Influence of Re addition to Ni/SiO2 catalyst on the reaction network and deactivation during hydrodeoxygenation of m-cresol

Article abstract of DOI:10.1016/j.cattod.2018.04.073

The reaction network of hydrodeoxygenation of m-cresol was investigated over Ni/SiO₂ and Ni-Re/SiO₂ catalysts at 300 °C and 1 atm H₂. m-Cresol conversion proceeds through three major primary pathways: phenyl ring hydrogenation to 3-methylcyclohexanone and 3-methylcyclohexanol, C[sbnd]C hydrogenolysis to CH₄ and phenol, and deoxygenation to toluene. Re addition promotes dehydrogenation of the initially formed methylcyclohexanone and methylcyclohexanol to surface intermediates followed by deoxygenation toward formation of toluene. C[sbnd]C hydrogenolysis happens on unsaturated compounds with phenyl ring, instead of saturated compounds, since dehydrogenation is requested prior to C[sbnd]C cleavage. The turnover frequency (TOF) for CH₄ formation follows the order of m-cresol < toluene << benzene, indicating that C[sbnd]C hydrogenolysis is accelerated after deoxygenation and/or demethylation. Increasing reaction temperature promotes both deoxygenation and hydrogenolysis. The toluene selectivity keeps low even at higher temperature of 340 °C on monometallic Ni catalyst due to successive C[sbnd]C hydrogenolysis to CH₄. Re addition not only promotes deoxygenation to toluene by providing active Ni-Re neighboring sites, but also inhibits the aromatics hydrogenolysis to CH₄, resulting in high toluene selectivity on bimetallic catalyst. Both catalysts show a similar deactivation trend and coke is speculated to be the major reason for deactivation.

Full text of DOI:10.1016/j.cattod.2018.04.073

Catalysis Today xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Influence of Re addition to Ni/SiO catalyst on the reaction network and
2
deactivation during hydrodeoxygenation of m-cresol
Feifei Yang , Hua Wang , Jinyu Han , Qingfeng Ge , Xinli Zhua,⁎
a
a
a
a,b
a
Collaborative Innovation Center of Chemical Science and Engineering, Key Laboratory for Green Chemical Technology of Ministry of Education, School of Chemical
Engineering and Technology, Tianjin University, Tianjin, 300072, China
b
Department of Chemistry and Biochemistry, Southern Illinois University, Carbondale, IL, 62901, United States
A R T I C L E I N F O
A B S T R A C T
Keywords:
2 2
The reaction network of hydrodeoxygenation of m-cresol was investigated over Ni/SiO and Ni-Re/SiO catalysts
Hydrodeoxygenation
m-cresol
Ni-Re bimetallic catalyst
Reaction network
CeC hydrogenolysis
Coke
2
at 300 °C and 1 atm H . m-Cresol conversion proceeds through three major primary pathways: phenyl ring
hydrogenation to 3-methylcyclohexanone and 3-methylcyclohexanol, CeC hydrogenolysis to CH and phenol,
4
and deoxygenation to toluene. Re addition promotes dehydrogenation of the initially formed methylcyclohex-
anone and methylcyclohexanol to surface intermediates followed by deoxygenation toward formation of to-
luene. CeC hydrogenolysis happens on unsaturated compounds with phenyl ring, instead of saturated com-
pounds, since dehydrogenation is requested prior to CeC cleavage. The turnover frequency (TOF) for CH
4
formation follows the order of m-cresol < toluene < < benzene, indicating that CeC hydrogenolysis is ac-
celerated after deoxygenation and/or demethylation. Increasing reaction temperature promotes both deox-
ygenation and hydrogenolysis. The toluene selectivity keeps low even at higher temperature of 340 °C on
monometallic Ni catalyst due to successive CeC hydrogenolysis to CH
ygenation to toluene by providing active Ni-Re neighboring sites, but also inhibits the aromatics hydrogenolysis
to CH , resulting in high toluene selectivity on bimetallic catalyst. Both catalysts show a similar deactivation
trend and coke is speculated to be the major reason for deactivation.
4
. Re addition not only promotes deox-
4
1. Introduction
processing in terms of energy consumption. Processing liquid phenolics
requires high pressures and the condensation step after pyrolysis is
detrimental for heat recovery of the entire process [17]. Therefore,
vapor phase HDO over base metal catalyst, such as Ni, seems to be a
choice [18–20]. However, the low activity of Ni, the high tendency for
CeC hydrogenolysis to CH especially at relative high temperatures
4
(> 300 °C), and the low stability are the major impediments for future
applications [20–22].
Fast pyrolysis of lignocellulosic biomass produces bio-oil, which has
abundant of phenolic compounds stemmed from lignin fraction of
biomass. It remains a challenge for upgrading of the crude bio-oil to
transportation fuels due to the high oxygen content. The oxygen in
phenolics is mainly existed in some functional groups like hydroxyl
3
group (eOH) and methoxyl group (eOCH ), which are connected to a
phenyl ring. Catalytic hydrodeoxygenation (HDO) is an effective way to
remove oxygen from phenolics [1–6].
There have reported a variety of catalysts, such as NiMoS, Pt, Ru,
and Fe, for HDO of phenolics operated at different reaction conditions
Some pioneering researches have been devoted to improving the
vapor phase HDO performance on Ni-based catalysts [18–20,22–26].
