Synergistic effects of Ru and Fe on titania-supported catalyst for enhanced anisole hydrodeoxygenation selectivity

Article abstract of DOI:10.1016/j.cattod.2017.08.025

The hydrodeoxygenation (HDO) of bio-oil is considered among the most promising techniques for upgrading pyrolysis bio-oils to biofuels. However, its commercial application is challenging because of the complexity of the reaction. In particular, minimizing the hydrogen consumption in HDO is a main concern. In this study, ruthenium supported on mesoporous TiO₂ (Ru/meso-TiO₂) catalysts were prepared with various Fe loadings for this reaction, using anisole as a model substrate. Analytical techniques, such as transmission electron microscopy, CO chemisorption, and X-ray photoelectron spectroscopy, revealed that the addition of Fe increased the number of oxygen vacancies, reaching a maximum when the Fe loading was 1 wt.%. H₂ consumption can be reduced by performing HDO reaction at high temperatures. The Ru/meso-TiO₂ without Fe converted anisole mainly to methoxycyclohexane, indicating that hydrogenation (HYD) is the dominant reaction pathway. The addition of Fe remarkably promotes the conversion and selectivity of the catalyst. The reaction path could be drastically changed from HYD to direct deoxygenation (DDO) by the addition of a proper amount of Fe. This change can be assigned to the synergism between Ru particles and Fe species on the surface of meso-TiO₂ support. The catalyst with 1 wt.% Fe exhibited the highest conversion and benzene selectivity due to its largest number of oxygen vacancies, indicating that it utilized a minimum amount of H₂.

Full text of DOI:10.1016/j.cattod.2017.08.025

Accepted Manuscript
Title: Synergistic Effects of Ru and Fe on Titania-Supported
Catalyst for Enhanced Anisole Hydrodeoxygenation
Selectivity
Authors: Tuan Ngoc Phan, Chang Hyun Ko
PII:
S0920-5861(17)30546-1
DOI:
Reference:
http://dx.doi.org/10.1016/j.cattod.2017.08.025
CATTOD 10968
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
31-5-2017
9-8-2017
12-8-2017
Please cite this article as: Tuan Ngoc Phan, Chang Hyun Ko, Synergistic Effects of
Ru and Fe on Titania-Supported Catalyst for Enhanced Anisole Hydrodeoxygenation
Selectivity, Catalysis Todayhttp://dx.doi.org/10.1016/j.cattod.2017.08.025
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.

Submitted to Catalysis Today on 31 May, 2017
Revised and Submitted to Catalysis Today on 09 August, 2017
Synergistic Effects of Ru and Fe on Titania-Supported
Catalyst for Enhanced Anisole Hydrodeoxygenation
Selectivity
Tuan Ngoc Phan, Chang Hyun Ko*
School of Chemical Engineering, Chonnam National University, Gwangju 61186, Republic of Korea
*Corresponding author: chko@jnu.ac.kr
1

Graphical abstarct
Highlights
 Hydrodeoxygenation of anisole studied as a model of upgrading bio-oil.
 Adding Fe to Ru-mesoporous TiO2 increased the number of oxygen vacancies.
 The Fe addition changed reaction path from hydrogenation to direct deoxygenation.
 The optimum Fe loading (1 wt.%) maximized benzene selectivity of anisole HDO.
Abstract
The hydrodeoxygenation (HDO) of bio-oil is considered among the most promising techniques for
upgrading pyrolysis bio-oils to biofuels. However, its commercial application is challenging because of the
complexity of the reaction. In particular, minimizing the hydrogen consumption in HDO is a main concern.
In this study, ruthenium supported on mesoporous TiO2 (Ru/meso-TiO2) catalysts were prepared with
various Fe loadings for this reaction, using anisole as a model substrate. Analytical techniques, such as
transmission electron microscopy, CO chemisorption, and X-ray photoelectron spectroscopy, revealed that
the addition of Fe increased the number of oxygen vacancies, reaching a maximum when the Fe loading
was 1 wt. %. H2 consumption can be reduced by performing HDO reaction at high temperatures. The
Ru/meso-TiO2 without Fe converted anisole mainly to methoxycyclohexane, indicating that hydrogenation
2

(HYD) is the dominant reaction pathway. The addition of Fe remarkably promotes the conversion and
selectivity of the catalyst. The reaction path could be drastically changed from HYD to direct deoxygenation
(DDO) by the addition of a proper amount of Fe. This change can be assigned to the synergism between
Ru particles and Fe species on the surface of meso-TiO2 support. The catalyst with 1 wt. % Fe exhibited
the highest conversion and benzene selectivity due to its largest number of oxygen vacancies, indicating
that it utilized a minimum amount of H2.
Keywords: Hydrodeoxygenation; Mesoporous TiO2; Ru-Fe Catalysts; CO Cleavage; Direct
Deoxygenation
3

