Boron nitride as an alternative support of Pd catalysts for the selective oxidation of lactose

Article abstract of DOI:10.1016/j.catcom.2012.10.007

The potential of boron nitride as innovative support for the selective oxidation of carbohydrates has been evaluated. Pd/h-BN catalysts as well as Pd/α-Al₂O₃ have been synthesized by two different methods for comparison: dry impregnation and deposition-precipitation. It is shown that BN is a suitable alternative to alumina and carbon for sugar oxidation in liquid phase. Very active and selective Pd/h-BN catalysts were obtained by the two synthetic methods under consideration.

Full text of DOI:10.1016/j.catcom.2012.10.007

Catalysis Communications 29 (2012) 170–174
Contents lists available at SciVerse ScienceDirect
Catalysis Communications
journal homepage: www.elsevier.com/locate/catcom
Short Communication
Boron nitride as an alternative support of Pd catalysts for the selective
oxidation of lactose
⁎
Nathalie Meyer, Kevin Bekaert, Damien Pirson, Michel Devillers, Sophie Hermans
Université Catholique de Louvain, Institut de la Matière Condensée et des Nanosciences (IMCN), Place Louis Pasteur 1/3, B-1348 Louvain-la-Neuve, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 21 July 2012
Received in revised form 13 September 2012
Accepted 10 October 2012
Available online 17 October 2012
The potential of boron nitride as innovative support for the selective oxidation of carbohydrates has been
evaluated. Pd/h-BN catalysts as well as Pd/α-Al2O3 have been synthesized by two different methods for com-
parison: dry impregnation and deposition–precipitation. It is shown that BN is a suitable alternative to alumi-
na and carbon for sugar oxidation in liquid phase. Very active and selective Pd/h-BN catalysts were obtained
by the two synthetic methods under consideration.
© 2012 Elsevier B.V. All rights reserved.
Keywords:
Boron nitride
Lactose
Palladium
Oxidation
Alumina
1. Introduction
acid finds applications in cosmetics and as an antioxidant [8,9]. It
is also the major constituent of organ preservation fluids used
Carbohydrates represent an important source of renewable raw ma-
terials because of their large availability, high framework complexity
and independency from fossil energies. The industrial chemo-catalytic
processes that convert sugars into specialty chemicals imply most
often fermentation or enzymatic steps. The use of supported catalysts
for these transformations is less widespread because of a lack of reliable
catalytic methods to transform carbohydrates into high added value
compounds. This is due to the fact that the transformation of sugars
gives rise to rapid deactivation of the catalysts. For example, lactose is
oxidized into 2-keto-lactobionic acid with a 5%Bi–5%Pt/C catalyst, but
the latter deactivates rapidly because of the formation of platinum
oxide at the catalyst surface [1]. Ideally, these transformations should
be carried out in water, the catalysts should be very stable to permit
recycling and the conversion and selectivity should be as high as
possible.
during transplantation [10]. The heterogeneous catalysts reported
so far to carry out this reaction are based on noble metals (Au, Pd,
Bi–Pd, Pt) and supported on Al2O3, SiO2, TiO2, carbon or zeolites
[1,8–14].
Boron nitride is isoelectronic to carbon and its hexagonal h-BN
form is isostructural to graphite with covalent bonds in the planes
and very weak attractions between the planes. Its strong mechanical
and corrosion resistance, its high stability against thermal shock,
chemical attacks [15] and oxidation [16–18] make BN a promising
catalyst support for all applications in which stability is a key point.
A further interest of BN as an alternative to carbon comes from the
possibility to characterize the BN-supported catalysts by vibrational
spectroscopy.
Examples of uses of BN as catalyst support are quite scarce in the
literature and are essentially restricted to processes requiring rela-
tively high temperatures. BN-supported noble metal catalysts have
been used to abate volatile organic compounds [19]. Pt/BN has been
applied in deep oxidation of methanol and benzene [20]. Postole et
al. studied the influence of the preparation methods of Pd/BN cata-
lysts on the activity for methane oxidation [16,21,22]. They also stud-
ied the characteristics of BN and BN-supported oxide systems and
their use in the reduction of NOx [23]. BN has also been used as sup-
port for the hydrogenation of crotonaldehyde with Pt–Fe catalysts
[24] or alkynes using Pd/BN catalysts [25] for example.
