An efficient bifunctional Ru-NbOPO4 catalyst for the hydrodeoxygenation of aromatic ethers, phenols and real bio-oil

Article abstract of DOI:10.1016/j.apcata.2019.117378

An efficient bifunctional NbOPO₄ supported Ru catalyst (Ru-NbOPO₄) was applied to the hydrodeoxygenation of aromatic ethers and phenols and the upgrading of bio-oil. Characterization results revealed that the Ru-NbOPO₄ catalyst possessed strong acidity, including Lewis and Br?nsted acids. The Lewis acid sites originated from the Nb[sbnd]O bonding structures, including slightly distorted octahedral NbO₆, regular tetrahedral NbO₄ and highly distorted octahedral NbO₆. In combination with the strong acidity of the Nb[sbnd]O species and excellent hydrogenation activity of the metallic Ru, the bifunctional Ru-NbOPO₄ catalyst exhibited an excellent catalytic activity in the hydrodeoxygenation of aromatic ethers and phenols with different structures, and even real bio-oil to alkanes. The hydrocarbon yield after real bio-oil upgradation was up to 88.2 %. Carbon deposition and enlargement of the Ru nanoparticles resulted in slight deactivation of the catalyst. The catalytic activity could be mostly recovered after being calcined and reduced.

Full text of DOI:10.1016/j.apcata.2019.117378

Journal Pre-proof
An efficient bifunctional Ru-NbOPO catalyst for the hydrodeoxygenation
4
of aromatic ethers, phenols and real bio-oil
Hongye Zhao, Xun Hu, Jianxiu Hao, Na Li, Keduan Zhi, Runxia He,
Yunfei Wang, Huacong Zhou, Quansheng Liu
PII:
S0926-860X(19)30533-2
DOI:
https://doi.org/10.1016/j.apcata.2019.117378
APCATA 117378
Reference:
To appear in:
Applied Catalysis A, General
Received Date:
Revised Date:
Accepted Date:
6 September 2019
12 December 2019
13 December 2019
Please cite this article as: Zhao H, Hu X, Hao J, Li N, Zhi K, He R, Wang Y, Zhou H, Liu Q, An
efficient bifunctional Ru-NbOPO catalyst for the hydrodeoxygenation of aromatic ethers,
4
phenols and real bio-oil, Applied Catalysis A, General (2019),
doi: https://doi.org/10.1016/j.apcata.2019.117378
This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. 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.
©
2019 Published by Elsevier.

An efficient bifunctional Ru-NbOPO
4
catalyst for the
hydrodeoxygenation of aromatic ethers, phenols and real bio-oil
a
b
a
a
a
a
Hongye Zhao , Xun Hu , Jianxiu Hao , Na Li , Keduan Zhi , Runxia He ,
c
a
a
Yunfei Wang , Huacong Zhou *, Quansheng Liu *
aCollege of Chemical Engineering, Inner Mongolia University of Technology; Inner Mongolia Key
Laboratory of High-Value Functional Utilization of Low Rank Carbon Resources, Hohhot 010051,
Inner Mongolia, China
bSchool of Material Science and Engineering, University of Jinan, Jinan, 250022, China.
cCollege of Chemical Engineering, Ordos Institute of Technology, Ordos 017000, China
*
Corresponding author: Tel: +86 0471 6575722;
Email addresses: hczhou@imut.edu.cn (H. Zhou), liuqs@imut.edu.cn (Q. Liu).
1
/ 42

Graphical Abstract
Bifunctional catalysis
Bio-oil
O
O
O
High quality liquid fuel
O
O
OH
OH
O
O
OH
O
O
Ru-NbOPO catalyst
4
100℃-170℃, 1MPa H2
Ether compounds
High added value products
An efficient bifunctional Ru-NbOPO
4
catalyst was designed for the hydrodeoxygenation of aromatic
ethers, phenols and real bio-oil into hydrocarbon.
Highlights

Combining the strong acidity of NbOPO
4
and excellent hydrogenation activity of the metallic
Ru, bifunctional Ru-NbOPO catalysts can catalyze the hydrodeoxygenation of real bio-oil.
4

Different acid varieties and strengths could be attributed to the different structures of the
Nb-O species.
2
/ 42


The product distribution can be tuned by varying the acidic strength and reaction conditions.
Abstract
An efficient bifunctional NbOPO
hydrodeoxygenation of aromatic ethers and phenols and the upgrading of bio-oil. Characterization
results revealed that the Ru-NbOPO catalyst possessed strong acidity, including Lewis and Brønsted
acids. The Lewis acid sites originated from the Nb-O bonding structures, including slightly distorted
octahedral NbO , regular tetrahedral NbO and highly distorted octahedral NbO . In combination
with the strong acidity of the Nb-O species and excellent hydrogenation activity of the metallic Ru,
the bifunctional Ru-NbOPO catalyst exhibited an excellent catalytic activity in the
4
supported Ru catalyst (Ru-NbOPO
4
) was applied to the
4
6
4
6
4
hydrodeoxygenation of aromatic ethers and phenols with different structures, and even real bio-oil to
alkanes. The hydrocarbon yield after real bio-oil upgradation was up to 88.2%. Carbon deposition
and enlargement of the Ru nanoparticles resulted in slight deactivation of the catalyst. The catalytic
activity could be mostly recovered after being calcined and reduced.
Keywords: Hydrodeoxygenation, Niobium phosphate, Lewis acid, Anisole, Bio-oil
1
Introduction
Being the only sustainable source of organic carbon, biomass is playing an increasingly
important role in our energy landscape. The conversion of renewable lignocellulosic biomass into
liquid fuels is particularly attractive but extremely challenging due to its recalcitrant nature [1].
3
/ 42