1. Introduction
The fast pyrolysis of lignocellulosic biomass has attracted significant attention as a promising
technique for producing bio-oil to alleviate the need for fossil fuels. However, the high oxygen
concentration in bio-oil (~20–40 wt. %) causes several undesirable properties, including high viscosity, low
energy density, and high acidity. Therefore, it must be upgraded for expanded applications [1-3].
Hydrodeoxygenation (HDO) is considered to the most feasible strategy to lower the oxygen content in bio-
oil, by direct removal of oxygen in the form of water [2, 4]. Phenol and substituted phenols such as anisole,
guaiacol, and cresol are normally used as model substrates of lignin-derived monomers, since they are
thermally stable liquids under hydropyrolysis conditions and contain oxygen functional groups (–OH and
–OCH3) commonly found in the bio-oils.
Catalytic HDO has been observed to take place via complex reaction pathways. Generally, the two
main routes of HDO for the model compounds are direct deoxygenation (DDO) which retains the aromatic
character, and hydrogenation (HYD) that saturates the aromatic ring [4, 5]. In order to minimize the
hydrogen consumption and maintain the octane number, many different types of catalysts have been
investigated for the HDO of model compounds to avoid saturating the aromatic ring [3, 4, 6]. Base metal
catalysts, such as Fe [7, 8], Ni [9-12], Co [11-13], Cu [12, 14], and W [8] have been shown to have strong
ability to cleave the C–O bonds in model compounds, and minimize the saturation of aromatic ring at high
reaction temperatures. Unfortunately, these metals are rapidly oxidized under hydrothermal conditions and
could not achieve high conversion under mild conditions. For example, Zhao et al. [8] performed HDO of
guaiacol over a series of transition metal phosphide catalysts composed of Ni, Co, W, Mo, and Fe. However,
they reported that the guaiacol conversion was drastically decreased to less than 30 % at 200 C. To address
this issue, noble metal catalysts such as Pt [15, 16], Pd [12], and Ru [17-19] were studied. Although these
catalysts result in considerable HDO conversion and stability under mild conditions, the DDO and HYD
pathways still occur side-by-side at comparable rates.
In light of these challenges, bimetallic catalysts that combine base and precious metals are promising
for obtaining high catalytic activity under mild reaction conditions and with minimum hydrogen
consumption. Among these metals, iron shows high selectivity for converting the model compounds into
aromatic hydrocarbons. Ruthenium displays relatively high catalytic activity but lower activity for aromatic
ring hydrogenation, compared to other noble metals [7, 18, 20]. The hydrogenating metal (Ru) forms H,
while the oxophilic metal (Fe) forms strong bond with the oxygen functionality of the reactant to facilitate
the cleavage of C–O bonds [21].
In addition to the active metal species, the catalyst support also plays a crucial role in controlling the
mechanism of HDO process. Metals on acid supports such as zeolites (HZSM-5, HY), silica (SBA-15, KIT-
6) and Al2O3 are commonly used as catalysts for HDO reaction [9, 18, 22-24]. Metals and acid sites have a
4

synergistic effect on enhancement of catalytic performance and the overall reaction rate [23, 24]. However,
these types of catalysts direct the bio-oil model compounds into saturated products and rapidly deactivated,
followed by losing their activity. For HDO reaction of m-cresol on supported Mo catalysts, SiO2 and Al2O3
were rapidly deactivated by forming coke, while ZrO2 and TiO2 supports mainly catalyzed the reaction via
DDO and improved the stability [25]. It is in good agreement with our previous work [26]. Ru on TiO2
support exhibits higher DDO selectivity than Ru on Al2O3 and SBA-15 support. Thus, reducible TiO2 is
now a prospective support, due to the high selectivity for DDO products [27, 28]. Recently, ordered
mesoporous metal-containing materials have been considered as catalyst supports, because of the facile
diffusion of the reactants into the pores and promotion of the active sites [29-31]. Lu et al. [29] proposed a
new method to synthesize mesoporous metal carbides (Mo2C and W2C). When used as catalysts for HDO
of anisole at atmospheric pressure and mild temperature (423–444 K), these mesoporous catalysts displayed
impressive selectivity toward the unsaturated product.
In this contribution, we focused on reducing the H2 consumption in the HDO process of anisole as
the model compound. A series of catalysts were prepared by combining metallic Ru and different amounts
of Fe species supported on meso-TiO2. The effect of reaction temperature on catalyst activity was also
investigated and optimized. X-ray photoelectron spectroscopy (XPS) analysis indicated that the number of
oxygen vacancies is closely related to the Fe loading. Consequently, the addition of Fe to the Ru/meso-
TiO2 catalyst produced higher activity for DDO of anisole.
2. Experimental section
2.1 Catalyst preparation
Mesoporous TiO2 support was prepared using a modified procedure described elsewhere [32, 33].
Typically, 10 mL of tetrabutylorthotitanate (Ti(OC4H9)4, TBOT) was dissolved in 100 mL distilled water
in 30 min. Then the solution was transferred to a Teflon-lined stainless steel autoclave and hydrothermally
treated at 200 C for 3 h. After naturally cooling to room temperature, the obtained solution was centrifuged
to collect the white precipitates. The collected solids were washed with distilled water several times, and
finally dried at 100 C overnight.
The Ru-Fe supported catalysts were prepared by an incipient impregnation method. Ruthenium (III)
chloride (RuCl3) and iron nitrate hexahydrate (Fe(NO3)3.9H2O) were used as the Ru and Fe sources,
respectively. The amounts of RuCl3 and Fe(NO3)3.9H2O were calculated to obtain 4.0 wt. % Ru and 1–5.0
wt. % Fe in the catalysts. After impregnating the as-synthesized meso-TiO2 support with Ru and Fe sources,
the catalyst was dried at 100 C for 12 h, followed by crushing, and sieving to obtain particles < 75 m in
size. The obtained catalysts were denoted as RuxFe, where x refers to the weight percent of Fe. The RuxFe
5