However, many compounds interesting for fine chemistry can be
produced from sugars [2,3]. For example, it is possible to oxidize glu-
cose into gluconic acid, a chelating biodegradable agent, with Bi–Pd/C
and Au–Pd/C catalysts [4–7].
The present work deals with the catalytic oxidation of lactose in
lactobionic acid using supported palladium catalysts. Lactobionic
⁎
Corresponding author. Tel.: +32 10 472810; fax: +32 10 472330.
E-mail address: sophie.hermans@uclouvain.be (S. Hermans).
1566-7367/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.catcom.2012.10.007

N. Meyer et al. / Catalysis Communications 29 (2012) 170–174
171
Table 1
In this work, the potential of boron nitride as a new support for lac-
Experimental conditions used for lactose oxidation.
tose selective oxidation was evaluated: Pd/h-BN catalysts have been
synthesized, and compared to Pd/α-Al2O3 catalysts. The influence of
several parameters, such as the synthesis method, the nature of the pre-
cursor or reducing agent, on the structural characteristics and the cata-
lytic performances has been investigated.
Parameter
Lactose oxidation
Starting substrate solution
Temperature
Stirring rate
500 mL lactose 10 mmol L−1
40 °C
1000 rpm
9
pH
Basic solution to fix the pH
Oxygen flow
Mass of catalyst
Duration of a catalytic test
KOH 0.1 mol L−1
0.5 L min−1
500 mg
4 h
2. Experimental
2.1. Starting materials
The supports were commercially available alumina (γ-Al2O3, Sigma-
Aldrich, SBET ≈173 m2g−1) and hexagonal boron nitride (h-BN, Sigma-
Aldrich, SBET≈24 m2g−1). A batch of the alumina support was calcined
in air at 1000 °C for 24 h to give α-Al2O3 (SBET≈17 m2g−1). The textur-
al properties of the supports are given in Supplementary data (Table S1).
The palladium precursors were sodium tetrachloropalladate(II)
(Na2PdCl4, Sigma-Aldrich, 99.99%) and diacetatobis(diethylamine)
palladium(II) (Pd(OAc)2(NHEt2)2) synthesized as described else-
where [26].
stirring was ensured by a mechanical stirrer (Heidolph RZR 2051 elec-
tronic). The experimental conditions used for lactose oxidation are
summarized in Table 1. The catalyst was recovered by filtration and
the filtrate was analyzed by HPLC. Because selectivity was found to
be 100% in all cases, the lactose conversion always equals the yield
in lactobionic acid. The other carbohydrates were tested in exactly
the same conditions and also gave 100% selectivity.
2.2. Catalyst synthesis
3. Results and discussion
The palladium loading of each catalyst prepared in this work is
5 wt.%.
3.1. Catalysts synthesized by dry impregnation (DI)
The catalytic results in lactose oxidation and XPS characterization
data of the Pd/α-Al2O3 and Pd/h-BN catalysts synthesized by dry im-
pregnation are displayed in Table 2.
2.2.1. Dry impregnation method (DI)
This method consists in using the amount of solvent needed to ob-
tain incipient wetness. This amount has been determined by visual
observation of the formation of a homogeneous wetted paste for 1 g
of each support: Vmethanol =700 μL for 1 g α-Al2O3 and 1.1 mL for
1 g h-BN. It corresponds in both cases to about ten times the porous
volume of the supports (pore volume of α-Al2O3 =0.1127 cm3/g
and of h-BN=0.1419 cm3/g). Pd(OAc)2(NHEt2)2 was dissolved in
the corresponding methanol volume and added dropwise to the sup-
port under constant stirring at room temperature. The catalyst was
then reduced by thermal decomposition under hydrogen flow in a tu-
bular oven (2 h, 200 °C). For some catalysts, the reduction step was
preceded by calcination (under air flow, 500 °C, 12 h).
The catalysts calcined in air before activation under H2 are more
active than without calcination. This observation can be correlated to
XRD and microscopy results, while XPS surface atomic intensity ratios
are similar. In the XRD analysis of the DI-1 catalyst (Fig. 1), the molec-
ular precursor is clearly visible after DI (Fig. 1(a)), PdO appears after
calcination (Fig. 1(b)), and metallic Pd after reduction (Fig. 1(c)).
