The Effect of Metal Type on Hydrodeoxygenation of Phenol Over Silica Supported Catalysts

Article abstract of DOI:10.1007/s10562-016-1815-5

Abstract: Different metals supported on SiO₂ were tested for the hydrodeoxygenation of phenol. For Pt/SiO₂, Pd/SiO₂ and Rh/SiO₂ catalysts, phenol is mainly tautomerized, followed by hydrogenation of the aromatic ring. The direct dehydroxylation of phenol followed by hydrogenolysis is favored over more oxophilic metals (Ru, Co and Ni). Graphical Abstract: [Figure not available: see fulltext.]

Full text of DOI:10.1007/s10562-016-1815-5

Catal Lett
DOI 10.1007/s10562-016-1815-5
The Effect of Metal Type on Hydrodeoxygenation of Phenol Over
Silica Supported Catalysts
1
,2
1
3
•
Camila A. Teles
Lisiane V. Mattos
•
Raimundo C. Rabelo-Neto
4
•
Jerusa R. de Lima
1,2
3
•
Daniel E. Resasco
•
Fabio B. Noronha
Received: 5 May 2016 / Accepted: 14 July 2016
Ó Springer Science+Business Media New York 2016
Abstract Different metals supported on SiO were tested
2
Graphical Abstract
for the hydrodeoxygenation of phenol. For Pt/SiO , Pd/
2
SiO and Rh/SiO catalysts, phenol is mainly tautomerized,
2
2
followed by hydrogenation of the aromatic ring. The direct
dehydroxylation of phenol followed by hydrogenolysis is
favored over more oxophilic metals (Ru, Co and Ni).
Keywords Phenol Á SiO Á Hydrodeoxygenation Á
2
Tautomerization Á Dehydroxylation Á Biomass Á Bio-oil
Electronic supplementary material The online version of this
article (doi:10.1007/s10562-016-1815-5) contains supplementary
material, which is available to authorized users.
1
Introduction
&
Fabio B. Noronha
fabio.bellot@int.gov.br
The bio-oil obtained from fast pyrolysis of lignocellulosic
biomass has great potential for the production of liquid
biofuels. However, the bio-oil contains a high amount of
oxygen-containing compounds such as phenol, cresol,
guaiacol, which are products from the decomposition of
lignin fraction of the lignocellulosic biomass. Due to the
high oxygen content, bio-oil exhibits a lower energy den-
sity compared to crude oil, high corrosivity and poor
thermal stability [1, 2]. Therefore, bio-oil has to be
upgraded to remove oxygen in order to be used as a fuel in
vehicles, which may be achieved by hydrodeoxygenation
(HDO) process [3]. Because of the complex composition of
1
2
Catalysis Division, National Institute of Technology,
Rio de Janeiro 20081-312, Brazil
Chemical Engineering Department, Military Institute of
Engineering, Pra c¸ a Gal. Tiburcio 80, Rio de Janeiro 22290-270,
Brazil
3
4
Chemical Engineering Department, Fluminense Federal
University, Rua Passo da P a´ tria 156, Niter o´ i 24210-240,
Brazil
School of Chemical, Biological, and Materials Engineering,
Center for Biomass Refining, The University of Oklahoma,
Norman, OK 73019, USA
123

C. A. Teles et al.
bio-oil, several studies investigate the HDO reaction of
model molecules representatives of specific fractions of the
lignocellulosic biomass such as guaiacol, phenol, cresol.
The reaction mechanism for HDO of phenolic com-
pounds such as phenol or cresol has been extensively
studied in the literature in order to design more efficient
catalysts for this process [4–16]. Three reaction pathways
have been proposed in the literature [17]. The first one is
based on the sequential hydrogenation (HYD) of the aro-
matic ring on metal sites followed by dehydration of
cyclohexanol on acid sites. Therefore, this reaction path-
way may take place on catalysts containing a support with
sufficient acidity to catalyze the dehydration of the alcohol
this case, the cresol molecule adsorbs perpendicular to the
surface, inhibiting the hydrogenation of the ring.
Experimental and theoretical calculations were used to
investigate the mechanism of m-cresol over Pt/SiO and
2
Ru/SiO catalysts in gas phase, at 573 K and atmospheric
2
pressure [21]. 3-methylcyclohexanone was the main pro-
duct formed over Pt/SiO catalyst whereas toluene and C –
2
1
C5 hydrocarbons were mainly produced over Ru/SiO2
catalyst. Density functional theory (DFT) methods were
used to investigate the m-cresol conversion over the Pt
(111) and Ru (0001) surfaces. The DFT results showed that
the reaction proceeds through the tautomerization route
over Pt (111) surface. However, the direct deoxygenation
of m-cresol is preferred instead of the tautomerization route
over Ru (0001) surface. This result was attributed to the
stronger oxophilicity of Ru, which leads to a high energy
barrier for the tautomerization pathway. In this case, an
unsaturated hydrocarbon species is formed during the
direct dehydroxylation of m-cresol, which undergoes
hydrogenolysis to C –C hydrocarbons. Hensley et al. [17]
(
cyclohexanol or 3-methylcyclohexanol) [18]. However,
dehydration does not occur to a significant extent over
silica or zirconia supported catalysts since these supports
do not have significant acidity [12]. The second one is the
so called direct deoxygenation (DDO) and involves the
cleavage of the C–O bond in the phenolic molecule. This
pathway requires a relatively high activation energy and
therefore, it may occurs depending on reaction temperature
and catalyst used [19]. The third mechanism is the tau-
tomerization of phenolic compound to a cyclohexadienone
intermediate (2,4-cyclohexadienone or 3,5-methylcyclo-
hexadienone) [10–12, 20, 21]. This intermediate may react
by two different routes. For example, in the HDO of phe-
nol, the hydrogenation of the ring, producing 2-cyclo-
hexen-1-one that is hydrogenated to cyclohexanone and
then cyclohexanol; or the hydrogenation of the carbonyl
group, leading to the formation of 2,4-cyclohexadienol,
which can in turn, convert to benzene. The differences in
the reaction pathway strongly depend on the type of metal
and support.
