Selective conversion of m-cresol to toluene over bimetallic Ni-Fe catalysts

Article abstract of DOI:10.1016/j.molcata.2013.09.029

The catalytic conversion of m-cresol in the presence of H₂ has been investigated on SiO₂-supported Ni, Fe, and bimetallic Ni-Fe catalysts at 300 °C and atmospheric pressure. Over the monometallic Ni catalyst, the dominant product is 3-methylcyclohexanone while 3-methylcyclohexanol and toluene appear in smaller amounts, even at high conversions. By contrast, on Fe and Ni-Fe bimetallic catalysts, the dominant product is toluene while the hydrogenation products (3-methylcyclohexanone and 3-methylcyclohexanol) are practically negligible in the entire range of conversions. To explain these differences, we have proposed a deoxygenation path that starts with the tautomerization of m-cresol to an unstable ketone intermediate (3-methyl-3,5-cyclohexadienone). The fate of this intermediate is determined by the ability of the catalyst to either hydrogenate the carbonyl group or the ring. The former would mostly occur on Fe and Ni-Fe catalysts that contain an oxophilic metal (Fe), while the latter would occur on Ni, which has a higher affinity for the aromatic ring. Hydrogenation of the carbonyl group produces a very reactive unsaturated alcohol (3-methyl-3,5-cyclohexadienol), which can be easily dehydrated to toluene. This would explain the high selectivity of Fe and Ni-Fe to toluene. By contrast, hydrogenation of the ring would result in 3-methylcyclohexanone, which can be further hydrogenated to 3-methylcyclohexanol. On supports that contain acid sites, which are active for dehydration, the formation of toluene would occur via dehydration of the alcohol and subsequent dehydrogenation. On the catalysts investigated in this work, dehydration of the corresponding alcohol does not occur, so the only path to toluene is via hydrogenation of the carbonyl of the unstable ketone intermediate. In addition, to the products mentioned above, xylenol is also observed in significant yields, which indicate that transalkylation of m-cresol is another reaction path occurring on these catalysts.

Full text of DOI:10.1016/j.molcata.2013.09.029

Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Selective conversion of m-cresol to toluene over bimetallic
Ni–Fe catalysts
Lei Nie, Priscilla M. de Souza1 , Fabio B. Noronha1,2, Wei An3,
,2
4
∗
Tawan Sooknoi , Daniel E. Resasco
School of Chemical, Biological, and Materials Engineering, The University of Oklahoma, Norman, OK 73019, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 12 May 2013
Received in revised form
The catalytic conversion of m-cresol in the presence of H has been investigated on SiO -supported Ni, Fe,
2
2
◦
and bimetallic Ni–Fe catalysts at 300 C and atmospheric pressure. Over the monometallic Ni catalyst, the
dominant product is 3-methylcyclohexanone while 3-methylcyclohexanol and toluene appear in smaller
amounts, even at high conversions. By contrast, on Fe and Ni–Fe bimetallic catalysts, the dominant prod-
uct is toluene while the hydrogenation products (3-methylcyclohexanone and 3-methylcyclohexanol)
are practically negligible in the entire range of conversions.
1
7 September 2013
Accepted 21 September 2013
Available online 12 October 2013
To explain these differences, we have proposed a deoxygenation path that starts with the tautomer-
ization of m-cresol to an unstable ketone intermediate (3-methyl-3,5-cyclohexadienone). The fate of this
intermediate is determined by the ability of the catalyst to either hydrogenate the carbonyl group or the
ring. The former would mostly occur on Fe and Ni–Fe catalysts that contain an oxophilic metal (Fe), while
the latter would occur on Ni, which has a higher affinity for the aromatic ring.
Hydrogenation of the carbonyl group produces a very reactive unsaturated alcohol (3-methyl-3,5-
cyclohexadienol), which can be easily dehydrated to toluene. This would explain the high selectivity of
Fe and Ni–Fe to toluene. By contrast, hydrogenation of the ring would result in 3-methylcyclohexanone,
which can be further hydrogenated to 3-methylcyclohexanol. On supports that contain acid sites, which
are active for dehydration, the formation of toluene would occur via dehydration of the alcohol and sub-
sequent dehydrogenation. On the catalysts investigated in this work, dehydration of the corresponding
alcohol does not occur, so the only path to toluene is via hydrogenation of the carbonyl of the unstable
ketone intermediate.
Keywords:
Ni–Fe/SiO2
Hydrodeoxygenation
Cresol
Biomass
Tautomerization
In addition, to the products mentioned above, xylenol is also observed in significant yields, which
indicate that transalkylation of m-cresol is another reaction path occurring on these catalysts.
©
2013 Elsevier B.V. All rights reserved.
1
. Introduction
chemicals from biomass. However, following the primary conver-
sion step, an effective catalytic upgrading step is needed in the
production of useful products [1–6]. As we have previously pointed
out [7], the upgrading process should not only minimize the oxy-
gen content in the product, but also maximize carbon retention. As
a result, conventional hydrotreating, the most common and effec-
tive method that has been tested for hydrodeoxygenation [8,9]
may not be the optimum path for upgrading of bio-oil with max-
imum carbon efficiency. We have recently investigated different
paths, including ketonization, aldol condensation, aromatization
and alkylation to accomplish the formation of C C bonds before
deoxygenation, which maximizes the liquid fuel yield and carbon
efficiency [10–12].
Phenolics are abundant components of the liquid product (bio-
oil) obtained from the fast pyrolysis of biomass. They derive
from both pyrolytic decomposition of lignin and condensa-
tion/aromatization of small oxygenates. They can represent an
important building block for potential production of fuels and
∗ Corresponding author. Tel.: +1 405 325 4370; fax: +1 405 325 5813.
E-mail address: resasco@ou.edu (D.E. Resasco).
Permanent address: Depto de Engenharia Química/Instituto Militar de Engen-
1
haria, Pra c¸ a Gal. Tiburcio, 80, CEP 22290-270, Rio de Janeiro, Brazil.
2
Permanent address: Instituto Nacional de Tecnologia (INT), Laboratório de
Several groups have investigated the use of sulfided catalysts
Catálise, sala 518, Av. Venezuela 82, CEP 20081-312, Rio de Janeiro, Brazil.
(e.g., CoMo, NiMo) typically used in conventional hydrotreating
3
Permanent address: Chemistry Department, Brookhaven National Laboratory,
for the deoxygenation of biomass-derived feedstocks [13]. These
catalysts require addition of H2S to remain stable and active; how-
ever, sulfidation is not required in the HDO of biomass-derived
P.O. Box 5000, Upton, NY 11973-5000, USA.