Lignocellulose, as the structured portion of woody biomass, is composed of cellulose (40–50 wt%),
hemicellulose (16–33 wt%) and lignin (15–30 wt%). Lignin is a heavily cross-linked polymer
containing three monomers of coumaryl, coniferyl and sinapyl alcohols [2, 3]. Methoxyl (-OCH )
3
and hydroxyl (-OH) are the major functional groups of lignin and lignin-derived compounds. Three
primary routes have been used to convert lignocellulosic material into liquid oil, including syngas
production by gasification [4], hydrolysis of biomass to produce sugar monomer units [5, 6] and
bio-oil production by flash pyrolysis or hydrothermal liquefaction [7]. Synthesic gas can be used to
produce hydrocarbons, methanol and other fuels [8]. Sugar monomer units [9, 10] and associated
lignin intermediates [11] have been selectively converted into targeted transportation fuels. The
hydrolyzation of biomass has achieved a leap from the initial use of model compounds [12, 13] to the
direct conversion of raw biomass to liquid fuels [14, 15]. Although the above studies have succeeded
in converting lignocellulosic biomass into liquid fuels or other high-valued chemicals, the selective
production of hydrocarbons with higher yield is still a challenge. Another strategy for obtaining
hydrocarbons efficiently is to link the biomass to the hydrocarbon products through a bio-oil
intermediate. The biopolymer in biomass can be broken down by flash pyrolysis and then used to
generate high concentrations of phenolic molecules known as bio-oil [16]. Bio-oil has a high
potential as a liquid fuel since it retains up to 70% of the initial energy of the starting biomass and
contains less nitrogen and sulphur than fossil fuels [17, 18]. Three types of C-O bonds, CAR-OCH
3
,
C
AR-OH and CAR-R, are frequently found in biomass-derived bio-oil. However, bio-oil cannot be
directly used as a liquid fuel owing to its high oxygen content, poor chemical stability and
corrosivity to the equipment [3, 19]. Therefore, bio-oil should be upgraded to improve its energy
4
/ 42

density and stability.
Catalytic hydrodeoxygenation is an effective bio-oil upgrading strategy, which can selectively
remove oxygen from raw bio-oil under moderate temperatures and hydrogen pressures in the
presence of efficient catalysts. Catalytic hydrodeoxygenation has been observed to take place via
complex reaction pathways. Generally, the two main routes for hydrodeoxygenation of model
compounds are direct deoxygenation (DDO) retaining the aromatic character and hydrogenation
(HYD) saturating the aromatic ring [20, 21]. A number of catalysts have been investigated for
hydrodeoxygenation, including metal sulfides [22, 23], noble metals [20, 24], transition metal
carbides [25, 26], phosphides [27, 28] and bifunctional catalysts [29-31]. Bifunctional catalysts can
conveniently provide both hydrogenation and deoxygenation sites on a single catalyst, converting the
oxygen-containing compounds directly to alkanes in one pot. Among the kinds of hydrogenated
metals, Ru has been regarded as a very effective species and is relatively cheaper than other noble
metals [32]. Materials with acidity such as ZrO
2
, TiO , molecular sieves and other acidic materials
2
have been widely used in the construction of bifunctional catalysts for deoxygenation [17, 33-35],
but relatively high reaction temperatures and catalyst deactivation are still major disadvantages that
need to be solved.
Niobium compounds and related materials have attracted great interest in recent years in the
field of catalysis, especially as solid acids [6, 12]. Several important reactions such as hydration,
hydrolysis and esterification can be readily catalyzed by Nb-based materials [36-39]. In addition,
Nb-based materials can also be used to construct bifunctional catalysts and obtain relatively good
deoxygenation effects. For example, Nb
2
5
O was applied in the deoxygenation of fatty acids,
5
/ 42

triglycerides and alcohol for the cleavage of the CAR-O bond [40, 41]. In order to improve the
deoxygenation performance of bifunctional catalysts, the acidity of the Nb-based support should be
further enhanced. Different treatment methods have been attempted to change the composition and
structure of niobium compounds, such as acid treatment [6] or introducing other metal ions to the
Nb-based system [12]. Bifunctional catalysts based on improved niobium compounds were also used
in the conversion of oxygen-containing substances, such as sorbitol and furan-based compounds, into
hydrocarbon fuels [9, 42, 43].
There have also been attempts in the direct hydrodeoxygenation of lignin and raw woody
biomass, and a relatively high hydrocarbon yield and selectivity were obtained [14, 15]. At present,
hydrodeoxidation of bio-oil to produce hydrocarbon fuels by various bifunctional catalysts has been
reported. For example, Lee et al. deoxidized a bio-oil model compound and bio-oil with Pt/HY as the
catalyst, and the results showed that the heating value of the bio-oil increased from 18.9 MJ/kg to
2
4.4 MJ/kg [44]. Shu et al. upgraded phenolic compounds and bio-oil using highly dispersed
Ru/TiO [45] and a Pt/TiO catalyst [46], and a series of value-added products composed of
2
2
hydrocarbons and alkylphenols were achieved. The Ru catalysts were used to upgrade fast pyrolysis
bio-oil to drop-in fuel [47]. Additionally, in-situ hydrogenation of model compounds and raw bio-oil
over the Raney Ni catalyst [48] and Ni/CMK-3 catalyst [49] were reported. Although many studies
have reported the advantages of bifunctional catalysts for the hydrogenation of model compounds
and raw bio-oil, there is a paucity of reports on the Nb-based bifunctional catalyst applied in real
bio-oil systems to produce hydrocarbons efficiently. In addition, the acidity of the Nb-based support
is important in the activity of the noble metals for hydrodeoxygenation of bio-substrates and NbO
x
6
/ 42

species were considered to play a key role in the cleavage of the CAR-O bond in the hydrolysis of
biomass [9, 14, 40-42, 50], but the structure of NbO species in Nb-based supports has not been
x
further analyzed in the previous works, and the relationship between the different structures of NbO
x
species and the deoxygenation effect still needs to be further studied.
In this work, a bifunctional Ru-NbOPO catalyst was designed for the hydrodeoxygenation of
4
the model compound anisole into cyclohexane (Scheme 1), and the effect of the catalyst structures
and reaction parameters on the hydrodeoxygenation performance was systematically investigated.
The catalytic activity of bifunctional Nb-based catalysts was synergistically affected by the
deoxygenation ability of the Nb-based supports and the hydrogenation ability of the Ru species. The
acidic property of the supports and the metallic property of the Ru species were revealed by various
characterizations. The acid source of the Nb-based catalyst and the effect of the Nb-O structure on
the acidity were studied in detail. After these, the catalytic pathway and the stability of Ru-NbOPO
4
were further studied. The bifunctional Ru-NbOPO catalyst was effective for the
4
hydrodeoxygenation of various aromatic ethers and phenolic compounds with different structures. In
addition to the model compound reactions, hydrodeoxygenation of real bio-oil into liquid alkanes
was also carried out to evaluate the feasibility of the Ru-NbOPO catalyst for practical applications.
4
7
/ 42

