Effective hydrodeoxygenation of lignin-derived phenols using bimetallic RuRe catalysts: Effect of carbon supports

Article abstract of DOI:10.1016/j.cattod.2017.07.027

We have previously shown that an activated carbon-supported ruthenium catalyst promoted with ReO_x (RuRe/AC) is highly active for the hydrodeoxygenation (HDO) of lignin-derived phenols (e.g., guaiacol). In this work, we have investigated the effect of carbon supports on the structure and HDO activity of bimetallic RuRe particles using three different carbon supports, i.e., activated carbon (AC), carbon black (Vulcan carbon, VC), multi-walled carbon nanotube (MWCNT). The MWCNT- and VC-supported catalysts show remarkably enhanced activity and hydrocarbon selectivity for the HDO of a range of phenolic molecules (i.e., guaiacol, eugenol, benzyl phenyl ether) compared to RuRe/AC. STEM-EDS and XPS analyses reveal that bimetallic RuRe particles are more common than monometallic Ru or Re particles in the VC- and MWCNT-supported catalysts, and hexavalent rhenium species are more easily reduced to tetravalent rhenium during the HDO reactions in these catalysts, suggesting that Ru and Re in close proximity are required for the efficient hydrogenolysis of phenols. The formation of bimetallic particles on the AC surface is likely hindered by high microporosity and high surface oxygen functionalities, both of which restrict the mobility of Re and Ru for assembly.

Full text of DOI:10.1016/j.cattod.2017.07.027

Accepted Manuscript
Title: Effective hydrodeoxygenation of lignin-derived phenols
using bimetallic RuRe catalysts: Effect of carbon supports
Authors: Kyung Bin Jung, Jinho Lee, Jeong-Myeong Ha,
Hyunjoo Lee, Dong Jin Suh, Chul-Ho Jun, Jungho Jae
PII:
S0920-5861(17)30519-9
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.cattod.2017.07.027
CATTOD 10941
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
12-6-2017
26-7-2017
29-7-2017
Please cite this article as: Kyung Bin Jung, Jinho Lee, Jeong-Myeong Ha, Hyunjoo
Lee, Dong Jin Suh, Chul-Ho Jun, Jungho Jae, Effective hydrodeoxygenation of lignin-
derived phenols using bimetallic RuRe catalysts: Effect of carbon supports, Catalysis
Todayhttp://dx.doi.org/10.1016/j.cattod.2017.07.027
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Effective hydrodeoxygenation of lignin-derived phenols using
bimetallic RuRe catalysts: Effect of carbon supports
Kyung Bin Junga,b, Jinho Leea,c, Jeong-Myeong Haa,d, Hyunjoo Leea,d, Dong Jin Suha,d,
Chul-Ho Junb, Jungho Jaea,d*
aClean Energy Research Center, Korea Institute of Science and Technology, Seoul 02792,
Republic of Korea
bDepartment of Chemistry, Yonsei University, Seoul 02841, Republic of Korea
cDepartment of Chemical Engineering, Hanyang University, Seoul 04763, Republic of Korea
dDivision of Energy & Environment Technology, KIST School, Korea University of Science
and Technology, Seoul 02792, Republic of Korea
*Corresponding author: Tel.: +82 2 958 5247; fax: +82 2 958 5209.
E-mail address: jjae@kist.re.kr (J. Jae).

Graphical abstract
Highlights
 Effect of carbon supports on HDO activity of RuRe bimetallic catalysts is
investigated.
 RuRe/MWCNT displays considerably higher activity for the HDO of phenols
than RuRe/AC.
 Ru and Re metals are mostly segregated on the surface of the low activity
RuRe/AC.
 High surface oxygen content of the AC support hinders the formation of
bimetallic particles.
 Ru and Re in close proximity are required for the efficient hydrogenolysis of
phenols.

Abstract
We have previously shown that an activated carbon-supported ruthenium catalyst promoted
with ReOx (RuRe/AC) is highly active for the hydrodeoxygenation (HDO) of lignin-derived
phenols (e.g., guaiacol). In this work, we have investigated the effect of carbon supports on
the structure and HDO activity of bimetallic RuRe particles using three different carbon
supports, i.e., activated carbon (AC), carbon black (Vulcan carbon, VC), multi-walled carbon
nanotube (MWCNT). The MWCNT- and VC-supported catalysts show remarkably enhanced
activity and hydrocarbon selectivity for the HDO of a range of phenolic molecules (i.e.,
guaiacol, eugenol, benzyl phenyl ether) compared to RuRe/AC. STEM-EDS and XPS
analyses reveal that bimetallic RuRe particles are more common than monometallic Ru or Re
particles in the VC- and MWCNT-supported catalysts, and hexavalent rhenium species are
more easily reduced to tetravalent rhenium during the HDO reactions in these catalysts,
suggesting that Ru and Re in close proximity are required for the efficient hydrogenolysis of
phenols. The formation of bimetallic particles on the AC surface is likely hindered by high
microporosity and high surface oxygen functionalities, both of which restrict the mobility of
Re and Ru for assembly.
Keywords: Lignin, Hydrodeoxygenation, Bimetallic, RuRe, Carbon