1
5
also suggested that the direct deoxygenation occurs on the
HDO of phenol over Fe (110) surface. However, the acti-
vation energy barrier and reaction energies for the C–O
bond cleavage are larger on Pd (111) and thus, this surface
is not active for deoxygenation. In both studies, the direct
deoxygenation occurred on the more oxophilic metal.
Different catalysts have been studied for HDO of phe-
nolic compounds such as phenol or cresol, including sup-
ported
noble
metals
(Pd,
Pt,
Rh,
Ru)
[1, 6, 10, 11, 16, 22–29] supported base metals (Ni,Co,-
Fe,Cu) [6, 30–33] and bimetallic catalysts [10, 26]. How-
ever, studies comparing the performance of a series of
catalysts with different metals are scarce and only few were
carried out in gas phase but in these cases, the HDO of
m-cresol was investigated [10, 29].
Nie et al. [10] studied the HDO of m-cresol in gas phase
on SiO -supported Ni, Fe, and bimetallic Ni–Fe catalysts at
2
5
73 K and atmospheric pressure. Over the monometallic
Therefore, there has been no systematic study for the
HDO of phenol reaction over a wide range of metals
under the same conditions, investigating the effect of the
type of the metal on reaction pathways. However, the
mechanism of HDO of phenol consists of a metallic
particle and a Lewis acid site of the support and the
reaction takes place in the metal-support interface. Then,
in order to obtain a better understanding of the effect of
the type of the metal on the HDO of phenol, it is funda-
mental the use of an inert support such as silica that does
not exhibit Lewis acid sites.
Ni catalyst, the dominant product was 3-methylcyclohex-
anone whereas toluene was the main product on Fe and Ni–
Fe bimetallic catalysts. The differences in product distri-
butions observed for Ni and Fe-containing catalysts were
explained in terms of a reaction pathway based in a tau-
tomerization step. According to this mechanism, m-cresol
is tautomerized to a highly unstable methylcyclohexa-
dienone intermediate that may react through two different
routes. In the case of Group VIII metal catalysts such as Ni,
Pt, or Pd that have a strong affinity for the aromatic ring,
the cresol molecule adsorbs in parallel to the surface and
they are highly active for ring hydrogenation. For oxophilic
metals such as Fe and its alloys, the strong interaction of
the oxygen from the carbonyl group of the keto tautomer
intermediate and the oxophilic metal (in this work repre-
sented by unreduced metal) promotes its hydrogenation. In
The goal of this work is to investigate the effect of the
metal type on the HDO of phenol over silica supported cat-
alysts. A systematic study was performed by carrying out the
phenol conversion over different metals (Pt, Pd, Rh, Ru, Ni
and Co) supported on SiO . Silica was used as an inert sup-
2
port that enables to investigate only the effect of the metal.
1
23

The Effect of Metal Type on Hydrodeoxygenation of Phenol…
2
Experimental
isopropyl alcohol and liquid nitrogen. Then, the catalyst
was passivated with a 5 % O /He mixture for 1 h at 209 K.
2
2
.1 Catalyst Synthesis
2.3 Catalytic Activity
SiO (Aldrich) was calcined under air flow at 1073 K for 5 h.
2
SiO supported metal (Pt, Pd, Rh, Ru, Ni, Co) catalysts were
2
The vapor-phase conversion of the phenol was carried
out in a fixed-bed quartz reactor, operating at atmo-
prepared by incipient wetness impregnation of the support
with aqueous solution of the respective precursor salts
spheric pressure of H and 573 K. Prior to reaction, the
2
[
Pt(NH ) (NO ) (Aldrich), Pd(NO ) Á2H O (Umicore),
3
4
3 2
3 2
2
catalyst was reduced in situ under pure hydrogen
RhCl ÁH O (Aldrich), Ru(NO)(NO ) (Alfa Aesar),
3
2
3 3
(
60 mL/min) at 773 K for 1 h. The catalysts were diluted
Ni(NO ) Á6H O (Acros) and Co (NO ) Á6H O (Acros)].
3
2
2
3 2
2
with inert material (SiC mass/catalyst mass = 3.0) to
After impregnation, the powder was dried in air at 293 K for
avoid hot-spot formation. The reactant mixture was
1
2 h and then calcined in air at 673 K for 3 h (2 K/min).
obtained by flowing H through the saturator containing
2
phenol, which was kept at the specific temperature
required to obtain the desired H /phenol molar ratio
2
2
.2 Catalyst Characterization
(
about 60). To avoid condensation, all lines were heated
at (523 K). The reaction products were analyzed by
GCMS Agilent Technologies 7890A, using HP-Innowax
capillary column and a flame-ionization detector (FID).