catalysts were ex-situ reduced in a tube furnace at 400 C for 3 h in H2 environment prior to introduce to
HDO reactor.
2.2 Characterization
High-angle X-ray diffraction (XRD) patterns were collected on a Rigaku D-MAX 3 instrument
equipped with a Cu Kβ X-ray source operating at 40 kV and 40 mA, in order to investigate the crystalline
structure of the catalysts. The Ru and Fe contents on the catalysts were identified by inductively coupled
plasma optical emission spectroscopy (ICP-OES) (OPTIMA8300, Perkin-Elmer, USA). High-resolution
transmission electron microscopy (HR-TEM, CM200, Philips, installed at Korea Basic Science Institute)
was conducted to understand the distribution of active phases in the pores of the support. The specific
surface area and the pore volume of meso-TiO2 support and Ru-Fe supported catalysts were determined via
nitrogen adsorption-desorption measurements using a Belsorp mini II system (BEL, Japan). Prior to N2
adsorption, the samples were degassed in vacuum at 100 C for 12 h. The dispersion of active sites was
calculated using CO pulse chemisorption (MicrotracBel, Japan) by passing 10% CO and 90% He gas
mixture through the samples at 50 C. The catalysts were pre-reduced at 400 C for 1 h in the gas flow of
10% H2 and 90% He before measurement. H2-temperature programmed reduction (H2-TPR) studies were
performed using a BEL-CAT (BEL, Japan) system from 50 to 600 C. The samples were pre-treated with
He at 200 C for 3 h, and then cooled to room temperature. Subsequently, a gas flow of 10% H2-90% He
was introduced through the sample, and the ambient temperature was increased to 600 C at a rate of 10 C
min-1. To study the electronic states of the RuxFe catalysts, ex-situ XPS data were collected on a VG
Multilab 2000 instrument equipped with an Al Kα X-ray monochromatic source. The binding energy was
calibrated according to the binding energy of C 1s at 284.8 eV.
2.3 Catalyst reaction test conditions
The catalytic activity for HDO of anisole over the catalysts was measured in a 250-mL stainless-steel
batch reactor. In a typical procedure, 1.0 g of catalyst and 40 mL of anisole solution (10.0 wt. % in decane)
were placed together in the reactor. The reactor was sealed and then degassed by purging with nitrogen.
Then, hydrogen gas was introduced into the reactor to obtain an initial pressure of 10 bar. The HDO reaction
was conducted at the temperature range of 200–350 C for 3 h under stirring at 1500 rpm. After cooling to
room temperature, the liquid-phase products were collected by centrifugation and analyzed using a gas
chromatographer (GC, Agilent Technology 6890) equipped with an HP-INNOWax column (30 m × 0.32
mm × 0.50 m) and a flame ionization detector (FID).
6

The metal dispersion, the conversion of anisole, and the product selectivity and yield are specified
as follows:
푀푒푡푎푙 푑푖푠푝푒푟푠푖표푛 % =
퐴푚표푢푛푡 표푓 퐶푂 푢푝푡푎푘푒
푀표푙푒푠 표푓 푚푒푡푎푙 푙표푎푑푖푛푔
( )
(1)
(2)
푀표푙푒푠 표푓 푎푛푖푠표푙푒 푟푒푎푐푡푒푑
( )
퐶표푛푣푒푟푠푖표푛 % =
× 100
푀표푙푒푠 표푓 푎푛푖푠표푙푒 푠푢푝푝푙푖푒푑
푀표푙푒푠 표푓 푝푟표푑푢푐푡 푖
푀표푙푒푠 표푓 푡표푡푎푙 푝푟표푑푢푐푡푠
( )
푆푒푙푒푐푡푖푣푖푡푦 표푓 푐표푚푝표푛푒푛푡 푖 % =
× 100
(3)
(4)
( )
푌푖푒푙푑 표푓 푝푟표푑푢푐푡 푖 % = 퐶표푛푣푒푟푠푖표푛 × 푆푒푙푒푐푡푖푣푖푡푦 표푓 푝푟표푑푢푐푡 푖 / 100
3. Results and discussion
3.1 Textural properties of the catalysts
A series of supported Ru-Fe catalysts were prepared with a fixed Ru loading (~ 4 wt. %) and Fe
contents from 1 to 5 wt. %. Figure 1 displays the N2 adsorption-desorption isotherms and the corresponding
pore size distributions. All the isotherms can be assigned as type IV with H2-type hysteresis loop, indicating
that mesoporous structure exists in these catalysts [34]. After Ru and Fe were added onto the meso-TiO2
support, the total pore volume was reduced gradually, as the active metal nanoparticles (Ru and Fe) were
deposited within the channel of support. The pure support has a narrow pore size distribution with a sharp
peak at 7 nm. In the case of RuxFe catalysts, Ru and Fe species partially occupy the void space in the
mesopores, which reduces the average pore size and broadens the pore size distribution. Nevertheless, even
at the high Fe loading of 5 wt. %, the catalysts still maintained their mesoporous structure. The textural
properties of the support and the catalysts obtained from N2 adsorption-desorption measurement are
summarized in Table 1. In addition to the total pore volume, the specific surface area of the RuxFe catalysts
calculated by the Brunauer-Emmett-Teller method was reduced slightly compared to the bare support.
The measured Ru and Fe loadings in the catalysts are displayed in Table 1. Their actual contents are
similar to the values calculated from the precursors. The CO uptake value and the dispersion of Ru
calculated from CO pulse chemisorption measurement are also shown in Table 1. The dispersion of Ru in
the RuxFe catalysts changed according to the level of Fe loading. Ru1Fe showed the highest CO uptake,
indicating that it contains the largest number of adsorption sites on the surface of support. However,
increasing the Fe content causes a significant decline in the metal dispersion. It is possible that, when huge
amounts of metal precursors were introduced into the pores of meso-TiO2 support at the same time, the
particle size of Ru and Fe species could increase drastically due to their agglomeration. Alternatively, the
decrease of CO uptake can be due to the changing surface properties of Ru particles, as the Ru particles
might be partially blocked by iron oxides (FeOx) [35].
7