Those works may be roughly divided into three aspects: adjusting the Ni
particle size [20,22,23]; optimizing the support (e.g. Ni/C, Ni/Al-MCM-
41) [25,26]; as well as adding an effective promoter (e.g. Fe, Re)
[19,24]. Combining Ni with an oxophilic metal, such as Fe, Re and Mo,
changes electronic structure and/or surface ensemble size, resulting in
greatly modified activity and selectivity [27]. For instance, Nie and
Resasco found that the reaction pathways on Ni-Fe bimetallic were
significantly different from those on pure Ni and pure Fe catalysts for
vapor phase conversion of m-cresol [19].
[
3,7]. Considering the economic benefits, it is important to develop
HDO catalysts with cheap active components [8–13]. On the other
hand, selective deoxygenation of phenolics to aromatics under mild
reaction conditions (low pressure and intermediate temperature) is
attractive with minimum hydrogen consumption [12–16]. These mild
reaction conditions are suitable for direct conversion of phenolics
produced from fast pyrolysis before their condensation to unstable li-
quids. The vapor phase HDO has an advantage over liquid phase
We recently reported that Ni-Re bimetallic catalyst is highly active
⁎
Corresponding author.
E-mail address: xinlizhu@tju.edu.cn (X. Zhu).
https://doi.org/10.1016/j.cattod.2018.04.073
Received 1 March 2018; Received in revised form 13 April 2018; Accepted 30 April 2018
0920-5861/ © 2018 Elsevier B.V. All rights reserved.
Please cite this article as: Yang, F., Catalysis Today (2018), https://doi.org/10.1016/j.cattod.2018.04.073

F. Yang et al.
Catalysis Today xxx (xxxx) xxx–xxx
Fig. 1. Effect of W/F on m-cresol conversion over 5Ni and 5Ni2.5Re (A), and major products distributions over 5Ni (B) and 5Ni2.5Re (C). Reaction conditions: T
300 °C, P = 1 atm, H /m-Cresol = 60, time on stream (TOS) = 0.5 h for each W/F.
=
2
for selective deoxygenation of m-cresol to toluene while greatly inhibits
CeC hydrogenolysis to CH at 300 °C and 1 atm H [24]. By detailed
2.2. Catalytic performance
4
2
characterization and density functional theory (DFT) calculation, we
found that Re addition reduces surface Ni ensemble size, which inhibits
the formation of the transition state for CeC hydrogenolysis. The in-
timate contact between Ni and Re provides an active deoxygenation site
with much lower activation energy. In addition, it facilitates desorption
of the deoxygenation product by lowering the electron density of Ni.
In this work, we investigated the conversion of possible inter-
mediates to establish the overall reaction network for m-cresol con-
The vapor-phase HDO of m-cresol was evaluated in a fixed-bed
quartz tube (6 mm o.d) reactor at atmospheric pressure, as detailed in
[24]. Briefly, in each run, the catalyst sample was charged at the center
of the reactor tube between two layers of quartz wool. The reaction
system was equipped with mass flow controllers for gases and a syringe
pump for the continuous injection of m-cresol or other liquid reactants.
All lines were heated at 230 °C to avoid any condensation. The catalyst
was reduced in-situ at 450 °C for 1 h in flowing H
perature was adjusted to the selected reaction temperature
(260–340 °C). The H /m-cresol molar ratio was kept at 60 for all runs.
2
, and then the tem-
version over 5 wt.% Ni/SiO
2 2
and 5 wt.% Ni-2.5 wt.% Re/SiO catalysts
at 300 °C and 1 atm H , and to understand the role of Re in controlling
2
2
product distributions. The results showed that aromatic compounds are
susceptible for the side reaction of CeC hydrogenolysis. Re addition
promotes selective deoxygenation to toluene, and also inhibits its fur-
The products were quantified online in a gas chromatograph (GC7890B,
Agilent), equipped with an Innowax capillary column (60 m) and a
flame ionized detector (FID). The effluent was trapped by methanol in
an ice-water bath and its components were identified by gas chroma-
tograph-mass spectrometer (GC–MS, Shimadzu QP2010s). A fresh cat-
alyst sample was used for each space time (W/F, weight of catalyst/
organic feed flow rate, gcat.h/gfeed (h)). The conversion and yield were
reported in molCarbon%.
ther hydrogenolysis to CH
4
, resulting in high deoxygenation to toluene
performance even at higher reaction temperature of 340 °C.
2. Experimental
2.1. Catalyst synthesis and characterization
3. Results and discussion
The catalysts used in this study were 5 wt.% Ni/SiO
2
(5Ni), 2.5 wt.%
3
3
.1. Reaction network for m-cresol HDO at 300 °C
Re/SiO
2
(2.5Re) and 5 wt.% Ni-2.5 wt.% Re/SiO (5Ni2.5Re).