Catalysts were evaluated at different W/F by varying the
catalyst amount in the range of 2.5–160 mg. The W/F is
defined as the ratio of catalyst mass (g) to organic feed
mass flow rate (g/h). The product yield and selectivity
for each product were calculated as follows:
The chemical composition of each sample was determined
on a Wavelength Dispersive X-Ray Fluorescence Spec-
trometer (WD-XRF) S8 Tiger (Bruker) with a rhodium
tube operated at 4 kW. The analyses were performed with
the samples (300 mg) in powder form using a semi-quan-
titative method (QUANT-EXPRES/Bruker). Specific sur-
face areas of the samples were measured on
a
Micromeritics ASAP 2020 analyzer by N adsorption at the
2
boiling temperature of liquid nitrogen. The X-ray powder
diffraction pattern of the calcined and reduced/passivated
samples were obtained with CuKa radiation (l = 15,406
A°) using a RIGAKU diffractometer. Data were collected
over the 2h range of 10° to 90° using a scan rate of 0.02°/
step and a scan time of 1 s/step. TPR experiments were
performed in a TPR/TPD 2900 Micromeritics system
equipped with a thermal conductivity detector (TCD). The
catalyst was pretreated at 673 K for 1 h under a flow of air
prior to the TPR experiment in order to remove adsorbed
species from the catalyst surface. The reducing mixture
mol of product produced
Yield ð%Þ ¼
 100
ð1Þ
ð2Þ
mol of phenol fed
mol of product produced
Selectivity ð%Þ ¼ mol of phenol consumed  100
The turnover frequency was calculated by using the
reaction rate for the formation deoxygenated products and
the metal dispersion measured by CO chemisorption and
XRD.
(
10.0 % H /N ) was passed through the sample (80-
2
3 Results and Discussion
3.1 Catalyst Characterization
2
5
00 mg) at a flow rate of 30 mL/min and the temperature
was increased to 1273 K at a heating rate of 10 K/min. The
metal dispersion of Pd, Pt, and Rh was measured by CO
chemisorption, using the dynamic adsorption method.
Before adsorption, the samples (50 mg) were reduced
The metal loading, the surface area and the metallic dis-
persion of the samples are reported in Table 1. The surface
area of Pt and Rh supported catalysts was similar to that
under pure H (60 ml/min) at 773 K for 1 h (10 K/min),
2
2
cooled to room temperature, and flushed in He for 30 min.
Then, 1 mL-pulses of 5 % CO in He were injected, until
saturation was reached, as monitored on a quadrupole mass
spectrometer (MKS Cirrus 200). For Ru, Ni and Co based
catalysts, the metal dispersion was obtained from the
diffractogram of reduced samples using the Scherrer
equation. The calcined samples were reduced under pure
hydrogen (30 mL/min) at 773 K for 1 h, purged under N2
at the same temperature for 30 min and cooled to 298 K.
The reactor was maintained at 209 K, using a mixture of
one of silica (171 m /g). For Pd, Ru, Ni and Co catalysts,
the surface area slightly decreased after addition of metal.
The metallic dispersion varied from 9 to 44 %.
X-ray diffraction patterns for all catalysts are shown in
Fig. 1. All samples exhibited a broad diffraction peak
corresponding to amorphous silica. For Pt/SiO
SiO , only the line characteristic of silica is observed,
and Rh/
2
2
whereas the diffractograms of Pd, Ru, Ni and Co based
catalysts showed the lines of the respective oxides (PdO,
NiO, Co O and RuO ). The absence of metal peaks in
3 4 2
123

C. A. Teles et al.
Table 1 Metal composition,
surface area, metal dispersion
and turnover frequency for
HDO of phenol
2
BET (m /g)
a
-3 -1
s )
Catalysts
wt% Me
Dispersion (%)
TOFHDO (910
Pt/SiO
2
0.95
0.55
1.0
166
155
167
143
143
143
44
15
28
15
9
15.7
7.2
Pd/SiO
2
Rh/SiO
Ru/SiO
2
2
22.6
24.1
6.1
3.8
Ni/SiO
Co/SiO
a
2
9.6
2
9.7
9
6.9
For Ru, Co and Ni catalysts, the dispersion was obtained using the diffractograms of reduced samples and
the Scherrer equation. Metallic dispersion was calculated by CO chemisorption for Pt, Pd and Rh supported
catalysts
Fig. 1 XRD patterns of a SiO
2
; b Pt/SiO
. (diamond) SiO
PdO; (filled diamond) NiO; (spade symbol) Co ; (claver symbol)
RuO
2
; c Pd/SiO
2 2
; d Rh/SiO ;
e Ru/SiO
2
; f Co/SiO
2
and g Ni/SiO
2
2
; (filled circle)
3 4
O
2
XRD patterns of Pt/SiO and Rh/SiO is likely due to the
2
2
high metallic dispersion rather than low metal loading
since Pd/SiO2 catalyst has the lowest metal loading
(
Table 1), but still shows a peak associated with PdO in
Fig. 2 TPR profiles of a SiO
e Ru/SiO ; f Co/SiO ; g Ni/SiO
2
2 2 2
; b Pt/SiO ; c Pd/SiO ; d Rh/SiO ;
XRD pattern.