Nb (C H O )
4 4 6 5
Dissolution
2
Dry
CTAB
NH ) HPO /H PO
4
1
05℃
H2O
(
4
2
4
3
NbPO@CTAB
CH3
O
NbPO@CTAB@H O
2
Reduction
H2
Impregnation
RuCI3
Ru-NbOPO4
RuCI -NbOPO4
NbOPO4
3
Scheme 1 Schematic diagram of the preparation of the Ru-NbOPO
of biomass-derived anisole into cyclohexane.
4
catalyst and its application in the deoxygenation
2
Experimental
2
.1 Materials
Anisole (99%), phenyl ether (99%), veratrole (99%), ethoxybenzene (99%), diphenyl ether
98%), p-cresol (99%), 3-methylanisole (99%), 4-allylanisole (98%), 2,6-dimethoxyphenol (98%),
-ethyl-2-methoxyphenol (98%), cyclohexane (99.5%), methylcyclohexane (99%), ethylcyclohexane
99%), propylcyclohexane (99%), dodecane (99%), ruthenium chloride (RuCl , 99.5%, Ru 47.0%),
, 99.9%) and hexadecy ltrimethyl ammonium
(
4
(
3
methoxycyclohexane (98.0%), niobium oxide (Nb
2
O
5
bromide (CTAB, 99%) were provided by J&K Scientific Ltd. Decane (98.5%), hexane (99%),
tartaric acid (99%), hydrofluoric acid (HF, 40%), ammonium hydroxide (NH ·H O, 99.99%),
diammonium hydrogen phosphate ((NH HPO , A.R) and phosphoric acid (H PO , A.R) were
provided by Sinopharm Chemical Reagent Co. Ltd. H (99.99%) was provided by the Hohhot
3
2
4
)
2
4
3
4
2
Hengsheng Scientific Instrument Department. All chemicals were directly used as received without
purification. The raw bio-oil used in this study was produced from the pyrolysis of cornstalk in a
8
/ 42

grinding pyrolysis reactor at 500°C. The preparation process was according to that reported in the
literature [51]. The elemental composition of the raw bio-oil is shown in Table 1.
Table 1 The elemental composition of raw bio-oil
Sample
C(%) N(%) H(%) S(%) O(%)
33.24
Raw bio-oil 57.42 0.44 8.43
0
2
.2 Catalyst preparation
The preparation method of the Nb-based supports is described in SI 2.2. The Ru-NbOPO
Ru-Nb (P) and Ru-Nb catalysts were prepared by the incipient wetness impregnation method
using a calculated amount of RuCl ·3H O as the Ru precursor. After impregnation, the obtained
materials were dried at 105°C for 4 h. Subsequently, RuCl -NbOPO , RuCl -Nb (P) and
RuCl -Nb were put into a hydrogenation kettle, and the internal air was replaced by nitrogen
three times in sequence. Finally, the kettle was filled with 4 MPa hydrogen and then heated to 150°C
4
,
2
O
5
2
O
5
3
2
3
4
3
2
O
5
3
2
O
5
3
+
0
for 3 h. After the process was done, the Ru in the sample was reduced to Ru , then the Ru-NbOPO
Ru-Nb (P) and Ru-Nb catalysts were obtained. Ru-Nb , using commercial Nb as support
Ru-Nb (com)), Co-NbOPO and Ru-SiO catalysts were also prepared as comparison samples by
4
,
2
O
5
2
O
5
2
O
5
2
O
5
(
2
O
5
4
2
the above method. Cyclohexane, methoxycyclohexane and cyclohexanol are denoted as CHE, MCH
and CHL, respectively, in the catalytic reaction section.
2
.3 Catalytic activity tests
The hydrodeoxygenation of anisole was performed in a 15 mL Teflon lined stainless steel
autoclave equipped with a magnetic stirrer. Decane was chosen as the solvent because of its good
ether (phenols) solubility and its inertia in the hydrodeoxygenation reaction [44]. Typically,
according to a molar ratio of the catalyst to anisole of 0.056, the catalyst Ru-NbOPO was first added
4
9
/ 42

into the reactor, and then, 1.1 g 10% (wt%) anisole solution with decane as the solvent was added.
After replacing the internal air with nitrogen three times, hydrogen was charged into the reactor to
the desired pressure, and the reactor was placed in a constant temperature air bath with stirring for
the desired reaction time. After the reaction, the reactor was cooled in cold water to quench the
reaction, and the organic phase was diluted with decane. The liquid products were analyzed
quantitatively by a gas chromatograph (GC, TECHCOMP GC7890 II) equipped with a flame
ionization detector using dodecane as the internal standard. Identification of the products and the
reactant was done using a GC-MS (Agilent 7890 series GC with a 5977 MS detector) equipped with
an HP-5MS UI capillary column as well as by comparing the retention times with respective
standards in GC traces.
In the reusability experiments, the catalyst was separated by centrifugation, washed with fresh
decane three times, and then reused for the next run without further reduction. Heterogeneity of
catalysts was detected by removing the solid catalysts from the reaction mixture, and the supernatant
was allowed to react to see if the product yield could further increase without the presence of the
solid catalysts.
In the conversion of real bio-oil experiments, 0.1 g raw bio-oil was first extracted by 3 g hexane
solvent and added into the autoclave as a substrate. A total of 0.1 g Ru-NbOPO catalyst was added
4
into the autoclave. After purging with hydrogen three times, the reactor was pressurized to 5 MPa at
room temperature and heated to 170°C for 10 h under stirring. The components of the liquid products
were analyzed by GC-MS, and the contents were calculated based on the peak areas.
1
0 / 42