2
Monometallic 5Ni and 2.5Re were prepared using an incipient wetness
impregnation method, and bimetallic 5Ni2.5Re was prepared by in-
cipient wetness co-impregnation method. The catalysts were char-
acterized by X-ray diffraction (XRD), Raman, Hydrogen temperature-
programmed reduction (H -TPR), high resolution transmission electron
2
microscopy (HRTEM), CO chemisorption, Fourier transform infrared
spectra of CO adsorption (CO-FTIR), X-ray photoelectron spectroscopy
.1.1. m-Cresol conversion at 300 °C
Fig. 1A shows m-cresol conversion as a function of W/F over 5Ni
and 5Ni2.5Re. To achieve ∼100% conversion, the W/F required for
Ni2.5Re (99%, 1.5 h) is ∼5 times shorter than that for 5Ni (93%, 7 h),
indicating that 5Ni2.5Re is much more active than 5Ni. As indicated in
24], there are five major primary products: toluene (Tol), 3-methyl-
cyclohexanone (MCHone), 3-methylcyclohexanol (MCHol), CH and
5
[
4
(
(
XPS), NH
3
and isopropylamine temperature-programmed desorption
phenol (Ph), corresponding to three parallel primary pathways: deox-
ygenation, hydrogenation and CeC hydrogenolysis (Scheme 1). Their
distributions over 5Ni and 5Ni2.5Re catalysts as a function of m-cresol
conversion are compared in Fig. 1B and C, respectively. Clear differ-
ences are evident: Tol yield is much higher but CH yield is much lower
4
on 5Ni2.5Re than those on 5Ni. This result indicates that Re addition
promotes selective deoxygenation to Tol and greatly inhibits CeC hy-
4
drogenolysis to CH .
The evolutions of minor products are shown in Fig. 2. These pro-
ducts include benzene (Ben), light hydrocarbons with carbon number
between 2–6 (LH), methylcyclohexane (MCHane), xylenols, p/o-cresol,
cyclohexanone (CHone) and cyclohexanol (CHol). Except for Ben and
NH -TPD and IPA-TPD), as detailed in the previous work [24]. The
3
dispersions of Ni measured by CO chemisorption are 1.55% and 4.47%
for 5Ni and 5Ni2.5Re, respectively. The acid densities measured by
NH
3
-TPD are 15.3 and 25.1 μmol/g for 5Ni and 5Ni2.5Re, respectively.
The coke deposition on the catalyst after reaction was evaluated by
temperature programmed oxidation (TPO) in a microreactor system
28], using flowing 5% O /He (30 mL/min) with a heating rate of
0 °C/min. The signals of H O (m/z = 18), CO (m/z = 44), and CO (m/
z = 28) were continuously monitored by a mass spectrometer (MKS).
[
1
2
2
2
2

F. Yang et al.
Catalysis Today xxx (xxxx) xxx–xxx
conversion, Re addition effectively promoted m-cresol conversion to
desirable deoxygenation products (increases from 54% to 74%) and
reduced C1-C6 (decreases from 38% to 16%).
3.1.2. MCHol and MCHone conversion
As suggested in Fig. 1, MCHone and MCHol appear to be inter-
mediate products. Their yields increase to a maximum value and then
decrease to zero. To verify this assumption, we compared the conver-
sion of MCHol and MCHone over 5Ni and 5Ni2.5Re, respectively
(
Fig. 4). They are either interconversion or dehydrogenation back to m-
cresol due to the lack of strong acid sites on both catalysts [21,24,28].
Note that their conversions are slowed down with increasing W/F. This
phenomenon may be ascribed to approaching the equilibrium limita-
tion between hydrogenation and dehydrogenation reactions [21,29].
The yield of deoxygenation product of Tol is rather low, indicating
deoxygenation is slower than hydrogenation/dehydrogenation reac-
tions and MCHone and MCHol are not the direct intermediates for Tol
formation. Fig. 5 compares the product distributions as a function of
conversion on these two catalysts. It is interesting to note that the yield
to m-cresol is higher on 5Ni2.5Re than that on 5Ni, for both MCHol
Scheme 1. Major reaction network of hydrodeoxygenation of m-cresol over 5Ni
and 5Ni2.5Re catalysts.
(
5
Fig. 5A) and MCHone (Fig. 5B) feeds. The results indicate that
Ni2.5Re has higher activity for dehydrogenation of MCHone/MCHol
than 5Ni. Recent theoretical calculation study [30] suggested that
partial hydrogenation of phenyl ring of phenol facilitates dehydrox-
ylation toward formation of benzene, while complete hydrogenation
products of cyclohexanol/cyclohexanone have higher barrier for de-
hydroxylation. Therefore, the enhanced dehydrogenation activity of
MCHone and MCHol on 5Ni2.5Re may facilitate the formation of such
partially hydrogenated intermediates on the catalyst surface. These
partially hydrogenated intermediates are readily dehydroxylated to-
ward formation of Tol. Indeed, the yield of Tol is higher on 5Ni2.5Re
than on 5Ni for both feeds (Fig. 5).
Fig. 2. Minor products distributions as a function of m-cresol conversion over
5Ni (A) and 5Ni2.5Re (B). Reaction conditions: T = 300 °C, P = 1 atm, H
2
/m-
Cresol = 60, TOS = 0.5 h for each W/F.
3
.1.3. The origin of CH
Note that even at very low m-cresol conversion level, the yield of
is similar to that of Ph (Fig. 1). The results imply that the molar
/Ph is about 6. Thus only a small portion of CH is formed
from demethylation (hydrogenolysis of Cphenyl-CH ) of m-cresol.