2
2
2
The TPR profiles of all samples are shown in Fig. 2.
Pt/SiO , Pd/SiO , Rh/SiO , Ru/SiO , and Ni/SiO cata-
2
2
2
2
2
lysts showed only one peak at 409, 332, 457, 521, and
37 K that corresponds to the reduction of the respective
oxides (PtO , PdO, Rh O , RuO , and NiO). The broad
3.2 HDO of Phenol over Me/SiO Catalysts
2
7
2
2
3
2
Figure 3 shows the phenol conversion and product yield at
573 K over Pt/SiO , Ru/SiO , and Ni/SiO as a function of
hydrogen consumption at high temperature observed for
2
2
2
Pd/SiO catalysts could be attributed to the reduction of
2
W/F. Silica support did not exhibited any activity even at
small PdO particles. For Co/SiO catalyst, two peaks are
2
the highest W/F used. Ni/SiO catalyst showed the highest
2
observed, which are due to the two step reduction of
activity. For Pt/SiO and Ni/SiO catalysts, cyclohexanone
2
2
cobalt oxide (Co O ? CoO ? Co) [34]. All samples
were completely reduced up to 773 K except for Co/SiO2
and cyclohexanol were the main products formed along the
entire W/F range, with small amounts of benzene and
methane (Ni catalyst). By contrast, methane was the
3
4
catalyst that still exhibited hydrogen consumption up to
8
83 K.
dominant product over Ru/SiO catalyst, with only small
2
1
23

The Effect of Metal Type on Hydrodeoxygenation of Phenol…
2 2 2
Fig. 3 Phenol conversion and yield of products for HDO of phenol as a function of W/F over Pt/SiO , Ru/SiO and Ni/SiO catalysts. Reaction
conditions: T = 573 K, P = 1 atm, and H /phenol molar ratio 60. ONE—cyclohexanone, OL—cyclohexanol
2
amounts of cyclohexanone, benzene and C –C hydrocar-
5
high but now, the formation of small amounts of benzene,
6
bons in the whole W/F range.
C –C hydrocarbons (n-hexane, 2-metyl-butane, cyclo-
5 6
The TOF for HDO was calculated taking into account
the data obtained under differential reaction conditions
hexene, cyclohexane) and methane is also observed. In the
case of Co/SiO catalyst, the selectivities to cyclohexanone
2
(
conversion around 10 %) and the results are reported in
and cyclohexanol significantly decreased whereas a sig-
nificant formation o-cresol (17.0 %) and small amounts of
C12 hydrocarbons is observed. The formation of C12
hydrocarbons such as biphenyl, cyclohexylbenzene, 2-cy-
clohexylphenol and 2-phenylphenol have been reported for
HDO of phenol over supported Pd catalysts [12]. Bicyclic
hydrocarbons were formed probably from the alkylation of
phenolic and aromatic rings by cyclohexanol or cyclo-
hexanone over Lewis acid sites [35]. The hydrogenolysis of
these C12 compounds over Lewis acid sites is likely
responsible for the formation of o-cresol [36]. In our work,
Table 1. The TOF varied significantly with the type of
metal. Rh/SiO catalyst was 3 fold more active than Pd, Ni
2
and Co-based catalysts. The following order of TOF was
observed: Pd/SiO2 & Ni/SiO2 & Co/SiO \ Pt/SiO \
Rh/SiO & Ru/SiO . This result suggests that the type of
2
2
2
2
metal strongly affects the deoxygenation activity.
The product distribution for HDO of phenol at 573 K
over all silica supported catalysts at similar levels of con-
version (around 10 %) is listed in Table 2. Cyclohexanone
and cyclohexanol were basically the only products formed
over Pt/SiO and Pd/SiO catalysts with small amounts of
SiO does not exhibit any measurable acidity. Therefore,
2
2
2
benzene. For Rh/SiO , Co/SiO and Ni/SiO catalysts, the
2
the Lewis acidity of Co/SiO catalyst was likely created by
2
2
2
selectivities to cyclohexanone and cyclohexanol are still
the presence of unreduced Co species after reduction at
123