4
Permanent address: King Mongkut’s Institute of Technology Ladkrabang, Cha-
longkrung Road, Ladkrabang, Bangkok 10520, Thailand.
1
381-1169/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2013.09.029

4
8
L. Nie et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55
feedstocks since hydrodesulfurization is not a necessary require-
ment. Thus, noble metal catalysts (e.g., Pt, Pd) can become attractive
since they are significantly more active for HDO than other cat-
alysts. For example, bifunctional Pt/Al O [14,15] and Pt/H-beta
zeolite [16] have been found to be active for the deoxygenation of
oxygenated aromatics. However, the high cost of noble metals may
be an impediment for practical applications. The search for alter-
native catalysts based on non-noble metals (e.g., Ni, Fe, Co, Ga) has
been the focus of several recent studies [17–22]. The combination
of metal sites (active for hydrogenation) with acid sites provided
by the support or added promoters (active for transalkylation and
dehydration) can optimize the catalyst resulting in improved deox-
ygenation activity and carbon efficiency.
understanding of the interactions of the hydroxyl group in pheno-
lics with a bimetallic catalyst such as Ni–Fe in which one of the met-
als exhibit a high oxophilicity and may not be completely reduced
under reaction conditions. The present study has allowed us to infer
a novel reaction pathway that includes some important conceptual
differences from those previously proposed in the literature.
2
3
2. Experimental
2.1. Catalyst synthesis and characterization
Monometallic catalysts Ni/SiO2 and Fe/SiO2 (5 wt% metal
loading) were prepared by incipient wetness impregnation of
The use of bimetallic catalysts adds another interesting aspect
since changes in electronic structure and/or surface ensemble size
due to the presence of adjacent atoms may greatly modify the
resulting selectivity [17,19]. In recent studies, Ni and Ni–Cu cat-
alysts have been evaluated for the HDO of guaiacol [23–25]. The
studies agree that the addition of Cu improves the catalyst per-
formance for HDO. Similarly, in a recent investigation of our own
group, Ni–Fe bimetallic catalysts were evaluated for the hydrogena-
tion and deoxygenation of furfural as a model compound of sugar
dehydration products. We found that the reaction pathways on
Ni–Fe bimetallics were dramatically different from those on pure
Ni or pure Fe catalysts. While pure Fe was inactive at the reac-
tion conditions investigated, it greatly altered the behavior of Ni
when added in 1:1 molar ratio. The high decarbonylation activity
typically observed on pure Ni was suppressed on the bimetallic cat-
the support (SiO , HiSil 233) with an aqueous solution of the
2
respective metal precursor: Ni(NO ) ·6H O (98%, Alfa Aesar) and
3
2
2
Fe(NO ) ·9H O (98% Sigma–Aldrich). As described in our previ-
3
3
2
ous work [26], the bimetallic Ni–Fe/SiO2 catalysts were prepared
by incipient wetness co-impregnation. In this case, the Ni loading
was kept constant at 5.0 wt.% on all samples, while the Fe load-
ing was varied from 2.0, 5.0 to 10 wt.%. As summarized in Table 1,
they are indicated with the approximate mass ratios as (2:1)Ni–Fe,
(
1:1)Ni–Fe, and (1:2)Ni–Fe/SiO , respectively.
2
XRD analysis was conducted on samples pre-reduced ex situ
◦
under pure H2 (100 ml/min) at 450 C for 1 h and passivated by
slow exposure to low O2 concentrations at room temperature,
before exposure to the atmosphere. The measurements were car-
ried out with a D8 Series II X-Ray Diffractometer (BRUKER AXS)
operated at 40 kV and 35 mA, using Cu K␣ monochromatic radiation
alyst, while the C O hydrogenation (at low temperatures) and C
hydrogenolysis (at high temperatures) were drastically enhanced
26].
A mechanism for the deoxygenation of phenolic compounds
O
◦
(
ꢀ = 0.154178 nm) in the 30–60 diffraction angle range.
The reducibility of the calcined samples was determined by tem-
perature programmed reduction (TPR). In these measurements,
[
2
3
0 mg of a sample was placed in a quartz reactor and heated at
0 C/min up to 500 C under a He flow of 20 ml/min, and held
that has been widely proposed in the literature [9,14,15,27] is the
two-step pathway, which involves an initial hydrogenation of the
◦
◦
at this temperature for 1 h. The reactor was then cooled down
3
aromatic ring followed by C(sp ) O bond cleavage via dehydration.
◦
to 30 C and the sample exposed to a stream of 5% H /Ar at a
2
This HDO pathway normally requires a bi-functional catalyst, with
a metal function that catalyzes hydrogenation/dehydrogenation of
the ring and an acid function that catalyzes dehydration. An alterna-
tive mechanism that has been suggested in several studies [27,28]
flow rate of 20 ml/min. Subsequently, the sample was heated to
◦
◦
8
00 C at a heating rate of 5 C/min. The variation in hydrogen
uptake was monitored on a TCD detector as a function of tem-
perature. The molar H2 uptake per gram of sample was quantified
from the peak area in the TPR profiles and calibrated with a CuO
standard. Morphology and size of the Ni–Fe clusters were char-
acterized by transmission electron microscopy (TEM, JEOL model
JEM-2100 LaB6). Before TEM analysis, the samples were reduced ex
2
3
does not involve the C(sp )-to-C(sp ) conversion. In analogy to the
direct desulfurization (DDS) path typically proposed in HDS cata-
2
lysts, a direct cleavage of the C(sp ) O bond has been claimed in
several reports and referred to as direct deoxygenation (DDO) [27].
Since such cleavage would require a very high activation energy,
one may expect that this path might only be possible at high tem-
peratures.
In the present contribution, we investigate the different prod-
uct distributions from the hydrodeoxygenation of m-cresol over Ni
and Ni–Fe catalyst. While the three cresol isomers only represent a
small fraction in the composition of a real bio-oil, they are present
in significant amounts in the product of primary upgrading pro-
cesses, such as catalytic pyrolysis or conversion of pyrolysis vapors
preceding condensation, after which a final dehydroxygenation
step is necessary. Nevertheless, we must emphasize that the main
objective of the present contribution is to expand the fundamental
◦
situ in pure H2 (100 ml/min) at 450 C for 1 h. The reduced samples
were then mixed with 2-propanol, sonicated, deposited onto the
TEM (Cu) grids, and dried. Average particle sizes for all the sam-
ples, as determined by TEM, are summarized in Table 1. The BET
surface area (Sg) was measured by conventional N2 physisorption
◦
on a Micromeritics ASAP 2010 unit, after evacuation at 350 C for
3
h.