The Ru and Fe species in the reduced catalysts were identified by XRD (Fig. S1). The addition of
Ru and Fe did not change the crystallographic properties of the support. Three peaks of metallic Ru were
observed at 2θ = 38.3°, 42.2°, and 44°; corresponding to the (100), (002), and (101) lattice planes,
respectively (JCPDS 06-0663). The small and broad peaks of Ru metal observed in all XRD patterns
indicate that the Ru particle size did not change when the Fe content was increased. Fe existed in the form
of FeOx on Ru1Fe and Ru3Fe catalysts, but appeared as metal in Ru5Fe.
The morphologies of the reduced catalysts with different Ru and Fe loadings were further
investigated by HR-TEM. In Fig. 2, it can be seen that the meso-TiO2 nanocrystallites are strongly
agglomerated, with the average particle size being around 20 nm. Mesopores are formed by the void spaces
between agglomerated particles. The small sphere-like particles with average size of 2–3 nm are indicated
as metallic Ru particles due to the high electron density of Ru atom, compare to those of TiO2 and FeOx
particles. The size of Ru particles did not increase significantly upon the addition of Fe, which is in good
agreement with the XRD data. Moreover, EDX analysis (Fig. S2, S3, and S4) further confirm that Ru and
Fe were uniformly distributed on the surface of the TiO2 support. With increasing the Fe content, the
number of uniformly distributed FeOx species increased, which is expected to increase the possibility to
block the surface of Ru particles, resulting in the low adsorption amount of CO. Based on the XRD and
TEM–EDX measurements, the CO chemisorption results indicate that the FeOx species may partially block
the Ru active sites, leading to alter the adsorption properties of Ru surface.
The H2-TPR profiles of the catalysts are illustrated in Fig. 3. For Ru0Fe, the lone peak at 280 °C was
attributed to the reduction of Ru precursors to metallic Ru species. Typically, the reduction peak of Ru
precursors on TiO2 support appears at lower temperatures (150–180 °C) [28, 36]. In our case, the higher
reduction temperature suggests that the Ru precursors strongly interact with meso-TiO2 support. When Fe
was co-impregnated with Ru into the support, the reduction peak of Ru precursors shifted to higher
temperatures (around 300 °C) and became sharper. It indicates either the presence of Ru-Fe interactions, or
that Ru interacts stronger with meso-TiO2 support in the presence of Fe. In the reduction profiles of the
catalysts containing Fe, the first two strong peaks between 100–200 °C, whose intensities increased with
increasing Fe amount, can be assigned as the reduction of Fe precursors [Fe(NO3)3.9H2O] to FeO1+x (0 < x
< 0.5) oxides. These temperatures are significantly lower than those for Fe2O3 and Fe3O4 reported in the
literature [37, 38]. The co-impregnation of a noble and a non-noble metal creates close interaction between
these two types of metal particles. Consequently, the reduction of the non-noble metal occurs at a lower
temperature while that of the noble metal happens at higher temperature [39]. According to H2-TPR
profiles, we conclude that the surface properties of metallic Ru particles were changed by Fe co-
impregnation.
8

