Supported nitrogen-modified Pd nanoparticles for the selective hydrogenation of 1-hexyne

Article abstract of DOI:10.1016/j.jcat.2011.01.003

2,2′-Bipyridine (bipy) and an imidazolium-functionalized bipy ligand were used in the synthesis of palladium nanoparticles in water and in ionic liquid. The function of such ligands was twofold: first, as stabilizing agents to prevent the agglomeration of the nanoparticles during synthesis, and second, to act as permanent catalyst modifiers. The N-modified Pd nanoparticles were subsequently deposited onto Carbon Nanofiber-based structured supports and tested in the liquid-phase hydrogenation of 1-hexyne. The catalysts were found to be significantly more selective (up to 98.5% at 25% conversion) than a reference catalyst with non-modified Pd nanoparticles on the same support (88%). Moreover, the high selectivity was maintained up to full conversion, and thus, the over-hydrogenation was suppressed due to a site-blocking effect of the N-containing ligands. The imidazolium-functionalized bipy ligand was found to interact more strongly with the nanocarbon support reinforcing Pd anchoring and reducing leaching into the liquid media.

Full text of DOI:10.1016/j.jcat.2011.01.003

Journal of Catalysis 279 (2011) 66–74
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Supported nitrogen-modified Pd nanoparticles for the selective hydrogenation
of 1-hexyne
⇑
Micaela Crespo-Quesada, Ryan R. Dykeman, Gabor Laurenczy, Paul J. Dyson, Lioubov Kiwi-Minsker
Institute of Chemical Sciences and Engineering, Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland
a r t i c l e i n f o
a b s t r a c t
2,20-Bipyridine (bipy) and an imidazolium-functionalized bipy ligand were used in the synthesis of palla-
dium nanoparticles in water and in ionic liquid. The function of such ligands was twofold: first, as stabiliz-
ing agents to prevent the agglomeration of the nanoparticles during synthesis, and second, to act as
permanent catalyst modifiers. The N-modified Pd nanoparticles were subsequently deposited onto Carbon
Nanofiber-based structured supports and tested in the liquid-phase hydrogenation of 1-hexyne. The cata-
lysts were found to be significantly more selective (up to 98.5% at 25% conversion) than a reference catalyst
with non-modified Pd nanoparticles on the same support (88%). Moreover, the high selectivity was main-
tained up to full conversion, and thus, the over-hydrogenation was suppressed due to a site-blocking effect
of the N-containing ligands. The imidazolium-functionalized bipy ligand was found to interact more
stronglywith the nanocarbon support reinforcing Pd anchoring and reducing leaching into the liquid media.
Ó 2011 Elsevier Inc. All rights reserved.
Article history:
Received 26 September 2010
Revised 8 December 2010
Accepted 4 January 2011
Available online 15 February 2011
Keywords:
Heterogeneous catalysis
Bipyridine ligands
Catalyst modifier
1-Hexyne hydrogenation
Palladium nanoparticles
Carbon nanofiber
Structured support
Ionic liquid
1. Introduction
from the coordinating ligand [4] that leads to a change in the
alkyne/alkene relative strength of adsorption. A decrease in
About 20% of the reactions used to produce fine chemicals and
pharmaceuticals are hydrogenations [1]. Supported Pd nanoparti-
cles (NPs) are known to selectively catalyze the hydrogenation of
alkynes to alkenes [2]. The yield of the alkene hydrogenation prod-
uct increases considerably when modifiers are added to the reac-
tion mixture. Recently, DFT calculations have been used to
explain the effect of CO, a commonly used modifier for gas-phase
hydrogenations, on the activity and selectivity of Pd in the hydro-
genation of acetylene [3].
In liquid-phase reactions, common modifiers include nitrogen-
containing ligands (ammonia, quinoline [4–7], pyridine, etc.) and
sulfur-containing compounds [8]. Phenanthroline and 2,20-bipyri-
dine (bipy) have also been used to stabilize Pd NPs with the former
(including alkylated derivatives) used for various partial hydroge-
nation reactions in both organic solvents [9,10] and ionic liquids
(ILs) [11].
alkene heat of adsorption favors its desorption and increases
the selectivity [4,12–14].
ꢀ Poisoning (site blocking) [8] via preferential adsorption of the
modifier compared to the alkene [5].
The molar ratio of nitrogen-containing ligand to substrate used
in hydrogenations ranges from 2 to ꢁ1500 [14–17], and the selec-
tivity toward the alkene increases with modifier concentration.
Activity, on the other hand, may either decrease or increase
depending on the metal dispersion. Indeed, the turnover frequency
(TOF) of the liquid-phase hydrogenation of 1-butyne was found to
increase by a factor of 5 over highly dispersed catalysts in the pres-
ence of piperidine, but this effect was found to be negligible at low
dispersions. Piperidine, being an electron-donating ligand, de-
creases the strong binding energy of alkynes on small metal parti-
cles, which increases the activity [14]. Increased activity was also
observed for phenylacetylene hydrogenation in the presence of
quinoline at high ligand:Pd molar ratios [17]. However, other stud-
ies report that a decrease in activity takes place at a base/Pd molar
ratio of ꢁ1500 due to site blocking [5].
The mechanism responsible for the increase in selectivity in the
presence of modifiers is reaction specific, but in general, it can be
due to:
ꢀ A ‘‘ligand’’ effect – a nucleophilic modifier increases the electron
Additives can be either added to the reaction mixture (reaction
modifiers) or introduced as a component of the catalyst (catalyst
modifiers). According to the latter approach, a 0.5% Pd/Al2O3 cata-
lyst treated with Zn acetate, pyridine, and KOH showed high selec-
tivity (up to 99.8%) in 2-methyl-3-butyn-2-ol hydrogenation in
density of the palladium surface through electron donation
⇑
Corresponding author. Address: EPFL-SB-ISIC-GGRC, Station 6, Switzerland.
Fax: +41 21 693 31 90.
E-mail address: lioubov.kiwi-minsker@epfl.ch (L. Kiwi-Minsker).