The surface chemistry of the catalysts was investigated by Dif-
fuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS)
of adsorbed pyridine. For these measurements, the catalyst powder
◦
◦
was loaded in the sample cup, reduced in situ at 400 C for 1 h under
a flow of H2 (30 ml/min) and cooled down to 100 C. Then, pyri-
Table 1
Characterization of catalysts from Ref. [26].
2
˚
Catalysts
Wt%Ni
Wt%Fe
BET (m /g)
H2 comsumption from TPR (mmol/gcat)
Diameter (nm)
Lattice constant (A)
XRD
DFT
Std.
Ni
5
5
5
5
0
0
2
5
10
5
126
130
115
124
128
1.01
1.48
2.20
–
11.2
10.0
10.0
9.6
3.53
3.57
3.58
3.58
2.87
3.52
–
3.55
–
3.52
–
(2:1)Ni–Fe
(1:1)Ni–Fe
(1:2)Ni–Fe
a
3.58a
–
Fe
–
19.1
–
2.87

L. Nie et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55
49
dine vapor was passed through the cell for approximately 30 min.
To eliminate the weakly bound pyridine, the sample was purged
◦
under H2 flow for 20 min at 100 C. Background and absorption
Fe/SiO2
−
1
spectra were recorded at a resolution of 4 cm , accumulating 256
scans.
(
1:2)Ni-Fe/SiO2
2.2. Catalytic activity measurements
(1:1)Ni-Fe/SiO2
The vapor-phase conversion of m-cresol in H over the Ni, Ni–Fe
2
and Fe catalysts was evaluated in a tubular quartz reactor at atmo-
spheric pressure. The flow reaction system is equipped with a mass
flow controller and a syringe pump for the continuous injection of
m-cresol. All lines were heated to avoid condensation of the com-
pounds. Pelletized catalyst (40–60 mesh) was packed in the reactor
between two layers of quartz wool and a preheating layer of glass
(2:1)Ni-Fe/SiO2
Ni/SiO2
4
0
42
44
46
θ (degree)
48
50
beads on top to improve the temperature uniformity. The catalyst
2
◦
was reduced in situ under a down-flow of 60 ml/min of H at 450 C
2
for 1 h, and then the temperature was reduced to the reaction tem-
Fig. 1. XRD patterns of monometallic Ni, Fe and bimetallic Ni–Fe catalysts pre-
reduced ex situ in hydrogen at 450 C for 1 h.
◦
perature 300 C. The H /m-cresol molar ratio was kept at 60:1 for
◦
2
all runs. The products were quantified by gas chromatography (GC
6
890, Agilent), using an Innowax capillary column and a flame ion-
ized detector (FID). The catalyst was evaluated under different W/F
0.058–0.464 h) varying the amount of catalyst (30–120 mg). The
3. Results and discussion
(
W/F is defined as the ratio of the catalyst mass (g) to organic feed
flow rate (g/h). The product yield and selectivity for each product
were calculated as follows:
3
.1. Catalyst characterization
Fig. 1 displays the XRD patterns of the catalyst samples after
H2 reduction. The monometallic Ni catalyst showed one peak
at 2ꢁ = 44.41 , which is ascribed to the Ni(1 1 1) reflection. The
mol of product
◦
Yield (%) =
× 100
mol of m-cresol fed
bimetallic catalysts exhibited a single peak around this region, but
shifted to lower angles with increasing Fe content. In agreement
with previous studies [19,26,36], these XRD data indicate the for-
mation of a Ni–Fe solid solution, rich in Ni. The shift to lower angles
with Fe content indicates an increase in the Fe concentration in
the solid solution. In addition to the peak corresponding to the
mol of product
mol of m-cresol consumed
Selectivity (%) =
2
.3. Computational method
Ni–Fe solid solution, the sample with the highest Fe concentra-
◦
DFT calculations were performed in the Vienna ab initio simula-
tion (1:2)Ni–Fe/SiO , exhibited a shoulder at 2ꢁ = 44.64 , which is
2
tion package (VASP) [29,30]. A spin-polarized GGA PBE functional
31], all-electron plane-wave basis sets with an energy cutoff of
00 eV, and a projector augmented wave (PAW) method [32,33]
ascribed to the ␣-Fe(1 1 0) reflection resulting from a fraction of
unalloyed Fe. The same result has been observed in previous XRD
studies conducted on a sample of Fe/Ni = 2 ratio, and the authors
assigned this double peak to unalloyed ␣-Fe or Fe-rich alloys with a
bcc structure [36]. The TPR profiles [26] also support the formation
of bimetallic clusters in the Fe–Ni catalysts. The characterization
results of the catalysts are summarized in Table 1; further details
about these catalysts can be found in our previous publication
[26].
