, 5600 MB, Eindhoven (Netherlands)
Fax : (+31)402476653
E-mail: j.c.schouten@tue.nl
[b] Dr. V. Engels, Dr. A. E. H. Wheatley
Department of Chemistry
ꢀ
90% in 120 min in the C O cross-coupling of phenoxide
University of Cambridge
Lensfield Road, Cambridge, CB2 1EW (UK)
E-mail: aehw2@cam.ac.uk
and 4-chloropyridine[7] (see Scheme 1). Commonly, substan-
[c] Prof. Dr. E. V. Rebrov
School of Chemistry and Chemical Engineering
Queenꢁs University Belfast
Stranmillis Road, Belfast, BT9 5AG (UK)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201102151.
Scheme 1. The Cu-catalyzed Ullmann ether synthesis of phenoxypyridine.
1800
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1800 – 1810

FULL PAPER
tially more expensive bromo- and iodo-substituted aromatics
have hitherto been required for efficient coupling to take
place.[8] The synthesis of diphenylether from phenol and bro-
mobenzene by using a copper catalyst was mainly developed
by Ullmann,[9] whereas Goldberg presented initial work on
catalytic arylamination using bromobenzene/2-aminobenzoic
acid and amidation using bromobenzene/2-phenolamide
(see Scheme 2).[10]
use of Cu0 nanoparticles in the coupling of phenol and iodo-
benzene at 50–608C.[20] Following the achievement of a 95%
yield in the first run (4 h) catalyst recycling by centrifugation
allowed yields of up to 65% to be maintained in the fourth
run (12 h), with the highest reaction rates being observed
for a mean particle size of (20ꢁ2) nm. However, inefficient
catalyst recycling limited the applicability of this method; a
significant drop in catalyst performance was reported after
four experimental cycles. Kim et al. studied a Cu2O nano-
cube catalyst with a mean cube size of (45ꢁ3) nm.[21] Cata-
lyst recovery by centrifugation resulted in excellent-to-quan-
titative yields after four runs, with particle morphology re-
maining largely unchanged. However, the potential of this
method was restricted due to high temperatures (T>1508C)
requiring high boiling point solvents, expensive bases (such
as Cs2CO3), and complications related to catalyst recovery.
To circumvent catalyst separation obstacles, Cu nanoparti-
cles can be deposited either onto a magnetic carrier or a
structured support. Investigations of the Ullmann syntheses
of diazodiarylethers on structured supports have been re-
ported by Knepper et al., who used triazene-functionalized
Scheme 2. Examples of Ullmann ether synthesis, Ullmann–Goldberg ami-
nation, and Goldberg amidation reactions.[22,26]
polyACTHUNRTGNEUNG(styrene) resins to couple various substituted phenols
with yields of up to 95%.[22] Deposition of the active metals
onto inorganic supports with a relatively low specific surface
area requires low metal loadings to minimize particle sinter-
ing. Additionally loss of the active metal due to leaching
may occur during operation of the catalytic reactor. In the
present work, novel synthetic protocols have been devel-
oped to obtain supported Cu and CuZn bimetallic catalysts
on structured titania supports coated onto glass beads. The
catalytic activity and stability of the Cu and bimetallic CuZn
catalysts on structured mesoporous TiO2 and non-porous
Subsequent developments in the field of Ullmann and
Goldberg chemistry have led to many industrial applica-
tions, such as the synthesis of intermediates in pharmaceuti-
cal, agrochemical, fine and polymer chemistry.[11] Other in-
dustrially interesting applications are the arylation reactions
of amines and alcohols under relatively mild reaction condi-
tions, which favor processing in industrial batch reactors.[12]
Coupling reactions in organic chemistry are traditionally
achieved using palladium catalysis and have received intense
attention since their discovery during the 1970s,[13] leading to
the award of the Nobel Prize in 2010. However, the high
cost of palladium led to the initiation of research into cop-
per(II) acetate catalyzed coupling reactions involving orga-
nometals, such as Pb,[14] Si,[15] Sn,[16] as well as organoboron
reagents and iodonium salts, as arylating agents.[17] The very
high specific surface area, uniform pore-size distribution in
the nanometer range, chemical stability and precisely tail-
ored pore morphology, which decrease residence time distri-
bution, make mesoporous materials very useful as support
matrices on which to load metal and metal oxide nanoparti-
cles (NPs).[18] Moreover, due to the strong interactions with
metal NPs such as Au, Ag, and Pt, nanostructured titania
has long been considered as a supporting matrix to prevent
aggregation and thus improve the dispersibility of NPs.[19]
ꢀ
TiO2 are compared for the Ullmann-type C O coupling of
potassium phenolate and 4-chloropyridine to give 4-phen-
ACHUTNGERNoNUG xyAHCTUNGTRNENpUGN yridine. The Cu oxidation state in the mono- and bi-
metallic catalysts has been investigated by 63Cu-NMR, XPS,
XRD, and TPO/TPR studies after different operation times.
Results and Discussion
Catalyst activity: In this paper we present various highly
active Cu-based heterogeneous catalysts as replacements for
common homogenous catalysts that use expensive ligands
ꢀ
for C O coupling reactions. This work follows previous
studies, in which the addition of bimetallic Cu0-based cata-
lysts showed much higher yields than monometallic cata-
lysts.[23] A list of catalysts, metal loadings, and experimental
conditions is given in Table 1. Prior to the experiments re-
ported, the catalysts were preconditioned in a solvent flow
for 12 hours.
Cu-catalyzed Ullmann coupling reactions: Copper catalysis
in organic reactions is mostly known from Ullmann-type
homo- and cross-coupling reactions of aromatic compounds.
However, hardly any progress has yet been made in terms
of rendering these heterogeneous catalysts sustainable in a
continuous process. Only a few examples in which Cu nano-
particles have been employed as catalysts in Ullmann reac-
tions have been reported. In 2007, Kidwai et al. reported the
Cu/mesoporous TiO2: The 4-phenoxypyridine yield from po-
tassium phenoxide and 4-chloropyridine (see reaction
Scheme in Figure 1) over the Cu/meso-TiO2 (0.2 mol% Cu
with respect to 4-chloropyridine) demonstrated increased
yields relative to previously published results (Figure 1).[23]
Chem. Eur. J. 2012, 18, 1800 – 1810
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1801

J. C. Schouten, A. E. H. Wheatley et al.
of the higher natural abundance of 63Cu and, there-
fore, its higher NMR sensitivity. Like 65Cu, 63Cu
nuclei have spin 3/2 and interact strongly with local
electric field gradients at the positions of the nuclei
arising from their asymmetric coordination in cop-
per(I) compounds. The latter can cause large quad-
rupolar line broadening, which cannot be averaged
out by MAS. This is the reason why there are rela-
tively few solid-state 63,65Cu NMR spectroscopic
studies and most of these have been carried out
without sample rotation.[24]
Table 1. Metal loadings for different supported systems as determined by ICP-OES.