3
Results and discussion
3
.1 Screening and characterization of the Nb-based catalysts
The catalytic performance of the Nb-based catalysts was tested using the hydrodeoxygenation
of anisole as a model reaction. As a control experiment, the effects of the support materials Nb
2
5
O ,
Nb (P) and NbOPO on the hydrodeoxygenation were first investigated. The test results showed
2
O
5
4
that there was no catalytic activity in the absence of Ru immobilization. As shown in Table 2, the
catalytic activities were directly affected by the catalysts employed. Typically, the anisole conversion
and cyclohexane yield of Ru-Nb
methoxycyclohexane was produced. Ru-Nb
and high methoxycyclohexane content. Ru-NbOPO
of cyclohexane, and the cyclohexane yield was much higher than those for Ru-Nb
Ru-Nb (P) and had no detectable oxygen-containing compounds under identical reaction
2
O
5
were relatively low, and a certain amount of
(P) gave high conversion but low cyclohexane yield
showed the highest activity for the production
and
2
O
5
4
2
O
5
2
O
5
conditions. It was also obvious that Co-NbOPO
deoxygenation activities.
4
, Ru-Nb
2
O
5
(com) and Ru-SiO exhibited poor
2
Table 2 Comparison of the Nb-based catalysts for deoxygenation of anisole a
O
O
OH
Ru-NbOPO4
CH OH
3
Entry
Catalysts
H2
Anisole
CHE
CHE
MCH
CHL
X
Anisole
Y
CHE
S
Y
MCH
S
MCH
(%)
(%)
(%)
(%)
(%)
1
2
Ru-Nb O
5
68.7
50.9
74.1
8.5
12.3
11 / 42

2
3
4
5
6
Ru-Nb
Ru-NbOPO
Ru-Nb (com)
Co-NbOPO
Ru-SiO
2
O
5
(P)
91.7
99.3
48.4
24.5
42.8
46.3
93.5
7.4
50.4
94.1
15.2
12.6
34.4
19.4
0
21.1
4
0
2
O
5
12.9
0.3
8.7
26.6
4
3.1
1.2
2
14.7
27.3
Reaction conditions: anisole: 0.1 g, catalyst: 0.1 g, reaction temperature: 170°C, H
time: 2 h.
2
pressure: 1 MPa and reaction
a
A small amount of cyclohexanol in the product was not quantified.
In order to determine the specific properties of the Nb-based catalysts, detailed characterizations
were conducted. Optical images of the as-obtained Nb-based catalysts are shown in Fig. S1. SEM
(
Fig. S2) and TEM (Fig. 1) results showed that the morphologies of Nb
2
O
5
, Nb
2
O
5
(P) and NbOPO
4
obviously changed from nanoparticles to nanosheets. The TEM also displayed that the Ru
nanoparticles were uniformly distributed on the nanosheet supports (Fig. 1ac) with an average
particle size decreasing from 4.5 nm on Ru-Nb
2
O
5
to 3.5 nm on Ru-NbOPO (Fig. 1df). EDS
4
showed that the molar ratios of P/Nb in the three catalysts changed significantly with the addition of
different kinds of phosphorus sources (Table S1). The molar ratio of P/Nb increased as the pH of the
phosphorus precursor decreased. Nitrogen adsorption-desorption isotherms showed that the supports
and corresponding catalysts were porous (Fig. S3). The porosity properties are summarized in Table
S2. It can be seen that the addition of phosphorus sources in the precursor solution decreased the
surface area in order from Nb
2
O
5
to NbOPO . After loading the Ru nanoparticle, the surface area and
4
pore volume decreased for all three catalysts.
1
2 / 42

a
b
c
d=4.5±1.4 nm
d=4.0±1.8 nm
d=3.5±0.9 nm
60
40
20
0
60
40
20
60
40
20
0
0
0
1
0
1
2
3
4
5
6
7
8
9
2
4
6
8
10
2
3
4
5
6
7
Size (nm)
Size (nm)
Size (nm)
Fig. 1 TEM images of (a) Ru-Nb
2
O
5
, (b) Ru-Nb
2
O
5
(P) and (c) Ru-NbOPO ; The particle size distributions of (d)
4
Ru-Nb , (e) Ru-Nb (P) and (f) Ru-NbOPO .
2
O
5
2
O
5
4
In order to understand the acidic properties of the Nb-based catalysts well, systematic studies of
-TPD (Fig. 2) and Py-FTIR (Fig. 3) were undertaken. Three main NH desorption peaks in the
regions of 100200°C (T ), 200500°C (T ) and 500800°C (T ) are shown in Fig. 2, which can be
assigned to the desorption of NH from weak, medium and strong acid sites, respectively [52]. The
distribution and the quantitative estimation of acidity are summarized in Table S3. It can be seen that
the medium and strong acids of Ru-NbOPO were much higher than those of Ru-Nb (P) and
Ru-Nb . This result could be attributed to the function of the acidic medium H PO in the
NH
3
3
1
2
3
3
4
2
O
5
2
O
5
3
4
synthesis. The Py-FTIR spectra were used to investigate the types of acidic sites. As shown in Fig. 3,
both Lewis acid and Brønsted acid sites could be identified for the three catalysts after evacuation at
different temperatures. The quantification of acidity at different temperatures (Table 3) showed that
the Lewis acid was much stronger than the Brønsted acid on each catalyst, and the Lewis acid of
Ru-NbOPO
4
was stronger than that in the other catalysts. The NH -TPD and Py-FTIR results
3
indicated that the addition of different phosphorus substances in the precursor solution influenced the
1
3 / 42

acidity of the Nb-based catalysts. The differences in acidity may be directly related to the structure of
the catalysts. Therefore, the structures of the catalyst were characterized in detail in the following
studies.
Ru-NbOPO
4
200°C
500°C
Ru-Nb O (P)
2
5
Ru-Nb O
2
5
0
200
400
600
800
Temperature (°C)
Fig. 2 NH -TPD profile of the Nb-based catalysts.
3
1
4 / 42