C. A. Teles et al.
Table 2 Product distribution
for HDO of phenol at 573 K and
atmospheric pressure over silica
supported metal catalysts
Pt/SiO
2
Pd/SiO
2
Rh/SiO
2
Ru/SiO
2
Ni/SiO
2
Co/SiO
2
W/F (h)
0.023
13.2
0.151
17.1
0.045
9.4
0.209
12.4
0.032
14.7
0.089
11.3
Conversion (%)
Selectivity (%)
ONE
a
88.0
11.0
0.8
–
93.7
3.0
3.3
–
85.3
4.2
3.1
7.2
–
19.4
3.4
5.8
8.7
62.3
0.4
–
63.6
20.0
4.8
35.9 (53.4)
12.0 (19.2)
11.7 (10.0)
5.0 (4.9)
OL
Benzene
C
5
–C
CH
12 hydrocarbons
o-cresol
6
hydrocarbons
1.2
4
–
–
7.2
13.6 (5.4)
4.8 (1.2)
C
0.2
–
–
0.2
–
1.2
–
2.0
17.0 (5.9)
C
nyl; cyclohexylbenzene; pentylbenzene; 2-phenylphenol
5 6
–C hydrocarbons—cyclohexane; cyclohexene; 2-methyl-butane; n-hexane. C12 hydrocarbons—biphe-
a
2
Selectivities for HDO of phenol over Co/SiO catalyst reduced at 873 K
7
73 K, as determined by TPR. In order to demonstrate the
ring-hydrogenated products, cyclohexanone and cyclo-
hexanol. Diffuse reflectance infrared Fourier transform
spectroscopy experiments revealed the participation of a
keto-tautomer intermediate (2,4-cyclohexadienone) in the
reaction. This intermediate can be hydrogenated in two
different pathways. For Pd/SiO and Pd/Al O catalysts,
role of the unreduced Co species on the reaction mecha-
nism, the HDO of phenol reaction was carried out over the
Co/SiO2 catalyst reduced at 873 K. According to TPR
profile, the cobalt oxide is completely reduced at this
temperature. In this case, the selectivity to bicyclic
hydrocarbons and o-cresol significantly decreased whereas
the formation of cyclohexanone increased, which agrees
very well with the proposed reaction pathway. Ru/SiO2
catalyst exhibits a high selectivity to methane (62.3 %) and
C –C hydrocarbons (8.7 %).
2
2 3
the ring is hydrogenated and cyclohexanone and cyclo-
hexanol are the main products formed. On the other hand,
the hydrogenation of the carbonyl group of the keto-in-
termediate tautomer leads to the formation of benzene via
rapid dehydration of the unstable cyclohexadienol inter-
5
6
Pt/SiO and Pd/SiO catalysts were also tested for HDO
2
mediate. This is observed in the case of Pd/ZrO catalyst.
2
2
of phenol at 623 K (Table 3). Increasing the reaction
temperature significantly increased the formation of ben-
zene and decreased the selectivity to cyclohexanone and
cyclohexanol.
These results demonstrated that the selectivity for HDO of
phenol can be controlled by using supports of varying
oxophilicity.
In the present work, the results obtained revealed that
the type of the metal also significantly affected product
distribution for HDO of phenol. Comparing the product
selectitivities obtained for HDO of phenol at 573 K over
silica supported catalysts, it is noticed the presence of two
different groups of catalysts. For Pt, Pd and Rh based
catalysts, the hydrogenation products (ONE and OL) were
mainly formed. In order to enhance the deoxygenation
activity of Pd and Pt-based catalysts, the HDO of phenol
was carried out at 623 K. Increasing the reaction temper-
ature significantly increased the formation of benzene and
decreased the selectivity to cyclohexanone and cyclohex-
anol (Table 3). According to our previous work on Pd/SiO2
catalyst, phenol is tautomerized to 2,4-cyclohexadienone
3
.3 Proposed Reaction Pathway
for Hydrodeoxygenation of Phenol
The reaction mechanism for HDO of phenol has been
extensively studied in the literature. Recently, we investi-
gated the performance of Pd catalysts supported on SiO2,
Al O and ZrO for the HDO of phenol in gas phase, at
2
3
2
5
73 K and 1 atm [12]. Pd supported on SiO and Al O
2 2 3
exhibited high selectivity to cyclohexanone, whereas Pd
supported on an oxophilic support such as ZrO favored the
2
selectivity towards benzene, reducing the formation of
Table 3 Products distribution
for HDO of phenol at 623 K and
atmospheric pressure
Catalysts
Conversion (%)
Selectivity (%)
Benzene
ONE
OOL
ANE-ENE
C –C6
5
Pt/SiO
2
9.2
4.8
24.8
44.6
61.6
52.7
8.0
2.7
1.8
–
3.4
0.0
Pd/SiO
2
1
23

The Effect of Metal Type on Hydrodeoxygenation of Phenol…
intermediate [12], followed by the hydrogenation of the
ring, producing cyclohexanone and then cyclohexanol. For
these catalysts, the sequential hydrogenation of the aro-
matic ring on metal sites followed by dehydration of
cyclohexanol to benzene on acid sites cannot occur because
silica does not exhibit sufficient acidity to catalyze the
dehydration of cyclohexanol.
cyclohexadienol. The reaction pathway depends on the
nature of metallic phase. The presence of an oxophilic
metal, such as the partially reduced Fe species present in
the NiFe/SiO catalysts promoted the hydrogenation of the
2
carbonyl group of the tautomer intermediate. On the other
hand, instead of the tautomerization route, the direct
dehydroxylation of m-cresol is favored over Ru based
catalyst. In this case, DFT calculations showed a higher
reaction energy and energy barrier for the hydrogenation of
the O in the carbonyl group of the keto tautomer due to a
stronger interaction of the O with Ru than with Pt. It was
proposed that the direct dehydroxylation produces a par-
tially unsaturated hydrocarbon species, which may be
hydrogenated to toluene or may undergo C–C bond
breaking, producing C –C hydrocarbons. Therefore, the
On the other hand, the more oxophilic metals such as
Ru, Co and Ni promoted the formation of C–C bond
hydrogenolysis products such as methane and C –C
5
6
hydrocarbons. The formation of hydrogenolysis products
for HDO of phenolic compounds has been scarcely
reported in the literature [21, 38], which is likely due to the
fact that the majority of the works were performed in liquid
phase.