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.01.003

M. Crespo-Quesada et al. / Journal of Catalysis 279 (2011) 66–74
67
Nomenclature
[BIHB]Br2 {4,40-bis[7-(2,3-dimethylimidazolium)heptyl]-2,20-
bipyridine}bromide
Symbols
Ci
D
concentration of a given species, i [mol Lꢂ1
]
[bmim][PF6] 1-butyl-3-methyl-imidazolium
phate
hexafluorophos-
dispersion of a nanoparticle [%]
ꢂ1
ki
apparent kinetic constant of the reaction i [L mol sꢂ1
]
Pd
[C2OHmim][BF4] 1-(2-hydroxyethyl)-3-methyl-imidazolium tet-
rafluoroborate
Ki
mPd
ri
V
X
adsorption constant of i [L molꢂ1
]
amount of active material in the catalyst [mol]
ꢂ1
[C3CNmim][Tf2N] 1-(3-cyanopropyl)-3-methyl-imidazolium
bis(trifluoromethane)sulfonimide
rate of the reaction i [mol mol sꢂ1
]
Pd
volume of liquid in the reaction [L]
conversion of 1-hexyne [%]
Bipy
CNF
SMF
SC
2,20-bipyridine
carbon nanofiber
sintered metal filter
supported catalyst
ethanol at a reduced hydrogen pressure [13]. A permanent modifi-
cation of Pd NPs with pyridine has been achieved by stabilizing
them inside poly-4-vinylpyridine block-copolymers – either poly-
styrene-block-poly-4-vinylpyridine [18,19] or poly(ethylene
oxide) – block-poly-2-vinylpyridine [20]. Both systems were stud-
ied in the hydrogenation of dehydrolinalool and showed 99% selec-
tivity, which was attributed to the electronic effects of the pyridine
functionality on the Pd surface. It has been noted that functional-
ized phenanthroline stabilizers, for example, substituted with sul-
fonate groups, can result in problems associated with the
immobilization of the resulting NPs. However, alkyl chains in the
3-position have been shown to decrease isomerization during the
partial hydrogenation of 2-hexyne [9,21]. Furthermore, bipy and
a dipyridyl amine functionalized with an imidazolium moiety have
been used to stabilize Rh and Pd NPs that effectively catalyzed the
hydrogenation of arenes and alkenes, respectively [22–24].
Sintered Metal Fibers (SMF) with a porosity of 80%, high
mechanical strength, and effective heat transfer properties are
advantageous structured supports. In order to increase the low
specific surface area of the SMFs, carbon nanofibers (CNF) were
grown on the surface, thus providing a tridimensional mesoporous
chemically inert graphite-like structure [25,26]. This composite
support can be subsequently oxidized to generate surface oxy-
gen-containing groups, thus enabling an active phase to be
strongly anchored.
performed in air unless otherwise specified. Reactions performed
under N2 were carried out using standard Schlenk-line techniques
or in a dry-box. Toluene was dried catalytically under nitrogen using
a solvent purification system manufactured by Innovative Technol-
ogy Inc, and water was purified with a Pure Water System supplied
by PG-Instruments. 4,40-Bis(7-bromoheptyl)-2,20-bipyridine [27],
[bmim][PF6] [28], [C2OHmim][BF4] [29], and [C3CNmim][Tf2N]
[30] were synthesized according to literature procedures.
2.2. Preparation of the Pd/CNF/SMF reference catalyst
Synthesis of the composite CNF/SMF support is detailed else-
where [26]. Briefly, after a calcination step of the SMFInconel to in-
crease surface roughness followed by a reduction step, ethane
decomposition in the presence of hydrogen is performed at
655 °C for 1 h rendering the SMFs homogeneously coated by a
CNF layer (ꢁ1
lm). The composites were then treated in boiling
hydrogen peroxide for 4 h in order to generate oxygen-containing
groups on the surface of the CNFs.
The method used to synthesize and immobilize the Pd NPs onto
the supports has been reported elsewhere [25]. Briefly, the acti-
vated CNF/SMF supports were immersed in a solution containing
a Pd precursor (Na2PdCl4) for 5 h at room temperature. Afterward,
they were dried, calcined, and reduced for 18 h, thus generating Pd
NPs with a mean diameter of ꢁ6.7 nm.
In the studies reported herein, an in-house-synthesized imi-
dazolium-tagged bipy ligand and bipy itself were used in the syn-
thesis of monodispersed Pd NPs with a dual aim: stabilizing the
nanoparticles by anchoring them to the support and inducing a
permanent modification of the surface. To the best of our knowl-
edge, the combination of a functionalized structured support and
a tailor-made ligand capable of strongly interacting in order to
minimize leaching while increasing selectivity has not been
previously reported. The resulting catalysts were tested in the
liquid-phase hydrogenation of 1-hexyne in n-heptane, and their
performance was compared to that of a reference Pd/CNF/SMF cat-
alyst, prepared by a conventional ion-exchange method [25] in the
absence of modifiers. The catalysts were further characterized by
scanning and transmission electron microscopy (SEM and TEM),
X-ray photoelectron spectroscopy (XPS), and CO-chemisorption
in order to rationalize the catalytic response observed.
2.3. N-modified Pd catalyst preparation
2.3.1. N-containing ligand synthesis
{4,40-Bis[7-(2,3-dimethylimidazolium)heptyl]-2,20-bipyridine}bro-
mide ([BIHB]Br2). Under an atmosphere of nitrogen, 4,40-bis(7-
bromoheptyl)-2,20-bipyridine (0.638 g, 1.25 mmol) and 1,2-
dimethylimidazole (0.804 g, 8.54 mmol) in toluene (50 mL) were
stirred under reflux for 20 h, after which a white solid formed.
The solid was removed by filtration and washed with ether
(20 mL), and residual solvent was removed in vacuo to give
[BIHB]Br2 (0.829 g, 94%) (see Scheme 1). Anal. Calc. for
C34H50Br2N6: C, 58.12%; H, 7.17%; N, 11.96%. Found: C, 58.18%; H,
7.45%; N, 11.86%. 1H NMR (400 MHz, D2O): d = 8.33 (d, J = 4.9 Hz,
2H), 7.73 (s, 2H), 7.22 (m, 2H+2H), 7.11 (d, J = 4.4 Hz, 2H), 3.93 (t,
J = 7.3 Hz, 4H), 3.67 (s, 6H), 2.46 (m, 4H+6H), 1.61 (m, 4H), 1.59
(m, 4H), 1.41, 1.11 ppm (M, 12H). 13C NMR (400 MHz, D2O):
d = 154.80, 154.32, 148.76, 143.95, 124.72, 122.15, 121.76,
120.61, 48.07, 34.64, 29.49, 28.94, 28.17, 28.11, 25.48, 8.93 ppm.