The electronic properties of the catalyst components were investigated by XPS. The spectra of Ru
3d and Fe 2p regions are displayed in Fig. S5 and S6, respectively. In the case of Ru0Fe and Ru1Fe, the Ru
3d region can be deconvoluted into two peaks: one at 280.9 eV and a smaller one at 282.3 eV, which can
be assigned to metallic Ru 3d5/2 and Ru3+ 3d5/2, respectively [40, 41]. In contrast, the Ru3+ peak disappeared
in the profiles of Ru3Fe and Ru5Fe catalysts. In the Fe 2p region (Fig. S3), the Ru-Fe catalysts showed
peaks attributable to Fe3+ [42]: a main peak of Fe3+ 2p3/2 at 711.2 eV and a Fe 2p1/2 peak at 725 eV. These
peaks shifted to lower binding energy in Ru1Fe (710.9 eV). Hence, the Fe species were partially reduced
into FeO1+x (0 < x < 0.5) with lower binding energy [43].
The XPS spectra of the Ti 2p region are shown in Fig. 4, and the deconvoluted data are summarized
in Table 2. The appearance of Fe species strongly affects the electronic properties of TiO2 support. In the
presence of Fe, the Ti 2p peaks of the catalysts were shifted to higher binding energy. A possible explanation
is that the introduced Fe species might partly diffuse into the crystal lattice of the meso-TiO2 support,
followed by the formation of Fe–O–Ti structure through the substitution of Ti4+ [44]. In the Fe–O–Ti
structure, electrons are more localized around the Fe ions than Ti, because of the higher electronegativity
of Fe. The reduced electron density on Ti sites means a higher binding energy of Ti 2p. Furthermore, the
Ti3+/Ti4+ ratio was changed by Fe loading, which can affect the formation of oxygen vacancy. The number
of Ti3+ was maximized when the Fe loading was 1 wt. % (Ru1Fe).
Figure 5 displays the deconvoluted core level of O 1s for the RuxFe catalysts. The spectra of all
samples were similar to each other. Essentially, four clearly different contributions were observed. The first
at ca. 529.8 eV was associated with crystal lattice oxygen O2- in the anatase TiO2 network. The second
component at 530.7 eV most probably corresponds to oxygen vacancy (Ox-, x < 2) on the surface [45]. The
O 1s contribution at 532 eV is related to oxygen in hydroxide ions (OH-) while the last one at 533 eV
corresponds to adsorbed water. The deconvoluted data of O 1s region are also summarized in Table 2. The
order of the Ox-/O2- atomic ratio is in agreement with that of the Ti3+/Ti4+ atomic ratio in the RuxFe catalysts.
The number of oxygen vacancies was maximized when the Fe loading was 1 wt. % (Ru1Fe). It indicates
that the oxygen vacancies were formed due to oxygen loss around Ti4+ in the anatase TiO2 matrix, which
also means that a portion of Ti4+ ions in TiO2 was reduced to Ti3+ in Ti2O3 as well. The area of oxygen
vacancies is enhanced via Fe doping, demonstrating that a larger amount of Ti4+ is reduced to Ti3+.
Alternatively, it can be attributed to the formation of the mixed oxide structure between Ti3+ and Fe ions in
the TiO2 matrix [46]. It is reported that the higher number of oxygen vacancies in the catalyst would be
favorable for HDO activity [25]. By adding a proper amount of Fe into the Ru/meso-TiO2 catalyst,
therefore, the number of oxygen vacancies could be maximized.
In summary, characterization of the catalysts reveals that the co-impregnation of Ru and Fe leads to
stronger metal-support interaction. The addition of Fe significantly changed the surface properties of meso-
9

TiO2 support, increasing the number of oxygen vacancies on the surface of support. Among these catalysts,
the Ru1Fe catalyst has the highest number of oxygen vacancies and Ti3+ defects, which can contribute to
its high performance for HDO of anisole.
3.2 Catalytic activity
The catalytic activity of Ru-Fe catalysts supported on meso-TiO2 was studied for the HDO of anisole.
Scheme 1 displays the possible reaction network. There are two major reaction pathways for HDO of
anisole: (1) direct deoxygenation (DDO) of Ar-O bond to form benzene, and (2) aromatic ring
hydrogenation to form methoxycyclohexane and its derivatives (cyclohexanol, cyclohexanone,
cyclohexene, and cyclohexane). Besides these two major reaction routes, the demethylation to produce
phenol and the ring-opening to n-hexane are considered as minor reaction pathways. In the mass production
of bio-fuel by HDO, the amount of hydrogen needed is a major concern. Thus, the ideal catalysts should
direct the HDO reaction toward the C-O bond cleavage, in order to reduce the oxygen level in pyrolysis
oils with minimized benzene ring saturation (i.e., reduced H2 consumption). When using anisole as the
starting material, only one mole of H2 is needed to obtain one mole of benzene, while three or four moles
of H2 are consumed to produce methoxycyclohexane or cyclohexane, respectively. Therefore, to minimize
the H2 consumption, benzene must be produced from anisole via DDO as much as possible.
The optimization of reaction conditions is one of the important strategies to minimize the H2
consumption. The catalytic performance of the Ru0Fe catalyst (without Fe) was carried out at the reaction
temperatures of 200–350 °C. As shown in Fig. 6, the temperature has a strong influence on the anisole
conversion and the product distribution. The Ru-supported meso-TiO2 catalyst is a potential candidate for
HDO of lignin model compounds, because of the high conversion at low temperature. Ru0Fe catalyzed
more than 60 % of anisole; mainly to benzene, methoxycyclohexane, cyclohexane, cyclohexanol, and a
trace amount of cyclohexene when the HDO reaction was conducted at 200 °C. When the reaction
temperature was raised to 350 °C, the conversion increased up to 98 %, indicating that the reaction was
almost complete. Chen et al. [47] reported the effect of reaction temperature on the activity for HDO of m-
cresol over silica-supported Ni, Pd, and Pt catalysts. These catalysts lost their activity at temperatures higher
than 300 °C due to the fast deactivation. In contrast, the Ru0Fe catalyst has high stability because of the
strong metal-support interaction.
Moreover, there is a gradual change in the product distribution with increasing reaction temperature.
At lower temperatures (200 and 250 °C), the major product is methoxycyclohexane, demonstrating that
HYD is the dominant pathway under this condition. It was reported that the hydrogenation of aromatic ring
naturally occurs at low temperature [48]. The reaction mechanism turned into the DDO pathway when the
reaction temperature was raised to 300 °C and 350 °C. Hence, the HYD reaction is strongly suppressed,
10