1. Introduction
The pyrolysis of lignocellulosic biomass, i.e., the thermal decomposition of biomass in
the absence of air, is a promising method for the production of a renewable liquid fuel called
bio-oil or pyrolysis oil [1-3]. However, bio-oil cannot be used as a high-grade fuel such as
gasoline and diesel due to its low calorific value, high acidity, and thermal instability caused
by its high oxygen content [1]. The hydrodeoxygenation (HDO) of biomass pyrolysis oils is
considered to be the most effective method for improving the quality of bio-oil by selectively
removing oxygen functionalities [4]. In this regard, research into the development of highly
active and stable catalysts for the HDO of pyrolysis oils has rapidly expanded in the last
decade [5]. Due to the diversity of molecules contained in the pyrolysis oils, lignin-derived
phenols (e.g., guaiacol, cresol, anisole) are commonly used as pyrolysis oil model compounds.
Specifically, guaiacol, 2-methoxyphenol, has been the most widely used reactant for the
design of catalysts to be used in the pyrolysis oil HDO process because it contains both
hydroxyl and methoxy groups, which are the representative functional groups in a pyrolysis
oil [5].
To date, the HDO of guaiacol has been studied with various catalysts, including noble
metals (Pt, Pd, Rh, Ru), transition metals (Ni, Fe), and metal sulfides (NiMoS) [5-10].
Although monometallic catalysts such as Pt and Ni show some HDO activity, the use of
metal-acid bifunctional catalysts (e.g., Rh/silica-alumina) or bimetallic catalysts (e.g., Pd-Fe,
Ni-Fe) was reported to enhance the deoxygenation of phenolic compounds remarkably due to
the synergy between the two active sites [11-18]. For instances, Lee et al. reported that the
use of BrØnsted acid catalysts such as silica-alumina as a support was essential for the
hydrogenolysis of the C‒OH bond in guaiacol over noble metal catalysts such as Rh and Pt

[11]. When Rh was supported on weakly acidic supports such as carbon, the major product
was a partially deoxygenated cyclohexanol, resulting from the C=C hydrogenation and
demethoxylation of guaiacol. However, the use of Rh supported on silica-alumina resulted in
the complete deoxygenation of guaiacol to cyclohexane due to the effective cleavage of the
C‒OH bond through acid-catalyzed dehydration followed by metal-catalyzed hydrogenation.
Several researchers have reported that this bifunctional pathway is more energetically
favorable than direct hydrogenolysis due to the high bond strength of Caromatic‒OH (468
kJ/mol) [19, 20]. Regarding bimetallic catalysts, Sun et al. reported that bimetallic PdFe/C
catalysts are significantly more active and selective toward the production of the
deoxygenated products such as benzene and toluene from guaiacol compared to
monometallic Pd/C [14]. The enhanced HDO activity of PdFe bimetallic catalysts was
explained in terms of the Pd-assisted H2 dissociation and stabilization of the metallic Fe
surface by Pd, which allows the direct deoxygenation of the C‒OH bond in phenol due to the
high oxophilicity of Fe. Similarly, Nie et al. reported that the bimetallic NiFe/SiO2 catalyst is
highly active for the HDO of m-cresol, 3-methylphenol, to toluene [16]. The modification of
the Ni surface by an oxophilic metal such as Fe induces repulsion onto the aromatic ring and
in turn results in the strong interaction with the carbonyl group, facilitating the formation of
toluene. In addition, our research group recently reported that the Re promotion of Ru/C
catalysts significantly enhances the rate of C‒O hydrogenolysis of guaiacol, resulting in the
complete deoxygenation to cyclohexane [21]. The high activity of the RuRe/C catalyst was
explained by the formation of acid sites from rhenium oxide species, while Ru provides metal
sites for hydrogenation. Re promoted noble metal catalysts (e.g., Pd) have also been shown to
be effective for the hydrogenation of various organic compounds such as amides and
carboxylic acids [22-25].
2

Overall, a combination of noble metals (e.g., Ru, Pd) with oxophilic promoters (e.g., Re,
Fe) is an effective approach for improving the HDO activity without the use of acidic
supports. Indeed, bimetallic catalysts have a significant advantage over metal-acid
bifunctional catalysts in terms of catalyst stability because sites of strong acidity such as
BrØnsted acid sites induce a higher rate of coke formation, resulting in the deactivation of the
catalyst [5]. Meanwhile, the effects of the support on the structure and HDO activity of these
bimetallic catalysts have rarely been investigated. Dwiatmoko et al. recently reported that the
physicochemical properties of carbon supports significantly affect the catalytic activity of the
monometallic Ru/C catalyst during the HDO of phenols by altering the Ru particle size and
the external surface area [10]. In this study, we investigated the effects of carbon supports on
the structure and HDO activity of bimetallic RuRe particles using three different carbon
supports, i.e., activated carbon (AC), carbon black (Vulcan carbon, VC), multi-walled carbon
nanotube (MWCNT).
2. Experimental
2.1. Materials
Guaiacol (99%) and eugenol (99%) were purchased from Sigma-Aldrich and benzyl
phenyl ether (BPE, 97%) and n-heptane (99%) were purchased from Alfa-Aesar. Metal
precursors of RuCl3∙xH2O and NH4ReO4 were purchased from Sigma-Aldrich. AC (Darco,
100 mesh particle size, powder) and MWCNT (carbon >98%, O.D. × L 6‒13 nm × 2.5‒20
μm) were purchased from Sigma-Aldrich and VC (Vulcan XC-72) was obtained from Cabot
Corp. (USA).
2.2. Catalyst preparation
3

Monometallic catalysts of 4 wt% Ru/C (C=AC, VC, MWCNT) and 3.64 wt% Re/C were
prepared by incipient wetness impregnation method. Prior to impregnation, the carbon
o
supports were sieved and dried in an oven at 105 C for 2 h. All supports were then
impregnated with aqueous solutions of RuCl3∙xH2O and NH4ReO4, respectively. The metal
impregnated carbons were dried in an oven at 105oC for 24 h. Bimetallic RuRe catalysts with
4 wt% Ru and 3.64 wt% Re content were prepared by successive impregnation of dried Ru/C
with an aqueous solution of NH4ReO4. This metal loading was corresponded to an atomic Ru :
Re ratio of 1 : 0.5, which was found to be the optimal composition of the RuRe catalyst in our
previous study [21]. Prior to use in experiments, the catalysts of Ru/C, Re/C, and RuRe/C
were dried in an oven at 105 ºC, reduced at 250 ºC for 3 h in a flowing mixture of H2 (5%)
and N2 (95%), and passivated at room temperature for 2 h in a flowing mixture of O2 (0.5%)
and N2 (99.5%).
Ru/AC catalysts promoted with Mo, W, and Fe were prepared by incipient wetness
impregnation of Ru/AC with aqueous solutions of H24Mo7N6O24, H42N10O42W12, and
Fe(NO3)3∙9H2O, respectively. All of the catalysts contained 4 wt% Ru with an atomic ratio of
Ru: M = 1 : 0.5 (M = Mo, W, Fe). After impregnation, the catalysts were dried in an oven at
105 ºC for 12 h, reduced at 250 ºC (for RuW/AC) and 300 ºC (for RuMo/AC and RuFe/AC)
for 3 h, and passivated at room temperature for 2 h.
2.3. Catalysts characterization
X-ray photoelectron spectroscopy (XPS) measurements were performed using a Theta
Probe AR-XPS System (Thermo Fisher Scientific, U.K) equipped with a Al Kα source
(1486.6 Ev). All binding energies were referenced to the C 1s peak at 284.6 eV. Prior to XPS
studies, the fresh RuRe/C catalysts were reduced at 250 ºC in flowing 5% H2 in N2, and
4