Catalyst
Conditions
Cu
loading
Support
amount
A
Overall
Cu
loading
Support
pore
size
A
[wt%][c]
A
[nm]
Cu/meso-TiO2
Cu/np-TiO2
Preconditioned
Spent[a]
2.0
1.8
11.5
2.0
2.8
2.1
0.21
0.03
2.0
0.60
0.04
0.02
0.00
0.08[e]
24ꢁ4
4.9ꢁ0.8
26ꢁ5
Preconditioned
Spent[a]
Cu66Zn34/meso-TiO2
Preconditioned
3.8
[a] 12 h flow experiment as described in the Experimental Section. [b] With respect to
Ti after preconditioning with the reaction mixture. [c] Weight percentage of films on
glass beads. [d] Total Cu/bead ratio. [e] Total copper content.
ꢀ
Figure 1. The Ullmann C O coupling as carried out by using a Cu/TiO2
catalyst. Yields obtained from Cu/meso-TiO2 using Ccat =0.2 mol% Cu
with respect to 4-chloropyridine at different temperatures and stirring
~
*
*
^
rates. Temperatures: 140 ( and ), 130 (&), 120 ( ), and 1108C ( and
Figure 2. MAS 63Cu NMR spectra of spent and fresh Cu/meso-TiO2, re-
spectively, before and after the reaction, as well as the spectra of solid
CuCl and Na2SO4, and the background spectrum of metallic copper in
the NMR coil, for comparison. Except the background spectrum, all
spectra are background corrected. 23Na and 63Cu NMR frequencies differ
by only d=2ꢃ103 ppm, which permits 23Na and 63Cu NMR signals to be
recorded in a combined manner.
~
~
~
*
*
^
). Stirring rates: 500 rpm ( and ) and 1500 rpm ( , &, , and ). Ini-
tial concentrations of potassium phenolate and 4-chloro
ACHTUNGTRENNUpGN yrACHTNURTEGNNUNGiACHUTGTNERNUGdN ine were
0.36 and 0.3 molLꢀ1, respectively.
A considerable amount of product (60%) was already pres-
ent in the reaction mixture after 5 min reaction time at
1408C. However, the maximum yield was similar to that re-
cently observed over CuZn NPs. It appears that strong prod-
uct adsorption on the catalyst surface severely limits the re-
action rate at longer reaction times. The Cu oxidation state
was determined by solid-state 63Cu NMR spectroscopy.
However, the amounts of Cu0 and Cu1+ could not be calcu-
lated due to there being a considerable contribution from
the background signal. Nevertheless, a qualitative signal sub-
traction was made by utilizing additional spinning side band
signals from the catalyst sample.
In contrast to many copper(I) compounds, the MAS 63Cu
NMR signal of Cu0 in metallic copper is relatively narrow,
as a result of the absence of electric field gradients at the
copper-atom positions in the fcc crystal structure. In the
present study we have, therefore, applied magic-angle spin-
ning 63Cu NMR spectroscopy as a facile tool with which to
monitor the fate of metallic copper in the catalyst particles.
The spectrum of the fresh catalyst contains a single signal at
the same position as the background signal of copper in the
NMR coil, which is at d=2334 ppm downfield with respect
to the NMR frequency of CuCl (Figure 2). This large chemi-
cal shift is a so-called Knight shift caused by the interaction
with the conduction electrons in copper metal. In the 63Cu
NMR spectrum of the spent catalyst Cu/meso-TiO2, no met-
allic copper signal is visible. Since elemental analysis indi-
Figure 2 shows the magic-angle spinning (MAS) 63Cu
NMR spectra of fresh and spent catalyst together with the
spectra of solid copper(I) chloride, metallic copper (from
the NMR coil), and solid sodium sulfate, for comparison.
Copper has two NMR-active isotopes: 63Cu and 65Cu. 63Cu
NMR spectroscopy was chosen in the present study because
1802
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1800 – 1810

New Cu-Based Catalysts Supported on TiO2 Films
FULL PAPER
cates that the spent catalyst still contains copper (see
Table 1), Cu0 has apparently been converted into Cu1+ or
Cu2+ during the four hour experimental run. Copper(II)
compounds are paramagnetic and so the NMR spectrum for
63,65Cu is invisible. In principle, copper(I) compounds are
NMR visible, but their usually large quadrupolar linewidth
generally prevents straightforward detection by use of MAS
NMR. The spent Cu/meso-TiO2 does show a single signal at
d=ꢀ2081 ppm. This is actually the 23Na NMR signal of a
sodium-containing impurity.1 The observed 23Na NMR
signal position is d=ꢀ13 ppm with respect to solid NaCl
(not shown), and coincides with that of solid Na2SO4
(Figure 2). The latter is used as a drying agent in the synthe-
sis of 4-chloropyridine (see the Experimental Section),[23]
one of the reactants in the catalytic test reaction.
Figure 3a and b show a porous titania structure from
SEM analysis before and after copper nanoparticle deposi-
tion, respectively. Importantly, the porous titania structure
remained essentially unaffected, even after a second calcina-
tion step subsequent to catalyst deposition.
Figure 4. HRTEM images of a) mesoporous-structured titania showing
the support and, b) the deposited CuZn nanoparticles in fresh CuZn/
meso-TiO2. c) Particle size distribution generated by measuring 200 parti-
cles (see the Supporting Information, Figure S3).
analysis of five spectra, over a surface of 500 nm2 the Cu/Ti
molar ratio was 2.7ꢁ0.2, which is close to the ratio of 2.5
found from a scan over a larger 250 mm2 area, confirming a
uniform catalyst loading on the support (see Table 1).
Figure 3. SEM image of silica beads coated with mesoporous titania (Cu/
meso-TiO2) a) before and b) after catalyst deposition.
Unsupported CuZn nanoparticles: We have recently report-
ed that bimetallic CuZn NPs showed higher product yields
compared with monometallic Cu NPs in the Ullmann etheri-
fication reaction.[23] As a result, bimetallic CuZn NPs were
selected for further investigation in this study. Figure 5a
shows 4-phenoxypyridine yield in the temperature range of
110–1408C and for varying stirring rates for a catalyst con-
centration of 1.2 mol% with respect to 4-chloropyridine.
Yields of 6 and 38% were observed after 300 s for this cata-
lyst at 1208C and 1408C, respectively.
EDX spectra were taken from the regions shown in Fig-
ure 3a and b. The data confirmed the abundant coverage of
copper upon catalyst deposition. It was concluded that there
was an increased copper loading on mesoporous titania
based on the integral ratio of the Cu Kb1 and Ti Kb1 emis-
sion lines, which were 0.42 and 0.25 for mesoporous
(2.1 wt%) and non-porous (0.2 wt%) titania supports, re-
spectively (see the Supporting Information and Table 1).
Figure 4a clearly reveals a highly porous titania matrix with
observed pore sizes of 5–50 nm proving a mesoporous sup-
port structure for this catalyst. The average pore size was
determined by HRTEM to be (24ꢁ4) nm for approximately
60% of the cross-sectional surface area on which the Cu
and CuZn nanoparticles were deposited (see Figure 4b).