2. Materials and methods
2.1. Materials
2.3.2. Pd nanoparticle synthesis and immobilization
All chemicals were purchased from commercial sources and
used without further purification. All manipulations were
Both bipy and [BIHB]Br2-stabilized Pd NPs were prepared using
the same procedure. In a typical experiment, Pd(CH3CO2)2 (5.6 mg,

68
M. Crespo-Quesada et al. / Journal of Catalysis 279 (2011) 66–74
N
Br
Br
N
Br
7
Br
N
Li
Li
Br
Br
7
N
4
N
N
n-BuLi, HNiPr2
THF , 78 to RT
7
7
toluene, Δ
N
N
N
N
N
N
37% Yield
N
N
85%Yield
Scheme 1. Synthetic route for the modified bipyridine stabilizer, [BIHB]Br2.
0.025 mmol), bipy (3.9 mg, 0.025 mmol) in 4 mL of either water or
[bmim][PF6] were stirred for 10 min and then placed in an ultra-
sonic bath for a further 10 min. The mixture was then placed in
an autoclave, pressurized to 10 bar under H2, and stirred for 1 h.
The reduction resulted in an opaque black solution, which was
then stirred for 1 h while exposed to air.
The Pd NPs obtained were separated by centrifugation, sus-
pended in fresh solvent, separated again by centrifugation, and
then redispersed in acetone. The nanoparticles were subsequently
deposited onto the activated CNF/SMF supports (see above) by
incipient wetness impregnation and dried for 15 h at 50 °C under
vacuum.
2.5. Characterization techniques
NMR spectra were recorded on either an Avance 400 MHz or a
400 MHz Bruker DRX instrument with 1H chemical shifts refer-
enced to residual solvent peaks.
X-ray photoelectron spectroscopy (XPS) data were collected by
an Axis Ultra instrument (Kratos analytical, Manchester, UK) under
ultra-high-vacuum condition (<10ꢂ8 Torr) and using a monochro-
matic Al K
tained at 150 W, and the emitted photoelectrons were sampled
from a square area of 750 ꢃ 350
m2. The photoelectron take-off
a X-ray source (1486.6 eV). The source power was main-
l
angle, between the surface and the direction in which the photo-
electrons were analyzed, was 90°. The analyzer pass energy was
80 eV for survey spectra (0–1000 eV) and 40 eV for high-resolution
spectra: Pd 3d5/2, O 1s, and C 1s. The adventitious carbon 1s peak
was calibrated at 284.5 eV and used as an internal standard to
compensate for any charging effects. Both curve fitting of the spec-
tra and quantification were performed with the CasaXPS software,
using relative sensitivity factors given by Kratos.
2.4. Hydrogenation reactions
Hydrogenations were carried out in a semi-batch stainless steel
reactor (250 mL autoclave, Büchi AG, Switzerland) equipped with a
heating jacket and a hydrogen supply system. The structured cata-
lyst was placed between two metal gauzes (2 ꢃ 8.5 cm) fixed on a
self-gassing hollow shaft, which was used as the stirrer. The reac-
tor was charged with 1-hexyne (4.13 g) and octane (2.82 g) and
filled to 200 mL with n-heptane. The reactor was flushed with N2
and set to the target temperature. The reactor was then flushed
with H2 and pressurized. The consumption of hydrogen was mon-
To determine the amount of Pd in the catalyst, the catalyst was
dissolved in hot nitric acid, and the sample was analyzed by atomic
absorption spectroscopy via Shimadzu AA-6650 spectrometer with
an air–acetylene flame.
Pd dispersion was measured by pulse adsorption of CO (3% CO
in He) carried out at 323 K in a Micromeritics AutoChem 2910. Be-
fore measurements, the catalyst was pretreated at 353 K in flows of
He (10 mL/min, STP), H2 (20 mL/min, STP), and He (10 mL/min,
STP). A stoichiometry of CO/Pd = 0.6 and a Pd surface density of
1.2 ꢃ 1019 atoms/m2 were used for these calculations.
itored using
Switzerland).
a Pressflow gas Controller (BPC-6002) (Büchi,
In addition, samples were periodically withdrawn from the
reactor via a sampling tube and analyzed by GC. GC analysis was
performed using an Auto System XL (Perkin Elmer) equipped with
TEM pictures were obtained using a Philips CM20 Electron
Microscope operating at 200 kV. SEM images were recorded on a
FEI XL-30 SFEG Sirion Electron Microscope equipped with a TLD
detector.
a 100 m Petrocol DH 0.25 mm capillary column with a 0.5
l coat-
ing at an oven temperature of 333 K. The pressure of the carrier gas
(He) was 280 kPa. The temperature of the Injector and the FID was
493 K, and n-octane was used as internal standard.
The solubility of H2 in [C3CNmim][Tf2N] was estimated using a
literature method [31]. At a pressure of 100 atm H2, 2.5 mL of the IL
was sealed in a sapphire NMR tube. After vigorous shaking, a 1H
NMR spectrum of the solution was recorded. The 1H NMR spectrum
was fitted with WINNMR and NMRICMA2.8/MATALAB programs
(nonlinear least squares fit, minimizing the difference between
the measured and calculated spectra to determine the spectral
parameters and integrals). The Henry’s constant for H2 was found
to be 510 MPa and the hydrogen concentration 1.28 mM at 1 atm
[32].
The conversion and selectivity were calculated from:
C01ꢂhexyne ꢂ C1ꢂhexyne
X1ꢂhexyne
¼
ð1Þ
ð2Þ
C01ꢂhexyne
Ci
P
S ¼
ꢄ 100
productsCi
The turnover frequency (TOF) was calculated as:
In order to measure solubility of the organic compounds in the
ILs, an excess of the compound (either n-hexane, 1-hexene or
1-hexyne) was added to a 3-mL screw cap vial containing a mag-
netic stir bar and the IL ([bmim][PF6], [C2OHmim][BF4] or
[C3CNmim][Tf2N]) and mixed at room temperature (20.0 °C) for
14 h, after which time the majority of the IL and a fraction of the
organic compound were transferred to a 5-mm NMR tube and al-
lowed to settle to ensure only the IL layer would be analyzed by
the receiver coil. Spectra were recorded at 25.1 °C without an inter-
nal reference. The resulting 1H NMR spectra were analyzed using
Bruker TOPSPIN 1.3 NMR software with deconvolution (fit type
r ½mol=molPd sꢅ
TOFg ½sꢂ1ꢅ ¼
ð3Þ
D
where D is the dispersion of the nanoparticle, defined as the ratio
between the number of surface atoms and the total number of
atoms of the nanoparticle, and r is the measured transformation
rate expressed in mols of 1-hexyne consumed per mol of Pd and
per unit of time.