[
4
were adopted. The Ni(1 1 1), Fe(1 1 0) and NiFe(1 1 1) surfaces
were modeled by a three-layer p(4 × 4) slab with the bottom
two layers fixed at their optimized bulk positions. The calculated
lattice constants are 3.522 A˚ for fcc Ni, 2.831 A˚ for bcc Fe, and
a = c = 3.553 A˚ b = 3.582 A˚ for fcc Ni Fe . The two successive slabs
2
2
were separated by a 18 A˚ vacuum region. Considering that the
surface composition may differ from that of the bulk alloy and
surface segregation may occur in a NiFe alloy [34], we also con-
sidered a model that contains “disordered” NiFe alloy surface,
which has the same Ni:Fe molar ratio of 1:1 as the perfect sur-
face, but is re-arranged to create “islands” of Ni and Fe. Such a
DRIFTS of adsorbed pyridine was conducted on the different
catalysts to compare their surface properties. The DRIFT spectra
obtained on the Ni, (1:2)Ni–Fe, and Fe/SiO2 catalysts after expo-
◦
sure to pyridine at atmospheric pressure and 100 C are reported
in Fig. 2, along with the spectrum obtained on pure SiO exposed
2
−
1
“
disordered” surface is energetically unfavorable compared to the
to pyridine. An intense band at 3741 cm
(not shown) corre-
perfect one, i.e., having energy penalty of 0.39 kcal/mol per sur-
face atom. The previous Mössbauer investigation [35] suggested
that certain phase separation could accompany particle growth in
NiFe alloy catalysts using the similar synthesis process as in this
work. The 3 × 3 × 1 k-points using the Monkhorst–Pack scheme
was used in the calculations. First-order Methfessel–Paxton smear-
ing of 0.2 eV was employed in the integration to speed up the
convergence. The conjugate gradient algorithm was used in the
sponding to the silanol O H stretching mode was observed for
all samples [37]. No absorption band was observed for the SiO2
support in the frequency region corresponding to adsorbed pyri-
−
1
dine (1620–1420 cm ). Therefore, it can be concluded that any
contribution to the spectra of pyridine interacting with hydrogen-
bonded or free silanol [38] are negligible under the conditions of
◦
the measurements (i.e., purged at 100 C). By contrast, the cata-
lysts do display pronounced absorption bands at 1445, 1577 and
−
4
−1
optimization. The convergence threshold was set 10 eV in total
1596 cm , which are characteristic of the vibrational modes 19b
−2
energy and 10 eV/ A˚ in force on each atom. The adsorption
and 8a of pyridine species interacting with coordinatively unsatu-
rated cations [39,40]. According to the relative band intensity for
comparable amounts of catalyst and surface areas, we can estimate
that the density of coordinatively unsaturated cations increases
in the order Ni/SiO < (1:2)Ni–Fe/SiO < Fe/SiO . The presence of
energy (Eads) is defined as Eads = E(cresol/slab) − Eslab – E(cresol),
where E(cresol/slab), Eslab, and E(cresol) are the total energy of
cresol/slab, clean surface, and gas-phase m-cresol in one supercell,
respectively.
2
2
2

5
0
L. Nie et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55
and transalkylation side reactions. Metallic Ni is a well-known
(
a)
b)
c)
hydrogenolysis catalyst, but transalkylation (disproportionation) is
typically catalyzed by acid sites. While the silica support provides
no significant acidity, a fraction of Ni may not be fully reduced,
as described above. The presence of coordinatively unsaturated
cations may act as Lewis acid sites [40], which might catalyze the
transalkylation that produces phenol and xylenol. On the other
hand, methane may arise from metal-catalyzed hydrogenolysis or
acid-catalyzed demethylation.
(
(
(d)
A remarkable difference in product distribution relative to
that obtained on Ni/SiO2 is observed on the iron-containing Fe
and Ni–Fe/SiO2 catalysts when compared under identical reaction
conditions. As summarized in Table 2, when the three differ-
◦
ent catalysts are compared at 300 C and W/F = 0.46 h a drastic
change in product selectivity is observed. While, as mentioned
above, 3-methylcyclohexanone is the dominant C7 compound over
Ni/SiO2 it is not formed either on Fe or Ni–Fe/SiO2 catalysts.
By contrast, on these catalyst the largely dominant product is
toluene (TOL). Interestingly, the physical mixture of Fe/SiO2 and
Ni/SiO2 catalysts behaves very similarly to the Ni/SiO2 catalyst
alone. That is, ring hydrogenation products were formed and 3-
methylcyclohexanone was the main one. Small amounts of toluene
and xylenol were observed on this mixture, while they were the
main products over the bimetallic catalysts, which may be due
1
600
1560
1520
1480
1440
-1
Wavenumber (cm )
◦
Fig. 2. DRIFT spectra of the pyridine chemisorption experiments at 100 C and atmo-
spheric pressure after pyridine chemisorptions for (a) pure SiO2, (b) Ni/SiO2, (c)
1:2)Ni–Fe/SiO2, (d) Fe/SiO2.
(
unreduced cations remaining on the surface of silica even after
high temperature reduction treatments has been demonstrated for
Fe/SiO2 [40] as well as for Ni/SiO2 [41].
to the rather low activity of Fe/SiO2 compared to Ni/SiO . Thus,
2
these results demonstrate that the typical behavior of unalloyed
Ni is not observed on the Ni–Fe bimetallic surface. As discussed
below, the observed difference in behavior may be due to surface
enrichment of Fe, or at least, the formation of Fe islands on the
surface, as suggested by the DFT results discussed in section 3.4
below.
The Fe-containing catalysts also produce xylenols, xylene, phe-
nol, and methane. In this case, again, the transalkylation reaction
responsible for the production of C8 products might be catalyzed
by the Lewis acid sites created by the presence of coordinatively
unsaturated cations of unreduced Fe and/or Ni.
3
.2. Catalytic activity and product distribution at varying
m-cresol conversion
The product distribution from the m-cresol conversion over
Ni/SiO2 at 300 C is shown in Fig. 3 as a function of W/F. Three
C7 compounds are obtained, 3-methylcyclohexan (ONE) that is
the main product, 3-methylcyclohexan (OL), and toluene (TOL). In
addition to the three C7 compounds, methane, phenol, and xylenol
are also obtained as products. They may arise from hydrogenolysis
◦
1
1
1
1
1
8
6
4
2
0
8
6
4
2
0
3.0
Methane
2.5
Conversion
ONE
Phenol
Xylenol
OL
2.0
TOL
1
.5
1.0
.5
0
0.0
0.0
0
.0
0.1
0.2
0.3
0.4
0.5
0.1
0.2
0.3
0.4
0.5
W/F (h)
W/F (h)
◦
Fig. 3. Product distribution of m-cresol over 5% Ni/SiO2 at 300 C as a function of W/F H2/feed molar ratio = 60. Pressure = 1 atm.
Table 2
◦
Conversion and product selectivity from the reaction of m-cresol over different catalysts at W/F = 0.46 h and 300 C. H2/feed molar ratio = 60. Pressure = 1 atm.
5
% Ni/SiO2
5% Fe/SiO2
8.8
5% Ni/SiO2 + 5% Fe/SiO2 (physical mixture)
12.6
5% Ni–5% Fe/SiO2
13.7
Conversion %
16.2
Selectivity %
Toluene
14.2
33.3
11.1
13.0
28.4
60.2
0.0
0.0
0.0
39.8
18.5
33.1
13.7
10.5
24.2
52.6
0.0
0.0
2.2
45.3
3
3
-Methylcyclohexanone
-Methylcyclohexanol
Methane
Transalkylation Products

L. Nie et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55
51
b 7
a
9
8
7
6
5
4
3
2
1
0
6
5
methane
xylene
phenol
xylenol
4
TOL
ONE
3
OL
minor
2
1
0
0
2
4
6
8
10
0
2
4
6
8
10
Fe loading (wt%)
Fe loading (wt%)
◦
Fig. 4. Yield of products of m-cresol over Ni–Fe bimetallic catalysts as a function of Fe loading at 300 C. H2/feed ration = 60. Pressure = 1 atm. W/F = 0.46 h. (a) C7 products:
-methyl-cyclohexanone (ONE); 3-methylcyclohexanol (OL); toluene (TOL); (b) transalkylation and hydrogenolysis products.