Ru-NbOPO
200°C
350°C
450°C
4
L
1
547
B+L
1
496
B
1
544
Ru-Nb O (P)
2
5
L
1
546
B+L
1
590
B
1
540
Ru-Nb O
2
5
L
543
1
B+L
B
534
1
487
1
1700
1600
Wavenumber(cm )
1500
11400
1300
-
Fig. 3 FTIR spectra of the Nb-based catalysts after pyridine adsorption and evacuation at different temperatures. B:
Brønsted acid sites; L: Lewis acid sites.
Table 3 Strength and distribution of the Lewis and Brønsted acids on the surface of the catalysts by FTIR of pyridine
adsorbed at different temperatures
-1
Evacuated
Py-FTIR (μmol g )
Catalysts
temperature
Brønsted
acid
Lewis
Total acid
(B+L)
338.78
174.71
54.36
The ratio of
B/L
(
°C)
00
350
acid
2
38.38
24.46
8.22
300.40
150.25
46.14
0.13
Ru-NbOPO
4
0.16
4
2
50
00
0.18
29.04
14.68
9.02
274.06
128.63
43.68
303.10
143.31
52.70
0.11
2
Ru-Nb O
5
(P)
350
0.11
4
2
50
00
0.21
18.73
9.55
180.54
115.04
41.23
199.26
124.58
41.81
0.10
2
Ru-Nb O
5
350
50
0.08
4
0.58
0.01
1
5 / 42

(A)
133.7
Ru-NbOPO
4
Ru-Nb O (P)
2
5
5
P 2p
Ru-Nb O
2
140
136
132
128
Binding Energy (eV)
(B)
(C)
Ru-NbOPO
Ru-NbOPO
4
4
206.5
07.0
07.4
2
O 1s
Ru-Nb O (P)
Ru-Nb O (P)
2
5
5
2
5
5
2
Ru-Nb O
2
Ru-Nb O
2
Nb 3d
5
/2
Nb 3d
3
/2
212
210
208
206
204
536
534
532
530
528
526
Binding Energy (eV)
Binding Energy (eV)
Fig. 4 The (A) P 2p, (B) Nb 3d and (C) O 1s spectra of the Nb-based catalysts.
The structures of the catalysts were further characterized by XPS, FTIR and Raman. The
chemical state of the P, Nb and O elements was investigated by XPS (Fig. 4). The P 2p of
Ru-NbOPO
to the surface metal phosphate species PO 3-
substance enhanced the strength of the Brønsted acid in the catalysts [55]. The Nb 3d5/2 of
Ru-Nb (206.5 eV), Ru-Nb (P) (207.0 eV) and Ru-NbOPO (207.4 eV) shifted towards higher
binding energies sequentially, which may be due to the variation in the oxidation states of the Nb
4
and Ru-Nb
2
O
5
(P) was at a similar binding energy of 133.7 eV, which could be assigned
4
(Fig. 4A) [53, 54]. The addition of a phosphorus
2
O
5
2
O
5
4
4
+
species (Fig. 4B). The fitting analysis of Nb 3d indicated that Nb in Ru-Nb
highest valency state Nb5+ in Ru-NbOPO (Fig. S4). The high oxidation state Nb5+ was more
deoxygenated than the low oxidation state Nb4+ due to its oxyphilicity [33]. The O 1s spectra
2
O
5
was oxidized to the
4
1
6 / 42

revealed the different chemical environments of oxygen (Fig. 4C). The O element mainly existed in
the binding form of a Nb-O interphase six-membered ring in Ru-Nb (Fig. S5) [56]. The niobium
oxygen coordinational octahedron with a binding form of Nb=O appeared in Ru-Nb (P), and this
form existed dominantly in Ru-NbOPO . The different binding forms of Nb with O provided the
basis for the acidity of the three catalysts.
2
O
5
2
O
5
4
The FTIR spectra of the Nb-based supports, the corresponding Ru catalysts, and the peak
attributions are shown in Fig. 5A and Table S4. All three Nb-based samples had vibration peaks at
−
1
6
30/582 cm , suggesting the Nb-O stretching vibration existed in the Nb-based samples [38, 57].
−
1
Ru-NbOPO showed a strong band at 986 cm due to the asymmetric vibration mode of the Nb=O
4
−
1
species [58, 59]. A weaker band appeared in Ru-Nb
2
O
5
(P) at 1005 cm but was lost in Nb
2
5
O . This
phenomenon was associated with the binding form of Nb and O in the catalysts. The bands around
−
1
−1
1
620 cm and 900 cm may be related to the bending vibrations of O-H groups [38, 58, 60]. Raman
spectroscopy further revealed the relationship between the binding structures and their corresponding
acidic properties, as shown in Fig. 5B. It can be seen that the peak shapes of the supports and
corresponding catalysts were essentially the same. From the summary of the relationship between the
structure and the Raman band frequency in Table S5, it can be seen that three different binding forms
−
1
−1
of Nb with O were present in Ru-Nb
2
O
5
(P) and Ru-NbOPO
4
at bands of 675 cm , 820 cm and 956
−
1
−1
cm [61-65]. An additional band was found at 450 cm on the Ru-NbOPO
4
catalyst, which was in
agreement with the stretching modes of the P-O group in the phosphate ion, indicating the P was
involved in the catalyst construction [66, 67]. Combined with the acidity measurements (NH -TPD
and Py-FTIR) and the structural analysis (XPS, FTIR and Raman), the origin of the acidic properties
3
1
7 / 42