For the leaching experiments, the catalysts were placed in the
stirrer cage and subjected to the same conditions as the reaction
runs except for hexyne in the liquid phase.

M. Crespo-Quesada et al. / Journal of Catalysis 279 (2011) 66–74
69
of Lorentzian) employed where necessary to obtain the appropriate
integration.
coating the surface of each fiber (of approximately 8
l
m in diame-
ter) of the SMFInconel (Fig. 1a). The white dots that can be observed
in the SEM image are Ni particles that are detached from the SMF
during the formation of the CNF and serve as catalysts for the for-
mation of the latter, as confirmed by EDX [26]. The TEM image pre-
sented in Fig. 1b shows monodispersed (ꢁ7 nm) Pd NPs that are
evenly distributed along a single CNF. The mean particle size was
also estimated via CO-chemisorption, yielding a value of 6.7 nm,
in good agreement with the TEM images. Fig. 1c shows a TEM im-
age of an isolated CNF. It can be appreciated that the Ni particle at
the tip of the fiber is covered by several layers of graphitic carbon.
The inset of the image shows a high resolution image of the graph-
ite-coated Ni nanoparticle. This layer of graphite ensures their
inertness.
3. Results and discussion
3.1. Catalysts used in this work
Table 1 lists the catalysts used throughout this work and pro-
vides a succinct description of the procedure used for their synthe-
sis; for more detailed descriptions see Sections 2.2 and 2.3.
3.2. Hydrogenation of 1-hexyne over the reference Pd/CNF/SMF
catalyst
Before comparing catalyst performances or deriving kinetic
expressions, it is necessary to verify whether the system operates
under kinetic regime. Both heat and internal and external mass
transfer influences were assessed for the reference catalyst. The
detailed discussion can be found in the Supporting Information.
Under the conditions used, the reaction was found to be under
the kinetic regime for the reference catalyst; being the most active
catalyst studied, this ensures a similar result for all the others.
The reference catalyst showed a high activity in the hydrogena-
tion of 1-hexyne with a TOF ꢁ200 sꢂ1 (see Scheme 2 for the reac-
tion network). The formation of n-hexane takes place not only in a
consecutive (path 2, Scheme 2) but also in a parallel way (path 3,
Scheme 2) since their formation is already detected at conversions
close to zero. 2-hexene isomers were almost undetectable until the
maximum yield of 1-hexene was reached. After the complete
The morphology of the reference catalyst comprises a network
of CNFs with an individual diameter of approximately 40–50 nm
Table 1
Catalysts used during this work.
Catalyst
Stabilizer
NP synthesis in IL coating
% Pda % IL
Reference
Reference
Reference
Reference
SILC(Bip)
SILC(Bih)
SC(Bip)
–
–
–
–
–
–
–
–
–
0.6b
–
[bmim][PF6]
[C2OHmim][BF4]
0.82b
0.48b
4.2
5.4
4.2
3.0
3.7
–
[C2CNmim][Tf2N] 0.62b
Bipy
[bmim][PF6]
[bmim][PF6]
[bmim][PF6]
–
–
4.0
3.9
3.2
2.0
[BIHB]Br2 [bmim][PF6]
Bipy H2O
[BIHB]Br2 H2O
SC(Bih)
–
a
Percentage expressed per mass of CNFs.
Percentage expressed per total mass of the catalyst.
b
Fig. 1. Morphology of the reference catalyst. (a) SEM image of the CNF/SMF support and (b) TEM showing well-dispersed Pd NPs anchored onto the CNF (c) TEM of a single
CNF where the graphitic layer protecting the Ni particle can be appreciated. The inset shows a HR-TEM image of the nanoparticle where the layers of C can be seen. The scale
bar in the inset corresponds to 10 nm.
Scheme 2. Reaction network for the hydrogenation of 1-hexyne.

70
M. Crespo-Quesada et al. / Journal of Catalysis 279 (2011) 66–74
Table 2
Adsorption and kinetic constants found for the reference catalyst and SILC(Bih).a
Adsorption constants (L molꢂ1
)
ꢂ1
Kinetic constants ðmol g sꢂ1
Þ
Pd
Reference
SILC(Bih)
Reference
SILC(Bih)
KY
KN
KE
K
23.1
–
0.8
35.9
9.6
0.1
k1
k2
k3
k4
2486.1
1162.3
208.8
11.1
2503.3
54.1
154.4
10.1
b
8.3
4.7
1 ꢃ 10ꢂ3
2 ꢃ 10ꢂ3
c
KH
a
Reaction conditions: 4.1 g of 1-hexyne, 2.82 g of octane as internal standard,
0.3 g of catalyst, 303 K, and 1.05 MPa of H2 in 200 mL of n-heptane.
b
Dimensionless.
This constant is the product KH CH and is dimensionless.
c
2
2
These eight Eqs. (4)–(11) can be solved simultaneously using
Berkeley’s Madonna software [34] with Rosenbrock’s method
[35]. The model parameters were estimated by fitting the simu-
lated curves to the experimental data giving the kinetic constants
k1–k4 and the adsorption constants KY, KE, KH, and K presented in
Table 2. It is worth noting that fitting such expressions makes
the result dependent on the starting values given to the constants.
In order to obtain reliable results, the starting values for each con-
stant are set and certain constraints applied, such as hydrogen
adsorption constant much smaller than the rest, no negative values
and higher values for k1 and KY. This kinetic model has been suc-
cessfully applied for other liquid-phase alkyne hydrogenations
over palladium catalysts [36,37]. Fig. 2 shows the prediction curves
obtained with the kinetic model proposed. A correlation coefficient
of 98.5% was achieved according to Eq. (12).
Fig. 2. Experimental points and kinetic curves modeled using Berkeley’s Madonna
software for the reference catalyst. d 1-Hexyne, j 1-hexene, N n-hexane, . 2-
hexene isomers. Reaction conditions: 4.1 g of 1-hexyne, 2.82 g of octane as internal
standard, 0.3 g of catalyst, 303 K, and 1.05 MPa of H2 in 200 mL of n-heptane.
consumption of 1-hexyne, 1-hexene was rapidly transformed into
hexane and its two isomers with a ratio of 1:1 (Fig. 2).