It is interesting to note that the yield of CH is rather low (< 0.5%)
4
LH, the others are rather low (< 2%) and therefore they are ignored for
further discussion.
To compare the overall HDO performances over 5Ni and 5Ni2.5Re,
the selectivity of desirable deoxygenation products (Tol, Ben and
CH
ratio of CH
4
4
4
3
MCHane), undesirable CH
4
and C2-C6 LH, and oxygenated products
4
(
MCHol, MCHone, CHone, CHol, Ph, p/o-cresol and xylenols) at three
during conversion of MCHone and MCHol conversion (Fig. 4), even at
higher conversion levels (> 60%). However, for m-cresol conversion
m-cresol conversion levels (4%, 48% and 94%) are reported in Fig. 3.
As expected, with the increasing of m-cresol conversion, the major
products changes from oxygenated type to deoxygenated type over both
catalysts. But the undesirable C1-C6 hydrocarbons become important
on 5Ni at high conversions. Compared to monometallic 5Ni at ∼94%
(Fig. 1), much higher yield of CH
CH is 1.9% even at a low conversion of m-cresol (11%) over 5Ni. Such
comparison may imply that phenyl ring or -CH connected to a phenyl
ring (unsaturated compounds with phenyl ring) is more susceptible for
4
is observed. For example, the yield of
4
3
4
CeC hydrogenolysis to CH than saturated products (MCHone, MCHol
and MCHane).
To verify that phenyl ring could produce CH
CHane), with elimination the effect of functional groups of eOH or
eCH , were fed as reactants individually to 5Ni and 5Ni2.5Re catalysts.
As shown in Fig. 6A and B, for feeding CHane, Ben is the major product
while CH is a minor one on both catalysts. In contrast, for feeding Ben,
CH (selectivity > 75%) is the dominating product while CHane is a
minor one (Fig. 6C and D). The results indicate that Ben is much more
susceptible than CHane for CeC hydrogenolysis to CH [31,32]. A
4
, Ben and cyclohexane
(
3
4
4
4
simple pseudo first order kinetic analysis could offer a clearer insight.
Table 1 shows the fitted first order reaction rate constants of 4 reac-
4 4
tions: CHane to Ben, CHane to CH , Ben to CH and Ben to CHane. On
both catalysts, k for CHane dehydrogenation to Ben is more than 20
1
times higher than k-1 for Ben hydrogenation to CHane, showing the
Fig. 3. Selectivity to different types of product at three m-cresol conversion
dominance of dehydrogenation over hydrogenation under current re-
levels (4%, 48% and 94%) over 5Ni and 5Ni2.5Re. Deoxygenation products
action conditions [21]. The k
significantly lower than k , and also lower than k
genolysis to CH . A reasonable explanation is that CHane is firstly de-
hydrogenated to Ben, and then Ben is hydrogenolyzed to CH , making
the k for apparent CHane hydrogenolysis to CH moderately lower
2
for CHane hydrogenolysis to CH
4
is
include Tol, Ben and MCHane; C1-C6 LH includes CH
Oxygenated products include MCHol, MCHone, CHone, CHol, Ph and xylenols.
Reaction conditions: T = 300 °C, P = 1 atm, H /m-Cresol = 60, TOS = 0.5 h,
adjusting m-cresol conversion level was achieved by varying W/F.
4
and C2-C6 LH;
for Ben hydro-
1
3
4
2
4
2
4
3

F. Yang et al.
Catalysis Today xxx (xxxx) xxx–xxx
Fig. 4. Product distributions of 3-methylcyclohexanol (MCHol) conversion over 5Ni (A) and 5Ni2.5Re (B); and 3-methylcyclohexanone (MCHone) conversion over
Ni (C) and 5Ni2.5Re (D) as a function of W/F. Reaction conditions: T = 300 °C, P = 1 atm, H /feed = 60, TOS = 0.5 h for each run.
5
2
Fig. 5. Product distributions as a function of 3-
methylcyclohexanol (MCHol) conversion over
Ni (open symbol and dash line) and 5Ni2.5Re
solid symbol and solid line) (A); Product dis-
5
(
tribution as a function of 3-methylcyclohex-
anone (MCHone) conversion over 5Ni (open
symbol and dash line) and 5Ni2.5Re (solid
symbol and solid line) (B). Reaction conditions
are the same as in Fig. 4.
than the k
with Desai et al, who showed that the actual rate of CeC hydrogenolysis
of CHane to CH was 0 over Ni/SiO with Ni particle size larger than
.5 nm [31]. Note that the dehydrogenation/hydrogenation activities
CHane to Ben or Ben to CHane) is nearly 4 times higher on 5Ni2.5Re
than those on 5Ni, while the CeC hydrogenolysis activity on the two
catalysts are similar. As a result, the CH yield on 5Ni2.5Re is much
3
for Ben hydrogenolysis to CH
4
. This is in good agreement
readily for hydrogenolysis to CH
Ph, Tol and Ben contribute to the formation of CH
drogenolysis of phenyl ring and demethylation (hydrogenolysis of
). To address which way is dominated, we compared the
formation turnover frequency (TOF) from different feeds (Ben, Tol
and m-cresol), as shown in Table 3. It should be noted that the hy-
drogenolysis of phenol is not studied here since it is a solid at room
temperature and therefore cannot be injected to the reactor directly
4
. Under current conditions, m-cresol,
4
by two ways: hy-
4
2
3
(
Cphenyl-CH
3
CH
4
4
lower than that on 5Ni for both Ben and CHane feeds (Fig. 6).