1
5
Chen et al. [37] studied the HDO of m-cresol in gas
phase at different temperatures over Pt/SiO , Pd/SiO and
oxophilicity of the metal surface can determine the reaction
pathway. For oxophilic metals such as Ru and Fe, the
deoxygenation is promoted whereas the hydrogenation is
favored over less oxophilic metals such as Pt, Pd or Ni. In
addition, this reaction mechanism may also explain the
formation of hydrogenolysis products for HDO of phenolic
compounds on some catalysts.
2
2
Ni/SiO catalysts. At 573 K, toluene and 3-methylcyclo-
2
hexanone were the only products formed over Pt- and Pd-
based catalysts whereas a significant formation of methane
and phenol was also observed for Ni/SiO catalyst. They
2
proposed that the direct deoxygenation to toluene and the
hydrogenation to methylcyclohexanone or methylcyclo-
hexanol are the main reactions over all catalysts but
products distribution varies depending on the metal.
According to them, the oxygen removal occurs through an
apparent direct hydrodeoxygenation that might involve the
direct hydrogenolysis of the C–O bond or the tautomer-
ization route. They ruled out the hydrogenation/deoxy-
genation route since the support has not enough acidity to
catalyze dehydration of 3-methylcyclohexanol. For Ni-
based catalyst, the hydrogenolysis of the C–C bond of the
methyl group of the m-cresol molecule produces methane
and phenol. However, they did not explain the reasons for
According to this mechanism, the stabilization of the
tautomer intermediate on the surface of these metals is
fundamental in determining the reaction pathway. A
higher stabilization of the tautomer intermediate on the
metal particle increases the energy barrier for the hydro-
genation of the carbonyl group, favoring the direct
deoxygenation. Therefore, the deoxygenation activity
depends on the affinity of the metals to bond with oxygen
from carbonyl group, as it was also proposed by Mor-
tensen et al. [6].
Mortensen et al. [6] studied the HDO of phenol in liquid
phase over Ru/C, Pd/C and Pt/C. Cyclohexane was the
main product formed over Ru/C catalyst whereas Pd/C and
Pt/C produced mainly cyclohexanol. They reported the
following order of activity for deoxygenation for HDO of
phenol: Ru/C [ Pd/C [ Pt/C. Based on the work of
Norskov et al. [38], the authors proposed a correlation
between the affinity of the metals to bind oxygen and their
deoxygenation activity. Therefore, the best performance to
deoxygenation of Ru-based catalyst was due to its strongest
binding energy with respect to oxygen than Pt or Pd.
In order to investigate the relationship between the
deoxygenation activity and the affinity of the metals to
bond with oxygen from carbonyl group, we took into
account the work of Lee et al. [39]. They studied the
hydrogenation of acetaldehyde to ethanol in liquid phase
over different metals (Pd, Pt, Rh, Ru, Ni, Co). Acetalde-
hyde may be bonded to the metal surface either through the
the higher hydrogenolysis activity on Ni/SiO catalyst.
2
A comprehensive study was performed about the HDO
of m-cresol in gas phase at 573 K over different silica
supported catalysts (Pd, Pt, Ru, Ni, Fe, NiFe) [10, 11, 21].
Hydrogenation products (3-methylcyclohexanone and
3
-methylcyclohexanol) were the main products formed
over Pt/SiO , Pd/SiO and Ni/SiO2 catalysts whereas
2
2
toluene was the dominant product on Fe and Ni–Fe
bimetallic catalysts. Ru/SiO catalyst showed significant
2
formation of toluene as well as CH and C –C hydrocar-
4
2
6
bons. The differences in product distributions were
explained in terms of a complex reaction pathway that
depended on the type of metal. For Pt/SiO , Pd/SiO , Ni/
2
2
SiO , Fe/SiO , NiFe/SiO catalysts, the reaction mecha-
2
2
2
nism involves the formation of a keto-tautomer interme-
diate (3-methyl-3,5-cyclohexadienone). This tautomer can
undergo the hydrogenation of the ring, producing
1
oxygen atom (g (O) configuration) or via the oxygen and
2
3
-methylcyclohexanone, or the hydrogenation of the car-
carbon atoms of the carbonyl group (g (C,O) configura-
tion) [40]. The adsorption configuration depends on the
bonyl group, leading the formation of 3-methyl-3,5-
123

C. A. Teles et al.
atomic oxygen calculated by Lee et al. [39] and the results
are shown in Fig. 4. The selectivity to deoxygenated
products increased in the same order of binding energies of
atomic oxygen. These results clearly demonstrated that
stronger bind of oxygen from carbonyl group to metal,
favors the formation of deoxygenated products.
A similar relationship could be obtained by plotting the
selectivity to deoxygenated products for HDO of phenol
and the Me–O bond dissociation energies [41]. There is a
reasonable correlation, taking into account the estimated
errors (Fig. S1). In addition, the trend was similar to that
one obtained considering the data from Lee et al. [39].