The average particle size, based on a count of 200 particles,
was found to be (7.0ꢁ0.6) nm and the particle size distribu-
tion is shown in Figure 4c. The presence of Cu, Zn and Ti in
the area shown in the above HRTEM images is confirmed
by EDX analysis. The Cu/Zn molar ratio is 2.0. From the
The experimental results can be accurately described by a
Langmuir–Hinshelwood-type equation with product inhibi-
tion:
kwcatCNCE
r ¼
ð1Þ
2
ð1 þ KNCN þ KECE þ KPCPÞ
Subscripts N, E, and P in Equation 1 refer to the nucleophile
(potassium phenolate), electrophile (4-chloropyridine), and
product (4-phenoxypyridine), respectively (r=rate in
molprod gcatꢀ1 sꢀ1, k=overall rate constant in L2 gꢀ2 molꢀ1 sꢀ1
and wcat =mass of catalyst in g). The kinetic parameters are
listed in Table 2 and the rate-model fit is shown in Figure 5b
for different catalyst concentrations.
1
NMR spectroscopic frequencies of 63Cu and 23Na differ by only 2ꢃ
103 ppm, so that 63Cu and 23Na NMR signals can be recorded within a
single NMR spectrum.
Chem. Eur. J. 2012, 18, 1800 – 1810
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1803

J. C. Schouten, A. E. H. Wheatley et al.
Figure 6. XRD spectra of nano-Cu (top) and nano-CuZn (bottom) prior
to heterogenization.
Figure 5. Yields of 4-phenoxypyridine by using unsupported bimetallic
CuZn nanoparticles at different temperatures and stirring rates using
C
cat =1.2 mol% with respect to 4-chloropyridine (a), and the observed
(&) and predicted yields (solid lines) at 1408C based on a Langmuir–
alloy formation has been followed by the oxidation driven
segregation of the elements. This observation is in agree-
ment with DFT and microscopic studies of the alloy system
by Greely and Hansen, respectively.[27] This segregation is
likely to have been enhanced by the exposure of the particle
to air in the course of an XRD measurement time of about
5.5 h.
*
*
Hinshelwood rate model (b). a) Temperatures: 140 ( and ), 130 (&),
~
~
~
^
*
120 ( ), and 1108C ( and ). Stirring rates: 500 rpm ( and ) and
~
*
^
1500 rpm ( , &, , and ). Initial concentrations of potassium phenolate
and 4-chloropyridine were 0.36 molLꢀ1 and 0.3 molLꢀ1, respectively.
ꢀ
Table 2. The kinetic parameters of the Cu-catalyzed Ullmann C O cou-
pling according to a Langmuir–Hinshelwood-type rate equation.
k [L2 gꢀ2 molꢀ1 sꢀ1
2.87ꢃ10ꢀ3
]
KN [Lmolꢀ1
]
KE [Lmolꢀ1
2.87ꢃ10ꢀ1
]
KP [Lmolꢀ1
]
Cu/non-porous TiO2: The 4-phenoxypyridine yield over the
Cu/np-TiO2 film (0.004 mol% Cu with respect to 4-chloro-
pyridine) was 25% higher at 1208C than the yield of the
same reaction using CuZn NPs (Figure 7). However, the
yield was similar at 1408C.
A low magnification SEM image of np-TiO2 films on
glass beads is shown in Figure 8a. A smooth surface of np-
TiO2 with a few cracks is seen after calcination at 6008C in
the larger magnification image Figure 8b. The formation of
cracks during calcination demonstrates that adhesion of the
titania film to the substrate is rather strong.
A SEM image of the Cu/np-TiO2 catalyst after calcination
at 4008C is shown in Figure 8c. One notable feature is that
the mean size of Cu NPs in the calcined films (14.3ꢁ
2.0 nm) was slightly increased relative to that of the unsup-
ported nanoparticles (9.6ꢁ1.0 nm).[6] However, as was ob-
served from TEM imaging and chemisorption analysis, the
1.26
9.57
At low conversion a rather constant reaction rate is ob-
served. As a result, the product yield is a linear function of
the reaction time and the reaction rate is independent of the
reactant and product concentrations. However, when a sub-
stantial amount of 4-phenoxypyridine is formed, adsorption
onto the active metal severely limits the reaction rate.
Figure 6 shows the XRD spectra of unsupported Cu and
CuZn catalyst. The [111], [200] and [220] refractions charac-
teristic of a face-centered cubic metal lattice are observed at
43, 50.5, and 74.58.[23,25] The oxide phases Cu2O and ZnO
are observed only in the bimetallic catalyst.[26] Comparison
with our previous results[6,23] underlines the stability of Cu
nanoparticles, whereas in case of CuZn nanoparticles, initial
1804
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1800 – 1810

New Cu-Based Catalysts Supported on TiO2 Films
FULL PAPER
ing the reaction rate doubled using unsupported CuZn NPs,
and this could be attributed to the large adsorption constant
of the product on the copper. Overall, the highest reaction
rates were found for the Cu/np-TiO2 catalyst. Relative to re-
cently published results by Engels et al.,[23] the lowest reac-
tion rate was found for unsupported CuZn NPs
(1.4 mmolprod mmolcatꢀ1 sꢀ1), which was almost one order of
magnitude lower than the reaction rate for Cu/meso-TiO2
(22.2 mmolprod mmolcatꢀ1 sꢀ1). The catalyst activities and acti-
vation energies for the different catalysts are shown in
Table 3.
Arrhenius plots (Figure 9) reveal that, at 110 and 1408C,
different stirring rates (of 500 and 1500 rpm) appear to have
no effect on the reaction constants for the different catalyst
systems. Activation energies were determined from the reac-
tion rate data taken at a reaction time of 120 s, such that the
highest conversion was below 8%. The activation energies
were found to be (75ꢁ2), (123ꢁ3), and (144ꢁ5) kJmolꢀ1
for Cu/np-TiO2, CuZn NPs and Cu/meso-TiO2, respectively.
The observed doubling of the activation energy for the
Figure 7. Yields obtained from Cu/np-TiO2 by using Ccat =0.004 mol%
Cu with respect to 4-chloropyridine at different temperatures and stirring
~
*
*
^
rates. Temperatures: 140 ( and ), 130 (&), 120 ( ), and 1108C ( and
~
~
~
*
*
^
). Stirring rates: 500 rpm ( and ) and 1500 rpm ( , &, , and ). Ini-
tial concentrations of potassium phenolate and 4-chloropyridine were
0.36 and 0.3 molLꢀ1, respectively.
mesoACTHUNRTGNEUNGporous titania system provides evidence for internal
diffusion limitations in the case of the Cu-catalyzed Ull-
ꢀ
mann-type C O coupling reaction when using a non-porous
titania support.
The Ullmann coupling catalyzed by the porous catalyst
obeys intrinsic kinetics, that is, constants for mass transfer
Figure 8. SEM images of the single layer TiO2-coated beads before
ACHTUNGTRENNUNG(a and b) and after copper deposition (c) for Cu/np-TiO2.
Figure 9. Arrhenius plots at a temperature range from 110 to 1408C and
stirring rates of 500 and 1500 rpm (only 110 and 1408C data shown for
experiments at 500 rpm) by using different Cu-based catalysts [Cu/meso-
*
catalyst dispersion was maintained unchanged. The thick-
ness of the titania film was estimated to be between 200 and
500 nm.
TiO2 (&;
E
act =144 kJmolꢀ1), unsupported CuZn NPs
( ; Eact =
123 kJmolꢀ1), and Cu/np-TiO2 ( ; Eact =75 kJmolꢀ1)].