A Langmuir–Hinshelwood mechanism was applied for model-
ing the kinetics assuming dissociative weak adsorption of hydro-
gen. For path 1 and 3 (Scheme 2), the second addition of
hydrogen was considered as the rate-determining step. For path
2, the first addition of hydrogen was considered as the rate-deter-
mining step. The adsorption equilibrium constants of the hydroge-
nated products are small compared to the initial acetylenic
compound. Therefore, the adsorption constants of n-hexane and
the alkene isomers, K, were considered to be equal. This assump-
tion, together with the weak hydrogen adsorption on Pt group met-
als [33], allows the following (simplified) equations to be deduced
(see the list of symbols, subscripts Y, E, A, and I refer to the alkyne,
alkene, alkane, and isomers, respectively):
2
3
2
ꢀ
ꢀ
R
ðx ꢂ xÞðy ꢂ yÞ
6
4
7
5
r2
¼
ð12Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
ꢀ
ꢀ
R
ðx ꢂ xÞ2Rðy ꢂ yÞ
The adsorption constant of 1-hexyne was found to be 2 orders
of magnitude higher than that of 1-hexene, in line with other re-
sults reported for hydrogenations over Pd [37,38]. Furthermore,
the adsorption constant of hydrogen was 5 orders of magnitude
smaller than that of the alkyne, thus validating the weak adsorp-
tion assumption made when developing the model. The kinetic
constant of the first reaction (path 1, Scheme 2) was only twice
as high as that for the over-hydrogenation (path 2, Scheme 2). This
is in agreement with the thermodynamic basis for the selectivity
proposed by Molnar et al. [12], i.e., the high selectivity toward
the alkene is due to a stronger adsorption of the alkyne, compared
to the alkene, rather than a higher kinetic constant of different
hydrogenation steps.
k1KY CY KHCH
r1
r2
r3
¼
¼
¼
ð4Þ
ð5Þ
ð6Þ
2
2
2
ð1 þ KY CY þ KECE þ KðCA þ CIÞÞ
k2KY CY KHCH
ð1 þ KY CY þ KECE þ KðCA þ CIÞÞ
k3KY CY KHCH
ð1 þ KY CY þ KECE þ KðCA þ CIÞÞ
k4KEC2
E
r4
¼
ð7Þ
3.3. Hydrogenation of 1-hexyne over the N-modified Pd catalysts
2
ð1 þ KY CY þ KECE þ KðCA þ CIÞÞ
3.3.1. N-modified Pd NPs
In order to ascertain an overall kinetic rate equation, the full set
Fig. 3 shows TEM images and the particle size distributions of
each sample of the bipy-stabilized Pd NPs synthesized in water
and in [bmim][PF6]. The solvent influences the mean diameter of
the NPs, but the ligand (either bipy or [BIHB]Br2) does not, with
approximately the same particle size distribution in the two sol-
vents independent of the ligand employed. Pd loading on CNFs
was found to be between 2% and 4%.
of differential equations describing the concentration variation of
all species participating in the reaction is required (subscripts 1,
2, 3, and 4 refer to the different reaction paths in Scheme 2):
dCY
¼ ðꢂr1 ꢂ r3Þ
¼ ðr1 ꢂ r2 ꢂ r4Þ
¼ ðr2 þ r3Þ
¼ r4
ð8Þ
ð9Þ
dt
dCE
dt
3.3.2. Catalytic effect of N-containing ligands: supported N-modified
Pd NPs synthesized in [bmim][PF6]
Fig. 4 shows the catalytic results for SILC(Bih) (see Table 1) in
the hydrogenation of 1-hexyne. The differences compared to the
reference catalyst (Fig. 2) are significant, since the over-hydrogena-
tion toward the alkane (path 2, Scheme 2) appears to be
dCA
dt
ð10Þ
ð11Þ
dCI
dt

M. Crespo-Quesada et al. / Journal of Catalysis 279 (2011) 66–74
71
Fig. 3. TEM images and (c) particle size distribution of the Pd NPs synthesized with bipy as stabilizer in (a) [bmim][PF6] and (b) water.
adsorption/desorption regime, 0.2 g of the catalyst was stirred in
200 mL of heptane containing 4.1 g of hexyne at 30 °C for 3 h.
The sample was studied by 1H NMR spectroscopy, and both bipy
and [BIHB]Br2 were below the detection limit. The concentration
may be thus estimated to be <3.75 lmol/L. This implies that in or-
der to react, the alkyne must temporarily displace the modifier
from the active site, which probably adopts a more packed adsorp-
tion conformation, which is reverted to the original mode of
adsorption once the alkyne has reacted. Therefore, the N-based li-
gands most probably increase the selectivity through a permanent
site-blocking effect.
Second, an electronic effect is also observed. As suggested in the
literature, a nucleophilic ligand increases the electron density of
Pd, thus leading to a change in the alkyne/alkene relative strength
of adsorption as well as a decrease in alkene heat of adsorption,
both of which were observed in the model (Table 2) [4,12–14].
A well-known example of Pd permanent modification aimed at
enhancing the selectivity is the addition of Pb to Pd in Lindlar’s
industrial catalyst [39,40]. In order to compare its effectiveness with
the catalytic results reported herein, Lindlar’s catalyst (5%Pd
3.5%Pb/CaCO3) was tested under the same reaction conditions (see
Supporting Information for details). The selectivity was found to
be only 95% at 25% conversion (against 98.5% over our catalyst),
but dropped quickly when approaching full conversion. The most
important difference is that the over-hydrogenation reaction was
not hindered with Lindlar’s catalyst, implying that a reaction mod-
ifier would have to be added to the reacting mixture in practical use.
Fig. 4. Experimental points and kinetic curves modeled using Berkeley’s Madonna
software for SILC(Bih). d 1-Hexyne, j 1-hexene, N n-hexane, . 2-hexene isomers.
Reaction conditions: 4.1 g of 1-hexyne, 2.82 g of octane as internal standard, 0.3 g of
catalyst, 303 K, and 1.05 MPa of H2 in 200 mL of n-heptane.
completely suppressed. Furthermore, both the initial selectivity
and the maximum yield of 1-hexene increased by 10%. A similar ef-
fect, albeit less pronounced, was observed with a S-based modifier
[36] and quinoline [37] added to the reaction mixture in the liquid-
phase hydrogenation of 2-methyl-3-butyn-2-ol.
