Natural zeolite supported Ni catalysts for hydrodeoxygenation of anisole

Article abstract of DOI:10.1039/d0gc04377j

Natural and synthetic (BEA, MOR) zeolite-supported nickel (～5 wt%) catalysts were prepared and employed for the hydrogenation of toluene and hydrodeoxygenation of anisole in a continuous-flow reactor. Ni/BEA and Ni/MOR display a higher level of metal dispersion and stronger metal-support interaction compared to the Ni/NZ and Ni/Escott catalysts, resulting in a higher concentration of charge-compensating Ni species and a larger high-temperature reduction peak. The Ni/BEA and Ni/MOR also present a significant mass of low-temperature desorbed H₂(centred at 150 °C) based on H₂-TPD, suggesting the H species are weakly adsorbed on small Ni clusters. In contrast, the H species were strongly adsorbed by the bulk Ni crystal over Ni/Escott and Ni/NZ, which were desorbed at maxima between 211 and 222 °C. We propose that the strongly adsorbed H species play a crucial role in the hydrogenation of toluene, leading to a significantly higher yield of methylcyclohexane over Ni/Escott and Ni/NZ compared to Ni/BEA and Ni/MOR. Both metal and acid sites are required in the hydrodeoxygenation of anisole. The strong Br?nsted acid sites and numerous smaller Ni species over Ni/BEA facilitated the transalkylation of anisole to phenol and methylanisole and subsequently hydrogenolysis of phenol to benzene, followed by the hydrogenation of benzene to cyclohexane.

Full text of DOI:10.1039/d0gc04377j

Volume 23
Number 13
7 July 2021
Pages 4623-4904
Green
Chemistry
Cutting-edge research for a greener sustainable future
rsc.li/greenchem
ISSN 1463-9262
PAPER
Michael Stockenhuber et al.
Natural zeolite supported Ni catalysts for
hydrodeoxygenation of anisole

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Natural zeolite supported Ni catalysts for
hydrodeoxygenation of anisole†‡
Cite this: Green Chem., 2021, 23,
4673
Penghui Yan,
Eric Kennedy and Michael Stockenhuber
*
Natural and synthetic (BEA, MOR) zeolite-supported nickel (∼5 wt%) catalysts were prepared and
employed for the hydrogenation of toluene and hydrodeoxygenation of anisole in a continuous-flow
reactor. Ni/BEA and Ni/MOR display a higher level of metal dispersion and stronger metal–support inter-
action compared to the Ni/NZ and Ni/Escott catalysts, resulting in a higher concentration of charge-com-
pensating Ni species and a larger high-temperature reduction peak. The Ni/BEA and Ni/MOR also present
a significant mass of low-temperature desorbed H2 (centred at 150 °C) based on H2-TPD, suggesting the
H species are weakly adsorbed on small Ni clusters. In contrast, the H species were strongly adsorbed by
the bulk Ni crystal over Ni/Escott and Ni/NZ, which were desorbed at maxima between 211 and 222 °C.
We propose that the strongly adsorbed H species play a crucial role in the hydrogenation of toluene,
leading to a significantly higher yield of methylcyclohexane over Ni/Escott and Ni/NZ compared to
Ni/BEA and Ni/MOR. Both metal and acid sites are required in the hydrodeoxygenation of anisole. The
strong Brønsted acid sites and numerous smaller Ni species over Ni/BEA facilitated the transalkylation of
anisole to phenol and methylanisole and subsequently hydrogenolysis of phenol to benzene, followed by
the hydrogenation of benzene to cyclohexane.
Received 28th December 2020,
Accepted 19th April 2021
DOI: 10.1039/d0gc04377j
rsc.li/greenchem
metals such as Pt, Pd and Ru, which are supported on carbon,
Al2O3, SiO2 and zeolites, are commonly employed in the HDO
1. Introduction
The depletion of readily accessible, traditional feedstocks of reactions.10–13 While these catalysts are reported to possess a
liquid fossil fuels and rapidly increasing demand for renew- higher activity compared to traditional HDO catalysts such as
able oil and green chemicals (such as green BTX), is creating NiMoS and CoMoS, their high cost restricts the widespread
strong incentive for the development of non-petrochemical commercial application of these catalysts.14 In addition, sul-
derived fuels and chemicals.1 Biocrude oils, which are derived fided catalysts are generally regarded less favourably as a result
from the pyrolysis or liquefaction of biomass, are considered of product contamination with sulfur. Non-noble metal cata-
to be suitable candidates due to the quantity of biomass avail- lysts such as metallic Ni, Fe, W and Mo or their combinations,
able around the world.2,3 However, bio-oils usually have a high supported on zeolites, are also the subject of investigations in
content of oxygen, leading to high acidity, reduced calorific HDO reactions, which generally exhibit
a high HDO
value, high viscosity and decreased thermal stability and thus activity.15–17 Ni–Ga alloy can suppress the hydrogenation of
cannot be directly used as transporting fuels.4,5 To address aromatic rings during the anisole hydrodeoxygenation, leading
these limitations, bio-oils upgrading is required to reduce the to an increased yield of benzene.18 Wang19 investigated the
oxygen content.