To investigate which molecule is more susceptible for demethyla-
without using solvent. Teles et al observed a CH
14.7% Ph conversion at 300 °C and 1 atm H [22], which suggests that
the phenyl ring of phenol is also readily for hydrogenolysis to CH . Over
both catalysts, the TOF for CH formation follows the order of m-
cresol < Tol < < Ben. To be specific, the values are 0.32, 0.079 and
4
selectivity of 7.2% at
tion to CH
as reactants (with elimination the effect of eOH group) individually to
follow the product distribution. As shown in Table 2, almost no CH is
4
, Tol, MCHane and methylcyclohexene (MCHene) were fed
2
4
4
4
detected for the feed of MCHene on either catalyst. MCHene is mainly
dehydrogenation to Tol or hydrogenation to MCHane, with the ratio of
−1
0.058 s for the feeds of Ben, Tol and m-cresol, respectively, on 5Ni. It
indicates that the rate determining step for Tol or m-cresol hydro-
dehydrogenation/hydrogenation of ∼5. It infers that the -CH
MCHene is not preferred for hydrogenolysis to CH . Furthermore, the
major product for the feed of MCHane is Tol, with small amount of CH
LH and Ben, consistent with the feed of CHane. However, for the feed of
Tol, the two major products are CH and Ben. It indicates that the
methyl group on aromatic ring is more susceptible than on saturated or
partial saturated cyclo-ring for demethylation to CH . This is expected
3
of
4
4
genolysis to CH is demethylation. The results indicate that CeC hy-
drogenolysis is accelerated after deoxygenation and/or demethylation.
In addition, based on the mechanism of CeC hydrogenolysis, one may
4
,
*
4
imagine that the partially unsaturated intermediates of C
7
H
7
and
*
C
6
H
4
OH species formed from deoxygenation and demethylation of m-
cresol could be even easier to be hydrogenolyzed to CH than the
corresponding Tol and Ph, due to free of H atom on one C atom in these
intermediates [36,37]. These factors resulted in the high yield of CH
even at low m-cresol conversion level (Fig. 1). We note that Re addition
significantly lowered CH formation TOF for all molecules (Table 3).
This is due to the reduced Ni ensemble size by Re dilution [24], which
4
4
since CeC hydrogenolysis requires completely losing all hydrogen
atoms on two adjacent C atoms to form three C-metal bonds for each C
atom before CeC cleavage [33–35]. The reaction network for the three
feeds conversion is summarized in Table 2.
4
4
As discussed above, it is clear that molecules with phenyl ring are
4

F. Yang et al.
Catalysis Today xxx (xxxx) xxx–xxx
Fig. 6. Product distributions of cyclohexane (CHane) conversion over 5Ni (A) and 5Ni2.5Re (B); and benzene (Ben) conversion over 5Ni (C) and 5Ni2.5Re (D).
Reaction conditions: T = 300 °C, P = 1 atm, H /feed = 60, TOS = 0.5 h for each run.
2
Table 1
Table 2
Reaction network for cyclohexane (CHane) and benzene (Ben) conversion and
the fitted pseudo first order reaction constants k
Product distributions of methylcyclohexene (MCHene), methylcyclohexane
(MCHane) and toluene (Tol) conversion over 5Ni and 5Ni2.5Re catalysts.
i
.
−1
h−1
)
Reaction
Fitted k
i
(mmolfeed
g
cat
5
Ni
5Ni2.5Re
Feed
MCHene
MCHane
Tol
k
k
k
k
1
CHane → Ben
4.9
1.8
2.8
0.2
17.8
2.6
3.8
Catalyst
W/F (h)
Conversion (%)
Product selectivity
5Ni
0.01
20.5
5Ni2.5Re
0.01
58.5
5Ni
0.2
10.2
5Ni2.5Re
0.2
22.1
5Ni
0.5
5.4
5Ni2.5Re
0.5
11.6
2
CHane → CH
Ben → CH
Ben → Chane
4
3
4
-1
0.8
(
CH
MCHane
Tol
Ben
%)
2
Reaction conditions: T = 300 °C, P = 1 atm, H /feed = 60, TOS = 0.5 h for
4
0.6
15.8
79.6
0
0.2
18.8
79.6
0
9.9
6.5
44.4
2.8
36.4
7.4
each run. The reaction data in Fig. 6 was fitted with the equation of -ln(1-
X) = k × (W/F), where X is the yield of certain product, W/F is space time, and
k is the rate constant.
79.4
10
0.7
86.5
5.5
1
47.2
2.4
46.7
6.4
C2-C6 LH
0.1
0.2
inhibits the formation of the transition state of CeC hydrogenolysis.
What’s more, the intimate interaction between Re and Ni reduce the d-
band occupancy of Ni [24], thus facilitating desorption of molecules
with phenyl ring (phenolics and aromatics).
Based on above studies, the major reaction network for m-cresol
HDO over 5Ni and 5Ni2.5Re at 300 °C is summarized in Scheme 1.
Reaction conditions: T = 300 °C, P = 1 atm, H
each run.