The same kinetic model as described above was used with the
exception that a new dimensionless constant, KN, was added in
the denominator of Eqs. (4)–(7) to take into account the permanent
adsorption of the N-containing ligands on the Pd NP surface:
3.3.2.1. Catalytic effect of ionic liquids. Supported Pd NPs embedded
in ILs have already been found useful for tailoring the selectivity in
several liquid-phase hydrogenations [41,42]. ILs exert an electronic
effect on embedded Pd NPs in the hydrogenation of acetylene caus-
ing an increase in the selectivity toward ethylene [43]. In order to
study the effect due solely to the IL in which the NPs were embed-
ded, the reference catalyst was covered with a layer of either
[bmim][PF6], [C2OHmim][BF4] or [C3CNmim][Tf2N] and tested in
the hydrogenation of 1-hexyne.
All catalysts were less active than the reference, and surpris-
ingly, the selectivity toward the target product did not improve
significantly. Nevertheless, the product distribution was affected
by the type of IL as seen in Fig. 5. With the use of [bmim][PF6],
selectivity toward 2-hexene isomers, particularly the trans-isomer,
was increased at the expense of the alkene compared to the
k1KY CY KHCH
r1
¼
ð13Þ
2
ðð1 þ KNÞ þ KY CY þ KECE þ KðCA þ CIÞÞ
A comparison of the constants is provided in Table 2. The results
shown suggest two possible explanations for the effect of the N-
containing ligands.
First, the initial coverage of the modifier (hN = 0.49) is slightly
higher than the maximum coverage of the substrate (hY = 0.46)
and is 2 orders of magnitude higher than the maximum coverage
of the desired product (hE = 2.4 ꢃ 10ꢂ3). In order to determine
whether the alkyne and the modifier were in
a dynamic

72
M. Crespo-Quesada et al. / Journal of Catalysis 279 (2011) 66–74
Fig. 5. Product distribution at 50% conversion for the IL-coated reference catalysts. Reaction conditions: 4.1 g of 1-hexyne, 2.82 g of octane as internal standard, 0.3 g of
catalyst, 303 K, and 1.05 MPa of H2 in 200 mL of n-heptane.
reference catalyst. Employing either [C2OHmim][BF4] or
[C3CNmim][Tf2N], on the other hand, showed an alkane:isomers
ratio significantly lower than that of the reference catalyst. 1H
NMR spectroscopy measurements revealed that the solubility of
1-hexyne, 1-hexene, and n-hexane is practically the same in
[bmim][PF6] and [C3CNmim][Tf2N] (see Table 4), hence eliminating
differences in solubility as being responsible for the differences ob-
served in the product distribution. Differences in H2 solubility in
ILs could also offer a possible explanation [41], but gas solubility
measurements showed no correlation between the two (Table 4).
A ligand site-blocking effect appears to be the most plausible
explanation for the change in product distribution since both hy-
droxyl and nitrile functionalities can conceivably interact with
the Pd surface, as is proposed for Pd NPs prepared in [C2OH-
mim][BF4] and [C3CNmim][Tf2N] reactions [30,44,45]. Presumably,
coverage of the Pd NP surface by the donor atoms of the IL is very
high and influences the reaction selectivity.
A similar behavior in the hydrogenation of 1-hexyne was found
regardless of the N-containing ligand used for synthesizing the Pd
NPs. The N-modified Pd NPs prepared in water demonstrated the
same initial selectivity toward 1-hexene (ꢁ98%). Moreover, the
reaction rate and TOF were in both cases significantly lower than
that of the reference catalyst.
XPS measurements show that the Pd NPs present different oxi-
dation states on their surfaces, as seen in Table 5 for Pd 3d5/2. All
three catalysts contain a peak between 335.2 and 335.4 eV, consis-
tent with metallic Pd. The reference catalyst shows a peak corre-
sponding to Pd2+ at 337.2 eV, which can be attributed to PdO.
The corresponding band of the two N-modified NPs was signifi-
cantly shifted to higher values (337.9–338.1 eV) corresponding to
Pd2+ ions coordinated with the N-containing ligands. Note,
Table 4
Determination of 1-hexyne, 2-hexene, n-hexane and H2 solubility in [bmim][PF6],
[C2OHmim][BF4] and [C3CNmim][Tf2N] by 1H NMR spectroscopy.
3.3.3. Catalytic effect of N-containing ligands: supported N-modified
Pd NPs prepared in water
Ionic liquid
Concentration (M)
a
1-Hexyne
1-Hexene
n-Hexane
H2 (mM)
The change in the catalytic behavior may be attributed to the N-
modified Pd NPs, and to confirm this hypothesis, the NPs were syn-
thesized in water, separated by centrifugation and redispersed in
acetone. The solution was then used for impregnating the CNF/
SMF composite supports (SC(Bip) and SC(Bih)), and the resulting
catalysts were tested under the same reaction conditions (see
Table 3).
[bmim][PF6]
[C3CNmim][Tf2N]
[C2OHmim][BF4]
1.24
0.85
0.17
0.23
0.18
<0.01
0.07
0.06
0.06
0.97b
0.47
1.28c
a
b
c
Measured at 10.1 MPa and adjusted for 0.101 MPa.
From Ref. [32].
From Ref. [29].
Table 5
Binding energies of Pd 3 5/2.
Table 3
Performance of the various catalysts tested in the hydrogenation of hexyne.a
ꢂ1
Catalyst
Pd 3d5/2 (eV)
Pd(0)
Reference
R ðmol mol sꢂ1
Þ
Catalyst
Sb (%)
TOF (sꢂ1
)
Pd
Pd(II)
Reference
SILC(Bip)
SILC(Bih)
SC(Bip)
88.0
97.0
94.0
98.4
97.8
47.0
4.3
11.0
1.7
223.8
29.5
60.8
11.5
7.6
Reference
SC(Bip)
SC(Bih)
Pd(0)
PdO
335.4
335.3
335.2
335.4
337.2
338.1
337.9
SC(Bih)
1.4
[53]
[54]
[55]
[56]
337.1
338.1
338.8
a
Reaction conditions: 4.1 g of 1-hexyne, 2.82 g of octane as internal standard,
0.3 g of catalyst, 303 K, and 1.05 MPa of H2 in 200 mL of n-heptane.
PdCl2(bipy)
Pd(C2H3O2)2
b
Selectivity toward 1-hexene at 25% conversion.

M. Crespo-Quesada et al. / Journal of Catalysis 279 (2011) 66–74
73
PdCl2(bipy) exhibits a binding energy of 338.1 eV. These results
show that the electronic properties of metallic Pd were not altered
by the presence of the N-containing ligands. However, Pd2+ species
present on the surface of the catalyst were indeed strongly affected
by the complexation of the N-containing ligands. The presence of
multiple oxidation states on the surface of catalysts is known to
influence their catalytic properties [46]; here, however, the specific
role of the N-modified Pd2+ species is not clear, but it is probable
that these species contribute to the differences between the Pd
NP catalysts.