3
Fig. 4 shows the effect of Fe in the catalyst on the product distri-
Fig. 6 shows the product selectivities as a function of m-cresol
◦
butions obtained at 300 C at a fixed W/F (0.46 h). It is interesting to
note that the yield of toluene increases with increasing Fe loading.
This clearly shows that the addition of Fe improves the deoxygen-
ation activity. However, the pure Fe catalyst exhibits a lower yield to
toluene, i.e., about 5% for the same conditions (see Table 2). The sec-
ond interesting trend is that the yields of 3-methylcyclohexanone
and 3-methylcyclohexanol decrease with increasing Fe loading.
Only a small fraction of 3-methylcyclohexanone is observed over
conversion. Over Ni/SiO , the selectivity to 3-methylcyclohexanone
2
is the highest, around 60%, while the selectivities to 3-
methylcyclohexanol and toluene are both around 20%. Over
(1:2)Ni–Fe/SiO , toluene is the only dominant one, whose selec-
2
tivity is almost 100% along all conversion. This again shows that
addition of Fe to Ni/SiO2 helps enhance the deoxygenation activ-
ity dramatically. The variation of activity with time on stream over
the Ni and Ni–Fe/SiO catalysts is compared in Fig. 7. While both of
2
(
2.5:1)Ni–Fe/SiO , but essentially zero on the other catalysts. This
them show some deactivation as a function of time, the bimetallic
catalyst seems to be slightly more stable.
2
result indicates that the addition of Fe inhibited the hydrogenation
of the ring.
Finally, as shown in Fig. 4b, the yields of xylenol (transalkylation
product) increase appreciably with increasing Fe loading. This may
be related to the increasing amount of Lewis acidity created by un-
reduced Fe, as discussed in Fig. 2. In line with these findings, a recent
study [23] on Ni and Ni–Cu catalysts used for the HDO of guaiacol
indicated that the selectivity for methyl-substitution compounds
increased with addition of Cu due to improvement of the acidic
properties on the catalyst surface.
Fig. 5 shows the product distributions for m-cresol conver-
sion on the (1:2)Ni–Fe/SiO2 as a function of W/F. The yields of
toluene and xylenol are the two major products and their yields
increase with W/F. By contrast, the hydrogenation products 3-
methylcyclohexanone and 3-methylcyclohexanol are negligible at
all W/F.
3.3. Using 3-methylcyclohexanone and 3-methylcyclohexanol as
feeds
Table 3 compares the product distributions obtained by feed-
ing 3-methylcyclohexanol (OL), 3-methylcyclohexanone (ONE), or
m-cresol over the monometallic and bimetallic catalysts. When
3-methylcyclohexanol was fed over the Ni/SiO2 catalyst, it was
quickly converted to 3-methylcyclohexanone. Similarly, when 3-
methylcyclohexanone was the feed, the corresponding alcohol was
obtained at high yield. These results indicate that this is a fast and
reversible reaction. In both cases, m-cresol was also observed as a
product, but toluene was not obtained. To get the same level of con-
version with m-cresol the W/F had to be raised by almost 20 times
since the rate of interconversion between 3-methylcyclohexanol
a 25
b 7
Methane
Conversion
Xylene
20
15
10
5
0
ONE
TOL
Minor
6
5
4
3
2
1
0
Phenol
Xylenol
0
.0
0.1
0.2
0.3
0.4
0.5
0.0
0.1
0.2
0.3
0.4
0.5
W/F (h)
W/F (h)
◦
Fig. 5. Product distribution of m-cresol over (1:2)Ni–Fe/SiO2 at 300 C as a function of W/F. H2/feed ration = 60. Pressure = 1 atm. (a) C7 products: 3-methylcyclohexanone
ONE); 3-methylcyclohexanol (OL); toluene (TOL); (b) transalkylation and hydrogenolysis products.
(

5
2
L. Nie et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55
a 70
b
100
90
80
70
60
50
40
30
20
10
0
ONE
OL
ONE
OL
TOL
60
TOL
5
0
0
4
30
2
0
0
0
1
0
2
4
6
8
10 12 14 16 18
0
2
4
6
8
10 12 14 16 18
Conversion of m-cresol (%)
Conversion of m-cresol (%)
◦
Fig. 6. C7 Product selectivity as a function of conversion of m-cresol at 300 C. a. Ni/SiO2; b. (1:2)Ni–Fe/SiO2. H2/feed molar ratio = 60. Pressure = 1 atm. C7 products: 3-
methylcyclohexanone (ONE); 3-methylcyclohexanol (OL); toluene (TOL).
Table 3
◦
Product distribution from different feed at 300 C. H2/feed molar ratio = 60. Pressure = 1 atm. C7 products: 3-methyl-cyclohexanone (ONE); 3-methylcyclohexanol (OL); toluene
(
TOL); methylcyclohexane (ENE).
Catalyst
Ni/SiO2
Feed
OL
0.02
ONE
0.02
m-Cresol
W/F (h)
Conversion (%)
Yield (%)
ONE
OL
m-Cresol
TOL
0.36
13.8
13.8
16.6
11.5
–
2.3
0
–
4.9
1.5
–
1.8
0
8.3
8.3
0
ENE
0
0
Transalkylation
0
0
5.6
(
1:2)Ni–Fe/SiO2
W/F (h)
Conversion (%)
Yield (%)
ONE
OL
m-Cresol
TOL
0.02
8.7
0.02
7.6
0.12
8.7
8.2
–
0.3
0
0.5
0
–
0
0
–
3.6
0
7.1
0.3
0
0.2
0
ENE
Transalkylation
5.1
SiO2
W/F
Conversion (%)
0.46
0.3
2
1
1
1
1
1
0
8
6
4
2
0
8
6
4
2
0
and 3-methylcyclohexanone is much faster than the conversion of
m-cresol to either of these compounds, which would suggest that
during the HDO of m-cresol the saturated alcohol and ketone are
pseudo-equilibrated.