2
/feed = 60, TOS = 0.5 h for
product distributions for m-cresol conversion. Lower temperature fa-
vors for phenyl ring hydrogenation to MCHol and MCHone, while
higher temperature promotes deoxygenation to Tol, but also makes
CeC hydrogenolysis to CH
selectivity to CH is 60% at 45% m-cresol conversion at 400 °C on
monometallic 5Ni [21], leading to lowered yield of aromatics. It is of
4
dominant on 5Ni [21]. For instance, the
3
.2. Effect of reaction temperature on HDO of m-cresol
4
As reported, reaction temperature has significant influence on
5

F. Yang et al.
Catalysis Today xxx (xxxx) xxx–xxx
Table 3
Table 4
Turnover frequency (TOF) of CH
4
formation from different feeds over 5Ni and
(s−1)
m-Cresol conversion and product distributions over monometallic 2.5Re cata-
lyst at different reaction temperatures.
5Ni2.5Re catalysts.
Feed
TOF of CH
Ni
4
Temperature (°C)
250
1.9
300
5.7
350
7.2
5
5Ni2.5Re
Conversion (%)
Product yield (%)
Tol
MCHone
MCHol
Ben
Tol
m-Cresol
0.32
0.15
1.0
0.1
0
4.8
0
0
6.1
0
0
0.079
0.058
0.047
0.033
CH
Ph
4
0
0
0.8
0
0
0.9
0.2
0
0.8
Reaction conditions: T = 300 °C, P = 1 atm, H
each run. Conversion level was < 10% by adjusting W/F.
2
/feed = 60, TOS = 0.5 h for
Xylenols and p/o-cresol
2
Reaction conditions: P = 1 atm, H /m-Cresol = 60, W/F = 1 h, TOS = 0.5 h.
different products show similar trend as the evolution of selectivities
Fig. 8). The apparent TOF of Tol on 5Ni keeps unchanged with in-
creasing temperature as the formed Tol could further be hydro-
genolyzed to CH and Ben at high rate. Re addition not only promotes
(
4
m-cresol deoxygenation to Tol by providing active Ni-Re neighboring
site for deoxygenation; but also inhibits CeC hydrogenolysis by redu-
cing ensemble size of Ni. Therefore, the high deoxygenation to toluene
performance is achieved on 5Ni2.5Re even at higher reaction tem-
peratures.
It should be noted that increasing reaction temperature also pro-
motes deoxygenation of m-cresol to Tol over monometallic 2.5Re, as
shown in Table 4. But the activity is much lower than those on 5Ni and
5Ni2.5Re. It further supports that the high performance of 5Ni2.5Re is
attributed to the synergistic effect between Ni and Re.
Fig. 7. Effect of reaction temperature on major products selectivities during
hydrodeoxygenation of m-cresol over 5Ni (A) and 5Ni2.5Re (B). Reaction
conditions: P = 1 atm, H /m-Cresol = 60, TOS = 0.5 h, the m-cresol conver-
2
3.3. Catalyst stability and deactivation
sions were between 6% and 11% by adjusting W/F.
Fig. 9 shows the stability tests of 5Ni and 5Ni2.5Re during m-cresol
conversion. To obtain a comparable initial conversion level over the
two catalysts, the W/F was 5 h for 5Ni and it was 0.8 h for 5Ni2.5Re.
Both catalysts show a similar deactivation trend. The deactivation is
fast at the initial 5 h on stream, and is slowed down with further in-
creasing reaction time. The deactivation is either due to the sintering of
Ni particles or the coke blocking the active site. As suggested in pre-
vious work [20], sintering is not significant under current mild reaction
condition. What’s more, when we changed the conversion level over
both catalysts by either deactivating catalyst or varying W/F, as shown
in Fig. 10, the yields of the major products follow the their respective
trend line as a function of m-cresol conversion. This comparison
Fig. 8. Effect of reaction temperature on the turnover frequencies (TOFs) of
different products during hydrodeoxygenation of m-cresol over 5Ni (open
symbol and dash line) and 5Ni2.5Re (solid symbol and solid line) catalysts.
Reaction conditions: P = 1 atm, H /m-Cresol = 60, TOS = 0.5 h, m-cresol
2
conversion were between 6%∼11% by adjusting W/F.
2
interest to study the catalytic performance of Re promoted Ni/SiO at
different reaction temperatures. As shown in Fig. 7, on both catalysts,
the selectivity to MCHol and MCHone decrease sharply with increasing
temperature, as hydrogenation reaction is thermodynamically un-
favored at higher temperatures [21]. Significant difference between the
two catalysts is that Tol selectivity keeps increasing from 10% to 50%
with temperature increasing from 260 to 340 °C on 5Ni2.5Re. By con-
trast, on 5Ni, it increases slightly from 10% to 17% and then decreases
Fig. 9. Effect of time on stream on m-cresol conversion over 5Ni and 5Ni2.5Re
catalysts. Reaction conditions: T = 300 °C, P = 1 atm, H /m-Cresol = 60. The
2
to 10%. On the other hand, CH
4
increases sharply from 0 to 32% on
W/F was adjusted for each catalyst to achieve similar initial conversion (W/F
= 5 h for 5Ni, W/F = 0.8 h for 5Ni2.5Re).