A supplementary effect of the N-modified stabilizers may be
due to the change of the surface chemistry of the carbon-based
supports. A useful tool to examine the nature of the oxygen groups
present on the surface of the CNF is XPS. The differences in binding
energy for various bonding states are quite small for electronega-
tive atoms. It is therefore convenient to measure the C1s signal.
Carbon atoms differ in their binding energy if they are bonded to
an oxygen atom (phenols, ethers, carbonyl groups) or to two oxy-
gen atoms (carboxyl and lactone groups) [47]. Their corresponding
signals appear as satellites on the high binding energy side of the
main C1s peak, as shown in Fig. S3 in the Supporting Information.
The C1s spectra can be resolved into six individual component
peaks comprising carbidic carbon (Peak I), graphitic carbon (Peak
II), phenolic or ether groups (Peak III), carbonyl groups (Peak IV),
A Langmuir–Hinshelwood model was found to be consistent
with the observed reaction kinetics and showed that the N-con-
tained ligands permanently adsorb on the Pd surface ‘‘blocking’’
the active sites for 1-hexene adsorption. Nonetheless, an electronic
effect from the N-coordinated Pd2+ species resulting in a change in
the adsorption strengths and kinetic constants also contributes to
the change in the catalytic performance.
The imidazolium-functionalized bipy ligand was found to lar-
gely reduce Pd leaching from the catalyst presumably due to a
stronger ionic interaction with the oxygenated surface groups
present on the CNF.
Acknowledgments
The Swiss National Science Foundation is acknowledged for
financial support. The authors also thank Nicolas Xanthopoulos
(EPFL-SB-CIME) for the XPS measurements.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2011.01.003.
References
carboxyl or ester groups (Peak V), and satellite peaks due to p–p*
transitions in aromatic rings (Peak VI). The corresponding binding
energies at which these peaks appear can be found elsewhere
[48,49].
[1] C.H. Bartholomew, R.J. Farrauto, Fundamentals of Industrial Catalytic
Processes, second ed., Wiley-Interscience, Hoboken, NJ, 2006.
[2] H.-U. Blaser, A. Indolese, A. Schnyder, H. Steiner, M. Studer, J. Mol. Catal. A 173
(2001) 3–18.
A qualitative inspection of the Fig. S3 shows an increase in the
number of carboxylic groups and thus of the acidity of the support
upon activation. The deposition of the N-modified Pd NPs affects
almost exclusively the carboxylic acid groups and suggests a sub-
sequent reduction in the acidity of the support. However, as the
vast majority of the surface oxygen groups lay beyond the detec-
tion limit of XPS [49], the effect of the addition of N-containing li-
gands on acidity must be minimal and the overall increase in
selectivity is, as expected, due to interactions of the N-containing
ligands with the Pd surface.
[3] M. Garcia-Mota, B. Bridier, J. Perez-Ramirez, N. Lopez, J. Catal. 273 (2010) 92–
102.
[4] T. Mallat, A. Baiker, Appl. Catal. A 200 (2000) 3–22.
[5] T.A. Nijhuis, G. van Koten, F. Kaptejn, J.A. Moulijn, Catal. Today 79-80 (2003)
315–321.
[6] R. Schlogl, K. Noack, H. Zbinden, A. Reller, Helv. Chim. Acta 70 (1987) 627–679.
[7] J. Yu, J.B. Spencer, Chem. Commun. (1998) 1103–1104.
[8] R. Tschan, M.M. Schubert, A. Baiker, W. Bonrath, H. Lansink-Rotgerink, Catal.
Lett. 75 (2001) 31–36.
[9] G. Schmid, S. Emde, V. Maihack, W. Meyer-Zaika, S. Peschel, J. Mol. Catal. A:
Chem. 107 (1996) 95–104.
[10] G. Schmid, M. Harms, J.O. Malm, J.O. Bovin, J. Van Ruitenbeck, H.W.
Zandbergen, W.T. Fu, J. Am. Chem. Soc. 115 (1993) 2046–2048.
[11] J. Huang, T. Jiang, B.X. Han, H.X. Gao, Y.H. Chang, G.Y. Zhao, W.Z. Wu, Chem.
Commun. (2003) 1654–1655.
[12] A. Molnar, A. Sarkany, M. Varga, J. Mol. Catal. A 173 (2001) 185–221.
[13] E.M. Sulman, Russ. Chem. Rev. 63 (1994) 923–936.
[14] J.P. Boitiaux, J. Cosyns, S. Vasudevan, Appl. Catal. 15 (1985) 317–326.
[15] J. Rajaram, A.P.S. Narula, H.P.S. Chawla, S. Dev, Tetrahedron 39 (1983) 2315–
2322.
[16] B.E. Green, C.S. Sass, L.T. Germinario, P.S. Wehner, B.L. Gustafson, J. Catal. 140
(1993) 406–417.
[17] M.A. Aramendia, V. Borau, C. Jimenez, J.M. Marinas, M.E. Sempere, F.J. Urbano,
Appl. Catal. 63 (1990) 375–389.
[18] L.M. Bronstein, D.M. Chernyshov, I.O. Volkov, M.G. Ezernitskaya, P.M. Valetsky,
V.G. Matveeva, E.M. Sulman, J. Catal. 196 (2000) 302–314.
[19] E. Sulman, Y. Bodrova, V. Matveeva, N. Semagina, L. Cerveny, V. Kurtc, L.
Bronstein, O. Platonova, P. Valetsky, Appl. Catal. A: Gen. 176 (1999) 75–81.
[20] N. Semagina, A. Bykov, E. Sulman, V. Matveeva, S. Sidorov, L. Dubrovina, P.
Valetsky, O. Kiselyova, A. Khokhlov, B. Stein, L. Bronstein, J. Mol. Catal. A 208
(2004) 273–284.
[21] G. Schmid, Chem. Rev. 92 (1992) 1709–1727.
[22] B. Léger, A. Denicourt-Nowicki, H. Olivier-Bourbigou, A. Roucoux, Tetrahedron
Lett. 50 (2009) 6531–6533.