Interestingly, the absence of toluene as a product when 3-
methylcyclohexanol was the feed indicates that dehydration does
not occur to any significant extent at the low W/F used (i.e., 0.02 h).
To further investigate the role of dehydration on the support, 3-
methylcyclohexanol was fed at the same W/F as used with m-cresol
Conversion on Ni
TOL yield on Ni
Conversion on Ni-Fe
TOL yield on Ni-Fe
(
i.e., W/F = 0.36 h) over pure SiO . As shown in Table 3, no conver-
2
sion was observed.
In addition, a significant fraction of the products from m-cresol
over Ni/SiO2 are due to transalkylation (xylenols, xylene, phenol).
However, this reaction did not occur when 3-methylcyclohexanol
or 3-methylcyclohexanone was used as a feed at W/F = 0.02 h.
One of the most interesting differences observed over the
bimetallic Ni–Fe/SiO2 catalyst was the appearance of the alcohol
dehydration product methylcyclohexene (ENE). This product was
particularly abundant when 3-methylcyclohexanol was used as
feed. The Lewis acid sites present on the bimetallic catalyst might
be responsible for the observed dehydration. However, even in this
case, no toluene was formed. This further proves that the dehy-
dration of 3-methylcyclohexanol cannot give toluene under this
0
20
40
60
80
100
120
140
TOS (min)
Fig. 7. Effect of time on stream on of conversion of m-cresol and toluene yield
over Ni/SiO2. (open symbol) and (1:2)Ni–Fe/SiO2 (solid symbol) at 300 C. H2/feed
ration = 60. Pressure = 1 atm. Toluene (TOL).
◦

L. Nie et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55
53
OH
Dehydration
TOL
C=O
Hydrogenation
Fe, Ni-Fe
OH
O
O
Tautomerization
C=C
Hydrogenation
Ni
Ni
OH
OL
ONE
Scheme 1. HDO reaction pathways over Ni, Fe, and Ni–Fe catalysts.
reaction conditions. Again, 3-methylcyclohexanone and m-cresol
were seen as products from 3-methylcyclohexanol.
aromatic ring and/or C C. On these metals aromatics and/or olefins
adsorb parallel to the surface and they are highly active for ring
and/or C C saturation [42–47]. When Group VIII metals are alloyed
with oxophilic metals such as Sn, Re, or Fe we may anticipate that
the interaction with the ring will be decreased while the interac-
tion with the carbonyl group will be enhanced. This is in fact the
case for ␣,␤-unsaturated aldehydes, for which Group VIII metals
result in saturation of the C C double bond, but bimetallics such
as Pt–Sn or Pt–Fe that contain oxophilic elements produce unsatu-
rated alcohols [48–52].
When feeding 3-methylcyclohexanone over the bimetallic
catalyst, the hydrogenation product 3-methylcyclohexanol and
dehydrogenation product m-cresol are detected. As in the previ-
ous case, methylcyclohexene is also formed from the dehydration
of methylcyclohexanol, but no toluene is detected. An interest-
ing question arises when comparing these trends. If m-cresol
is readily formed when feeding either 3-methylcyclohexanol or
3
-methylcyclohexanone, why is toluene not formed by ring dehy-
drogenation from methylcyclohexene? If m-cresol were produced
by ring dehydrogenation of 3-methylcyclohexanol the dehydro-
genation of methylcyclohexene to toluene would be much easier.
As discussed below, the reason for this discrepancy is that m-cresol
is not directly produced from 3-methylcyclohexanol, but rather
from 3-methylcyclohexanone.
We may expect that this chemoselective hydrogenation of the
carbonyl group may occur in our case for the conversion of m-cresol
over Fe-containing catalysts.
Therefore, in case (a), which is typical of Ni, 3-
methylcyclohexanone becomes dominant (Table 3). This product
can be further hydrogenated to 3-methylcyclohexanol. As men-
tioned above, if the support contains enough acidity to catalyze
the dehydration of the saturated alcohol 3-methylcyclohexanol,
toluene could be obtained after dehydration/dehydrogenation, as
previously proposed for the conversion of phenol over Ni/SiO2
[53] and cresol over Pt catalysts [22]. As reported previously
[9,14,22], over a bi-functional catalyst containing acid and metal
functions, the deoxygenation of phenolic compounds proceeds
via hydrogenation of the aromatic ring to the corresponding
alcohols followed by dehydration to the olefin then dehydro-
genated back to aromatic. However, over our Ni/SiO2 catalyst
under the reaction conditions investigated, dehydration does
not occur to a significant extent. As shown in Table 3, when we
fed 3-methylcyclohexanol over Ni/SiO2, we did not observe any
formation of toluene or methylcyclohexene. Only dehydrogenation
of 3-methylcyclohexanol to 3-methylcyclohexanone and cresol
was seen. This indicates that toluene does not derive from the
dehydration/dehydrogenation of the 3-methylcyclohexanol.
In case (b), since hydrogenation of the aromatic ring is sup-
pressed on the Fe-containing catalysts, the deoxygenation of
m-cresol requires the participation of its keto tautomer methyl-
cyclohexadienone intermediate. Once the carbonyl group of such
species is hydrogenated, the resulting methylcyclohexadienol can
be readily dehydrated to toluene, which is driven by aromatic sta-
bilization. The electrophilic character of unreduced metal (either Fe
or Ni) might help stabilize the tautomer intermediate, thus enhanc-
ing the deoxygenation activity [54]. Due to its oxophilicity, Fe is
3.4. Mechanistic implications of the observed product
distributions
The significant differences observed between the m-cresol con-
version over Ni or over Fe-containing catalysts may be explained
in terms of a reaction pathway that we summarize in Scheme 1,
which can account for all the results and product trends reported
above.
First, we propose that over Ni/SiO2 at least two parallel path-
ways may exist. One is the direct hydrogenation of the aromatic
ring to 3-methylcyclohexanol, which could be dehydrated and then
dehydrogenated to toluene. However, a second (indirect) path can
also exist and this might be crucial for the highly selective pro-
duction of toluene. It is known that m-cresol can tautomerize to a
highly unstable methyl cyclohexadienone. While this intermediate
cannot be detected under reaction conditions, we may anticipate
that depending on the catalyst, it could result in:
(
a) methyl cyclohexanone, when the double bonds in the ring are
hydrogenated, or
(
b) methyl cyclohexadienol, when the C O group is hydrogenated.