5Ni, but mildly increases from 0 to 19% on 5Ni2.5Re. The TOFs of those
6

F. Yang et al.
Catalysis Today xxx (xxxx) xxx–xxx
Fig. 10. Major products yield as a function of m-cresol conversion over 5Ni (A) and 5Ni2.5Re (B). Different conversions were achieved by adjusting W/F (solid
symbols) or by deactivating catalyst (open symbols). The reaction conditions for deactivation were the same as Fig. 9; the data for W/F series are reproduced from
Fig. 1.
suggests that deactivation of catalysts reduces the numbers of active
sites without altering the reaction chemistry [4], and thus coke should
be the major reason for deactivation.
2.4% gcarbon/gcatalyst, respectively, which is consistent with stability
trend in Fig. 9. The slightly higher amount of coke on 5Ni2.5Re may be
related to its higher acid density (25.1 μmol/g for 5Ni2.5Re and
15.3 μmol/g for 5Ni). It should be noted that these two catalysts were
operated at very different space times. During the same period of re-
action, significantly more m-cresol feed was converted on per mass of
5Ni2.5Re than that on 5Ni. Therefore, much less coke is formed on
5Ni2.5Re than that on 5Ni in terms of the same amount of converted m-
cresol.
TPO of spent catalysts was performed to quantify the coke formed
during HDO of m-cresol. As shown in Fig. 11, 5Ni shows a major CO
2
evolution peak at 548 °C, while it shifts to lower temperature of 506 °C
on 5Ni2.5Re. The results imply that coke formed on 5Ni2.5Re is hy-
drogen richer than that on 5Ni [38]. There are two reasons may account
for this shift. First, Re addition reduces the d-band occupancy of Ni, and
thus reduces the adsorption energy of m-cresol, which facilitates des-
orption of coke precursors (aromatic compounds) and therefore retards
the formation of hydrogen poor coke. Second, 5Ni is highly active for
CeC hydrogenolysis. The carbon species on the Ni surface during CeC
hydrogenolysis may polymerize to graphite-like carbon (hydrogen poor
coke combusted at higher temperature) if it is not hydrogenated effi-
ciently [38]. Whereas on 5Ni2.5Re, deoxygenation to Tol is the major
4. Conclusion
The reaction network for m-cresol hydrodeoxygenation has been
investigated over 5Ni and 5Ni2.5Re at 300 °C and 1 atm H
mainly deoxygenation to toluene, hydrogenation to 3-methylcyclohex-
anone and 3-methylcyclohexanol, as well as hydrogenolysis to CH and
phenol under differential reaction conditions. CH is mainly derived
from the molecules with phenyl ring by demethylation and hydro-
genolysis of phenyl ring. The CH formation rates from different mo-
lecules follow the order of m-cresol < toluene < < benzene, which is
responsible for the much higher molar CH yield than phenol even at
2
. m-Cresol is
4
reaction and CeC hydrogenolysis to CH
4
is the minor one, leading to
4
reduced graphite-like coke. In addition, O atom of phenolics is readily
adsorbed on Re species [24]. These strongly adsorbed species may ac-
count for the hydrogen rich coke combusted at lower temperatures. The
quantified amounts of carbon on 5Ni and 5Ni2.5Re catalysts are 2.1 and
4
4
low m-cresol conversion level. Increasing reaction temperature pro-
motes m-cresol deoxygenation to toluene, but the toluene selectivity
and TOF keep low on monometallic 5Ni as toluene is readily further
hydrogenolyzed to CH
rate, promotes m-cresol deoxygenation to toluene and inhibits CeC
hydrogenolysis to CH , thus leading to high deoxygenation perfor-
4
. Re addition increases the intrinsic reaction
4
mance to toluene even at high reaction temperature. In addition, both
catalysts show a similar deactivation trend and coke is speculated to be
the major reason for deactivation.
Acknowledgements
The authors thank the support from the National Natural Science
Foundation of China (21676194 and 21576204) and the Ministry of
Education of China for Program of New Century Excellent Talents in
University (NCET-12-0407).
References
Fig. 11. TPO profiles of spent 5Ni and 5Ni2.5Re catalysts after HDO of m-
cresol. Reaction condition is shown in Fig. 9. Note that the profiles have been
normalized by weight of catalyst.
[1] P.M. Mortensen, J.D. Grunwaldt, P.A. Jensen, K.G. Knudsen, A.D. Jensen, Appl.
Catal. A 407 (2011) 1–19.
7

F. Yang et al.
Catalysis Today xxx (xxxx) xxx–xxx
[20] F. Yang, H. Wang, J. Han, Q. Ge, X. Zhu, Catal. Today (2018), http://dx.doi.org/10.
1016/j.cattod.2018.02.035 In press.
[21] C. Chen, G. Chen, F. Yang, H. Wang, J. Han, Q. Ge, X. Zhu, Chem. Eng. Sci. 135
(2015) 145–154.
[22] C.A. Teles, R.C. Rabelo Neto, J.R. de Lima, L.V. Mattos, D.E. Resasco, F.B. Noronha,
Catal. Lett. 146 (2016) 1848–1857.
[23] F. Yang, D. Liu, Y. Zhao, H. Wang, J. Han, Q. Ge, X. Zhu, ACS Catal. 8 (2018)
1672–1682.