while the DDO reaction is strongly enhanced by increasing the reaction temperature. The benzene
selectivity reaches a maximum value of 80 % at 300 °C, which is 2-fold higher than that at 200 °C. Omotoso
et al. reported that the activity of TiO2-supported catalysts for DDO reaction of model compound is strongly
related to the oxygen defect sites on the surface of TiO2 support [49]. Higher reaction temperatures could
enhance the formation of oxygen vacancies, resulting in higher DDO selectivity. The DDO of anisole to
benzene is essentially more favorable at higher reaction temperature. Hence, the H2 consumption can be
reduced by controlling the reaction condition.
Another method to enhance the DDO selectivity is promoting Ru/meso-TiO2 catalyst by second
metal doping. Subsequently, Ga, and In were introduced into the Ru/meso-TiO2 catalyst to compare to Fe
for their effects on the HDO of anisole. According to Fig. S7, the combinations of Ru-Ga and Ru-In are not
effective. Ru-Ga/meso-TiO2 catalyst exhibited low activity for HDO reaction; while Ru-In/meso-TiO2
mainly catalyzed the HDO of anisole toward HYD pathway, producing methoxycyclohexane and
cyclohexanone as the main products. In the case of Fe, the anisole conversion and product distribution were
drastically affected by the Fe loading, as can be seen in Fig. 7. It demonstrates that Fe plays a crucial role
in controlling the HDO mechanism of anisole. Generally, the combination of Ru and Fe leads to a higher
selectivity of benzene and a lower selectivity of methoxycyclohexane, indicating that the DDO pathway is
the main reaction pathway over Ru-Fe/meso-TiO2 catalysts. With low Fe loading, Ru1Fe enhanced both
the anisole conversion and benzene selectivity. However, excess Fe species (Ru3Fe and Ru5Fe catalysts)
might agglomerate together and cover the active Ru sites, leading to a drop in benzene selectivity and an
increase in the yield of by-products. Li et al. [39] and Li et al. [43] proposed that bimetallic Ru-Fe was
more efficient for C-O bond scission than monometallic catalysts for the hydrogenolysis of glycerol to
glycols and of carboxylic acids to alcohols, respectively. Furthermore, they found that the Fe species were
crucial for the activity and stability of Ru-Fe bimetallic catalyst.
The higher DDO selectivity over the Ru-Fe/meso-TiO2 catalysts can be related to the number of
oxygen vacancies, as illustrated in Scheme 2. We propose that Ru sites on the meso-TiO2 surface are mainly
responsible for the hydrogen adsorption, followed by hydrogen dissociation to form an active hydrogen-
enriched surface. The oxygen atom in anisole might be adsorbed on the oxygen vacancies on the surface of
meso-TiO2 support. Then, the cleavage of C–O occurred, leading to the DDO of anisole. A higher number
of oxygen vacancies would be favorable for the HDO of lignin model compounds via the DDO of oxygen-
containing groups without ring saturation. Among these catalysts, Ru1Fe showed the largest number of
oxygen vacancies, resulting in the highest conversion and benzene selectivity.
Moreover, the increased selectivity for C–O scission products can also be attributed to the oxophilic
Fe sites. The reactant (anisole) might be absorbed on the Fe sites through its oxygen functionality (methoxy
group) due to the strong metal-oxygen bonds. Recently, Hong et al. reported that phenolic compounds are
11

easily adsorbed on the Fe sites, subsequently distorting the C–O bond in the former [50]. Fe sites act as an
anchor for the reactant, allowing surrounding sites to donate hydrogen for DDO. However, when the Fe
content is increased up to 5 wt. %, metallic Ru sites may be shielded by Fe sites, which prevents the contact
of Ru particles with H2. Consequently, the H2 dissociation is suppressed, leading to reduced anisole
conversion and benzene selectivity.
Figure 8 displays the reaction temperature dependence for the Ru1Fe catalyst. Similar to the data of
Ru0Fe (Figure 6), the anisole conversion increased with increasing reaction temperature, with the complete
HDO of anisole obtained at 350 °C. The drastic drop of benzene selectivity at 350 °C could be due to further
hydrogenation of benzene to form cyclohexane in the excess hydrogen environment after all the anisole
was reacted. Over the Ru1Fe catalyst, the maximum benzene selectivity, which is more than 80%, was
obtained at 250 °C. This percentage is much higher than that over Ru0Fe catalyst (64% at 300 °C). It
indicates that the consumption of H2 and the required reaction temperature are both reduced when using the
Ru1Fe catalyst. The synergic catalytic effect derived from Ru-Fe interaction is shown to deliver good
performance for the HDO of anisole under mild reaction conditions. By controlling the reaction conditions
of HDO and Fe content, the H2 consumption can be minimized via the DDO pathway.
4. Conclusions
The addition of Fe to the Ru/meso-TiO2 catalyst significantly changed the surface properties of this
catalyst and its catalytic performance for the HDO of anisole. Anisole was catalyzed to react via the HYD
pathway over monometallic Ru/TiO2 catalyst, but via the DDO pathway over Ru-Fe/meso-TiO2. The Ru1Fe
catalyst exhibited the highest anisole conversion and DDO selectivity, making it a promising candidate for
HDO of lignin-derived model compounds. The enhanced performance with Fe is attributed to the increased
number of oxygen vacancies on the surface of TiO2 support. In this catalyst, small Ru particles dissociate
H2 while the oxophilic Fe sites increase the interaction of the oxygen-containing functional groups with the
support surface. As a consequence of the synergy between Ru and Fe on meso-TiO2, the C–O bond cleavage
is enhanced without ring hydrogenation, resulting in minimized H2 consumption.
Acknowledgments
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea
government (Ministry of Science, ICT & Future Planning) (No. 2016R1A4A1012224). This research was
also supported by the Nano-Material Technology Development Program through the National Research
Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future Planning (NRF-
2015M3A7B4050493).
12