The last diffractogram (d) refers to the same catalyst activated with-
out prior calcination and shows thinner Pd peaks than in Fig. 1(c), in-
dicating larger crystallites. SEM analyses of DI-1C point in the same
direction and reveal smaller Pd particles at the surface of the support
and a more homogeneous distribution in the case of the pre-calcined
sample (Fig. S3(a) and (b)). The same observation can be made by
XRD (Fig. 2) and SEM (Fig. S3) analyses of BN-supported catalysts.
It also appears that the catalysts prepared on boron nitride are
much more active in the oxidation reaction than their homologues
supported on alumina. Experimental Pd/support XPS intensity ratios
were higher than the calculated values for BN, indicating better sur-
face accessibility to active sites for the sugar transformation. SEM im-
ages of the BN-supported catalysts (Fig. S3) showed metallic Pd
particles to be smaller and better spread on BN than on alumina. Be-
cause BN is more hydrophobic than alumina, the former takes more
time to sedimentate in CH3OH than in water. This probably increases
the interactions between the support and the precursor improving
the Pd dispersion on the surface in the final catalyst.
2.2.2. Deposition–precipitation method (DP)
Pd/α-Al2O3 and Pd/h-BN catalysts were prepared by a deposition–
precipitation method following a literature procedure [11]: 5 g of
support dispersed in 100 mL of a Na2CO3 2.5 wt.% aqueous solution
were stirred 15 min at room temperature. The calculated amount of
Na2PdCl4 or Pd(OAc)2(NHEt2)2 dissolved in 50 mL of distilled water
was added dropwise to the suspension within 30 min. After 15 addi-
tional minutes of stirring, 5 g of NaBH4 in 50 mL of distilled water
were added. The suspension was stirred for 1 h at room temperature.
Finally, the catalyst was filtered, washed with distilled water and
dried for 12 h at 80 °C. The filtrates were analyzed by atomic absorp-
tion spectrometry and no losses of Pd were observed. In addition, all
the solid samples were analyzed by ICP and the Pd wt.% was found
to be 4.37–4.71 in all cases except one (DP-8, Pd=3.85 wt.%).
The catalysts were characterized after activation by XPS, SEM,
TEM, powder XRD and Raman spectroscopy. In situ Raman measure-
ments were carried out for some catalysts prepared by DI to study the
activation step (see Supplementary data S2 for more details).
Table 2
Catalysts prepared by dry impregnation: Pd(OAc)2(NHEt2)2 as precursor.
Catalyst
Name
XPS
Lactobionic acid
yield (%) at t=4 h
Calc.
Exp.
Pd/Al (×100)
2.4
2.4
2.3. Oxidation of lactose
DI-1
Pd/α-Al2O3
Pd/α-Al2O3
1.7
1.6
11.7
29.5
DI-1Ca
All catalytic tests were performed in a thermostatized double-
walled glass reactor. The pH of the lactose solution was measured con-
tinuously by a combined AgCl/Ag Beckman electrode. An automatic
titration device Metrohm 842 Titrando was used to neutralize the
acids formed over time with KOH (Riedel-de-Haën, ≥85%). Constant
Pd/B (×100)
1.2
1.2
DI-2
Pd/h-BN
Pd/h-BN
2.2
2.5
27.4
48.6
DI-2Ca
a
Catalysts calcined under air flow at 500 °C for 12 h before reduction.

172
N. Meyer et al. / Catalysis Communications 29 (2012) 170–174
Pd
Pd
25°C
50°C
70°C
80°C
85°C
90°C
95°C
100°C
110°C
120°C
130°C
α-Al2O3
Pd
Pd
(d)
(c)
(b)
(a)
PdO
PdO
precursor
0
10
20
30
40
50
60
70
80
4000
3600
3200
2800
2400
2000
2-theta (°)
Raman shift (cm-1)
Fig. 1. XRD diffractograms of Pd/α-Al2O3 catalysts: (a) after DI, (b) after calcination, (c)
after reduction, and (d) activated without prior calcination.
Fig. 3. DI-2: Raman spectra of Pd(OAc)2(NHEt2)2/h-BN heated in situ under N2 flow.
interaction is better in water. The catalysts prepared on α-Al2O3 with
two different precursors but activated with NaBH4 (DP-3 and DP-4) ex-
hibit a higher activity and are characterized by much higher XPS Pd/Al
ratios, indicating more palladium available for the sugar at the surface
of the support. SEM images of DP-3 and DP-4 show particle sizes com-
prised between 15 and 30 nm (Fig. S7). XRD analyses of these two sys-
tems are quite similar with very broad peaks for Pd.