The preference in hydrogenating the ring or the carbonyl group
strongly depends on the catalyst used. For example, Group VIII
metal catalysts such as Ni, Pt, or Pd have a strong affinity for the

5
4
L. Nie et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55
Fig. 8. Optimized adsorption structures of m-cresol on (a) Ni(1 1 1) surface, (b) perfect NiFe(1 1 1) surface, (c) patched NiFe(1 1 1) surface and (d) Fe(1 1 0) surface. (e) Gas
phase. C OH bond lengths are labeled in each molecule.
more efficient in linking to the O of the carbonyl than Ni, which in
turn has a higher affinity for the aromatic ring. Hence, case (b) is
largely promoted over Fe containing catalysts resulting in enhanced
deoxygenation activity. It must be emphasized that although the
products from this pathway would appear to result from a direct
from the analysis of the reaction products. That is, the interaction
with the O of the C O is strongest on the Fe(1 1 0) surface, followed
by the NiFe(1 1 1) alloy and negligible or rather repulsive on the
Ni(1 1 1) surface. If the C O bond elongation is used as an indicator
for the M O affinity a clear trend is observed, i.e., 1.408 A˚ > 1.399
2
C(sp ) O cleavage, i.e., DDO, this is not the case. This is a path-
(1.396) > 1.379 > 1.376 for the Fe, Ni–Fe, Ni, and gas-phase, respec-
tively.
way that requires a relatively low activation energy and is more
◦
likely to occur under the reaction conditions investigated (300 C,
Note that the calculated binding energy of m-cresol on
Fe(1 1 0) (−23.6 kcal/mol) is even higher than that on Ni(1 1 1)
(−15.5 kcal/mol) or the bulk-terminated alloy NiFe(1 1 1)
(−16.5 kcal/mol). However, in the case of Fe(1 1 0) the enhanced
atmospheric H ).
2
As reported in our previous study on furfural and benzaldehyde
[
26], the addition of Fe to the monometallic Ni catalyst, suppresses
hydrogenation of the ring. The reason for this activity drop may be
that Fe causes repulsion to the ␲ electron system of the ring. As
shown in the DFT calculations below, this repulsion is particularly
pronounced when the Ni–Fe particles show a non-uniform distri-
bution, forming Fe patches on the surface. Thus, it is then expected
that Ni–Fe and Fe catalysts will have a poor aromatic ring hydro-
genation activity compared to Ni catalysts [3,18,55]. By contrast,
the presence of Fe promotes the hydrogenation of the C O group.
adsorption energy is due to
a strong interaction with the
oxygenated group rather than the ring.
4. Conclusion
A novel type of catalyst has been identified that shows high
selectivity for hydrodeoxygenation of phenolic compounds. It is
demonstrated that when the catalyst contains an oxophilic metal,
such as Fe, the chemoselective hydrogenation of a keto tautomer
intermediate can be greatly enhanced, resulting in a cyclodiene
alcohol that is readily dehydrated to the aromatic hydrocarbon,
favored by the stabilization provided by the aromaticity of the prod-
uct. This mechanism does not involve the partial hydrogenation
3.5. DFT calculation
We have calculated the energetics of the adsorption of m-
cresol on Ni(1 1 1), Fe(1 1 0) and NiFe(1 1 1) alloy surface, which
can provide information on how m-cresol molecules interact with
the metal surfaces. As shown in Fig. 8, the m-cresol prefers a flat
adsorption configuration on Ni(1 1 1) Fe(1 1 0) and bulk-terminated
NiFe(1 1 1) alloy surface, but a slanted configuration on a patchy
NiFe(1 1 1) alloy surface due to repulsion of the phenyl ring. The
reason for the major difference is in line what was proposed above
2
of the ring nor a direct C(sp ) O cleavage, as has been previously
proposed. By contrast, when Fe is not present, the Ni-containing
catalyst preferentially interacts with the aromatic ring and pro-
motes its saturation, producing 3-methyl-cyclohexanone and
3-methylcyclohexanol. On the catalysts investigated here, these
saturated rings are not able to produce toluene via dehydration/

L. Nie et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 47–55
[22] C. Zhao, S. Kasakov, J. He, J.A. Lercher, J. Catal. 296 (2012) 12–23.
55
dehydrogenation. In addition to these C7 products, some transalky-
[
[
[
23] X. Zhang, T. Wang, L. Ma, Q. Zhang, Y. Yu, Q. Liu, Catal. Commun. 33 (2013)
lation products are also observed. It is proposed that they arise
from Lewis acids associated with the incomplete reduction of Fe
and/or Ni cations.
15–19.
24] M.V. Bykova, D.Y. Ermakov, V.V. Kaichev, O.A. Bulavchenko, A.A. Saraev, M.Y.
Lebedev, V.A. Yakovlev, Appl. Catal. B 113–114 (2012) 296–307.
25] V.A. Yakovlev, S.A. Khromova, O.V. Sherstyuk, V.O. Dundich, D.Y. Ermakov, V.M.
Novopashina, M.Y. Lebedev, O. Bulavchenko, V.N. Parmon, Catal. Today 144
Acknowledgments
(
2009) 362–366.
[
[
26] S. Sitthisa, W. An, D.E. Resasco, J. Catal. 284 (2011) 90–101.
27] M. Badawi, J.-F. Paul, E. Payen, Y. Romero, F. Richard, S. Brunet, A. Popov, E.
Kondratieva, J.-P. Gilson, L. Mariey, A. Travert, F. Maugé, Oil Gas Sci. Technol.
Support from the National Science Foundation (NSF EPSCoR
814361) and the Department of Energy (DE-FG36GO88064) is
0
(
2013), http://dx.doi.org/10.2516/ogst/2012041.
greatly appreciated. Allocation of computing time provided by the
OU Supercomputing Center for Education and Research (OSCER)
at the University of Oklahoma is acknowledged. We thank CAPES,
Brazil for a scholarship for one of us (PMS).
[28] E.O. Odebunmi, D.F. Ollis, J. Catal. 80 (1983) 56–64.
[29] G. Kresse, J. Hafner, Phys. Rev. B 48 (17) (1993) 13115–13118.
[
[
30] G. Kresse, J. Furthmuller, Phys. Rev. B 54 (16) (1996) 11169–11186.
31] J.P. Perdew, K. Burke, Phys. Rev. Lett. 77 (18) (1996) 3865–3868.
[32] P.E. Blochl, Phys. Rev. B 50 (24) (1994) 17953–17979.
[
[
[
[
33] G. Kresse, D. Joubert, Phys. Rev. B 59 (3) (1999) 1758–1775.