HDO of anisole over Al2O3 supported NiMo, MoP and Ru cata-
Hydrodeoxygenation (HDO), which employs bifunctional lysts. NiMo with moderate hydrogenation activity enhanced
catalysts and hydrogen gas cofeed to remove oxygen from feed- the transalkylation reaction via a partial hydrogenation reac-
stocks, is considered to be an effective method for biocrude oil tion. The transalkylation and hydrogenolysis of anisole to BTX
upgrading.6,7 Bifunctional catalysts, including supports and was also facilitated by zeolites (ZSM-5, BEA) supported Ni–Fe
active metals, play a pivotal role in the HDO process.8,9 Noble alloy and low pressure.20 The pore size of the supports plays
an important role in HDO reactions. Gamliel21 observed Ni
supported on modified mesoporous USY zeolite was more
effective for HDO of anisole than the Ni/USY catalyst with
higher micropore volume.
Chemical Engineering, School of Engineering, University of Newcastle, Callaghan,
NSW, 2308, Australia. E-mail: michael.stockenhuber@newcastle.edu.au
†This paper is dedicated to the memory of Mr Greg Stephen from Zeolites
Natural, untreated zeolites, which are abundant and in-
Australia, a lifelong supporter of Australian zeolite science.
expensive compared to synthetic zeolites (such as BEA, MFI),
are rarely applied in the HDO reactions. The natural zeolites
‡Electronic supplementary information (ESI) available. See DOI: 10.1039/
d0gc04377j
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 4673–4684 | 4673

View Article Online
Paper
Green Chemistry
present crystalline hydrated aluminosilicates with framework chemisorption of carbon monoxide was performed at 35 °C
structures such as clinoptilolites and chabazites.22 In addition, over a pressure range between 30 and 90 mbar, and the total
the natural zeolites are considered “green” as they are not number of gas adsorbed was calculated based on the ideal gas
manufactured unlike synthetic zeolites, which require hazar- law. Based on the volume of chemisorbed CO, the metal dis-
dous substances (such as organic structure-directing agents) persion (D) and average metal particles size (d) were calculated
during the synthesis.23
using eqn(1) and (2).30,31
Natural zeolites, which can be activated by water or ion
exchange, have been used for hydrocarbon cracking or isomer-
ization reactions.24–26 Due to their typical characteristics,
especially their high content of acid sites and desirable tex-
tural properties,22,27 natural zeolites have potential application
in the upgrading of biocrude oil. In addition, natural zeolites
usually contain elevated quantities of Fe species which play an
important role in C–O cleavage,28,29 and therefore it is
expected that the natural zeolites supported Ni catalysts could
exhibit high activity towards the biocrude oil deoxygenation.
In this study, two natural zeolites (NZ and Escott) and two
synthetic (BEA and MOR) zeolites supporting nickel (∼5 wt%)
were investigated. The properties of natural zeolites and cata-
lysts were studied by XRF, FTIR, TPD, TPR, XRD, TEM, XPS,
and chemisorption. Meanwhile, hydrogenation of toluene and
HDO of anisole were performed to examine the hydrogenation
and HDO activities of catalysts.
NS
Nt
SF ꢁ MW ꢁ n ꢁ 102
W wt%
D ¼
ꢀ 100% ¼
ð1Þ
ð2Þ
φ ꢁ MW ꢁ σ
D ꢁ ρ ꢁ NA
dðnmÞ ¼
where Ns is the moles of accessible metal, Nt represents moles
of total metal content, SF is stoichiometry factor (CO/Ni = 1/1),
Mw is formula weight of the particle (g mol−1), n is the mole of
active gas adsorbed, wt% represents the weight% of metal in
sample, W is the total weight of sample, σ is the atomic cross-
sectional area of metal (m2 per atom), NA is the Avogadro
number, φ is a constant which reflects particle shape (e.g. φ is
6 for spheres and cubes), and ρ is the metal particle density
(g cm−3).
The X-ray diffraction patterns of catalysts were obtained in a
Phillips X′pert Pro diffractometer with Cu Kα radiation. The
XRD reflections were recorded in the range of 2θ = 5–90°.