The samples prepared on boron nitride also reveal the influence of
the solvent: the activity of DP-6 is almost twice that of DP-5. In a
water/methanol mixture, BN gives rise to a homogeneous suspension
which increases the contact between the precursor and the support,
contrary to water alone. This is confirmed by an XPS Pd/B ratio higher
for DP-6 than for DP-5.
In the BN-supported catalysts reduced with NaBH4 (DP-7 and
DP-8), it comes out that the chlorinated precursor has a negative in-
fluence on the activity. A neutral precursor seems better adapted for
the synthesis of supported boron nitride catalysts.
The most competitive BN-based catalyst in this table is DP-8 pre-
pared from Pd(OAc)2(NHEt2)2 and activated with NaBH4. It is the
only one that contained slightly less Pd (3.85 wt.%) as determined
by ICP analysis of the solid. XRD analysis shows very broad Pd peaks
meaning small Pd crystallites (Fig. 4), while TEM images show parti-
cle sizes in the range of 2–30 nm (Fig. 5). This result can be correlated
with literature observations quoting that having a particle size
around 3 nm makes the catalyst more active and selective for sugar
oxidation [2,27].
The samples obtained via this preparation method can be ana-
lyzed by in situ Raman spectroscopy in order to study the activation
step. The Raman spectrum of the precursor shows a few sharp
peaks for Pd(OAc)2(NHEt2)2, while that of the alumina is featureless
(Fig. S4(a) and (b)). The spectrum of h-BN shows one main peak at
1400 cm−1 (Fig. S4(c)). The spectrum of the DI-1 catalyst before re-
duction is clearly the superimposition of the complex and the support
(Fig. S5(a)). Concerning the catalyst DI-2 (Fig. S5(b)), the Raman shift
range 4000–2000 cm−1 was selected to monitor the precursor peaks
useful to study the activation step. The latter was achieved (Fig. 3) by
heating the catalyst under nitrogen, and it was found that the activa-
tion corresponds to a ligand abstraction phenomenon leading to the
reduction of Pd(II) to Pd(0) as confirmed by XPS. The same study
was carried out for the Al2O3-supported catalyst, leading to the
same observation (Fig. S6).
3.2. Catalysts synthesized by deposition–precipitation (DP)
Catalytic performances of catalysts prepared by deposition–
precipitation are illustrated in Table 3. It comes out that Pd/Al2O3 cata-
lysts reduced with formalin (DP-1 and DP-2) are less active than their
homologues reduced with NaBH4 (DP-3 and DP-4). In addition, a sol-
vent effect can be observed when these two catalysts are analyzed
one step further. Indeed, the DP-2 catalyst prepared in the H2O/
CH3OH mixture is slightly less active than the DP-1 catalyst prepared
in water. Alumina, being more hydrophilic, is better dispersed in
water than in a water/methanol mixture. Hence, the precursor/support
DP-5 and DP-6 synthesized with the same precursor and the same
reducing agent in respectively water and in a mixture of water/
CH3OH are less active than DP-8 probably because of the presence
= h-BN
Pd
(d)
Table 3
Catalysts prepared by deposition–precipitation method (DP).
Pd
Catalyst Precursor
Name
Reducing
agent
XPS
Calc. Exp.
Lactobionic
acid yield (%)
at t=4 h
(c)
PdO
PdO
(b)
Pd/Al (×100)
DP-1
DP-2a
DP-3
DP-4
Pd(OAc)2(NHEt2)2 Pd/α-Al2O3 Formalin
Pd(OAc)2(NHEt2)2 Pd/α-Al2O3 Formalin
Na2PdCl4
2.4
2.4
2.5
2.4
5.9
6.4
11
12
10.7
7.8
42.9
38.1
Pd/α-Al2O3 NaBH4
precursor
(a)
Pd(OAc)2(NHEt2)2 Pd/α-Al2O3 NaBH4
*
*
Pd/B (x100)
DP-5
DP-6a
DP-7
DP-8
Pd(OAc)2(NHEt2)2 Pd/h-BN
Pd(OAc)2(NHEt2)2 Pd/h-BN
Formalin
Formalin
NaBH4
1.2
1.2
1.2
1.2
1.2
14.3
26.8
21
0
10
20
30
40
50
60
70
80
2.0
2.7
2.3
2-theta (°)
Na2PdCl4
Pd/h-BN
Pd(OAc)2(NHEt2)2 Pd/h-BN
NaBH4
64
Fig. 2. XRD diffractograms of Pd/h-BN catalysts: (a) after DI, (b) after calcination, (c)
after reduction, and (d) activated without prior calcination.