for the Nb-based catalysts was further investigated (Fig. S6). It could be inferred that the acid sites
originated from their specific molecular structures. The slightly distorted octahedral NbO [61, 62],
regular tetrahedral NbO [63, 64] and highly distorted NbO octahedra [63, 65] served as
coordinatively unsaturated surface acid sites resulting in Lewis acids. Brønsted acids may be
attributed to the presence of Nb-OH and P-OH groups on the surface [55]. For Ru-NbOPO , the
structures of the regular tetrahedral NbO and highly distorted NbO octahedra were present in large
quantities, resulting in strong Lewis acidity. In Ru-Nb (P), comparatively, these two structures
were much less prevalent than that of Ru-NbOPO , resulting in a medium Lewis acidity. As to
Ru-Nb , only the slightly distorted surface NbO octahedra existed, leading to the lowest Lewis
acidity among the three catalysts. A large number of phosphorus species effectively enhanced the
strength of the Brønsted acid, such as in Ru-NbOPO . The combination of strong Lewis acids with
6
4
6
4
4
6
2
O
5
4
2
O
5
6
4
strong Brønsted acids made the total acidity of the Ru-NbOPO
4
catalyst far exceed the other two
catalysts (Table S3).
(A)
(B)
820
Ru-NbOPO
Nb O
2
4
5
Nb O (P)
2 5
9
56
Ru-Nb O
Ru-Nb2O5(P)
Ru-Nb O
2
5
675
Ru-Nb O (P)
2
5
450
2
5
1
620
1215
1005
630
5
82
NbOPO
4
NbOPO
Nb O (P)
4
2 5
900
Nb O
2 5
Ru-NbOPO
4
986
1
800 1600 1400 1200 1000 800 600 400
200
400
600
800 1000 1200 1400
Raman shift
-
1
Wavenumber (cm )
Fig. 5 (A) FTIR and (B) Raman spectra of the Nb-based supports and catalysts.
The properties of the metallic Ru species determined the hydrogenation performance of the
1
8 / 42

catalysts. As the XPS results show in Fig. 6, the Ru 3d spectra of RuCl (Fig. 6a) and unreduced Ru
3
catalysts (Fig. 6b) indicated that both unsupported Ru species and Ru species immobilized on
different supports existed in similar cationic states before reduction treatment. After hydrogen
reduction (Fig. 6c), the position shifted to the lower binding energy, suggesting the cationic state was
successfully reduced to the metallic state. The fitting analysis of the Ru 3d spectra (Fig. S7) indicated
0
that the Ru species was present on the surface of the three catalysts [68]. The Ru 3d5/2 peaks of
Ru-Nb
interaction between the active Ru and the Nb-based supports [56]. The existing states of the Ru
species on different supports were further investigated by means of H -TPR (Fig. 7). Since the
2
O
5
(P) and Ru-NbOPO
4
shifted to higher values, implying that there were different degrees of
2
Nb-based supports were almost irreducible below 500°C, the reduction peaks in the temperature
range of 80500°C should come from the reduction of the cationic ruthenium species [69].
Unsupported RuCl
3
gave a hydrogen reduction peak around 157°C, and the reducibility of the Ru
>Ru-Nb (P)>Ru-Nb
Fig. 7), correlating well with the catalytic activity. Compared with the other two catalysts, the
species on different supports was observed in the order of Ru-NbOPO
4
2
O
5
2
O
5
(
Ru-NbOPO supported Ru species was more reducible, which may be related to the smaller particle
4
size, the uniform distribution of Ru nanoparticles on the NbOPO
between the Ru species and supports [70].
4
support, and the interaction
1
9 / 42

Ru-NbOPO
4
(
c)
C1s+Ru 3d3/2
Ru-Nb O (P)
2
5
5
Ru-Nb O
2
Ru 3d5/2
Ru0
2
79.7
RuCl -NbOPO
4
(
(
b)
a)
284.6
3
RuCl -Nb O (P)
3
2
5
5
RuCl -Nb O
3
2
Ru3+
81.8
2
RuCl
3
Ru3+
82.2
2
288 286 284 282 280 278
Binding Energy (eV)
Fig. 6 The Ru 3d spectra of (a) RuCl
3
and the Nb-based catalysts, (b) before and (c) after H reduction.
2
3
84
6
63
5
41
1
36
RuCl -NbOPO4
3
NbOPO4
RuCl -Nb O (P)
3
2
5
Nb O (P)
2
5
4
42
5
66
5
18
1
46
RuCl -Nb O
5
3
2
Nb O5
2
5
08
1
66
1
57
RuCl3
0
200
400
600
800
Temperature (C)
Fig. 7 H
2
-TPR of RuCl
3
and the Nb-based supports and catalysts.
The Ru loading in the catalysts showed distinct impacts on the catalytic performance of anisole.
2
0 / 42

As shown in Fig. S8A, with Ru loading changed from 1 wt% to 10 wt% at 140°C, the anisole
conversion and cyclohexane yield increased accordingly, while the methoxycyclohexane yield
decreased. However, methoxycyclohexane could not be completely converted to cyclohexane even if
the Ru loading increased to 10 wt%. Increasing the reaction temperature had an obvious effect on
improving the selectivity of cyclohexane. When the reaction temperature increased to 170°C (Fig.
S8B), the cyclohexane yield reached up to 95.5% and the methoxycyclohexane reduced to an
undetectable level when Ru loading was 5%. The conversion performance was not further improved
by increasing the Ru loading. Therefore, a Ru loading of 5% was considered to be the optimal
condition and used in further studies. This may be due to the fact that synergistic catalysis can only
be achieved if the surface of the catalyst provides a sufficient amount of both active metal and acid
centers. When the Ru loading was less than 5%, few metal sites were provided leading to a poor
catalytic performance. However, when the amount of Ru was more than 5%, the catalytic activity
decreased due to the increase in the Ru nanoparticle size. Different metal loading also had significant
influence on the production of hydrocarbons in other substrate systems [71]. Except for the Ru
loading quantities, the particle size of the Ru nanoparticles on the NbOPO support may also affect
4
the catalytic activity. It is well known that the particle size of noble metals plays an important role in
the catalytic activities and product distributions [72-74]. According to the experimental results, the
size distribution of Ru nanoparticles for the three catalysts was different, and the performances of the
catalysts had obvious differences. As the Ru particle size decreased from 4.5 nm on Ru-Nb
nm on Ru-NbOPO , the conversion of anisole increased gradually from 68.7% on Ru-Nb
on Ru-NbOPO . Therefore, it could be postulated that Ru nanoparticles with smaller particle sizes
2
O
5
to 3.5
4
2
O
5
to 99.3%
4
2
1 / 42