Acid site concentration and the relative strength of acid
sites on each catalyst were determined by temperature-pro-
grammed desorption (TPD) of ammonia, using an apparatus
described in detail elsewhere.32 The heat of desorption was cal-
2. Experimental
2.1. Catalyst preparation
Two natural zeolites NZ (Si/Al = 5.0) and Escott (Si/Al = 5.7) culated based on the procedure published by Redhead:33
were supplied by Zeolite Australia Limited (234 Escott Road,
ΔEdes
RTmax
ν1Tmax
Werris Creek, NSW, Australia). H-BEA (Si/Al = 12.5) and
H-MOR (Si/Al = 10) zeolites were obtained from Zeolyst
International (Conshohocken, US). The Escott and NZ (8 g) raw
materials were ion exchanged with 200 mL of ammonium
nitrate (NH4NO3) (0.5 M) at 60 °C for 24 h to generate acid
sites. This ion exchange procedure was repeated 3 times. The
wet solids were then filtrated and dried at 120 °C overnight
and calcined at 500 °C for 4 h. A pre-determined mass of
nickel nitrate hexahydrate (Ni(NO3)2·6H2O) was dissolved in
appropriate distilled water and then impregnated on NZ,
Escott, BEA and MOR supports. The final mixtures were dried
at 50 °C for 12 h, followed by calcination in air at 500 °C for
4 h. The term ‘pre-reduced’ represents the catalysts which were
reduced by H2 (50 mL min−1) at 500 °C for 3.5 h.
¼ ln
ꢂ 3:64
ð3Þ
β
where ΔEdes is the activation energy of desorption, R is the
ideal gas constant. The frequency factor is estimated to be con-
stant, ν1 = 1013 −1. Tmax is the temperature at which the de-
s
sorption rate is a maximum, β is the heating rate.
H2-TPD experiments were carried out in the same apparatus
as the NH3-TPD measurements. 0.2 g calcined catalyst was
reduced in a pure H2 (50 mL min−1) at 500 °C for 3.5 h. The
pre-reduced sample was vacuumed for 1 h at 500 °C then
cooled to 50 °C and saturated with 10 mbar H2. H2 desorption
patterns were obtained from 50 °C to 450 °C at a heating rate
of 10 °C min−1. The desorbed H2 was detected by a mass
spectrometer.
Transmission electron microscopy (TEM) imaging was
carried out in a JEOL JEM-1200EXII TEM (200 kV) microscope,
2.2. Catalyst characterisation
Nitrogen adsorption–desorption was carried out at −196 °C in equipped with an EDXS spectrometer. Samples were prepared
a Micromeritics Gemini surface area analyser to determine the using an ethanol dispersion method and placed on a copper
surface area and pore properties of catalysts. Prior to nitrogen grid.
adsorption, the samples were degassed using Micromeritics
Temperature-programmed reduction (TPR) was performed
Vac Prep 061 sample preparation device. The t-plot and in a custom-built instrument8 to determine the reducibility of
Barrett–Joyner–Halenda (BJH) models were used to estimate catalyst. The catalyst was first activated in air (20 mL min−1) at
micro and mesopore volumes, as well as the pore size distri- 500 °C for 60 min to remove water and other impurities, fol-
bution. The extent of metal dispersion of each catalyst was lowed by cooling to 50 °C. Reduction was carried out using a
investigated by volumetric chemisorption. All samples were 2.04 vol% H2/Ar flow (50 mL min−1) from 50 °C up to 900 °C
pre-treated in pure H2 (100 mL min−1) at 450 °C for 2 h. The with a heating rate of 10 °C min−1. The hydrogen consumption
4674 | Green Chem., 2021, 23, 4673–4684
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Paper
was recorded by TCD. A trap containing dry ice was placed balanced, 2000 ppm methane, 2000 ppm ethane, 2000 ppm
between the reactor and TCD to remove water. ethene, 2000 ppm propane, 2000 ppm propene, 2000 ppm
FTIR spectra were recorded with a Bruker Tensor 27 FTIR n-butane, 2000 ppm isobutene) was used as calibration gas for
spectrometer. Self-supporting wafers were prepared by appli- the Micro-GC.
cation of a pressure of 5 ton cm−2. The pressed wafer was
placed in a sample holder in the centre of a furnace located in
Toluene conversionðXtolueneÞ ¼
ð4Þ
ð5Þ
ð6Þ
ð7Þ
toluene in feed ꢂ toluene in products
the FTIR cell. Prior to the adsorption experiment, the sample
was first off-line reduced in pure H2 (50 mL min−1) at 500 °C
for 3.5 h. The pre-reduced sample was then ground and
pressed to prepare a wafer, followed by activation at 500 °C by
three repeated injections of hydrogen (10 mbar) per evacuation
cycles and held at 500 °C for 1 h under vacuum (10−7 mbar).
The IR spectra of adsorbed NH3 were recorded at 150 °C with a
NH3 pressure range between 1.0 × 10−4 mbar and 10 mbar. For
quantitative analysis of the number of catalytically active sites,
the intensities of bands of coordinatively bound ammonia
ꢀ 100%
ꢀ 100%
toluene in feed
Anisole conversionðXanisoleÞ ¼
anisole in feed ꢂ ansiole in products
anisole in feed
Product yieldðYÞ ¼
carbon mass of product
ꢀ 100%
carbon mass of feed
Product selectivityðSÞ ¼
carbon mass of product
carbon mass ofðinitial feed ꢂ feed after reactionÞ
(1575–1649 cm−1) and Brønsted acid sites (1340–1553 cm−1
)
were quantified using the respective extinction coefficients:
ꢀ 100%
0.11 cm2 μmol−1 for the NH4 band and 0.026 cm2 μmol−1 for
+
the NH3-L band.34 The IR spectra of adsorbed H2 were
recorded at 50 °C with a H2 pressure range between 1.0 × 10−3
mbar and 10 mbar. H2 and NH3 were desorbed from the
adsorption temperature to 500 °C at a heating rate of 10 °C
The conversion of toluene and anisole was calculated using
eqn (4) and (5). Product yield and selectivity were determined
by eqn (6) and (7). Product yield was calculated on the basis of
the carbon mass.
min−1
.