References
[1] A. J. R. Hensley, Y. Hong, R. Zhang, H. Zhang, J. Sun, Y. Wang, J. S. McEwen, ACS Catal. 4 (2014)
3381-3392.
[2] N. Arun, R. V. Sharma, A. K. Dalai, Renew. Sustain. Engergy Rev. 48 (2015) 240-255.
[3] G. W. Huber, S. Iborra, A. Coma, Chem. Rev. 106 (2006) 4044-4098.
[4] 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.
[5] F. E. Massoth, P. Politzer, M. C. Concha, J. S. Murray, J. Jakowski, J. Simons, J. Phys. Chem., B 110
(2006) 14283-14291.
[6] D. C. Elliott, Energy & Fuel 21 (2007) 1792-1815.
[7] R. N. Olcese, M. Bettahar, D. Petitjean, B. Malaman, F. Giovanella, A. Dufour, Appl. Catal. B 115-
116 (2012) 63-73.
[8] H. Y. Zhao, D. Li, P. Bui, S. T. Oyama, Appl. Catal. A 391 (2011) 305-310.
[9] Y. Yang, C. O. Hermández, V. A. P. O’Shea, P. Pizarro, J. M. Coronado, A. P. Serrano, Appl Catal. B
145 (2014) 91-100.
[10] Y. Han, D. N. McIlroy, A. G. McDonald, J. Analytic. Appl. Pyrol. 117 (2016) 94-105.
[11] N. T. T. Tran, Y. Uemura, S. Chowdhury, A. Ramli, Appl. Catal. A 512 (2016) 93-100.
[12] V. G. Deshmane, S. L. Owen, R. Y. Abrokwah, D. Kuila, J. Molec. Catal. A 408 (2015) 202-213.
[13] T. Mochizuki, S. Y. Chen, M. Toba, Y. Yoshimura, Appl. Catal. B 146 (2014) 237-243.
[14] K. L. Deutsch, B. H. Shanks, Appl. Catal. A 144 (2012) 447-448.
[15] H. Ohta, H. Kobayashi, K. Hara, A. Fukuoka, Chem. Comm. 47 (2011) 12209-12211.
[16] Y. Wang, T. He, K. Liu, J. Wu, Y. Fang, Bioresour. Technol. 108 (2012) 280-284.
[17] G. Yao, G. Wu, W. Dai, N. Guan, L. Li, Fuel 150 (2015) 175-183.
[18] W. Zhang, J. Chen, R. Liu, S. Wang, L. Chen, K. Li, ACS Sustain. Chem. Eng. 2(2014) 683-691.
[19] A. A. Dwiatmoko, S. Lee, H. C. Ham, J. W. Choi, D. J. Suh, J. M. Ha, ACS Catal. 5 (2015) 433-437.
[20] J. Sun, A. A. Karim, H. Zhang, L. Kovarik, X.S. Li, A.J. Hensley, J.S. McEwen, Y. Wang, J. Catal.
306 (2013) 47-57.
[21] A. M. Robinson, J. E. Hensley, J. W. Medlin, ACS Catal. 6 (2016) 5026-5043.
[22] W. Song, Y. Liu, E. Baráth, C. Zhao, J. A. Lercher, Green Chem. 17 (2015) 1204-1218.
[23] A. J. Foster, P. T. M. Do, R. F. Lobo, Top Catal. 55 (2012) 118-128.
[24] H. Lee, H. Kim, M. J. Yu, C. H. Ko, J. K. Jeon, J. Jae, S. H. Park, S. C. Jung, Y. K. Park, Sci. Rep. 6
(2016) 28765-28772.
13