~
The
supported
catalysts
showed higher initial rates and
product yields than the unsup-
ported catalysts (see Figures 1,
5, and 7) and Cu/np-TiO2 dem-
onstrated higher activity at low
reaction temperatures. Howev-
er, the influence of catalyst
loading on reaction rates was
most strongly observed for the
unsupported CuZn NPs. At a
Table 3. Reaction rates and activation energies (Eact) for different Cu-based catalysts.
Catalyst type
T
[8C]
Cu loading
[wt%]
[mol%][a]
Initial rate
Eact
[kJmolꢀ1
TON[c]
Product
yield [%]
G
G
[molgcatꢀ1 sꢀ1
]
G
]
[molprod molcat sꢀ1
]
CuZn NPs[b]
Cu/meso-TiO2
Cu/np-TiO2
120
140
120
140
120
140
–
–
1.8
1.8
–
12ꢃ10ꢀ1
12ꢃ10ꢀ1
2ꢃ10ꢀ3
2ꢃ10ꢀ3
4ꢃ10ꢀ5
4ꢃ10ꢀ5
0.002
0.011
0.014
0.104
1.715
4.55
123ꢁ3
–
0.1
0.6
0.9
6.6
109
289
41.9
80.3
47.6
83.8
63.7
81.7
144ꢁ5
–
75ꢁ2
2.0
[a] Relative to 4-chloropyridine. [b] For experimental data on CuZn NPs, see Engels et al., ref. [23]. [c] TON=
fourfold increase in metal load- turnover number.
Chem. Eur. J. 2012, 18, 1800 – 1810
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1805

J. C. Schouten, A. E. H. Wheatley et al.
are much smaller than time constants for reaction. As
shown in Figure 1, Figure 5, and Figure 7, decreasing the
stirring rate from 1500 to 500 rpm has no effect on the cata-
lyst activity. This confirms the absence of external mass-
transfer limitations, as would be expected in a liquid-solid
reaction in which the reactants are completely dissolved in
the solvent. The higher product yields achieved over the Cu/
meso-TiO2 are due to an increased surface area facilitating a
higher catalyst loading. The spent Cu/meso-TiO2 and Cu/np-
TiO2 films were analyzed using SEM-EDX after experi-
ments at 1408C (Figure 10a and 10b, respectively). The
EDX signal for Ti remained unchanged before and after re-
action, confirming good adhesion of the mesoporous titania
film. Meanwhile, a 10% reduction in the signal for Cu was
observed in the spent meso-TiO2, which indicated only
minor Cu losses due to leaching.
ergies of 1021.4 eV (Zn 2p), 932.9 eV (Cu 2p), and 458.7, as
well as 465.2 eV (Ti 2p) were observed.
The Cu 2p3/2 peak at 932.9 eV indicated the presence of
metallic copper in the bimetallic CuZn catalyst. The appear-
ance of a second Cu 2p3/2 component at 934 eV accompa-
nied by strong shake-up signal between 938 and 945 eV indi-
cates the formation of copper(II) oxide. Comparison of
monometallic copper and bimetallic CuZn catalysts indi-
cates that copper is stabilized against oxidation by the pres-
ence of zinc, probably through alloy formation. The mono-
metallic copper nanoparticles are already partially oxidized
after impregnation onto the titania support and are com-
pletely oxidized after reaction, as was also confirmed by the
63Cu NMR study. In the presence of zinc, copper remained
in the metallic state after impregnation and even after 4 h
on-stream; however, no Cu0 was observed after a reaction
time of 12 h. For the origins of this effect, both electronic
and kinetic mechanisms can be considered.[28] Whereas in
the first case, Zn is likely to act as a sacrificial anode
through its oxidation to ZnO, the formation of core/shell
particles and their consequences for oxygen diffusion also
needs to be considered. XPS results for the unsupported
CuZn nanoparticles show the presence of metallic copper,
likely through the formation of alloy phases and/or a solid
solution. In the course of the catalytic reaction, Zn is likely
to segregate to the Cu particle surface. This phenomenon
leads to a Cu/ZnO core/shell system by subsequent oxida-
tion of Zn and, eventually, to deactivation. The formation of
such a ZnO shell may lead to a considerable diffusion barri-
er to oxygen reaching the Cu surface. Consequently, oxida-
tion of copper is inhibited, yet at the same time the accessi-
bility of the reactants to the active copper surface also de-
creases, which is reflected in the conversion plateau. On the
other hand, the subsequent complete separation of the shell
from the core will lead to the formation of morphologically
Figure 10. SEM analysis of spent Cu catalyst on mesoporous (a) and non-
porous (b) titania, respectively, after a reaction time of 120 min in a
stirred batch reactor at 1500 rpm.
Catalyst deactivation and stability: The catalyst activity ex-
periments demonstrate that the best catalyst support is the
mesoporous titania (Cu/meso-TiO2). For this reason the fol-
lowing deactivation and stability study has mostly focused
on Cu oxidation and Cu/meso-TiO2 leaching from the glass
beads. XPS measurements were performed on unsupported
Cu and CuZn NPs, Cu/meso-TiO2 and CuZn/meso-TiO2.
The supported catalysts were tested for different processing
times in a dimethylacetamide (DMA) flow of 5 mLminꢀ1.
The liquid viscosity was increased by dissolving potassium
phenolate, while maintaining the temperature at 1408C. Of
particular note, the preliminary flow experiments estab-
lished that the mesoporous titania coating on the glass sup-
port appeared to be stable against leaching for 12 h under
flow operation. However, ICP measurements demonstrated
19% titania loss for the non-porous titania film after 4 h.
Moreover, when Cu/meso-TiO2 was tested, copper losses ap-
peared to be approximately 17% in the first 4 h, whereas
the titania was relatively unaffected. As a result of a rela-
tively stable catalyst support system, activity of the Cu/
meso-TiO2 catalyst could be maintained at tolerably high
level for three runs of 4 h each.
Figure 11. XPS analysis of mesoporous titania coated with Cu and CuZn
NPs at different reaction times (glass beads loaded into a fixed-bed reac-
tor).
X-ray photoelectron spectroscopy: The results of the XPS
measurements are shown in Figure 11. Peaks at binding en-
1806
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1800 – 1810

New Cu-Based Catalysts Supported on TiO2 Films
FULL PAPER
undefined ZnO agglomerates distinct from the former Cu
core, which may remain largely intact as a new Cu nanopar-
ticle that is prone to oxide formation. The resulting decrease
in the specific surface area of Zn due to ZnO particle
growth and the accompanying decrease of surface-to-
volume ratio may then be the cause for the observed reduc-
tion in the Zn 2p peak intensity. However, carbon deposi-
tion during the reaction represents a more likely reason for
the simultaneous depletion of Cu, Zn, and Ti signals. This is
best seen from a quantitative XPS analysis based on con-
stant oxygen content in the carbon holey film substrate (see
Experimental Section and the Supporting Information, Fig-
ure S9). Sodium and nitrogen impurities in the XPS data are
attributed to artefacts residual from the copper nanoparticle
synthesis.