passivated in flowing 0.5% O2 in N2. The spent RuRe/C catalysts recovered through filtration
were directly subjected to XPS studies without any pretreatment. All samples were exposed
to air before measurements.
CO-chemisorption measurements were performed using a pulsed injection of 10%
CO/He with a BETCAT-M instrument (BEL JAPAN, Inc). The change in CO concentration
of the effluent gas was monitored using a thermal conductivity detector (TCD). Prior to
analysis, catalyst samples were reduced in flowing H2 for 2 h at 250 ºC. The sample was then
cooled to 50 ºC in flowing He and further treated at 50 ºC for 30 min to remove adsorbed H2.
The contents of Ru and Re in the carbon supports were determined by inductively
coupled plasma‒optical emission spectrometer (ICP‒OES) analysis (ICAP 6000 series,
Thermo, USA). Catalyst samples were pretreated in concentrated nitric acid at 100 °C for 24
h to leach the metals from the carbon supports. Temperature programmed reduction (TPR)
studies were carried out using a BELCAT-B instrument (BEL Japan, Inc). TPR was
conducted under 5% H2/Ar flow from 30 ºC to 900 ºC with a heating rate of 10 ºC /min.
Before TPR measurements, catalyst samples were pretreated under 5% O2/He flow at 130 ºC
for 1 h. The Brunauer-Emmett-Teller (BET) surface areas and porosity of the carbon supports
and catalysts were determined from the N2 adsorption-desorption isotherms obtained using a
BELSORP-mini II instrument (BEL Japan, Inc.). The micropore volume and average pore
diameter of the samples were calculated by t-plot and BJH methods, respectively.
In order to investigate the distribution of Ru and Re on the carbon support, scanning
transmission electron microscopy (STEM) analyses were performed with energy dispersive
X-ray (EDS) mapping using a Talos F200X TEM. Lattice morphology and particle size
distribution of the catalysts were examined by high-resolution TEM analyses using a FEI
Titan 80-300 TEM. The reduced and passivated catalysts were used for TEM studies. Prior to
5

analysis, catalyst samples were sonicated in ethanol for 30 min.
Temperature programmed desorption analyses of the carbon supports were conducted
using a Belcat-B instrument coupled with TCD and mass spectrometry (TPD-TCD/MS). All
samples were initially pretreated at 150 ºC under a He flow for 1 h. Subsequently, the TPD
measurements were performed under a He flow over the temperature range, 100‒900 ºC, with
gas detection through TCD and MS (CO (m/z = 28), CO2 (m/z = 44), and H2O (m/z = 18)).
2.4. Catalytic experiments
The reaction was carried out in a 150 mL batch autoclave equipped with a stirrer and an
electric furnace. For a typical reaction, n-heptane (24 mL), guaiacol (0.24 g), and catalyst
(0.05 g) were charged into the reactor, and the reactor was sealed. Next, the reactor was
purged with three times with N2 followed by pressurizing with H2 to 20 bar. The reactor was
then heated to the desired temperature (i.e., 200‒240 ºC) with a ramping rate of 10 ºC/min,
while the reaction mixture was stirred at 300 rpm with an impeller. The reactions were
performed for 1‒4 h. After the desired reaction time, the reactor was cooled using an air
cooling fan and the liquid products were collected, filtered, and analyzed. The identification
of liquid products was performed by using GC–MS (Agilent 7890/5975), and each product
was quantified using a gas chromatography (Agilent 7890B) equipped with a flame ionization
detector (FID) by creating calibration curves using standard chemicals. Conversion was
defined as the moles of the feed that reacted (determined from GC analysis) divided by the
initial moles of the feed loaded in the reactor. Product selectivity was defined as the moles of
the product divided by the moles of the feed that reacted. However, in the case of benzyl
phenyl ether reactant, product selectivity was calculated on a molar carbon basis (the moles
of carbon in the product divided by the moles of carbon in the feed that reacted).
6

3. Results and Discussion
3.1. Effect of oxophilic promoter on Ru/AC catalysts
Initial experiments focused on the investigation of the activity of the Ru/AC catalysts
promoted with several oxophilic metals (i.e., Re, W, Mo, Fe) for guaiacol conversion, as
shown in Table 1. For all of the catalysts, a M : Ru molar ratio (M: oxophilic metal) was fixed
to 0.5. The reactions were carried out at 200 ºC under 20 bar H2 in a short reaction time of 1 h.
The catalysis results showed that the HDO activity and selectivity of RuRe/AC are higher
than those of the Ru/AC catalysts promoted with Mo, Fe, and W. The conversion of guaiacol
decreased in the order RuRe (57.7%) > RuW (55.5%) > RuMo (37.4%) > RuFe (11.1%), and
the selectivity to the completely deoxygenated cyclohexane decreased in the order RuRe
(15.2%) >> RuW (8.8%) > RuFe (7.1%) > RuMo (5.5%). These results indicate that Re is the
most suitable oxophilic promoter for enhancing the HDO activity of Ru/AC catalyst. The
effect of Re/Ru ratio on the activity of RuRe/AC catalyst is also shown in Table 1. The
conversion of guaiacol increased with increasing Ru/Re ratio in the catalyst, but the
selectivity to cyclohexane decreased with increasing Ru/Re ratio. Overall, the RuRe/AC
catalyst with the Re : Ru atomic ratio of 0.5 showed the highest selectivity to cyclohexane
with moderate conversion of guaiacol. Thus, this atomic ratio was chosen for further study.
3.2. Catalytic performances of bimetallic RuRe catalysts supported on different carbons
The guaiacol conversion and product selectivity obtained over RuRe catalysts supported
on different carbons are displayed in Fig. 1. The reactions were carried out at 200 ºC under 20
bar H2 in a short reaction time of 1 h. The results showed that the HDO activity of
RuRe/MWCNT and RuRe/VC are remarkably high relative to that of RuRe/AC, despite the
7