However, the relationship between the binding energy of
oxygen from the carbonyl group and the selectivity to
deoxygenated products seems to represent more effectively
our proposal, which is based on the adsorption of oxygen
atom from the carbonyl group of the tautomer formed (2,4-
cyclohexadienone). This is likely the reason for the best
fitting of the selectivity to deoxygenated products of our
work with the data reported by Lee et al. [39].
Fig. 4 Selectivity to deoxygenated products as a function of the
binding energies of atomic oxygen calculated by Lee et al. [39]
type of the metal. Therefore, they used DFT to calculate
the binding energies of atomic C and O on the different
metal surfaces. Oxygen binds more strongly to Ru than the
other metals. The following binding energies of atomic
Therefore, a complete HDO of phenol reaction pathways
over silica supported metal catalysts is proposed in
Scheme 1, representing the main reaction route for each
metal.
oxygen
were
-5.58) \ Ni \ (-5.40) \ Rh
-4.48) \ Pt (-4.20).
The selectivity to deoxygenated products obtained in our
observed:
Ru
(-5.90) \ Co
(-5.04) \ Pd
(
(
Besides the effect of the type of the metal, it could be
argued that the changes in product distribution are due to
the different metal dispersion of the catalysts studied in this
work was plotted as a function of the binding energies of
Scheme 1 Reaction pathways
for HDO of phenol over silica
supported metal catalysts
1
23

The Effect of Metal Type on Hydrodeoxygenation of Phenol…
Table 4 Products distribution
for HDO of phenol at 573 K and
atmospheric pressure
Catalysts
Dispersion (%)
Conversion (%)
Selectivity (%)
Benzene
ONE
OL
CH4
C -C6
5
Ni/SiO
Ni/SiO
2
2
5.4
9.0
10.1
12.0
5.0
4.9
3.1
5.4
6.0
69.6
71.3
30.4
30.5
23.8
23.0
6.7
1.2
2.0
0.3
–
Ru/SiO
Ru/SiO
2
2
6.0
45.7
50.2
12.0
9.0
15.0
4.8
4.0
work. However, the effect of the metal dispersion for HDO
of phenol reaction is controversial in the literature
For Pt/SiO , Pd/SiO and Rh/SiO2 catalysts, phenol is
2 2
mainly tautomerized, followed by hydrogenation of the C–
C bond of the tautomer intermediate formed, producing
cyclohexanone and cycloexanol. By contrast, the direct
dehydroxylation of phenol followed by hydrogenolysis
might also occur over more oxophilic metals such as Ru,
Co and Ni. This is supported by a correlation between
deoxygenation activity and metals affinity to bond with
oxygen.
[
12, 25, 42]. Newman et al. [25] and Mortensen et al. [42]
reported an influence of metal particle size on the HDO of
phenol in liquid phase over Ru and Ni supported catalysts,
respectively. However, the effect of Pd particle size on the
product distribution for HDO of phenol in gas phase over
Pd/ZrO catalysts with different Pd dispersions was pre-
2
viously investigated [12]. Both catalysts exhibited similar
selectivities indicating that metal dispersion does not sig-
nificantly affect the product distribution in the HDO of
phenol.
Acknowledgments Camila A. Teles and Jerusa R. de Lima thank
CAPES (Coordena c¸ a˜ o de Aperfei c¸ oamento de Pessoal de Ensino
Superior) for the scholarship received.
Therefore, the HDO of phenol in gas phase at 573 K was
carried out over Ni/SiO and Ru/SiO catalysts with dif-
2
2
ferent metal dispersion in order to investigate the effect of
metal particle size on product distribution. The selectivity
obtained at low phenol conversion is listed in Table 4. For
References
1
2
3
. Wang Y, Wu J, Wang S (2013) RSC Adv 3:12635–12640
. Huber GW, Iborra S, Corma A (2006) Chem Rev 106:4044–4098
. Mortensen PM, Grunwaldt JD, Jensen PA, Knudsen KG, Jensen
AD (2011) Appl Catal A 407:1–19
Ni/SiO catalyst, decreasing Ni dispersion from 9 to 5 %
2
did not change the selectivity to all products. When Co/
SiO is compared with the Ni/SiO catalyst with the same
2
2
4
5
. Zanuttini MS, Dalla Costa BO, Querini CA, Peralta MA (2014)
Appl Catal A 482:352–361
. Massoth FE, Politzer P, Concha MC, Murray JS, Jakowski J,
Simons J (2006) J Phys Chem B 110:14283–14291
dispersion (9 %), it is clear that the differences in product
distribution are due to the type of the metal. Ru/SiO cat-
2
alyst exhibited a product distribution quite different from
the other catalysts. In particular, it was noticed a significant
formation of hydrogenolysis products such as methane and
C –C hydrocarbons. However, the selectivity to
6. Mortensen PM, Grunwaldt JD, Jensen PA, Jensen AD (2013)
ACS Catal 3:1774–1785
7
8
9
. Liu H, Jiang T, Han B, Liang S, Zhou Y (2009) Science
26:1250–1252
5
6
3
hydrogenolysis products was approximately the same for
. Velu S, Kapoor MP, Inagaki S, Suzuki K (2003) Appl Catal A
245:317–331
. Hong DY, Miller SJ, Agrawal PK, Jones CW (2010) Chem
Commun 46:1038–1040
both Ru/SiO catalysts with different metal dispersion. In
2
addition, Ru/SiO and Pd/SiO catalysts have the same
2
2
metal dispersion but the selectivities are quite different,
which demonstrated the key role of the type of the metal on
catalyst performance.