The two peaks in the TPO curve for Cu/meso-TiO2 signify
the occurrence of two separate oxidation steps and provides
the temperatures at which they occur. The first oxidation to
CuO occurs at 230–2508C and is followed by a second pro-
cess at 280–2908C. These data are in good agreement with
experiments performed using a commercial CuO/SiO2 cata-
lyst (see the Supporting Information).
The reduction cycle for Cu/meso-TiO2 reveals features in
three temperature ranges. At 2258C all the copper present
at the exterior surface of the film was reduced to Cu0. This
was followed by two subsequent reductions in the tempera-
ture range 310–3708C, corresponding to the reduction of
CuO species at the interior surface of the porous film. Due
to particle size distribution and dispersion throughout the
mesoporous titania film, different reduction processes occur.
Indeed, it can be postulated[29] that the wide temperature
range over which reduction occurs is due to the high disper-
sion of copper nanoparticles in the mesoporous titania film.
The combined results of BET-surface measurements by
physisorption and TEM imaging showed a high dispersion
of the copper nanoparticles in the mesoporous titania
matrix. These findings are supported by the fact that in the
case of Cu/np-TiO2, a single copper reduction peak appears
over a broad temperature range from about 3508C.[30] In
this latter case, the copper is uniformly distributed over the
thickness of the titania film, which also explains the low ac-
cessibility of copper and therefore considerable mass-trans-
fer limitations and decreased apparent EAct. Cu loadings of
Cu/np-TiO2 and Cu/meso-TiO2 were determined by measur-
ing the total amount of hydrogen consumed after complete
oxidation to CuO. The graphical uptake of hydrogen used to
determine the Cu loadings is depicted in Figure 13 and is
Temperature-programmed oxidation/reduction: Tempera-
ture-programmed oxidation/reduction (TPO/R) experiments
were carried out to determine the amount of oxygen and hy-
drogen uptake by the Cu and the temperature at which a
spontaneous surface uptake of these gases occurs. This, in
turn, allowed the oxidation state and the reduction tempera-
ture for the regeneration from spent catalyst to be deter-
mined. Figure 12 shows the results for both the tempera-
Figure 12. TPO signals (dashed line) and TPR signals (solid line) for
copper on mesoporous (Cu/meso-TiO2) and non-porous (Cu/np-TiO2) ti-
tania.
ture-programmed oxidation (dashed lines) and reduction
(black lines) of Cu/meso-TiO2 and Cu/np-TiO2. Since the
first step in these experiments consists of an oxidative treat-
ment at varying temperatures from 350 to 4508C, the
dashed lines provide an indication of the original oxidation
state of the copper. The oxygen uptake of Cu/meso-TiO2
shows that the copper is predominantly present as Cu0.
These data are consistent with the XPS and 63Cu NMR spec-
troscopic results. In contrast, the copper in Cu/np-TiO2 ap-
pears to be predominantly present in the +2 state, since no
oxidation feature appears in the TPO experiment.
Figure 13. TPR results for the determination of copper loading by using
integrated values of hydrogen uptake.
summarized in Table 4. A fresh sample of Cu/meso-TiO2
shows a single reduction peak at 3218C due to the abundant
presence of copper inside the porous film, whereas the spent
catalyst reveals several reduction peaks probably due to
copper migration to the outer film surface. For Cu/np-TiO2
a broad reduction peak was observed at (higher) tempera-
tures ranging from 4358C to 6008C. For CuZn/meso-TiO2,
Chem. Eur. J. 2012, 18, 1800 – 1810
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1807

J. C. Schouten, A. E. H. Wheatley et al.
Table 4. Integrated H2-uptake values from a TPR-study providing the
copper loading for the Cu-based heterogeneous catalysts.
Cu/meso-TiO2 films oxidation of Cu was observed. The
presence of Zn appeared to be crucial for the stabilization
of metallic Cu. Accordingly, XPS, 63Cu NMR spectroscopy,
and TPO/TPR data showed the formation of Cu2+ in the
spent titania-supported catalysts, which is in line with the re-
sults of Newton and Lai[32] and this represents the main
source of catalyst deactivation in the Cu/meso-TiO2 films.
Support leaching was found to be the main reason for cata-
lyst deactivation in the case of the non-porous TiO2 films.
This observation underlines the significant potential for fur-
ther catalyst development, exploiting tailored metal-support
interactions between copper and the nanoscopic oxide after
improving adhesion of the TiO2 support to a structured sub-
strate.
Catalyst type
T [oC]
Metal loading [wt%]
Oxidation to Reduction to Fresh
Spent
Cu
Cu1+ Cu2+ Cu0
Zn0
Cu
Zn
Cu/meso-TiO2
Cu/np-TiO2
CuZn/meso-TiO2
224
–
227
282
–
–
321
435
352
–
–
0.31
0.38
0.12
–
–
0.29
0.22
0.26
–
[a]
[a]
671[b]
[a] A single oxidation peak was observed. [b] Although the reverse pro-
cess (zinc hydrolysis) requires high temperatures (19008C), melting of
the catalyst was observed, indicating Zn0 formation.
two peaks appeared at 352 and 6718C. The first peak clearly
originates from the formation of zero-valent copper, where-
as the latter can be attributed to the partial reduction of the
zinc oxide species.[31] This latter conclusion is consistent with
the observation of sintering of the catalyst due to the melt-
ing of metallic Zn at 419.58C. As compared to the metal
loadings found from ICP-OES measurements (shown in
Table 1) these TPR data (Table 4) afforded corresponding
results. The catalyst stability was studied at a reaction time
sufficient for the production of 0.1 kg of 4-phenoxypyridine.
Experiments were carried in a packed-bed reactor at a di-
Experimental Section
Catalyst preparation: Synthesis of catalyst nanoparticles: A polyvinylpyr-
rolidone (PVP; MWACTHNUTRGNEUNG
(average)=40,000)-stabilized suspension of Cu0 and
CuZn (2.5 mgCu mLꢀ1) nanoparticles was prepared by using a technique
described in detail by Engels et al.[23]
Synthesis of non-porous titania support: A total of 20 g of glass beads
(E&R Chemicals & Equipment B.V., sieved diameter 200–300 mm) was
cleaned using acetone and dispersed in dry isopropanol (200 mL) in a
rotary evaporator in a nitrogen atmosphere. The dry beads were stirred
overnight in aqueous sulfuric acid (2.5m, Sigma–Aldrich) prior to support
ACHTUNGTRENNUNGmethylACHTUNGTRENNUNG
acetamide flow of 5 mLminꢀ1 in multiple runs of 12 h
at 1408C. A Cu loading of 2 wt% was observed in both
spent Cu/np-TiO2 and Cu/meso-TiO2 catalysts, which corre-
spond to active metal losses due to leaching of 82 and 10%,
respectively (see Table 1).
The Ti/Si ratio was reduced by 25% in the spent Cu/
meso-TiO2 catalysts, which confirms a rather good adhesion
of the mesoporous titania films to the glass surface as com-
pared with that of Cu/np-TiO2 (80% Ti loss). The low Cu/Ti
ratio, measured in the spent Cu/np-TiO2 catalyst after 12 h
on-stream, suggests that most of the Cu was located uni-
formly over the TiO2 film and leached simultaneously with
it.
coating.