identical loading amounts of Ru and Re in all of the catalysts. The conversion of guaiacol
over RuRe/AC was only 57.7%, while those over RuRe/MWCNT and RuRe/VC were 99.4%
and 99.6%, respectively. The selectivity to cyclohexane decreased in the order
RuRe/MWCNT (40.4%) > RuRe/VC (33.9%) > RuRe/AC (15.2%), indicating that the
hydrogenolysis of the C‒OH bond is more facile in the MWCNT and VC supports. The other
products detected were cyclohexanol, phenol, and 2-methoxycyclohexanol, which resulted
from the C=C hydrogenation and demethoxylation of guaiacol. Cyclohexanol showed the
highest selectivity in all of the catalysts, indicating that the cleavage of the C‒OH bond is the
slowest step in the overall HDO reaction.
Product yields as a function of time in the HDO of guaiacol are plotted in Fig. 2. The
reaction temperature was raised from 200 to 240 ºC to promote the deoxygenation reaction.
The RuRe/MWCNT displayed a significantly faster hydrogenolysis rate than that of
RuRe/AC; the complete deoxygenation of guaiacol to cyclohexane (~80% yield) was
achieved over RuRe/MWCNT within 1 h, whereas over RuRe/AC, it took more time (>4 h)
for the completion of the reaction. The RuRe/VC also showed a fast hydrogenolysis rate,
reaching almost complete conversion to cyclohexane within 1 h, but its rate was slightly
lower than that of RuRe/MWCNT. In addition to cyclohexane, C3-C6 alkanes such as
propane, butane, pentane, and hexane were also formed as side products during the course of
the reaction. These straight-chain alkanes may arise from the ring-opening of cyclohexanol or
cyclohexane followed by cracking. Compared to the reactions at 200 ºC, the yields of these
alkane products were significantly higher at the reaction temperature of 240 ºC, indicating
that a high temperature leads to the undesired ring-opening reaction. The most active catalyst
of RuRe/MWCNT showed the highest formation rate of alkanes, followed by RuRe/VC and
RuRe/AC.
8

3.3. Comparison of RuRe/MWCNT with its monometallic counterparts
Fig. 3 shows the catalysis results of Ru/MWCNT, Re/MWCNT, and their physical
mixture for the HDO of guaiacol. The conversion of guaiacol over Ru/MWCNT was high
(96.3%), but the selectivity to cyclohexane was only 8.3%, which is significantly lower than
that of RuRe/MWCNT (40.4%). This result suggest that the presence of Re is essential for the
effective hydrogenolysis of cyclohexanol to cyclohexane. Meanwhile, Re/MWCNT exhibited
very low and negligible HDO activity (<2% conversion), indicating that Re alone is not a
good HDO catalyst. When a physical mixture of Ru/MWCNT and Re/MWCNT was
employed as a catalyst, the conversion of guaiacol was 99.5% and the selectivity to
cyclohexane was increased slightly to 12.2%, which was significantly lower than that over
the bimetallic RuRe/MWCNT. This indicates that Ru and Re must be in close proximity to
maximize the rate of C-O hydrogenolysis for cyclohexanol. This is consistent with multiple
studies from other researchers who showed that noble metals such as Pd and Ru should be in
close proximity to Re sites for the efficient hydrogenolysis of biomass-derived polyols and
phenols [21, 26-28].
3.4. HDO of other phenolic derivatives over RuRe/MWCNT and RuRe/AC
The catalytic performances of the RuRe catalysts supported on MWCNT and AC were
also compared for the HDO of other phenolic compounds contained in lignin pyrolysis oils.
Table 2 shows the catalysis results of RuRe/MWCNT and RuRe/AC for the HDO of benzyl
phenyl ether (BPE) and eugenol at 200 ºC for 1 h. The conversion of BPE over
RuRe/MWCNT (99.7%) was much higher than that over RuRe/AC (9.7%). The selectivity to
9

the fully deoxygenated products of cyclohexane and methylcyclohexane was also higher over
RuRe/MWCNT compared to RuRe/AC. This result indicates that RuRe supported on
MWCNT is more active for the hydrogenolysis of the α-O-4 aryl-ether bond than RuRe/AC.
For the HDO of eugenol, RuRe/MWCNT also exhibited higher HDO activity than RuRe/AC;
the conversion of eugenol over RuRe/MWCNT and RuRe/AC were 99.4% and 31.1%,
respectively. In addition, RuRe/MWCNT exhibited considerably higher selectivity to fully
deoxygenated propylcyclohexane (63%) compared to RuRe/AC (11.1%), indicating that
RuRe/MWCNT is a more active hydrogenolysis catalyst for bulky molecular structures.
Overall, the RuRe catalyst supported on MWCNT demonstrates superior hydrogenolysis
activity to RuRe/AC for a range of lignin-derived phenolic compounds, including guaiacol.
3.5. Characterization of carbon supports and RuRe/C catalysts
Table 3 shows the textural properties of carbon supports and RuRe catalysts supported
on carbons as measured by N2-physisorption. AC had a BET surface area more than three
times higher than those of the MWCNT and VC supports due to its high microporosity; the
micropore volume of AC was 0.30 cm3/g, whereas those of MWCNT and VC were only
0.007 and 0.04 cm3/g, respectively. Meanwhile, MWCNT had a significantly larger mesopore
volume than those of the AC and VC supports; the mesopore volume of MWCNT was 3.21
cm3/g, whereas those of AC and VC were 0.50 and 0.56 cm3/g, respectively. The pore size
distribution calculated by the BJH method revealed that MWCNT had mesopores 10-100 nm
in size, while AC had a relatively narrow pore size distribution with an average mesopore
diameter of 3.2 nm. The high mesoporosity of MWCNT likely arose from the interparticle
spaces between the nanotubes. After metal impregnation, the surface areas and pore volumes
of the carbons decreased greatly due to the pore blockage and surface coverage by the metals.
The CO chemisorption results of the supported RuRe catalysts are also displayed in Table 3.
10