1
0. Nie L, De Souza PM, Noronha FB, An W, Sooknoi T, Resasco
DE (2013) J Mol Catal A 388:47–55
11. De Souza PM, Nie L, Borges LEP, Noronha FB, Resasco DE
2014) Catal Lett 144:2005–2011
(
1
1
1
2. De Souza PM, Rabelo-Neto RC, Borges LEP, Jacobs J, Davis
BH, Sooknoi T, Resasco DE, Noronha FB (2015) ACS Catal
4
Conclusions
5
:1318–1329
3. De Souza PM, Rabelo-Neto RC, Borges LEP, Jacobs G, Davis
BH, Graham UM, Resasco DE, Noronha FB (2015) ACS Catal
The effect of metal type for the HDO of phenol in gas
phase was investigated over silica supported catalysts. Pt/
5
:7385–7398
4. Senol OI, Ryymin EM, Viljava TR, Krause AOI (2007) J Mol
Catal A 277:107–112
SiO , Pd/SiO and Rh/SiO catalysts favor the formation of
2
2
2
hydrogenated products (cyclohexanone and cycloexanol),
whereas Co/SiO , Ni/SiO and mainly Ru/SiO catalysts
15. Sun Q, Chen G, Wang H, Liu X, Han J, Ge Q, Zhu X (2016)
ChemCatChem 8:551–561
2
2
2
1
1
6. Foster AJ, Do PTM, Lobo RF (2012) Top Catal 55:118–128
7. Hensley AJR, Wang Y, McEwen JS (2015) ACS Catal 5:523–536
exhibits significant formation of hydrogenolysis products
C –C hydrocarbons and methane). Two different reaction
(
5
6
18. Zhao C, Kasakov S, He J, Lercher JA (2012) J Catal 296:12–23
19. Furimsky E (2000) Appl Catal A 199:147–190
pathways take place depending on the type of the metal.
123

C. A. Teles et al.
2
2
0. Nie L, Resasco DE (2014) J Catal 317:22–29
1. Tan Q, Wang G, Nie L, Dinse A, Buda C, Shabaker J, Resasco
DE (2015) ACS Catal 11:6271–6283
2. Echeandia S, Pawelec B, Barrio VL, Arias PL, Cambra JF,
Loricera CV, Fierro JLG (2014) Fuel 117:1061–1073
3. Zhao C, He J, Lemonidou AA, Li X, Lercher J (2011) J Catal
32. Zhang X, Wang T, Ma L, Zhang Q, Huang X, Yu Y (2013) Appl
Energy 112:533–538
33. Guvenatam B, Kursun O, Heeres EHJ, Pidko EA, Hensen EJM
(2013) Catal Today 233:83–91
34. Noronha FB, Perez CA, Schmal M, Frety R (1999) Phys Chem
Chem Phys 1:2861–2867
35. Nimmanwudipong T, Runnebaum RC, Tay K, Block DE, Gates
BC (2011) Catal Lett 141:1072–1078
36. Kallury RKMR, Tidwell TT, Boocock DGB, Chow DHL (1984)
Can J Chem 62:2540–2545
37. Chen C, Chen G, Yang F, Wang H, Han J, Ge Q, Zhu X (2015)
Chem Eng Sci 135:145–154
38. Norskov JK, Rossmeisl J, Logadottir A, Lindqvist L (2004) J
Phys Chem B 108:17886–17892
39. Lee J, Xu Y, Huber GW (2013) App. Catal. B: Environ.
140:98–107
2
2
2
2
2
80:8–16
4. Horacek J, St’ a´ vov a´ G, Kelbichov a´ V, Kubicka D (2013) Catal
Today 204:38–45
5. Newman C, Zhou X, Goundie B, Ghampson IT, Pollock RA,
Ross Z, Wheeler MC, Meulenberg R, Austin RN, Frederick BG
(
2014) Appl Catal A 477:64–74
6. Do PTM, Foster AJ, Chen J, Lobo RF (2012) Green Chem
4:1388–1398
7. Wan H, Chaudhari RV, Subramaniam B (2012) Top Catal
5:129–139
2
2
2
2
1
5
8. Zhu X, Nie L, Lobban LL, Mallinson RG, Resasco DE (2014)
Energy Fuels 28:4104–4111
9. Hong Y, Zhang H, Sun J, Ayman KM, Hensely AJR, Gu M,
Engelhard MH, McEwen J-S, Wang Y (2014) ACS Catal
40. Ferrin P, Simonetti D, Kandoi S, Kunkes E, Dumesic JA, Nors-
kov JK, Mavrikakis M (2009) J Am Chem Soc 131:5809–5815
41. Bond dissociation energies. https://labs.chem.ucsb.edu/zakarian/
armen/11—bonddissociationenergy.pdf. Accessed 13 July 2016
42. Mortensen PM, Grunwaldt JD, Jensen PA, Jensen AD (2016)
Catal Today 277:297
4
:3335–3345
3
3
0. Shin EJ, Keane MA (2000) Ind Eng Chem Res 39:883–892
1. Zhao C, Camaioni DM, Lercher JA (2012) J Catal 288:92–103
1
23