A total amount of 0.75 g of titanium(IV) tetraethoxide
(Ti(OEt)4, 99.99 wt.%, Sigma–Aldrich) mixed with an aqueous HNO3 so-
AHCTUNGTRENNUNG
lution (65 wt.%, Fluka) in isopropanol was added to the dry glass beads
and vigorously stirred for 1 h. The isopropanol was then removed under
reduced pressure at 608C. The resulting coated beads were dried over-
night in a stove at 808C and subsequently calcined at 6008C for 8 h
(108Cminꢀ1 heating and cooling).
Synthesis of mesoporous titania support: Glass bead preparation was
done similarly to that described for the non-porous titania support. A ti-
tania-precursor sol with a composition of 1 TiACTHNUTRGENUG(N OEt)4:0.006 Pluronic:40
ethanol:1.3 H2O:0.1 trifluoroacetic acid was prepared as described else-
where.[33] Pluronic F127, (EO100PO65EO100, EO=ethylene oxide, PO=
propylene oxide, BASF) was used as surfactant. After dissolution of the
surfactant in ethanol, followed by the addition of water and trifluoroace-
tic acid, the titanium(IV) tetraisopropoxide (TiACTHNURTGNENG(U OiPr)4, 99.99%, Fluka)
Conclusion
was added dropwise while stirring. The resulting mixture was left to age
for 24 h after a stirring period of 8 h at room temperature. A desired
amount of the solution was slowly added to a glass-filter vessel filled with
an observable single layer of glass beads. The solution withdrawal rate
was set at 2 mmminꢀ1 using a flow controlled dip-coater. The resulting
films were maintained in a glove box at a relative humidity of 80% for
30 h before the residual surfactant was removed by burning at 2508C for
Two novel heterogeneous copper-based, nanostructured cat-
alysts have been synthesized by using non-porous TiO2 and
mesoporous TiO2 films deposited on glass beads as the
structuring substrate. The catalytic activity of these films,
after impregnation with Cu or CuZn nanoparticles, have
4 h. Finally the films were calcined at 3808C (heating rate of 18Cminꢀ1
)
ꢀ
been studied in the copper-catalyzed Ullmann C O coupling
and a residual pressure of 10 mbar for 8 h.
reaction of 4-phenoxypyridine and potassium phenolate in a
batch-wise operated stirred tank reactor. The highest prod-
uct yield of 84% was obtained in 90 min by using catalysts
comprising 2 wt% Cu/meso-TiO2 films with mean pore sizes
of (24ꢁ4) nm. After 12 h preconditioning, the observation
of a Cu loading of 1.8 wt% demonstrated that leaching of
Cu does not play a significant role in mesoporous titania
support systems. Bimetallic CuZn nanoparticles with a Cu/
Zn molar ratio of 2 and a mean size of 4 nm were also de-
posited onto mesoporous TiO2 films. The Cu0 state could be
preserved in these catalysts, whereas for the monometallic
Catalyst impregnation onto titania support: The titania films were impreg-
nated with a colloidal suspension (metal concentration: 0.25 mgmLꢀ1) of
Cu or CuZn nanoparticles. The impregnated films were dried at 808C.
This procedure was repeated twice. Subsequently, the coated beads were
calcined for 2 h at 4508C (unless otherwise mentioned) after a heating
rate of 18Cminꢀ1 firstly at a residual pressure of 10 mbar and later at at-
mospheric pressure to remove residual organic compounds. Finally, the
catalyst was reduced for 6 h under
a
hydrogen flow of 1 mLminꢀ1
(4 vol% H2 in N2) at 3508C (108Cminꢀ1).
Catalyst characterization: XRD spectra of PVP-protected copper NPs
were recorded on a Rçntgen PW3040/60 XPert PRO powder X-ray dif-
fractometer with a high resolution PW3373/00 CuLFF (unmonochromat-
1808
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1800 – 1810

New Cu-Based Catalysts Supported on TiO2 Films
FULL PAPER
ed) tube at l=1.5404 ꢄ (CuKa) in the range 2V=5–808 and with scan-
ning speed of 0.23 8minꢀ1 (CuZn) and 0.098minꢀ1 (Cu). The powder sam-
ples were prepared by solvent evaporation from the colloidal suspensions
deposited on a glass plate (PW3071/60Bracket).
quots. The data were compared with literature data and with previously
acquired, qualitative GC-MS measurements.
Catalyst surface area, pore sizes, and metal dispersion were determined
by physisorption and chemisorption techniques (see the Supporting Infor-
mation). The catalytic activity is presented in terms of amount of the re-
action product obtained per unit time per catalyst weight (not accounting
XPS data were obtained with
a Kratos AXIS Ultra spectrometer
equipped with a monochromatic Al Ka X-ray source and a delay-line de-
tector (DLD). Spectra were obtained by using an aluminium anode
(Al Ka=1486.6 eV) operating at 150 W. For survey and region scans, con-
stant pass energies of 160 and 40 eV were used, respectively. The back-
ground pressure was 2ꢃ10ꢀ9 mbar. Samples were prepared in a glovebox
(d= <10 ppm O2) and transported in a closed sample holder for oxygen-
free XPS analysis.
for the weight of glass beads). The initial reaction rate (rinit
) in
molprod gcatꢀ1 sꢀ1 was determined from the product concentration after a
reaction time of 5 min (Equation 2).
C ꢂ VL
ð2Þ
rinit
¼
Dt ꢂ wcat
HRTEM images were recorded using a FEI Tecnai G2 Sphera transmis-
sion electron microscope at an acceleration voltage of 200 kV. For TEM
analyses, 250 mg of the solid catalyst was pulverized and suspended in
ethanol. A volume of 30 mL of the suspension was dip-coated onto a 200
mesh molybdenum grid (carbon holey film) and subsequently dried at
ambient temperature and pressure. Support surface morphology was ana-
lyzed on a FEI Quanta series FEG 3D G2 SEM using an acceleration
voltage of 5 kV and magnifications of between 5,000x–100,000x, provid-
ing a maximum lateral resolution of 50 nm2. Catalyst elemental composi-
tions were elucidated by EDX analysis at a spot size of 50 nm2 at an in-
teraction-volume of 100 mm.
In which C is the concentration of 4-phenoxypyridine in mol Lꢀ1after re-
action time of 5 min, VL is the solvent volume in L, Dt is the time interval
(300 s) in s, and w is the catalyst weight in g.
The catalytic activity was measured for the three catalyst types: the un-
supported CuZn nanoparticles (CuZn NPs), Cu supported on non-porous
titania (Cu/np-TiO2) and Cu supported on structured mesoporous titania
(Cu/meso-TiO2). The catalyst deactivation was studied in a fixed-bed re-
actor at different processing times.