[24] F. Yang, D. Liu, H. Wang, X. Liu, J. Han, Q. Ge, X. Zhu, J. Catal. 349 (2017) 84–97.
[25] Y. Yang, C. Ochoa-Hernández, V.A. de la Peña O’Shea, P. Pizarro, J.M. Coronado,
D.P. Serrano, Appl. Catal. B 145 (2014) 91–100.
[
[
2] A.V. Bridgwater, G.V.C. Peacocke, Renew. Sustain. Energy Rev. 4 (2000) 1–73.
3] M. Saidi, F. Samimi, D. Karimipourfard, T. Nimmanwudipong, B.C. Gates,
M.R. Rahimpour, Energy Environ. Sci. 7 (2014) 103–129.
[
[
4] X. Zhu, L.L. Lobban, R.G. Mallinson, D.E. Resasco, J. Catal. 281 (2011) 21–29.
5] A.J. Ragauskas, G.T. Beckham, M.J. Biddy, R. Chandra, F. Chen, M.F. Davis,
B.H. Davison, R.A. Dixon, P. Gilna, M. Keller, P. Langan, A.K. Naskar, J.N. Saddler,
T.J. Tschaplinski, G.A. Tuskan, C.E. Wyman, Science 344 (2014) 1246843.
6] J. Zakzeski, P.C.A. Bruijnincx, A.L. Jongerius, B.M. Weckhuysen, Chem. Rev. 110
[
[
(2010) 3552–3599.
7] Q. Bu, H. Lei, A.H. Zacher, L. Wang, S. Ren, J. Liang, Y. Wei, Y. Liu, J. Tang,
Q. Zhang, R. Ruan, Bioresour. Technol. 124 (2012) 470–477.
[
8] J.G. Dickinson, P.E. Savage, ACS Catal. 4 (2014) 2605–2615.
[26] N.T.T. Tran, Y. Uemura, S. Chowdhury, A. Ramli, Appl. Catal. A 512 (2016)
93–100.
[9] P.M. Mortensen, J.D. Grunwaldt, P.A. Jensen, A.D. Jensen, ACS Catal. 3 (2013)
1
774–1785.
[27] D. Hibbitts, Q. Tan, M. Neurock, J. Catal. 315 (2014) 48–58.
[28] Q. Sun, G. Chen, H. Wang, X. Liu, J. Han, Q. Ge, X. Zhu, ChemCatChem 8 (2016)
551–561.
[
[
[
10] C. Zhao, S. Kasakov, J. He, J.A. Lercher, J. Catal. 296 (2012) 12–23.
11] J.S. Moon, E.G. Kim, Y.K. Lee, J. Catal. 311 (2014) 144–152.
12] T. Prasomsri, T. Nimmanwudipong, Y. Román-Leshkov, Energy Environ. Sci. 6
[29] L. Nie, D.E. Resasco, J. Catal. 317 (2014) 22–29.
(2013) 1732–1738.
[30] D. Liu, G. Li, F. Yang, H. Wang, J. Han, X. Zhu, Q. Ge, J. Phys. Chem. C 121 (2017)
[13] R.N. Olcese, M. Bettahar, D. Petitjean, B. Malaman, F. Giovanella, A. Dufour, Appl.
12249–12260.
Catal. B 115-116 (2012) 63–73.
[31] P.H. Desai, J.T. Richardson, J. Catal. 98 (1986) 392–400.
[32] W.B.V. Scheppingen, M.L. Cieplik, R. Louw, Eur. J. Org. Chem. (2001) 2101–2106.
[33] D.D. Hibbitts, D.W. Flaherty, E. Iglesia, ACS Catal. 6 (2016) 469–482.
[34] Z.J. Zhao, L.V. Moskaleva, N. Rösch, J. Catal. 285 (2012) 124–133.
[35] D.W. Flaherty, D.D. Hibbitts, E. Iglesia, J. Am. Chem. Soc. 136 (2014) 9664–9676.
[36] Q. Tan, G. Wang, L. Nie, A. Dinse, C. Buda, J. Shabaker, D.E. Resasco, ACS Catal. 5
(2015) 6271–6283.
[37] C.J. Weststrate, J.W. Bakker, A.C. Gluhoi, W. Ludwig, B.E. Nieuwenhuys, Surf. Sci.
601 (2007) 748–756.
[38] C.H. Bartholomew, Catal. Rev. Sci. Eng. 24 (1982) 67–112.
[
[
14] J.C. Hicks, J. Phys. Chem. Lett. 2 (2011) 2280–2287.
15] J. Sun, A.M. Karim, H. Zhang, L. Kovarik, X.S. Li, A.J. Hensley, J.S. McEwen,
Y. Wang, J. Catal. 306 (2013) 47–57.
[
16] X. Zhu, L. Nie, L.L. Lobban, R.G. Mallinson, D.E. Resasco, Energy Fuels 28 (2014)
4
104–4111.
[
[
[
17] G.W. Huber, S. Lborra, A. Corma, Chem. Rev. 106 (2006) 4044–4098.
18] E.J. Shin, M.A. Keane, Ind. Eng. Chem. Res. 39 (2000) 883–892.
19] L. Nie, P.M. de Souza, F.B. Noronha, W. An, T. Sooknoi, D.E. Resasco, J. Mol. Catal.
A 388-389 (2014) 47–55.
8