2.3. Catalytic reaction and product analysis
3. Results and discussion
3.1. Characterisation results
The hydrodeoxygenation reactions were tested in a continuous-
flow reactor. Catalyst was charged in the centre of a tubular
reactor and held in place with quartz wool. A thermocouple The physicochemical properties of supports and pre-reduced
was placed in the middle of the catalyst bed to monitor the catalysts are summarised in Table 1, with nitrogen adsorption/
reaction temperature. The catalysts were reduced at 500 °C for desorption isotherms displayed in Fig. S1.‡ It is noted that the
3.5 h at a flow rate of 50 mL min−1 pure H2 before the reactivity natural zeolites possess lower pore volume and surface area
test. After catalyst pre-treatment, the temperature was adjusted compared to HBEA. The NZ zeolite shows a higher micropore
to reaction temperature and the total hydrogen pressure was surface area compared to Escott, indicating a higher number
increased to 4 MPa. The reactant (anisole or toluene) was of micropores contained by NZ zeolite. The HMOR exhibits the
introduced into the reactor with a HPLC pump.
highest micropore surface area and a similar mesopore surface
The liquid products from HDO of anisole were analysed by area with NZ zeolite. Due to the low concentration of pores,
an Agilent GC-6980 MS via a Rtx-200 MS (30 m length, Escott and NZ supported Ni catalysts display a significantly
0.25 mm ID, and 0.5 µm film thickness) column. Helium was lower metal dispersion and larger Ni particle distribution com-
used as carrier gas with a sample injection volume of 1 µL and pared to Ni/BEA and Ni/MOR. In addition, NZ and Escott zeo-
a split ratio of 100 : 1 applied for all samples. Injector and lites contain a low amount of Fe (∼1.0%).
detector were maintained at 220 °C and 285 °C, respectively.
The X-ray diffraction patterns of supports and pre-reduced
The initial oven temperature (40 °C) was held for 5 min, catalysts are shown in Fig. 1. The metallic Ni (1 1 1), (2 0 0)
increasing to 250 °C at a rate of 10 °C min−1. The filament and (2 2 0) crystal reflections were observed, indicating NiO
detector was turned off during the elution of the injection was converted to Ni species after reduction. Besides, Ni/BEA
solvent (CH2Cl2). The liquid products (methanol, methyl-cyclo- exhibits a larger FWHM compared to Ni/MOR, Ni/NZ and
hexane, cyclohexane, cyclohexene, cyclohexanol, 1,1′-bicyco- Ni/Escott, suggesting the generation of smaller Ni species in
hexyl, 2-methoxycyclohexane, anisole, toluene, and phenol) Ni/BEA, which is consistent with the results obtained from
were calibrated by standard chemicals from Sigma-Aldrich, chemisorption (as summarised in Table 1). The heulandite,
other complex coupling products (notably cyclohexyl-anisole SiO2 and potassium aluminium silicate reflections were also
and methylcyclohexyl-anisole), which have similar structures observed in Escott natural zeolite. In contrast, chabazite, mor-
with 1,1′-bicyclohexyl, were calibrated using the same response denite, SiO2 and sanidine diffraction patterns are detected in
factor as 1,1′-bicyclohexyl. Gaseous products from the continu- NZ natural zeolite. A high intensity of SiO2 reflections was
ous-reactor were analysed via an online Varian 490-GC micro observed in the Escott zeolite compared to NZ, suggesting a
gas chromatograph equipped with two columns (10 m PPQ; high concentration of SiO2 is present in the Escott zeolite
15 m 5CB), and a mixture of gaseous components (helium which leads to a lower micropore volume due to the lack of
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 4673–4684 | 4675

View Article Online
Paper
Green Chemistry
Table 1 Physicochemical properties of the supports and catalysts used in this study
Ni loadinga
(wt%)
Fe contenta
(wt%)
S-micro
S-meso
Vmicro
Vmeso
dmeso
(nm)
Dd
(%)
dNi-chem .