Magic-angle-spinning (MAS) 63Cu NMR spectra were recorded at room
temperature on a Bruker DMX500 spectrometer, operating at a frequen-
cy of 132.57 MHz. A 4 mm MAS probe head was used with a sample ro-
tation rate of 6 kHz. To suppress baseline artifacts resulting from probe
ringing, 63Cu NMR spectra were recorded by use of a three-pulse pulse
sequence q+x–qꢁx–t–qy–d1 with short pulses of 1 ms pulse length (corre-
sponding to an excitation angle of 188), an interpulse delay of 0.1 ms, and
an interscan delay of 100 ms. Typically 8192 scans were recorded. The
radio-frequency carrier position was chosen close to the resonance fre-
quency of microcrystalline copper(I) chloride powder. Following the con-
vention in several solid-state 63,65Cu NMR studies,[24] solid CuCl was also
used as external reference for the chemical shift d=0 ppm. The 63Cu
NMR spectrum of the NMR coil alone (without sample) consisted of a
single d=90 ppm wide signal (at half height) without MAS sidebands,
and this was subtracted from the MAS NMR spectra of the catalysts as
background correction. The fact that the MAS NMR signal of Cu0 in the
fresh catalyst was narrower and had spinning sidebands, made an unam-
biguous subtraction possible.
Catalyst pretreatment: Prior to experiments, the catalysts were pretreated
in a flow of dimethylacetamide (5 mLminꢀ1) for 4, 8, or 12 h (the same
duration as the subsequent experiment). In catalyst stability experiments,
potassium phenolate was dissolved in the solvent. In the oxidation stabili-
ty experiments, 4-chloropyridine was added to the solvent at 1408C. De-
tails on the experimental setup are provided in the the Supporting Infor-
mation.
Derivation of kinetic model: Experiments were carried out over the tem-
perature range 110–1408C by using metallic copper (99%, Aldrich). The
catalyst loading (1–10 mol% with respect to 4-chloroACHTUNGTNERpUNNG yrACHUNTRTGENNUNGiACHUTGTNRENNUGdine), initial re-
actant concentrations (potassium phenoxide 0.3–0.75 molLꢀ1 and 4-chlor-
opyridine 0.2–0.5 molLꢀ1), initial product concentrations (4-phenoxypyri-
dine 0.05–0.25 molLꢀ1) and the crown ether concentration (1–10 mol%
of 18-crown-6 ether with respect to potassium phenoxide) were varied.
The mole ratios of potassium phenoxide and 4-chloropyridine were
varied from 0.5 to 4.[25]
The metal loading was determined by ICP-OES on a SPECTRO CIR-
OSCCD spectrometer.
Acknowledgements
For temperature-programmed oxidation a quartz glass tube (outer diame-
ter: 6 mm, inner diameter: 3 mm) was filled with a known amount of sup-
ported catalyst, sealed, and pretreated in a convection oven under hydro-
gen flow for 6 h at 3508C. After cooling, an oxygen flow (4 vol% in He)
was applied at a flow rate of 8 mLminꢀ1 and heated to 3008C at a rate of
18Cminꢀ1. During the temperature desorption step, a hydrogen flow
(4 vol% in N2) was passed over the catalyst bed from RT to 4008C at a
rate of 108Cminꢀ1. A commercial copper catalyst (BTS-catalyst 30 wt%
CuO/SiO2, 125–250 mm, R3–11, lot nr. 0849, BASF) was used as reference
standard.
We thank the Dutch Technology Foundation STW (project MEMFiCS
GSPT-07974), DSM Research, Friesland Campina, LioniX, IMM (Ger-
many), Milestone s.r.l. (Italy) for their financial and in-kind support. We
also acknowledge the European Commission (project NoE EXCELL
NMP3-CT-2005–515703) and The Royal Society (International Joint Proj-
ect 2008/R4 ꢁꢁSmart Structured Multiphase Reactors for Process Intensifi-
cationꢁꢁ). Thanks go to Prof. Dr. Bert W. Meijer and Dr. Jef A. J. M.
Vekemans for the analytical equipment and fruitful discussions.
Catalytic activity measurements: Potassium phenolate and 4-chloropyri-
dine were separately prepared as described elsewhere.[23] A mixture con-
taining an accurate amount of 4-chloropyridine (0.75 mol), potassium
phenolate (0.90 mol), and 18-crown-6 as appropriate (ca. 0.01 molequiv
with respect to phenolate), was stirred in a baffled glass reactor (75 mL)
loaded with DMA (15–25 mL) at 458C until dissolution was complete.
After heating the reaction mixture to the desired temperature, an appro-
priate amount of catalyst was added to the reactor, which was mechani-
cally stirred at 500 or 1500 rpm. The procedure was carried out in an
argon atmosphere and the resulting slurry was heated. Due to the differ-
ent synthetic protocols different copper loadings resulted; therefore the
amount of catalyst in the reactor was adjusted for each catalyst type to
provide similar Cu metal loadings. The solvent temperature during the
reaction was measured using a thermocouple. The yield of 4-phenoxypyr-
idine was determined by recording the 1H NMR spectra of reaction ali-
[1] a) J. M. Campelo, D. Luna, R. Luque, A. A. Romero, Chem
Chem 2009, 2, 18–45; b) R. Ferrando, J. Jellinek, R. L. Johnston,
Chem. Rev. 2008, 108, 845–910.
[2] C. N. R. Rao, G. U. Kulkarni, P. J. Thomas, P. P. Edwards, Chem.
Soc. Rev. 2000, 29, 27–35.
[3] R. Ferrando, A. Fortunelli, R. L. Johnston, Phys. Chem. Chem.
Phys. 2008, 10, 640–649.
[4] W. J. Zhang, Y. Liu, R. G. Cao, Z. H. Li, Y. H. Zhang, Y. Tang, K. N.
Fan, J. Am. Chem. Soc. 2008, 130, 15581–15588.
[5] a) M. Mavrikakis, B. Hammer, J. K. Nørskov, Phys. Rev. Lett. 1998,
81, 2819–2822; b) E. V. Rebrov, A. Berenguer-Murcia, H. E. Skel-
ton, B. F. G. Johnson, A. E. H. Wheatley, J. C. Schouten, Lab Chip
2009, 9, 503–506.
[6] V. Engels, F. Benaskar, D. A. Jefferson, B. F. G. Johnson, A. E. H.
Wheatley, Dalton Trans. 2010, 39, 6496–6502.
ACHTUNGTRENNUNGSus-
AHCTUNGTRENNUNG
Chem. Eur. J. 2012, 18, 1800 – 1810
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1809

J. C. Schouten, A. E. H. Wheatley et al.
[7] a) J. K. Nørskov, T. Bligaard, J. Rossmeisl, C. H. Christensen, Nat.
Chem. 2009, 1, 37–46; b) G. A. Somorjai, C. J. Kliewer, React. Kinet.
Catal. Lett. 2009, 55, 191–208.
c) Z. R. Ismagilov, E. V. Matus, A. M. Yakutova, L. N. Protasova,
I. Z. Ismagilov, M. A. Kerzhentsev, E. V. Rebrov, J. C. Schouten,
Catal. Today 2009, 147, S81–S86.
[8] E. Sperotto, G. P. M. van Klink, G. van Koten, J. G. de Vries, Dalton
Trans. 2010, 39, 10338–10351.
[20] M. Kidwai, N. K. Mishra, V. Bansal, A. Kumar, S. Mozumdar, Tetra-
hedron Lett. 2007, 48, 8883–8887.
[9] F. Ullmann, P. Sponagel, Ber. Dtsch. Chem. Ges. 1905, 38, 2211–
2212.
[21] J. Y. Kim, J. C. Park, A. Kim, A. Y. Kim, H. J. Lee, H. Song, K. H.