[25] M. Shetty, K. Murugappan, T. Prasomsri, W. H. Green, Y. Román-Leshkov, J. Catal. 331 (2015) 86-
97.
[26] T. N. Phan, Y. K, Park, I. G. Lee, C. H. Ko, Appl. Catal. A 544, 2017, 84-93.
[27] M. B. Griffin, G. A. Ferguson, D. A. Ruddy, M. J. Biddy, G. T. Beckham, J. A. Schaidle, ACS Catal.
6 (2016) 2715-2727.
[28] S. Boonyasuwat, T. Omotoso, D. E. Resasco, S. P. Crossley, Catal. Lett. 143 (2013) 783-791.
[29] Q. Lu, C. J. Chen, W. Luc, J. G. Chen, A. Bhan, F. Jiao, ACS Catal. 6 (2016) 3506-3514.
[30] L. Zhou, M. Chen, Y. Wang, Y. Su, X. Yang, C. Chen, J. Xu, Appl. Catal. A 475 (2014) 347-354.
[31] O. Sorsa, H. Romar, U. Lassi, J. Kallio, Electrochim. Acta 230 (2017) 49-57.
[32] J. Yu, G. Wang, B. Cheng, M. Zhou, Appl. Catal. B 69 (2007) 171-180.
[33] S. Pavasupree, J. Jitputti, S. Ngáminlapasathian, S. Yoshikawa, Mater. Res. Bull. 43 (2008) 149-157.
[34] M. Kruk, M. Jaroniec, Chem. Mater. 13 (2001) 3169-3183.
[35] J. Yang, S. Li, L. Zhang, X. Liu, J. Wang, X. Pan, N. Li, A. Wang, Y. Cong, X. Wang, T. Zhang,
Appl. Catal. B 201 (2017) 266-277.
[36] J. M. G. Carballo, E. Finocchio, S. García, S. Rojas, M. Ojeda, G. Busca, J. L. G. Fierro, Catal. Sci.
Tech. 1 (2011) 1013-1023.
[37] E. E. Unmuth, L. H. Schwartz, J. B. Butt, J. Catal. 63 (1980) 404-414.
[38] H. Y. Lin, Y. W. Chen, C. Li, Thermochim. Acta 400 (2003) 61-67.
[39] W. Li, L. Ye, P. Long, J. Chen, H. Ariga, K. Asakura, Y. Yuan, RSC Adv. 4 (2014) 29072-29082.
[40] W. Song, Z. Chen, W. Lai, I. Rodríguez-Ramos, X. Yi, W. Weng, W. Fang, App. Catal. A. 540
(2017) 21-30.
[41] C. Sui, T. Zhang, Y. Dong, F. Yuan, X. Niu, Y. Zhu, Molec. Catal. 435 (2017) 174-181.
[42] H. Fang, J. Zheng, X. Luo, J. Du, A. Roldan, S. Leoni, Y. Yuan, Appl. Catal. A 529 (2017) 20-31.
[43] B. Li, J. Wang, Y. Yuan, H. Ariga, S. Takakusagi, K. Asakura, ACS Catal. 1 (2011) 1521-1528.
[44] J. Li, J. Xu, WW. L. Dai, H. Li, K. Fan, Appl. Catal. B 85 (2009) 162-170.
[45] W. Feng, G. Wu, L. Li, N. Guan, Green Chem. 13 (2011) 3265-3272.
[46] B. Bharti, S. Kumar, H. N. Lee, R. Kumar, Sci. Rep. 6 (2016) 32355.
[47] C. Chen, G. Chen, F. Yang, H. W. Wang, J. Han, Q. Ge, X. Zhu, Chem. Eng. Sci. 135 (2015) 145-
154.
[48] C. Newman, X. Zhou, B. Goundie, I. T. Ghampson, R. A. Pollock, Z. Ross, M. C. Wheeler, R. W.
Meulenberg, R. N. Austin, B. G. Frederick, App. Catal. A 477 (2014) 64-74.
[49] T. Omotoso, S. Boonyasuwat, S. P. Crossley, Green Chem. 16 (2014) 645-652.
14

[50] Y. Hong, H. Zhang, J. Sun, K. M. Ayman, A. J. R. Hensley, M. Gu, M. H. Engelhard, J. S. McEwen,
Y. Wang, ACS Catal. 4 (2014) 3335-3345.
15

Figure caption
Fig. 1. N2 adsorption-desorption isotherms and pore size distributions (inset) of support and catalysts.
Fig. 2. Representative HR-TEM images and corresponding particle size distribution (inset) of (a) Ru0Fe,
(b) Ru1Fe, (c) Ru3Fe, and (d) Ru5Fe.
Fig. 3. H2-TPR profiles of RuxFe catalysts.
Fig. 4. XPS spectra of the RuxFe catalysts in the Ti 2p region.
Fig. 5. XPS spectra of the RuxFe catalysts in the O 1s region.
Fig. 6. Temperature dependence of the Ru0Fe catalyst. Reaction conditions: 10 wt.% anisole (40 mL),
catalyst (1.0 g), 10 bar H2 for 3 h.
Fig. 7. Conversion and product distribution of the RuxFe catalysts. Reaction conditions: 10 wt.% anisole
(40 mL), catalyst (1.0 g), 200 °C, and 10 bar H2 for 3 h.
Fig. 8. Reaction temperature dependence of the Ru1Fe catalyst. Reaction conditions: 10 wt.% anisole (40
mL), catalyst (1.0 g), 10 bar H2 for 3 h.
Scheme 1. Reaction network for the HDO of anisole.
Scheme 2. Reaction mechanism for the HDO of anisole over the Ru1Fe catalyst.
16

Figure 1
17

Figure 2
18

Figure 3
19

Figure 4
20

Figure 5
21

Figure 6
22

Figure 7
23

Figure 8
24

Scheme 1
25

Scheme 2
26

Table 1. Textural characterization of supported Ru-Fe bimetallic catalysts.
CO uptake
Ru loading
(wt. %)
Fe loading
(wt. %)
Specific surface
area (m2/g)
Total pore
volume (cm3/g)
0.33
Pore size
(nm)
Dispersion
(%)
Catalysts
(mol/g)
Meso-TiO2
Ru0Fe
-
-
175
7.5
-
-
3.60
3.51
3.24
3.25
-
1.16
3.29
5.94
164
158
158
159
0.32
7.7
6.8
6.2
6.3
224
263
198
93.4
62.8
47.4
21.7
6.7
Ru1Fe
0.27
Ru3Fe
0.24
Ru5Fe
0.25
Table 2. XPS experiments: atomic % of species for O 1s and Ti 2p.
O 1s BE eV (at. %)
Catalysts
Ti 2p BE eV (at. %)
O2-
Oxygen vacancy
16.1
OH-
Adsorbed water
2.1
Ti4+
Ti3+
Ru0Fe
Ru1Fe
Ru3Fe
Ru5Fe
46.8
27.0
41.8
38.2
7.4
16.5
8.2
19.3
10.9
17.2
18.5
7.4
29.3
21.0
17.7
3.8
0.7
6.4
11.2
8.9
8.6
7.1
27

28