(nm)
b
c
b
c
c
e
Catalyst
(m2 g−1
)
(m2 g−1
)
(cm3 g−1
)
(cm3 g−1
)
Escott
NZ
HBEA
HMOR
Ni/Escott
Ni/NZ
Ni/BEA
Ni/MOR
0
0
0
0
4.5
4.8
5.3
5.2
1.0
1.0
0
10
16
29
175
30
15
28
162
25
0.005
0.09
0.15
0.18
0
0.09
0.15
0.18
0.03
0.05
0.35
0.05
0.03
0.03
0.30
0.04
8.3
7.7
9.0
6.4
9.1
7.3
8.1
5.7
—
—
—
—
4.6
6.8
13.7
9.7
—
—
—
—
13.5
9.0
4.5
6.4
177
269
349
4
167
286
335
0
1.0
1.0
0
0
a By XRF analysis. b From N2 adsorption measurements (t-plot). c From N2 adsorption measurements (BJH method), dmeso = the average diameter
of mesopores. d Metal dispersion, based on CO-chemisorption (eqn (1)). e Average Ni particle size, based on chemisorption (eqn (2)).
Fig. 1 XRD of supports and pre-reduced Ni/Escott, Ni/NZ, Ni/BEA and Ni/MOR catalysts.
micropores in SiO2. NZ zeolite was composed of a high con- species possess a strong interaction with BEA and MOR which
centration of chabazite and mordenite zeolites, which contain allows the electrons transfer from Ni to supports.
a significantly higher concentration of micropores (as shown
in Table 1).
The temperature-programmed desorption of ammonia
(NH3-TPD) and ammonia FTIR spectra for the pre-reduced cat-
HR-TEM experiments were carried out to further investigate alysts are displayed in Fig. 4. As shown in Fig. 4a, Ni/MOR exhi-
the morphology and metal size distribution of pre-reduced Ni/ bits the highest concentration of acid sites, followed by Ni/NZ,
Escott, Ni/NZ and Ni/BEA with results presented in Fig. 2. The Ni/BEA and Ni/Escott. The Ni/MOR and Ni/BEA displayed the
grey areas are the image from the zeolite supports and the majority of NH3 desorption occurs around 405 °C and 285 °C,
black dots indicate the presence of Ni particles. The average respectively, while the natural zeolites (Fig. S2‡) and catalysts
particle size distribution of Ni/Escott and Ni/NZ was found to presented
a
lower desorption temperature (∼250 °C),
range from 6–31 nm. In contrast, the average particle size dis- suggesting the natural zeolites contain weak acid sites with the
tribution of Ni/BEA ranges from 3–11 nm, indicating a signifi- heats of desorption summarised in Table S1.‡ The main de-
cant increase in Ni species distribution.
sorption peak of catalysts can be ascribed to the Brønsted acid
Ni 2p3/2 spectra of the pre-reduced samples were examined sites. To confirm this assumption, IR spectra obtained from
by X-ray photoelectron spectroscopy (XPS) with the profiles adsorption of NH3 on Ni/NZ, Ni/Escott, Ni/BEA and Ni/MOR
presented in Fig. 3. The position of Ni 2p XPS spectra can be were recorded with results displayed in Fig. 4b. The adsorption
influenced by the local chemical environment and shift to of NH3 on all samples exhibits two main bands in the range of
higher binding energy (lower kinetic energy due to strong 1370–1507 cm−1 and 1563–1700 cm−1. The band centred
+
Coulomb interaction of the photoelectron) when the charge around 1450 cm−1 accounts for the vibrations of NH4 , which
density of the Ni decreases.35 The Ni 2p3/2 binding energy of was formed by NH3 molecules reacting with Brønsted acid
Ni/Escott and Ni/NZ was observed at 853.6 eV and 853.4 eV sites and coordinated to the zeolite framework.37,38 Based on
respectively, which is attributed to Ni2+.36 Compared to band areas, the Ni/MOR catalyst possessed the highest concen-
Ni/Escott, the Ni 2p3/2 binding energy of Ni/BEA and Ni/MOR tration of Brønsted acid sites, followed by the Ni/BEA, Ni/NZ
displays a positive shift by 1.2 eV, indicating the smaller Ni and Ni/Escott. The Ni/Escott exhibits the lowest quantity of
4676 | Green Chem., 2021, 23, 4673–4684
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Paper
Fig. 2 TEM images and particle distributions of pre-reduced Ni/Escott, Ni/NZ and Ni/BEA. Note the differing scales used in the HR-TEM Ni/BEA
image (10 nm) compared to the HR-TEM Ni/Escott and Ni/NZ images (20 nm).
which could be attributed to the re-oxidization of part metallic
Ni species during the wafer preparation, forming Lewis cat-
ionic Ni sites.8
The differential NH3-FTIR band areas extracted from
1370–1507 cm−1 (Brønsted acid sites) of Ni/NZ and Ni/BEA as a
function of temperature are plotted in Fig. 4c and d respect-
+
ively. It shows the majority of NH4 vibrations decrease at
246 °C and 290 °C for Ni/NZ and Ni/BEA, respectively, relating
to the main peaks in NH3-TPD (Fig. 4d), suggesting the main
NH3 desorption peaks in NH3-TPD profiles are attributed to
the Brønsted acid sites.