are more conducive for hydrodeoxygenation of CAR-O bonds, which is consistent with the above
literature [72-74]. In conclusion, the catalytic activity of bifunctional Nb-based catalysts is directly
affected by the Ru loading and particle size.
The performances of the prepared Nb-based catalysts were compared with other analogous
hydrodeoxygenation catalysts reported in the literature (Table S6). It can be seen that the conversion
and the final cyclohexane yield were nearly at the same level for Ru-Nb
loaded on other acidic supports (Entries 47, 1518, 20, 22, 23, 29, 30), while the cyclohexane yield
for Ru-NbOPO was much higher than all other catalysts. Compared with other analogous catalysts,
the reaction temperature of the Ru-NbOPO catalyst was significantly reduced to 170°C (Entries
30), and the required hydrogen pressure was reduced to 1 MPa (Entries 47, 1224, 2830). The
2
O
5
(P) and Ru-Nb
2
5
O as Ru
4
4
4
Ru-NbOPO catalyst prepared in this work exhibited a higher TOF value than most other analogous
4
catalysts (Entries 48, 1120, 22, 25, 2730) and was only lower than a few catalysts (Entries 10, 21,
3, 24, 26). These results indicate that the prepared Ru-NbOPO catalyst is highly efficient for the
deoxygenation of anisole under mild conditions. The relatively high catalytic activity of the
Ru-NbOPO catalyst could be attributed to the strong acidity produced by the Nb-O species and the
2
4
4
excellent hydrogenation ability of the Ru nanoparticles. The synergistic promotion of these two
effects provided more active sites to break CAR-O bonds and hydrogenation of hydrocarbons, and the
porosity of the catalyst promoted the mass transfer and thus reduced the hydrogen pressure and
shortened the reaction time.
3
.2 Effect of the reaction conditions
The effect of catalyst dosage on the conversion of anisole was then studied (Fig. 8). It can be
2
2 / 42

seen that both the anisole conversion and the cyclohexane yield increased significantly as the molar
ratio of the catalyst to anisole increased in the range of 0.0120.107. However, the
methoxycyclohexane yield first increased and then decreased for all three catalysts. When the molar
ratio of the catalyst to anisole increased to 0.056, the cyclohexane yield of Ru-NbOPO
highest, reaching 93.5%, and was much higher than Ru-Nb (P) and Ru-Nb . At the same molar
ratio of catalyst to anisole, the anisole conversion for Ru-Nb (P) was significantly higher than that
for Ru-Nb , and the methoxycyclohexane yield also increased obviously. These results may be due
to the relatively higher acidity of Ru-Nb (P) than Ru-Nb , enhancing the deoxygenation activity.
For Ru-NbOPO , when the molar ratio of the catalyst to anisole was 0.056, the anisole conversion
4
was the
2
O
5
2
O
5
2
O
5
2
O
5
2
O
5
2
O
5
4
and cyclohexane yield reached a maximum, and methoxycyclohexane was no longer generated.
Further increasing the ratio only slightly improved the conversion and yield, which indicated a molar
ratio of catalyst to anisole of 0.056 was sufficient for the complete deoxygenation of anisole under
the studied reaction conditions.
2
3 / 42

1
1
1
00 Ru-NbOPO4
80
60
40
20
0
Conv.
CHE Yield
CHE Sel.
MCH Yield
MCH Sel.
00
Ru-Nb O (P)
2
5
80
60
40
20
0
00
Ru-Nb O
2
5
80
60
40
20
0
0.02 0.04 0.06 0.08 0.10
Molar ratio of catalyst to anisole
Fig. 8 Effect of the molar ratio of catalyst to anisole on the deoxygenation performance. Reaction conditions: anisole:
.1 g, decane: 1.0 g, reaction temperature: 170°C, H pressure: 1 MPa and reaction time: 2 h.
0
2
The effects of reaction temperature on the hydrodeoxygenation of anisole over Nb-based
catalysts were also investigated (Fig. 9). Reaction temperature plays an essential role in the
hydrodeoxygenation process. The anisole conversion and cyclohexane yield of the three catalysts
were significantly improved with an increase of temperature. On Ru-NbOPO at 120°C, the
4
cyclohexane yield was 39.9% with 13.7% methoxycyclohexane also being created, which came from
a coexistence of the DDO and HYD reaction routes. Increasing the temperature from 120°C to
1
70°C, the cyclohexane yield increased accordingly and the methoxycyclohexane yield decreased
remarkably, which strongly suggested the direct cleavage of aromatic C-O bonds in favor of the
DDO route at high temperature. Notably, at 170°C, the cyclohexane yield reached the maximum
value of 93.5%, while methoxycyclohexane was no longer generated. Clearly, increasing temperature
2
4 / 42

accelerated the conversion of anisole and reduced the generation of methoxycyclohexane and thus
improved the cyclohexane yield. This trend could also be seen on Ru-Nb (P) and Ru-Nb
Further increasing the tempreture to 190°C, the conversion performance on Ru-NbOPO changed
little while that on Ru-Nb (P) and Ru-Nb continued to grow. It can be concluded that
Ru-NbOPO is a powerful catalyst for hydrodeoxygenation, and a reaction temperature of 170°C is
capable of highiy selective conversion of anisole into cyclohexane.
2
O
5
2
5
O .
4
2
O
5
O
2 5
4
1
1
1
00 Ru-NbOPO4
8
6
4
2
0
0
0
0
0
Conv.
CHE Yield
CHE Sel.
MCH Yield
MCH Sel.
00
Ru-Nb O (P)
2
5
8
6
4
2
0
0
0
0
0
00
Ru-Nb O
2
5
80
60
40
20
0
120 130 140 150 160 170 180 190
Temperature (C)
Fig. 9 Effect of the reaction temperature on deoxygenation of anisole. Reaction conditions: anisole: 0.1g, decane: 1.0
g, catalyst: 0.1g, H pressure: 1 MPa and reaction time: 2 h.
2
Besides reaction temperature, H
cyclohexane, as shown in Fig. 10. The H
product distribution. For Ru-Nb , Ru-Nb
the maximum value when the H pressure was 0.5 MPa, 0.75 MPa and 1 MPa, respectively. Further
2
pressure is another key factor for the high yield of
pressure affected both the anisole conversion and the
(P) and Ru-NbOPO , the cyclohexane yield reached
2
2
O
5
2
O
5
4
2
2
5 / 42