RuRe/AC exhibited the highest metal dispersion (30.6%) followed by RuRe/VC (11.4%) and
RuRe/MWCNT (8.6%).
Fig. 4 shows the H2-TPR profiles of Ru/AC, Re/AC, and RuRe catalysts supported on
different carbons. The TPR curves of Re/AC and Ru/AC showed a peak at 300 ºC and 90 ºC,
respectively, which can correspondingly be assigned to the reduction of rhenium oxide and
ruthenium oxide. Meanwhile, all of the RuRe catalysts supported on carbons showed a single
reduction peak at 150‒180 ºC, which is higher than that of Ru/AC but much lower than that
of Re/AC. This indicates that the presence of Re hinders the reduction of ruthenium oxide,
whereas the reducibility of the Re species is increased by the presence of Ru due to the close
contact between Ru and Re on the carbon support. Multiple studies have shown that in
bimetallic catalysts containing noble metals such as Ru and Pd, noble metals can increase the
reducibility of other transition metals such as Re due to the hydrogen transfer from their
reduced species [29, 30]. The peak temperature of the RuRe catalysts decreased in the order
of RuRe/AC (180 ºC) > RuRe/MWCNT (160 ºC) > RuRe/VC (150 ºC). This result suggests
that activated hydrogen on the Ru surface is more easily transferred to rhenium oxide in the
RuRe/VC and RuRe/MWCNT than RuRe/AC due to the closer proximity of Re to Ru. The
amounts of H2 consumption of the catalysts during the H2-TPR analysis are also shown in
Table S1. The amounts of H2 consumption of the RuRe catalysts decreased in the order of
RuRe/MWCNT > RuRe/VC > RuRe/AC, indicating that the degree of oxidation of the two
metals is highest in the RuRe/MWCNT.
Fig. 5 shows the XPS spectra of the Re 4f levels of the RuRe catalysts before and after
the reactions. Kim et al. showed that the active phase of the RuRe catalyst for the HDO of
guaiacol is rhenium oxide (ReOx), together with metallic Ru [21]. Multiple studies have
shown that ReOx can act as acid sites for promoting C‒OH bond cleavage through
11

dehydration (e.g., glycerol) [28, 31, 32]. The XPS results showed that the surface rhenium
species of all of the fresh RuRe catalysts exist as a mixture of ReO2, ReO3, and Re2O7.
Regarding the relative distribution of the ReOx species, ReO3 was the major species in all of
the catalysts and the fresh RuRe/AC contained less ReO2 compared to the RuRe/MWCNT
and RuRe/VC. After the reaction at 200 ºC, RuRe/MWCNT and RuRe/VC showed a large
increase in the amount of ReO2, while the valence of the Re species of the RuRe/AC
remained nearly unchanged. This result suggests that Ru and Re on the MWCNT and VC
supports are in closer proximity than those on the AC support, by which Re species are easily
reduced to +4. The increase in the amount of tetravalent Re after the reaction is especially
large in RuRe/MWCNT, the most active HDO catalyst, which suggests that the valance of Re
can be an important factor for the improved catalytic activity.
Fig. 6 shows the HR-TEM images and metal particle size distributions of RuRe/AC and
RuRe/MWCNT. The average particle sizes of RuRe/AC and RuRe/MWCNT were 1.1±0.3
nm and 3.0±0.6 nm, respectively. The STEM analysis of RuRe/VC revealed that the average
particle size of RuRe/VC was 2.1±0.4 nm (Fig. S1). The average metal particle size measured
by TEM decreased in the order of RuRe/MWCNT > RuRe/VC > RuRe/AC, which was
consistent with the trend of metal dispersion measured by CO chemisorption. The
compositions of the metal particles were also analyzed by STEM-EDS, and the number
distribution of the particle compositions for over 30 selected particles in the three catalysts
were compared, as shown in Fig. 7. All of the catalysts displayed broad composition
distributions with a large number of monometallic Ru nanoparticles. In the AC-support
catalyst, Ru-only particles were more common than RuRe bimetallic particles, suggesting
that Ru and Re component metals are mostly segregated on the surface of carbon. In contrast,
RuRe catalysts supported on VC and MWCNT contained a relatively large number of RuRe
12

bimetallic particles. The majority of metal particles were present in a bimetallic state,
especially for the VC-supported catalyst.
Table 4 shows the composition distributions of RuRe catalysts supported on three
carbons measured by ICP-OES, STEM-EDS, and XPS. The Ru : Re molar ratios of the
catalysts determined by ICP were nearly identical (1:0.5) and were very close to the design
target ratio (1:0.5). In contrast, the average Ru : Re EDS ratios of RuRe/MWCNT, RuRe/VC,
and RuRe/AC were 1:0.4, 1:0.23, and 1:0.18, which are significantly different from both their
ICP and target ratios. It should be noted that the EDS ratios in Table 4 are the average values
of all of the metal particles in the STEM images and can differ from those in Fig. 7, which
were obtained from certain larger particles. The difference between the EDS and ICP ratios is
likely caused by the tiny Re-rich particles, which are too small to be detected in the STEM,
below ca. 1 nm; the Ru : Re EDS ratio of RuRe/MWCNT with a larger particle size (1:0.4) is
closer to the ratio determined by the ICP analysis, and the Re content in the catalyst decreases
with a decrease in the metal particle size. Overall, this result suggests that Re is mainly
segregated from Ru on the AC support and exists in the form of very small particles.
In addition, the average surface Ru : Re ratios determined by XPS differed greatly
among the catalysts (see Table 4). The surface Ru : Re ratio decreased in the order of
RuRe/MWCNT (1:1.6) > RuRe/VC (1.1) > RuRe/AC (1:0.75). This result indicates that Re is
more enriched on the surface of RuRe/MWCNT than on those of RuRe/VC and RuRe/AC.
Moreover, the surface Re/Ru ratio was directly correlated with the HDO activity trend; the
selectivity toward cyclohexane in the HDO of guaiacol decreased in the order of
RuRe/MWCNT (41.0%) > RuRe/VC (32.6%) > RuRe/AC (16.2%). This result suggests that
Re is the main active site in the hydrogenolysis of the C-OH bond of guaiacol.
13