H2-TPR profiles of catalysts are displayed in Fig. 5. Two main
reduction peaks are detected over all catalysts. The low-tempera-
ture H2 consumption peak (centred around 412–440 °C), which
is the dominant reduction peak observed over Ni/Escott catalyst,
can be assigned to the reduction of bulk Ni nanoparticles pre-
sented on the zeolite external surface and some mesopores.8,41
In our previous studies, the low-temperature reduction peak
Fig. 3 Ni 2p3/2 spectra of pre-reduced Ni/Escott, Ni/NZ, Ni/BEA and Ni/
MOR catalysts.
appeared in catalysts with an average Ni particle size larger than
1.6 nm.8 The low-temperature H2 consumption zone (∼400 °C)
was also observed over ZSM-5 and SiO2 supported Ni,42,43 which
contributed to the reduction of NiO nanoparticles possessed
acid sites which could be attributed to its lower concentration weak interaction with support.
of micropores as compared to Ni/MOR, Ni/NZ and Ni/BEA. The
In contrast, the Ni/BEA and Ni/MOR exhibit the main high-
band centred at 1620 cm−1 can be ascribed to symmetric temperature reduction peak (562 °C, 541 °C) relating to
bending vibrations of NH3 coordinated to Lewis acid sites smaller Ni particles that are characterised by strong metal–
(Al3+, Ni2+).39,40 In addition, it is noted the intensity of support interaction.15,43 In addition, the reduction tempera-
vibrations at 1620 cm−1 over pre-reduced Ni/BEA was signifi- ture for the smaller Ni species over Ni/MOR, Ni/BEA and Ni/NZ
cantly enhanced compared to the parent HBEA (Fig. S3‡), catalysts is higher than that of Ni/Escott, which could be attrib-
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 4673–4684 | 4677

View Article Online
Paper
Green Chemistry
Fig. 4 (a) NH3-TPD profiles of pre-reduced samples; (b) infrared spectra were obtained from adsorption of ammonia in pre-reduced Ni/Escott,
Ni/NZ, Ni/BEA and Ni/MOR, samples contacted with 10 mbar pressure of NH3 at 150 °C then were vacuumed; the spectra were normalised for wafer
thickness using overtones of lattice vibrations between 1556 cm−1 and 2105 cm−1
;
differential NH3-FTIR peak area loss of (c) Ni/NZ
(1400–1500 cm−1) and (d) Ni/BEA (1400–1500 cm−1) as a function of temperature (10 °C min−1).
peak around 562 °C even without Ni loading (Fig. S4‡), attribu-
ted to the reduction of Fe oxides contained in natural zeolites
as indicated by XRF.
H2-TPD and H2-FTIR experiments were carried out in order to
investigate the H2 desorption activity of catalyst, with profiles dis-
played in Fig. 6. The original supports do not exhibit desorbed
H2 as shown in Fig. S5,‡ indicating the Fe atoms presented in
the natural zeolites do not possess H2 adsorption activity.
All catalysts present a high concentration of desorbed H2
following Ni loading and subsequent reduction to metallic
species (as indicated in Fig. 6a). The H2-TPD profiles were
fitted by mixed Gaussian–Lorentzian function with the concen-
tration of desorbed H2 over various temperature ranges sum-
marised in Table 2. Ni/BEA and Ni/MOR exhibit a primary H2
desorption peak around 150 °C with a small peak around
212–226 °C. In contrast, a large H2 desorption peak centred
around 211 °C and a small low-temperature desorption peak
(154 °C) are observed over Ni/Escott. The second derivative
function was applied in differentiating the H2-TPD profiles of
the pre-reduced catalysts (as shown in Fig. S6‡). In line with
fitted profiles of the H2-TPD, the Ni/BEA and Ni/MOR display
Fig. 5 H2-TPR profiles of calcined Ni/BEA, Ni/MOR, Ni/NZ and
Ni/Escott catalysts.
uted to the Ni species being formed in the micropores of the majority of desorbed H2 at temperature around 143 °C,
Ni/MOR, Ni/BEA and Ni/NZ engendering a stronger interaction corresponding to the main desorption fitted peak at 150 °C
with support. In contrast, the Ni/Escott has larger pores and (Fig. 6a). Ni/NZ show two major peaks located at 139 °C and
bigger Ni particles, which may reduce the strength of the 231 °C respectively, relating to the two desorption peaks at
metal–support interaction. Escott and NZ both exhibit a small 147 °C and 222 °C in Fig. 6a.
4678 | Green Chem., 2021, 23, 4673–4684
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Paper
Fig. 6 (a) H2-TPD and (b) H2-FTIR profiles of pre-reduced catalysts, the spectra were normalised for wafer thickness using overtones of lattice
vibrations between 1556 cm−1 and 2105 cm−1 in the first spectrum without gas adsorption; (c) the relationship of surface Ni (Ni loading × dispersion)
and the peak area of Ni–H vibrations (peak area was derived from 1500–1700 cm−1).