a
Synthesized in H2O–Na2CO3+CH3OH 50/50.

N. Meyer et al. / Catalysis Communications 29 (2012) 170–174
173
: h-BN
γ-Al2O3 and it indeed displayed half the activity of α-Al2O3 and se-
lectivity below 100%.
Pd/h-BN was also tested in maltose and glucose oxidation. In both
reactions, Pd/h-BN was active: yields of 43% and 55% in maltobionic
acid and gluconic acid, respectively, were obtained after a 4 h reac-
tion. The selectivity was 100% in both cases.
XPS analyses carried out after every catalytic test confirmed that
there was no change in the Pd(0) oxidation state (Fig. S10). Pd losses
during catalytic tests were always b1% as verified by ICP on catalysis
filtrates. The recyclability of the catalysts was also studied. For both
the BN and α-Al2O3-supported catalysts, there is a loss of about 50%
of activity in the second run but no further decrease in the next run.
Pd
0
10
20
30
40
50
60
70
80
2-theta (°)
4. Conclusions
Fig. 4. XRD diffractogram of DP-8: Pd/BN catalyst.
We have shown that BN is a promising alternative support to re-
place alumina and carbon for sugar oxidation in the liquid phase. Lac-
tose was the most studied but BN could also be used as support for
other sugar conversions. It is a robust and non-porous material,
which seems advantageous to avoid internal diffusion limitations
when carrying catalysis in water. In addition, the absence of surface
hydroxyl groups is expected to limit side reactions which should
guarantee high selectivity and stabilizes the metal in its reduced
form. Finally, it allows spectroscopic in situ characterization. Very ac-
tive Pd/h-BN catalysts were obtained by either dry impregnation or
deposition–precipitation.
of bigger Pd aggregates, as evidenced by TEM (Fig. S8). It is also due to
the reducing agent since NaBH4 is more efficient than formaldehyde.
Indeed, the Pd XPS spectrum of DP-8 shows only Pd(0), which is
claimed in the literature as the active phase for sugar oxidation [2],
while for DP-5 and DP-6, XPS 3d peak can be decomposed in two dou-
blets for Pd(II) and Pd(0) (Fig. S9).
Except for DP-7, there is a correlation between XPS results and the
yield in lactobionic acid for the BN-supported catalysts. Indeed, the
higher the Pd/B ratio, the higher the activity in this reaction.
The preparation method has also a clear influence on the catalytic
performances. On alumina as well as on BN, the catalysts are more ac-
tive when the deposition–precipitation method is used instead of dry
impregnation. The former method permits to obtain smaller Pd parti-
cles at the surface of the support as shown by microscopy and XRD
analyses. However, when a catalyst prepared by DI is calcined and
then activated under hydrogen, its catalytic performances are greatly
increased and reach the activity of its counterpart prepared by DP.
Mirescu and Pruesse showed that very active and selective Au/
alumina catalysts were obtained for the oxidation of lactose in
lactobionic acid. However, when palladium is supported on alumina
[11] or on carbon [14], it gives a lower selectivity. In our case, when
palladium is supported on boron nitride, 100% selectivity is reached.
We believe that the high selectivity of BN-supported catalysts is due
to the absence of surface functional groups, the small surface area,
the absence of micropores and probably also to the particle sizes
obtained (which, if too small, promote overoxidation). A control ex-
periment was carried out by preparing a catalyst by DP on a porous
Acknowledgments
The authors wish to thank the Fonds de la Recherche Scientifique
(FRS-FNRS) with the assistance of the Fédération Wallonie-Bruxelles
and the Belgian National Lottery, as well as the Université Catholique de
Louvain for funding. This work was also partially funded by the Belgian
State (Belgian Science Policy, IAP Project INANOMAT No. P6/17). We are
grateful as well to Jean-François Statsijns, Michel Genet and Pascale Lipnik
for technical assistance and useful discussions.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://
dx.doi.org/10.1016/j.catcom.2012.10.007.