3.6. Reusability of RuRe/MWCNT and RuRe/VC
The reusability of the high activity RuRe/MWCNT and RuRe/VC catalysts was also
investigated. For this purpose, the spent catalysts from the reactions of guaiacol were
recovered through filtration and reused under the same reaction conditions. Overall, the
recycled catalysts showed the comparable HDO activity to the fresh catalysts (Table S2).
Interestingly, the recycled RuRe/MWCNT even displayed a higher selectivity to cyclohexane
than the fresh RuRe/MWCNT (55.4% vs 40.4%). Characterization of the spent catalysts by
CO chemisorption revealed that the metal dispersion of the RuRe/MWCNT increased
significantly from 8.6% to 26% after the reaction, whereas that of the RuRe/VC remained
relatively unchanged (Table S2). The STEM-EDS analysis of the spent RuRe/MWCNT also
revealed that the bimetallic RuRe particles were more common in the spent catalyst
compared to the fresh catalyst (Fig. S2). Taken together, this indicates that metal migration
occurs in the RuRe/MWCNT catalyst during the reaction, leading to better dispersion and
formation of bimetallic RuRe particles. In addition, ICP analyses of the product solutions
confirmed that no leaching of Re occurred during the reactions. Therefore, these results
demonstrate that the RuRe catalysts supported on MWCNT and VC are reusable.
3.7. Discussion
Bimetallic RuRe catalysts exhibited enhanced activity and selectivity for the HDO of
lignin-derived phenolic molecules when Ru and Re were loaded onto MWCNT and VC
supports. Extensive characterization of the carbon supports and the bimetallic catalysts
suggests that carbon supports affect the HDO activity of bimetallic RuRe catalysts in two
ways. First, the presence of micropores in the AC support can make the bulky phenolic
reactants less accessible to the metal sites, leading to lower activity (i.e., conversion) of
14

RuRe/AC. As shown in Table 3, the ratio of the micropore to the mesopore volumes of AC is
much higher than those of the MWCNT and VC supports. Although the higher surface area of
AC induced by micropores resulted in the higher metal dispersion of RuRe/AC (30%), the
metal sites located in the micropores would be inaccessible for the reactants, thus suppressing
the catalytic activity. Moreover, the activity differences between RuRe/AC and
RuRe/MWCNT were more significant in the HDO of the bulky reactants, such as BPE; the
BPE conversion over RuRe/AC and RuRe/MWCNT were 9.7% and 99.7%, respectively. This
result clearly indicates that the activity differences are linked to the differences in the pore
structure of the carbon supports (i.e., pore diffusivity).
Second, the carbon supports affect the formation of bimetallic particles, which is
necessary for the efficient hydrogenolysis of the C-OH bond of guaiacol. Characterization of
the catalysts by STEM-EDS revealed that bimetallic RuRe particles are more common than
monometallic Ru or Re particles in the VC- and MWCNT-supported catalysts. Several
researchers have reported that in noble metal catalysts promoted with ReOx (e.g., RhRe/C and
PdRe/CeO2), the active sites for the selective hydrogenolysis of polyols and ethers are ReOx
adjacent to Rh or Pd atoms, which can activate and transfer H2 to neighboring rhenium
species [26, 27, 33]. Tomishige et al. suggested that the dehydroxylation reaction of polyols
proceeds by the tetra/hexavalent redox cycle of rhenium oxide followed by hydrogenation
[26]. The XPS results also revealed that a large amount of hexavalent rhenium is reduced to
tetravalent rhenium in the MWCNT- and VC-supported catalysts during the HDO reaction,
whereas there is very little change in the Re valance of the RuRe/AC. Combining the results
from STEM-EDS and XPS, it can be concluded that RuRe/MWCNT and RuRe/VC have
greater numbers of bimetallic active sites than RuRe/AC, leading to higher selectivity
towards the completely deoxygenated products (e.g., cyclohexane). In addition, between
15

RuRe/MWCNT and RuRe/VC, RuRe/MWCNT displayed higher selectivity to cyclohexane
than RuRe/VC. This may be due to the higher surface concentration of Re species in the
RuRe/MWCNT, as shown in Table 4. It is likely that the surfaces of the bimetallic particles in
the RuRe/MWCNT are more enriched with Re, leading to the higher rate of C-O
hydrogenolysis.
Differences in the metal distribution on the carbon surfaces may be related to the
properties of the carbons. Karanjkar et al. suggested that the mobility of Re on the carbon
surface is restricted by oxygen-containing functional groups due to its strong oxophilicity
[27]. In order to form bimetallic RuRe particles, two metal components should migrate across
the carbon surface during the reduction process. Hence, the formation of bimetallic particles
on the carbon surface can be hindered by the presence of surface oxygen functionalities.
Accordingly, TPD analyses of the three carbons were carried out to measure the quantity of
oxygen functionalities, as shown in Fig. S3. The results showed that the amounts of CO and
CO2 desorbed from AC were significantly greater than those from MWCNT and VC,
indicating that the oxygen functionalities of AC exceeded those of MWCNT and VC.
Therefore, the mobility of Re on the AC support can be largely restricted by these oxygen
functionalities and this, in turn, can lead to the formation of fewer bimetallic particles,
leaving many segregated Ru or Re nanoparticles. In contrast, for MWCNT and VC with
much lower oxygen functionalities, the facile migration of Re on the carbon surface can
facilitate the assembly of Ru and Re, leading to the preferential formation of a bimetallic
structure, which is required for the efficient hydrogenolysis of lignin-derived phenols.
4. Conclusions
In conclusion, the MWCNT- and VC-supported RuRe catalysts displayed considerably
16