1500–1700 cm−1) is displayed in Fig. 6c. The concentration of
Ni–H vibrations increases with the increasing surface Ni,
Table 2 Relative quantity of desorbed H2 in different temperature
rangesa
suggesting the Ni–H vibrations are formed mainly depending
Low temperature
High temperature
on the number of surface Ni sites. While Ni/NZ and Ni/Escott
presented a higher amount of high-temperature desorbed H2,
a lower concentration of Ni–H vibrations was observed,
suggesting the H species that were strongly adsorbed by the
bulk Ni did not form the invisible Ni–H vibrations. Since the H
absorbed in bulk Ni is desorbed at very low temperature
(<0 °C),45,48 the high-temperature desorbed H2 could be attrib-
uted to the layer of chemisorbed hydrogen between the Ni par-
ticles and support or remote from the Ni clusters, i.e., hydro-
gen species on the support.45,49 While Ni/MOR exhibits a lower
dispersion and larger average Ni particle size compared to Ni/
BEA, a lower concentration of high-temperature desorbed H2
(as shown in Table 2) was observed, which could be attributed
to the partial blockage of two-dimensional micropores in Ni/
MOR.
(∼150 °C) (a.u.)
(∼210 °C) (a.u.)
Ni/Escott
Ni/NZ
Ni/BEA
Ni/MOR
84
316
226
77
214
424
300
41
a Based on the peak area in H2-TPD profiles (Fig. 6a); the peaks were
fitted by mixed Gaussian–Lorentzian function; 0.2 g pre-reduced
samples were used in the experiments.
The low-temperature desorption peak could be attributed to
hydrogen species weakly chemisorbed on the surface of small
Ni species.44–46 A significantly higher concentration of low-
temperature desorbed H2 was observed in Ni/BEA and Ni/MOR
compared to Ni/NZ and Ni/Escott, which could be attributed to
the higher concentration of surface Ni species. While Ni/Escott
exhibits a lower concentration of desorbed H2 in the low-temp-
erature region, it shows a significantly higher concentration of
3.2. Results of hydrogenation and HDO experiments
3.2.1. Hydrogenation of toluene. Hydrogenation of toluene
high-temperature desorbed H2 compared to Ni/BEA. Besides, over the prepared catalysts was carried out to investigate the
the Ni/NZ displays a higher low-temperature H2 desorption catalyst hydrogenation activity, with performance results sum-
peak area compared to Ni/Escott, which can be attributed to marised in Table 3. The hydrogenation activity of catalyst
the high amount of micropores in NZ facilitating the for- mainly depends on Ni properties (such as metal dispersion
mation of small Ni species which weakly chemisorb the H and particle size), and catalysts with a high metal dispersion
species on the metal surface.
commonly exhibit an improved hydrogenation activity.50,51
The surface H species are also characterised by H2-FTIR as Ni/Escott, however, owing to the lowest Ni dispersion, exhibits
presented in Fig. 6b. The IR band at 1620 cm−1 observed in the highest yield of methylcyclohexane, followed by Ni/NZ, Ni/
pre-reduced Ni catalysts (absent in calcined catalysts) could be BEA and Ni/MOR, indicating a higher hydrogenation activity
attributed to the dissociative chemisorbed H species, e.g. possessed by natural zeolite supported Ni catalysts.
Ni–H.44,47 The Ni/BEA displays the highest amount of surface
While Ni/BEA exhibits a significantly higher concentration
Ni–H vibrations, followed by Ni/MOR, Ni/NZ and Ni/Escott, of surface Ni compared to Ni/NZ and Ni/Escott, a lower methyl-
which is well in line with the peak area sequence of low-temp- cyclohexane formation rate was observed. Ni/BEA and Ni/MOR
erature desorbed H2 (as shown in Fig. 6a). The relationship of show a high concentration of desorbed H2 as well as Ni/NZ
surface Ni concentration (Ni loading × dispersion) and the and Ni/Escott (as presented in Fig. 6 and Table 2), however,
intensity of Ni–H vibrations (peak area derived from the majority of H species were weakly adsorbed on Ni surface
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 4673–4684 | 4679

View Article Online
Paper
Green Chemistry
Table 3 Conversion and yield of hydrogenation of toluenea
sented in Fig. 7 to correlate the desorbed H2 and the reaction
rate. A similar (normalised) methylcyclohexane formation rate
was observed over Ni/Escott, Ni/NZ, Ni/BEA and Ni/MOR when
the methylcyclohexane formation rate was normalised by the
high-temperature desorbed H2 sites, indicating the strong
adsorption/desorption H sites on the larger Ni nanoparticles
play a crucial role in the hydrogenation reactions.