References
[1] A. Abbadi, K.F. Gotlieb, J.B.M. Meiberg, H. Van Bekkum, Applied Catalysis A: Gen-
eral 156 (1997) 105–115.
[2] M. Besson, P. Gallezot, Catalysis Today 57 (2000) 127–141.
[3] P. Gallezot, Catalysis Today 121 (2007) 76–91.
[4] A. Heinen, J. Peters, H. Van Bekkum, Carbohydrate Research 328 (2000) 449–457.
[5] M. Wenkin, P. Ruiz, B. Delmon, M. Devillers, Journal of Molecular Catalysis A:
Chemical 180 (2002) 141–159.
[6] S. Hermans, A. Deffernez, M. Devillers, Applied Catalysis A: General 395 (2011)
19–27.
[7] Y. Önal, S. Schimpf, P. Claus, Journal of Catalysis 223 (2004) 122–133.
[8] A.V. Tokarev, D.Y. Murzin, Journal of Molecular Catalysis A: Chemical 255 (2006)
199–208.
[9] S. Karski, I. Witonska, Journal of Molecular Catalysis A: Chemical 245 (2006)
225–230.
[10] J.H. Southard, F.O. Belzer, Annual Review of Medicine 46 (1995) 235–247.
[11] A. Mirescu, U. Pruesse, Applied Catalysis B: Environmental 70 (2007) 644–652.
[12] A.V. Tokarev, D.Y. Murzin, Microporous and Mesoporous Materials 113 (2008)
122–131.
[13] E.V. Murzina, D.Y. Murzin, Catalysis Today 131 (2008) 385–392.
[14] J. Kuusisto, A.V. Tokarev, E.V. Murzina, M.U. Roslund, J.-P. Mikkola, D.Y. Murzin, T.
Salmi, Catalysis Today 121 (2007) 92–99.
[15] R.T. Paine, C.K. Narula, Chemical Reviews 90 (1990) 73–91.
[16] G. Postole, M. Caldararu, N.I. Ionescu, B. Bonnetot, A. Auroux, C. Guimon,
Thermochimica Acta 434 (2005) 150–157.
[17] J.A. Perdigon-Melon, A. Auroux, J.M. Guil, B. Bonnetot, Studies in Surface Science
and Catalysis 143 (2000) 227–237.
[18] G. Postole, A. Gervasini, A. Auroux, B. Bonnetot, Chemical Engineering Transac-
tions 4 (2004) 271–278.
Fig. 5. TEM image of DP-8: Pd/h-BN catalyst.
[19] J.C.S. Wu, C.Y. Chen, D. Lin, Catalysis Letters 102 (2005) 223–227.

174
N. Meyer et al. / Catalysis Communications 29 (2012) 170–174
[20] C.A. Lin, J.C.S. Wu, J.W. Pan, C.T. Yeh, Journal of Catalysis 210 (2002) 39–45.
[21] G. Postole, A. Gervasini, C. Guimon, A. Auroux, B. Bennetot, Journal of Physical
Chemistry 110 (2006) 12572–12580.
[22] G. Postole, A. Gervasini, A. Auroux, B. Bonnetot, Materials Science Forum 518
(2006) 203–210.
[25] Y. Yabe, T. Yamada, S. Nagata, Y. Sawama, Y. Monchugi, H. Sajiki, Advanced Syn-
thesis and Catalysis 354 (2012) 1264–1268.
[26] S. Hermans, M. Wenkin, M. Devillers, Journal of Molecular Catalysis A: Chemical
136 (1998) 59–68.
[27] P.M. Arvela, E.V. Murzina, B. Campo, T. Heikkilä, A. Leino, K. Kordas, D. Wolf, A.V.
Tokarev, D.Y. Murzin, Research on Chemical Intermediates 36 (2010) 423–442.
[23] G. Postole, A. Gervasini, M. Caldararu, B. Bonnetot, A. Auroux, Applied Catalysis A:
General 325 (2007) 227–236.
[24] J.C.S. Wu, T.-S. Cheng, C.-L. Lai, Applied Catalysis A: General 314 (2006) 233–239.