higher activity and hydrocarbon selectivity for the HDO of lignin-derived phenolic molecules
compared to the AC-supported catalyst, despite their idential metal loadings (an atomic Ru :
Re ratio of 1:0.5). Characterization of the carbon supports and the bimetallic catalysts by N2-
physisorption, H2-TPR, STEM-EDS, XPS, and TPD suggests that the pore structure and
surface oxygen functionalities of the carbon supports can be the two factors that influece the
HDO activity and selectivity of the catalysts. The presence of micropores in the AC support
makes the phenolic reactants, especially with bulky molecular structures, less accessible to
the metal sites, leading to lower activity (i.e., conversion) of RuRe/AC. Higher surface
oxygen content of the AC support also hinders the formation of bimetallic particles, which is
necessary for the efficient hydrogenolysis of C-OH bond of guaiacol, due to the strong
oxophilicity of Re. Bimetallic RuRe particles are more common than monometallic Ru
particles in the VC- and MWCNT-supported catalysts, and hexavalent rhenium species are
more easily reduced to tetravalent rhenium during the HDO reactions, suggesting the closer
proximity of Re to Ru in these catalysts. Between RuRe/MWCNT and RuRe/VC,
RuRe/MWCNT displayed higher selectivity to cyclohexane than RuRe/VC due to the higher
surface concentration of Re species.
Acknowledgements
This work was supported by the New & Renewable Energy Core Technology Program
of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) granted
financial resource from the Ministry of Trade, Industry & Energy, Republic of Korea (No.
20153030101580). This research was also supported by the National Research Council of
Science & Technology (NST) grant by the Korea government (MSIP) (No. CAP-11-04-
17

KIST).
References
[1]
[2]
G.W. Huber, S. Iborra, A. Corma, Chem. Rev. 106 (2006) 4044-4098.
J.H. Jae, G.A. Tompsett, Y.C. Lin, T.R. Carlson, J.C. Shen, T.Y. Zhang, B. Yang, C.E.
Wyman, W.C. Conner, G.W. Huber, Energy Environ. Sci. 3 (2010).
[3]
[4]
T.R. Carlson, Y.-T. Cheng, J. Jae, G.W. Huber, Energy Environ. Sci. 4 (2011) 145-161.
D.C. Elliott, H. Wang, M. Rover, L. Whitmer, R. Smith, R. Brown, ACS Sustainable
Chem. Eng. 3 (2015) 892-902.
[5]
[6]
[7]
[8]
[9]
M. Saidi, F. Samimi, D. Karimipourfard, T. Nimmanwudipong, B.C. Gates, M.R.
Rahimpour, Energy Environ. Sci. 7 (2014) 103-129.
A. Gutierrez, R.K. Kaila, M.L. Honkela, R. Slioor, A.O.I. Krause, Catal. Today 147
(2009) 239-246.
T. Nimmanwudipong, R.C. Runnebaum, D.E. Block, B.C. Gates, Energy & Fuels 25
(2011) 3417-3427.
R. Olcese, M. Bettahar, D. Petitjean, B. Malaman, F. Giovanella, A. Dufour, Appl.
Catal., B 115 (2012) 63-73.
H. Zhao, D. Li, P. Bui, S. Oyama, Appl. Catal., A 391 (2011) 305-310.
[10] A.A. Dwiatmoko, L. Zhou, I. Kim, J.-W. Choi, D.J. Suh, J.-M. Ha, Catal. Today 265
(2016) 192-198.
[11] C.R. Lee, J.S. Yoon, Y.-W. Suh, J.-W. Choi, J.-M. Ha, D.J. Suh, Y.-K. Park, Catal.
Commun. 17 (2012) 54-58.
[12] C. Zhao, J.A. Lercher, ChemCatChem 4 (2012) 64-68.
[13] W. Song, Y. Liu, E. Baráth, C. Zhao, J.A. Lercher, Green Chem. 17 (2015) 1204-1218.
[14] 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.
[15] Y. Hong, H. Zhang, J. Sun, K.M. Ayman, A.J. Hensley, M. Gu, M.H. Engelhard, J.-S.
McEwen, Y. Wang, ACS Catal. 4 (2014) 3335-3345.
[16] L. Nie, P.M. de Souza, F.B. Noronha, W. An, T. Sooknoi, D.E. Resasco, J. Mol. Catal.
A: Chem. 388 (2014) 47-55.
[17] P.T. Do, A.J. Foster, J. Chen, R.F. Lobo, Green Chem. 14 (2012) 1388-1397.
[18] H. Lee, H. Kim, M.J. Yu, C.H. Ko, J.-K. Jeon, J. Jae, S.H. Park, S.-C. Jung, Y.-K.
Park, Scientific Reports 6 (2016) 28765.
[19] A.J. Foster, P.T. Do, R.F. Lobo, Top. Catal. 55 (2012) 118-128.
[20] C. Zhao, J. He, A.A. Lemonidou, X. Li, J.A. Lercher, J. Catal. 280 (2011) 8-16.
[21] M. Kim, J.-M. Ha, K.-Y. Lee, J. Jae, Catal. Commun. 86 (2016) 113-118.
18