Park, Eur. J. Inorg. Chem. 2009, 4219–4223.
[10] I. Goldberg, Ber. Dtsch. Chem. Ges. 1906, 39, 1691–1692.
[11] a) F. Benaskar, V. Engels, N. Patil, E. V. Rebrov, J. Meuldijk, V.
Hessel, L. A. Hulshof, D. A. Jefferson, J. C. Schouten, A. E. H.
Wheatley, Tetrahedron Lett. 2010, 51, 248–251.
[22] K. Knepper, M. E. P. Lormann, S. Brase, J. Comb. Chem. 2004, 6,
460–463.
[23] V. Engels, F. Benaskar, N. Patil, E. V. Rebrov, V. Hessel, L. A. Hul-
shof, D. A. Jefferson, J. A. J. M. Vekemans, S. Karwal, J. C. Schout-
en, A. E. H. Wheatley, Org. Process Res. Dev. 2010, 14, 644–649.
[24] a) J. A. Tang, B. D. Ellis, T. H. Warren, J. V. Hanna, C. L. B. Mac-
donald, R. W. Schurko, J. Am. Chem. Soc. 2007, 129, 13049–13065;
b) D. Rusanova, K. J. Pike, R. Dupree, J. V. Hanna, O. N. Antzutkin,
I. Persson, W. Forsling, Inorg. Chim. Acta 2006, 359, 3903–3910;
c) B. E. G. Lucier, J. A. Tang, R. W. Schurko, G. A. Bowmaker, P. C.
Healy, J. V. Hanna, J. Phys. Chem. C 2010, 114, 7949–7962.
[25] a) A. Casitas, I. Nikolaos, M. George, M. Costas, X. Ribas, Dalton
Trans. 2011, 40, 8796–8799; b) F. Benaskar, V. Engels, N. G. Patil,
M. Chaibi, E. V. Rebrov, J. Meuldijk, V. Hessel, L. A. Hulshof,
A. E. H. Wheatley, J. C. Schouten, unpublished results.
[26] The International Centre for Diffraction Data (ICDD), Newtown
Square, PA, 2010.
[27] a) J. Greeley, A. A. Gokhale, J. Kreuser, J. A. Dumesic, H. Topsoe,
N.-Y. Topsoe, M. Mavrikakis, J. Catal. 2003, 213, 63–72; b) P. L.
Hansen, J. B. Wagner, S. Helveg, J. R. Rostrup-Nielsen, B. S. Clau-
sen, H. Topsoe, Science 2002, 295, 2053–2055.
[12] a) J. F. Hartwig, Angew. Chem. 1998, 110, 2154–2177; Angew. Chem.
Int. Ed. 1998, 37, 2046–2067; b) J. F. Hartwig, Acc. Chem. Res. 1998,
31, 852–860; c) A. F. Littke, G. C. Fu, Angew. Chem. 2002, 114,
4350–4386; Angew. Chem. Int. Ed. 2002, 41, 4176–4211; d) A. R.
Muci, S. L. Buchwald, Top. Curr. Chem. 2002, 219, 131–209;
e) Handbook of organopalladium chemistry for organic synthesis
(Eds.: E.-i. Negishi, A. de Meijere), Wiley-Interscience, New York,
2002; f) D. Prim, J. M. Campagne, D. Joseph, B. Andrioletti, Tetrahe-
dron 2002, 58, 2041–2075; g) J. P. Wolfe, S. Wagaw, J.-F. Marcoux,
S. L. Buchwald, Acc. Chem. Res. 1998, 31, 805–818; h) B. H. Yang,
S. L. Buchwald, J. Organomet. Chem. 1999, 576, 125–146.
[13] J. F. Hartwig, M. S. Driver, F. E. Goodson, B. C. Hamann, J. Louie,
G. Mann, Abstr. Am. Chem. Soc. 1998, 216, U449–U449.
[14] a) P. Lopez-Alvarado, C. Avendano, J. C. Menendez, J. Org. Chem.
1995, 60, 5678–5682; b) D. H. R. Barton, J. P. Finet, Pure Appl.
Chem. 1987, 59, 937–946.
[15] P. Y. S. Lam, S. Deudon, K. M. Averill, R. Li, M. Y. He, P. DeShong,
C. G. Clark, J. Am. Chem. Soc. 2000, 122, 7600–7601.
[16] a) P. Y. S. Lam, G. Vincent, D. Bonne, C. G. Clark, Tetrahedron Lett.
2002, 43, 3091–3094; b) P. Y. S. Lam, C. G. Clark, S. Saubern, J.
Adams, K. M. Averill, D. M. T. Chan, A. Combs, Synlett 2000, 5,
674–676.
[28] N. S. McIntyre, S. Sunder, D. W. Shoesmith, F. W. Stanchell, J. Vac.
Sci. Technol. 1981, 18, 714–721.
[29] a) K. V. R. Chary, G. V. Sagar, C. S. Srikanth, V. V. Rao, J. Phys.
Chem. B 2007, 111, 543–550; b) M.-F. Luo, P. Fang, M. He, Y.-L.
Xie, J. Mol. Catal. A: Chem. 2005, 239, 243–248.
[17] a) D. M. T. Chan, K. L. Monaco, R. P. Wang, M. P. Winters, Tetrahe-
dron Lett. 1998, 39, 2933–2933; b) D. A. Evans, J. L. Katz, T. R.
West, Tetrahedron Lett. 1998, 39, 2937–2937; c) P. Y. S. Lam, C. G.
Clark, S. Saubern, J. Adams, M. P. Winters, D. M. T. Chan, A.
Combs, Tetrahedron Lett. 1998, 39, 2941–2944.
[18] a) J. Hassan, M. Sevignon, C. Gozzi, E. Schulz, M. Lemaire, Chem.
Rev. 2002, 102, 1359–1469; b) S. V. Ley, A. W. Thomas, Angew.
Chem. 2003, 115, 5558–5607; Angew. Chem. Int. Ed. 2003, 42, 5400–
5449; c) F. Monnier, M. Taillefer, Angew. Chem. 2009, 121, 7088–
7105; Angew. Chem. Int. Ed. 2009, 48, 6954–6971.
[30] J. W. Evans, M. S. Wainwright, A. J. Bridgewater, D. J. Young, Appl.
Catal. 1983, 7, 75–83.
[31] A. Steinfeld, Int. J. Hydrogen Energy 2002, 27, 611–619.
[32] a) M. A. Newton, Chem. Soc. Rev. 2008, 37, 2644–2657; b) F. J. Lai,
W. N. Su, L. S. Sarma, D. G. Liu, C. A. Hsieh, J. F. Lee, B. J. Hwang,
Chem. Eur. J. 2010, 16, 4602–4611.
[33] L. N. Protasova, E. V. Rebrov, T. S. Glazneva, A. Berenguer-Murcia,
Z. R. Ismagilov, J. C. Schouten, J. Catal. 2010, 271, 161–169.
[19] a) M. S. Chen, D. W. Goodman, Science 2004, 306, 252–255; b) M.
Jakob, H. Levanon, P. V. Kamat, Nano Lett. 2003, 3, 353–358;
Received: July 13, 2011
Published online: January 3, 2012
1810
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 1800 – 1810