increasing the H
yield decreased accordingly. On Ru-NbOPO
was no methoxycyclohexane generated. Increasing the H
2
pressure promoted the generation of methoxycyclohexane, and the cyclohexane
4
, when the H
2
pressure was less than 1.0 MPa, there
pressure to 3 MPa, a small amount of
2
methoxycyclohexane appeared. GC-MS was used for identifying the product composition in the
reactant solution catalyzed by Ru-NbOPO , as shown in Fig. S9. The typical GC-MS spectra showed
that there was no methoxycyclohexane formation when the hydrogen pressure was 1 MPa (Fig. S9A).
In contrast, a certain amount of methoxycyclohexane was detected in the product when the H
pressure was 3 MPa (Fig. S9B). The results of the GC-MS further confirmed that H first saturated
the aromatic ring, and the subsequent deoxygenation process was achieved at high H pressure,
which attributed to the high HYD activity of the Ru active sites [75, 76]. Nevertheless, low H
4
2
2
2
2
pressure will hinder the hydrogenation of the aromatic rings. Accompanied by the decrease of
hydrogenation of the aromatic ring at low hydrogen pressure, the DDO pathway became the
dominant reaction owing to the oxytropic performance of the strong acidity [3, 77]. The DDO and
HYD coexisted in the Ru-Nb
Ru-Nb system (Fig. S11). Therefore, a combination of the appropriate temperature and H
pressure conditions can be used to tailor the reaction product distribution.
2
O (P) catalyzed system (Fig. S10), while only HYD presented in the
5
2
O
5
2
2
6 / 42

Ru-NbOPO
4
100
8
6
4
2
0
0
0
0
0
Conv.
CHE Yield
CHE Sel.
MCH Yield
MCH Sel.
1
00 Ru-Nb O (P)
2
5
8
6
4
2
0
0
0
0
0
1
00 Ru-Nb O
2 5
80
60
40
20
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
H (MPa)
2
Fig. 10 Effect of the H pressure on the deoxygenation of anisole. Reaction conditions: anisole: 0.1 g, decane:1.0 g,
2
catalyst: 0.1 g, reaction temperature: 170°C and reaction time: 2 h.
The influence of the reaction time at different temperatures was also investigated (Fig. 11). It
can be seen that the conversion of Ru-Nb
2
O
5
(P) was much higher than that of Ru-Nb
2
5
O , while the
difference in the yield was relatively small at the same reaction time. After a 3 h reaction, the
cyclohexane yield was still lower than 60%, and the methoxycyclohexane remained on the two
catalysts. As for Ru-NbOPO , the reaction proceeded quickly within the first 0.5 h, and no
4
methoxycyclohexane formed. Under these experimental conditions, anisole could be completely
converted with a high cyclohexane yield of 93.5% within 2 h with a high selectivity of 94.1%,
indicating the high efficiency of the catalyst. The performance of Ru-NbOPO under a lower
4
temperature of 100°C was also investigated. As shown in Fig. S12, the cyclohexane yield in the
initial reaction stage was relatively low but was enhanced to 90.9% when reaction time was
2
7 / 42

prolonged to 14 h, but there was still 5.0% methoxycyclohexane that could not be deoxygenated.
These results imply that Ru-NbOPO can also remove oxygen at a very low temperature.
4
1
1
1
00 Ru-NbOPO4
80
60
40
20
0
Conv.
CHE Yield
CHE Sel.
MCH Yield
MCH Sel.
00 Ru-Nb O (P)
2
5
8
6
4
2
0
0
0
0
0
00
Ru-Nb O
2
5
80
60
40
20
0
0.5
1.0
1.5
2.0
2.5
3.0
Reaction time (h)
Fig. 11 Effect of the reaction time on the deoxygenation of anisole. Reaction conditions: anisole: 0.1 g, decane: 1.0 g,
catalyst: 0.1 g, H pressure: 1 MPa and reaction temperature: 170°C.
2
3
.3 Reaction pathway for the hydrodeoxygenation of anisole
The results obtained above showed that the catalytic activity and product distribution were
significantly affected by the catalyst properties and reaction conditions. According to the above
results, two different reaction pathways took place during the hydrodeoxygenation of anisole: DDO
and HYD, as shown in Scheme 2. More specifically, in the DDO route, Lewis acid sites were
abundantly distributed on the surface of Ru-NbOPO . The lone pair electrons of oxygen in the
4
methoxy group of anisole would be adsorbed strongly on the Lewis acid site Nb cations [78, 79],
resulting in the activation of the CAR-O bond. The activated CAR-O bond cleaved heterolytically, and
2
8 / 42

then the positively charged phenyl group could combine with the active hydrogen originating from
the dissociation of hydrogen promoted by Ru sites, forming the intermediate benzene. Subsequently,
benzene was converted into cyclohexane via hydrogenation. The product distribution on Ru-NbOPO
Fig. S9A) showed that only the DDO product dimethyl ether formed, but no other
oxygen-containing substances appeared. The Lewis acid sites had strong oxyphilicity [15, 40, 80].
Comparatively, for Ru-Nb , the route of HYD was the main reaction pathway. The metallic Ru
4
(
2
O
5
accepted the delocalized aromatic ring electron, resulting in the activation of the ring, which allowed
the nucleophilic addition reaction to occur through the adsorbed H attacking the aromatic rings,
completing the conversion of anisole to methoxycyclohexane [78, 81]. Then, the
methoxycyclohexane was converted to cyclohexanol and further to cyclohexane via dehydration over
the Brønsted acid sites. It was found that no DDO product dimethyl ether was produced (Fig. S11).
The generation of the HYD route may result from the insufficient acidity of the Lewis acid on the
Ru-Nb
and methoxycyclohexane were present in the products. It can be concluded that both the DDO and
HYD pathways were active during the course of anisole hydrodeoxygenation on Ru-Nb (P).
Lower H pressure favored DDO, while higher H pressure was conducive to HYD. Therefore, the
2
O
5
catalyst. From the typical GC-MS spectra of Ru-Nb
2
O
5
(P) (Fig. S10), both dimethyl ether
2
O
5
2
2
superior performance of the Nb-based material supported Ru catalysts for hydrodeoxygenation of
anisole was probably due to the oxophilic sites represented by Nb cations and the efficient
hydrogenation of metallic Ru.
2
9 / 42