[22] G. Beamson, A.J. Papworth, C. Philipps, A.M. Smith, R. Whyman, J. Catal. 278 (2011)
228-238.
[23] M. Stein, B. Breit, Angew. Chem. Int. Ed. 52 (2013) 2231-2234.
[24] Y. Takeda, M. Tamura, Y. Nakagawa, K. Okumura, K. Tomishige, Catal. Sci. Tech. 6
(2016) 5668-5683.
[25] T. Toyao, S. Siddiki, A.S. Touchy, W. Onodera, K. Kon, Y. Morita, T. Kamachi, K.
Yoshizawa, K.i. Shimizu, Chem. Eur. J. 23 (2017) 1001-1006.
[26] N. Ota, M. Tamura, Y. Nakagawa, K. Okumura, K. Tomishige, ACS Catal. 6 (2016)
3213-3226.
[27] P.U. Karanjkar, S.P. Burt, X. Chen, K.J. Barnett, M.R. Ball, M.D. Kumbhalkar, X.
Wang, J.B. Miller, I. Hermans, J.A. Dumesic, Catal. Sci. Tech. 6 (2016) 7841-7851.
[28] M. Chia, Y.J. Pagan-Torres, D. Hibbitts, Q. Tan, H.N. Pham, A.K. Datye, M. Neurock,
R.J. Davis, J.A. Dumesic, JACS 133 (2011) 12675-12689.
[29] K.H. Kang, U.G. Hong, Y. Bang, J.H. Choi, J.K. Kim, J.K. Lee, S.J. Han, I.K. Song,
Appl. Catal., A 490 (2015) 153-162.
[30] J.K. Kim, J.K. Lee, K.H. Kang, J.C. Song, I.K. Song, Appl. Catal., A 498 (2015) 142-
149.
[31] L. Ma, D. He, Top. Catal. 52 (2009) 834-844.
[32] Y.T. Kim, J.A. Dumesic, G.W. Huber, J. Catal. 304 (2013) 72-85.
[33] S. Tazawa, N. Ota, M. Tamura, Y. Nakagawa, K. Okumura, K. Tomishige, ACS Catal.
6 (2016) 6393-6397.
19

Fig. 1. Guaiacol conversion and product selectivity at 200 ºC for 1 h over RuRe/AC,
RuRe/MWCNT, and RuRe/VC.
RuRe/AC
RuRe/VC
RuRe/MWCNT
Fig. 2. Product yields as a function of time at 240 ºC over RuRe/MWCNT, RuRe/VC, and
RuRe/AC. Others include propane, butane, pentane and hexane. Time 0 indicates the time
when the reactor reached the desired temperature (240 ºC).
20

Fig. 3. Guaiacol conversion and product selectivity at 200 ºC for 1 h over Ru/MWCNT,
Re/MWCNT, and their physical mixture.
Fig. 4. H2-TPR profiles of Ru/AC, Re/AC, and RuRe catalysts supported on different carbons.
Before TPR measurements, the catalysts were treated under 5% O2/He flow at 403 K for 1 h.
21

Fig. 5. XPS spectra of Re 4f of fresh and spent RuRe catalysts supported on different carbons.
Fig. 6. HR-TEM images and particle size distriubtions of RuRe/AC and RuRe/MWCNT.
22

RuRe/AC
RuRe/VC
RuRe/MWCNT
23

Fig. 7. STEM images and EDS mapping images of RuRe catalysts supported on different
carbons.
24

Table 1. Effect of oxophilic promoter on the activity of Ru/AC catalyst for guaiacol conversion. Reaction conditions: T = 200 ºC, Pinitial = 20
bar H2, and t = 1 h.
Metal content (%)
Product selectivity (%)
Conversion
(%)
Catalyst
Ru
Ma
Atomic ratio
(Ru:M)
2-Methoxy
cyclohexanol
Cyclohexane
Cyclohexanol Phenol
Othersb
(wt%)
(wt%)
RuW/AC
RuMo/AC
RuFe/AC
4
3.64
1.90
1.11
2.4
1 : 0.5
1 : 0.5
1 : 0.5
1 : 0.25
1 : 0.5
1 : 1
55.5
37.4
11.1
86.6
57.7
27.7
8.8
47.9
43.9
34.1
64.4
56.2
36.0
4.3
33.3
35.5
23.8
19.7
19.0
17.2
5.7
4
5.5
4.7
10.0
2.7
4
7.1
32.3
0.5
5.3
4
12.7
15.2
9.7
2.8
RuRe/AC
3.64
5.0
2.8
6.2
2.7
13.3
22.0
aM : Metal (W, Mo, Fe, Re)
bOthers include hexane, pentane, and butane.
25

Table 2. HDO of other lignin model compounds, eugenol and BPE, over RuRe/AC and
RuRe/MWCNT at 200 ºC for 1 h.
RuRe/
RuRe/
RuRe/AC
RuRe/AC
MWCNT
MWCNT
Eugenol
BPE
-
-
-
-
Conversion (%)
Conversion (%)
Selectivity (%)
Cyclohexane
31.1
99.4
9.7
99.7
Selectivity (%)
Propyl
cyclohexane
11.1
7.4
63.0
0
2
24
48
6
Propyl
benzene
Methyl
cyclohexane
44
4-propyl
cyclohexanol
Dicyclohexyl-
methane
20.1
20.1
16.3
33.8
0
7.5
46.5
3-propyl
methoxy-
benzene
Cyclohexanol
22
4-propyl
phenol
3.2
Table 3. Textural properties of carbon supports and RuRe catalysts supported on carbon.
Pore
SBET
VTotal
VMeso
VMicro
Metal dispersion
Sample
diameter (nm) (m2 g-1)
(cm3 g-1) (cm3 g-1) (cm3 g-1) (%)
Activated carbon (AC)
3.24
977
289
0.79
3.22
0.50
3.21
0.29
-
-
Multi-walled carbon
nanotube (MWCNT)
10-100
0.007
Vulcan carbon (VC)
RuRe/AC
20-100
4.2
224
366
215
109
0.60
0.29
1.24
0.45
0.56
0.19
1.24
0.45
0.04
-
0.1
30.6
8.6
11.4
RuRe/MWCNT
RuRe/VC
10-100
20-100
-
-
26