34] A.V. Ruban, H.L. Skriver, J.K. Norskov, Phys. Rev. B 59 (24) (1999) 15990–16000.
35] G.B. Raupp, W.N. Delgass, J. Catal. 58 (3) (1979) 337–347.
References
36] L. Wang, D. Li, M. Koik, S. Koso, Y. Nakagawa, Y. Xu, K. Tomishige, Appl. Catal. A
[1] P.M. Mortensen, J.-D. Grunwaldt, P.A. Jensen, K.G. Knudsen, A.D. Jensen, Appl.
392 (2011) 248–255.
Catal. A: Gen. 407 (2011) 1–19.
[
37] A.P. Legrand, H. Hommel, A. Tuel, A. Vidal, H. Balard, E. Papirer, P. Levitz, M.
Czernichowski, R. Erre, H. Van Damme, J.P. Gallas, J.F. Hemidy, J.C. Lavalley, O.
Barres, A. Burneau, Y. Grillet, Adv. Colloid Interface Sci. 33 (1990) 91.
38] H.E. Bergna, W.O. Roberts, Surfactant Science Series, vol. 131, 2006.
39] L. Oliviero, A. Vimont, J.-C. Lavalley, F.R. Sarria, M. Gaillard, F. Mauge, Phys.
Chem. Chem. Phys. 7 (2005) 1861–1869.
[
[
2] G.W. Huber, S. Iborra, A. Corma, Chem. Rev. 106 (2006) 4044–4098.
3] R. Olcese, M.M. Bettahar, B. Malaman, J. Ghanbaja, L. Tibavizco, D. Petitjean, A.
Dufour, Appl. Catal. B 129 (2013) 528–538.
[
[
[
[
[
4] D.M. Alonso, S.G. Wettstein, J.A. Dumesic, Chem. Soc. Rev. 41 (2012) 8075–
8
098.
5] P.M. Mortensen, J.D. Grunwaldt, P.A. Jensen, K.G. Knudsen, A.D. Jensen, Appl.
Catal. B 407 (2011) 1–19.
6] S. Echeandia, P.L. Arias, V.L. Barrio, B. Pawelec, J.L.G. Fierro, Appl. Catal. B 101
[
[
40] G. Connell, A. Dumesic, J. Catal. 105 (1987) 285–298.
41] K. Hadjiivanov, M. Mihaylov, D. Klissurski, P. Stefanov, N. Abadjieva, E. Vas-
sileva, L. Mintchevz, J. Catal. 185 (1999) 314–323.
(
2010) 1–12.
[
42] P.N. Rylander, Catalytic Hydrogenation Over Platinum Metals, Academic Press,
New York, 1967, pp. 309.
[
[
7] D.E. Resasco, J. Phys. Chem. Lett. 2 (2011) 2294–2295.
8] A. Ausavasukhi, Y. Huang, A.T. To, T. Sooknoi, D.E. Resasco, J. Catal. 290 (2012)
[
[
43] G.C. Bond, Platinum Metals Rev. 12 (3) (1968) 100–105.
44] M.L. Honkela, J. Bjork, M. Persson, Phys. Chem. Chem. Phys. 14 (2012)
9
0–100.
[9] E. Furimsky, Appl. Catal. A 199 (2000) 147–190.
5849–5854.
[
[
[
[
[
[
10] P.A. Zapata, J. Faria, M.P. Ruiz, D.E. Resasco, Top. Catal. 55 (2012) 38–52.
11] L. Nie, D.E. Resasco, Appl. Catal. A 447–448 (2012) 14–21.
12] T.N. Pham, D. Shi, T. Sooknoi, D.E. Resasco, J. Catal. 295 (2012) 169–178.
13] D.C. Elliott, Energy Fuels 21 (2007) 1792–1815.
14] A.J. Foster, P.T.M. Do, R.F. Lobo, Top. Catal. 55 (2012) 118–128.
15] C. Zhao, Y. Kou, A.A. Lemonidou, X.B. Li, J.A. Lercher, Chem. Commun. 46 (2010)
[
[
45] L. Delle Site, A. Alavi, C.F. Abrams, Phys. Rev. B 67 (2003) 193406.
46] M.L. Honkela, J. Bjork, M. Perssonbd, Phys. Chem. Chem. Phys. 14 (2012)
5849–5854.
[
[
47] S.D. Lin, M.A. Vannice, J. Catal. 143 (1993) 554–562.
48] J. Beccat, J.C. Bertolini, Y. Gauthier, J. Massardier, P. Ruiz, J. Catal. 126 (1990)
451.
4
12–414.
[
[
49] T. Birchem, C.M. Pradier, Y. Berthier, G. Cordier, J. Catal. 161 (1996) 68.
50] A. Borgna, B.G. Anderson, A.M. Saib, H. Bluhm, M. Havecker, A. Knop-Gericke,
A.E.T. Kuiper, Y. Tamminga, J.W.J. Niemantsverdriet, Phys. Chem. B 108 (2004)
[
[
[
16] X. Zhu, L.L. Lobban, R.G. Mallinson, D.E. Resasco, J. Catal. 281 (2011) 21–29.
17] J.H. Sinfelt, Catal. Today 29 (1973) 308.
18] R. Olcese, M.M. Bettahar, D. Petijean, B. Malaman, F. Giovanella, A. Dufour, Appl.
Catal. B 115–116 (2012) 63–73.
17905.
[
[
[
[
51] T.B.L.W. Marinelli, S. Naubuurs, V. Ponec, J. Catal. 151 (1995) 431.
52] T.B.L.W. Marinelli, V. Ponec, J. Catal. 156 (1995) 51.
53] E. Shin, M.A. Keane, Ind. Eng. Chem. Res. 39 (2000) 883–892.
54] G.V. Smith, F. Notheisz, Heterogeneous Catalysis in Organic Chemistry, Aca-
demic Press, 1999.
[
[
19] A. Ausavasukhi, T. Sooknoi, D.E. Resasco, J. Catal. 268 (2009) 68–78.
20] A. Ausavasukhi1, Y. Huang, A.T. To, T. Sooknoi, D.E. Resasco, J. Catal. 290 (2012)
9
0–100.
21] A. Cho, J. Shin, A. Takagaki, R. Kikuchi, T.S. Oyama, Top. Catal. 55 (2012)
69–980.
[
9
[
55] K.J. Yoon, M.A. Vannice, J. Catal. 82 (1983) 457–468.