3.4. Resistance to leaching
Leaching of active Pd species from the support is one of the
main causes for deactivation in three-phase hydrogenations [50–
52]. To compare the level of leaching between the bipy and
[BIHB]Br2 modified catalysts, two halves of the same disk of CNF/
SMF support were impregnated with two N-modified NPs. Both
were stirred in 200 mL of heptane at 30 °C, 1.05 MPa of H2,
2000 rpm for 1 h. The analysis of solvent phase revealed that the
[BIHB]Br2 (9% leaching) modified catalyst was more stable than
the bipy modified system (14% leaching, implying 55% more leach-
ing in the latter), presumably due to a stronger interaction with the
oxygenated surface groups present on the CNF. Ionic interactions
between surface-bound Oꢂ atoms and the cationic [BIHB]2+ ligands
are probably responsible for the higher stability of the system.
[23] Y. Hu, Y.Y. Yu, Z.S. Hou, H. Li, X.G. Zhao, B. Feng, Adv. Synth. Catal. 350 (2008)
2077–2085.
[24] Y. Hu, H.M. Yang, Y.C. Zhang, Z.S. Hou, X.R. Wang, Y.X. Qiao, H. Li, B. Feng, Q.F.
Huang, Catal. Commun. 10 (2009) 1903–1907.
4. Conclusions
[25] P. Tribolet, L. Kiwi-Minsker, Catal. Today 105 (2005) 337–343.
[26] P. Tribolet, L. Kiwi-Minsker, Catal. Today 102–103 (2005) 15–22.
[27] N. Terasaki, T. Akiyama, S. Yamada, Chem. Lett. (2000) 668–669.
[28] J. Dupont, C.S. Consorti, P.A.Z. Suarez, R.F.d. Sousa, Org. Synth. 79 (2004) 236.
[29] X. Yang, N. Yan, Z.F. Fei, R.M. Crespo-Quesada, G. Laurenczy, L. Kiwi-Minsker, Y.
Kou, Y.D. Li, P.J. Dyson, Inorg. Chem. 47 (2008) 7444–7446.
[30] Y.G. Cui, I. Biondi, M. Chaubey, X. Yang, Z.F. Fei, R. Scopelliti, C.G. Hartinger,
Y.D. Li, C. Chiappe, P.J. Dyson, Phys. Chem. Chem. Phys. 12 (2010) 1834–1841.
[31] D.B. Zhao, Z.F. Fei, R. Scopelliti, P.J. Dyson, Inorg. Chem. 43 (2004) 2197–2205.
[32] P.J. Dyson, G. Laurenczy, C.A. Ohlin, J. Vallance, T. Welton, Chem. Commun.
(2003) 2418–2419.
N-modified Pd-nanoparticles immobilized on CNF/SMF sup-
ports were synthesized and studied in the selective hydrogenation
of 1-hexyne. The catalysts were found to be significantly more
selective (up to 98.5% at 25% conversion) than a reference catalyst
with non-modified Pd nanoparticles on the same support (88%).
Moreover, the high selectivity was maintained up to full
conversion.

74
M. Crespo-Quesada et al. / Journal of Catalysis 279 (2011) 66–74
[33] U.K. Singh, M. Albert Vannice, J. Catal. 191 (2000) 165–180.
[34] R.I. Macey, G.F. Oster, Berkeley MadonnaTM (1997–2001) p.
[35] H.H. Rosenbrock, C. Storey, Computational Techniques for Chemical Engineers,
Pergamon Press, Oxford, 1966.
[44] N. Yan, X. Yang, Z.F. Fei, Y.D. Li, Y. Kou, P.J. Dyson, Organometallics 28 (2009)
937–939.
[45] L. Zhou, L. Wang, Synthesis – Stuttgart (2006) 2653–2658.
[46] G.A. Somorjai, Science 227 (1985) 902–908.
[36] M. Crespo-Quesada, M. Grasemann, N. Semagina, A. Renken, L. Kiwi-Minsker,
Catal. Today 147 (2009) 247–254.
[37] N. Semagina, M. Grasemann, N. Xanthopoulos, A. Renken, L. Kiwi-Minsker, J.
Catal. 251 (2007) 213–222.
[38] A. Bruehwiler, N. Semagina, M. Grasemann, A. Renken, L. Kiwi-Minsker, A.
Saaler, H. Lehmann, W. Bonrath, F. Roessler, Ind. Eng. Chem. Res. 47 (2008)
6862–6869.
[39] H. Lindlar, Helv. Chim. Acta XXXV, Fasciculus II (1952) 446–450.
[40] R. Schlogl, K. Noack, H. Zbinden, A. Reller, Helv. Chim. Acta 70 (1987)
627–679.
[47] H.P. Boehm, Carbon 40 (2002) 145–149.
[48] S. Biniak, G. Szymanski, J. Siedlewski, A. Swiatkowski, Carbon 35 (1997) 1799–
1810.
[49] Z.R. Yue, W. Jiang, L. Wang, S.D. Gardner, C.U. Pittman, Carbon 37 (1999) 1785–
1796.
[50] M. Besson, P. Gallezot, Catal. Today 81 (2003) 547–559.
[51] J. Panpranot, O. Tangjitwattakorn, P. Praserthdam, J.G. Goodwin Jr., Appl. Catal.
A: Gen. 292 (2005) 322–327.
[52] J.G. Ulan, W.F. Maier, J. Mol. Catal. 54 (1989) 243–261.
[53] G. Johansson, J. Hedman, A. Berndtsson, M. Klasson, R. Nilsson, J. Electron
Spectrosc. Relat. Phenom. 2 (1973) 295–317.
[54] T.H. Fleisch, G.J. Mains, J. Phys. Chem. 90 (1986) 5317–5320.
[55] V.I. Nefedov, I.A. Zakharov, Moiseev II, M.A. Poraikos, M.N. Vargafti, A.P. Belov,
Zh. Neorg. Khim. 18 (1973) 3264–3268.
[41] J.P. Mikkola, P. Virtanen, H. Karhu, T. Salmi, D.Y. Murzin, Green Chem. 8 (2006)
197–205.
[42] J.P.T. Mikkola, P.P. Virtanen, K. Kordas, H. Karhu, T.O. Salmi, Appl. Catal. A –
Gen. 328 (2007) 68–76.
[43] M. Ruta, G. Laurenczy, P.J. Dyson, L. Kiwi-Minsker, J. Phys. Chem. C 112 (2008)
17814–17819.
[56] V.I. Nefedov, Y.V. Salyn, Moiseev II, A.P. Sadovskii, A.S. Berenbljum, A.G.
Knizhnik, S.L. Mund, Inorg. Chim. Acta 35 (1979) L343–L344.