In addition, toluene was not activated by pre-reduced sup-
ports, suggesting the acid sites and Fe species cannot catalyse
the hydrogenation of toluene, and thus Ni species are the
active sites.
b
X-toluene
Y-methylcyclohexane
Y-others
c
Catalyst
(%)
(%)
(%)
r-methylcyclohexane
Ni/Escott
Ni/NZ
Ni/BEA
Ni/MOR
Escott
NZ
21.8
15.2
9.2
4.5
0
0
0
20.9
13.2
4.8
2.2
0
0
0
0.4
1.7
3.8
2.1
0
0
0
6.5
4.1
1.4
0.6
0
0
0
BEA
MOR
0
0
0
0
a Reaction conditions: Catalyst (0.05 g), WHSV (3.05 min−1); PH
2
(4.0 MPa), temperature (220 °C), H2 flow rate (150 mL min−1), liquid
The pore size of supports has a significant impact on the
flow rate (0.153 g min−1). b Others include cyclohexene, ethyl-cyclopen- activity and stability of catalysts in the HDO reactions.52,53
tane, (methylcyclohexyl)-toluene. c r-methylcyclohexane = methylcyclohexane
Catalysts with high concentration of mesopores sometimes
formation rate per gram catalyst per minute, and the rate is an average
over 20–40 min, unit (10−3 molmethylcyclohexane gcat−1 min−1).
provide an enhanced activity for large-size molecules present in
biocrude oils due to a less pronounced diffusion limitation.6,13
In this study, Ni/MOR and Ni/NZ catalysts display a high con-
centration of micropores and few mesopores, while a similar
which were desorbed at a lower temperature (∼150 °C).
Therefore, Ni/BEA and Ni/MOR display a lower methyl-
cyclohexane formation rate at reaction temperature of 220 °C
as the concentration of active H species on charge compensat-
ing Ni was low under reaction conditions. As a result, the low-
temperature H species weakly adsorbed by small Ni particles
play a limited role in the hydrogenation of toluene. In com-
parison, the Ni/Escott displays the majority of desorbed H2 at
a higher temperature (∼211 °C), the active H species remained
on the large Ni particles during the reaction and facilitated the
hydrogenation reactions.
TOF-high T desorbed H was detected over all catalysts, indicating
2
there was no significant diffusion limit for toluene in the micro-
pores. The internal pore diameter of MOR is 6.7 Å,6 which is
larger than the kinetic diameter of toluene (5.68–5.85 Å),54–56
therefore, the entrance and diffusion of toluene in the micro-
pores of MOR is less restricted.
3.2.2. Hydrodeoxygenation of anisole. The natural zeolites
supported Ni catalysts possess a higher hydrogenation activity
compared to Ni/BEA in the hydrogenation of toluene. The
HDO reactions in contrast require metal sites as well as acid
sites for hydrogenation and deoxygenation. Anisole (a model
compound of biocrude oil) was employed to investigate the
HDO activity of catalysts, with performance results summar-
ised in Table 4.
The rate of formation of methylcyclohexane, based on the
number of H2 desorbed sites (low temperature, high-tempera-
ture, and the sum of both) (as summarised in Table 2), is pre-
While methylcyclohexane formation rate over Ni/Escott in
hydrogenation of toluene is 4.6 and 10.8 times higher than
that of Ni/BEA and Ni/MOR, respectively, the cyclohexane for-
mation rate of Ni/Escott in HDO of anisole is only 2.2 and 3.4
times higher than Ni/BEA and Ni/MOR, respectively, indicating
the high concentration of Brønsted acid sites in Ni/BEA and
Ni/MOR facilitates the HDO reaction. The normalised cyclo-
hexane formation rates based on the peak area of high-temp-
erature desorbed H2 (TOF-high T desorbed H ) and concentration
2
of Brønsted acid sites (TOF-B) are depicted in Fig. 8. The Ni/
Escott exhibits the highest TOF-B, followed by Ni/NZ, Ni/BEA
and Ni/MOR, however, this order is reversed when cyclohexane
production rate is normalised by the peak area of high-temp-
erature desorbed H2, suggesting acid and metal sites both
promote the HDO of anisole to cyclohexane.
A significantly enhanced selectivity of methylanisole was
observed in Ni/BEA as compared to Ni/Escott, Ni/NZ and Ni/
MOR. The formation of methylanisole can be attributed to the
Fig. 7 Normalised methylcyclohexane formation rate for Ni/Escott, Ni/
NZ, Ni/BEA and Ni/MOR; TOF-low T desorbed H2 = methylcyclohexane transalkylation of anisole via acid sites, as reported by
literature20,57 and confirmed by the performance results of
pure HBEA (Table S2,‡ phenol and methylanisole were the
formation rate based on the number of low-temperature H2 desorbed
sites, TOF-high T desorbed H = methylcyclohexane formation rate based on
2
the number of high-temperature H2 desorbed sites, TOF-total desorbed H
2
main products). The surface Ni was also reported to facilitate
= methylcyclohexane formation rate based on the sum of low-tempera-
ture and high-temperature H2 desorbed sites, (the number of H2 des-
the transalkylation.19 As a result, the transalkylation of anisole
orbed sites was summarised in Table 2).
to methylanisole and phenol was promoted by the strong
4680 | Green Chem., 2021, 23, 4673–4684
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Paper
Fig. 8 Normalised cyclohexane formation rate for Ni/Escott, Ni/NZ, Ni/
BEA and Ni/MOR catalysts; TOF-B = cyclohexane formation rate based
on the concentration of Brønsted acid sites, TOF-high
=
desorbed H2
T
cyclohexane formation rate based on the peak area of high-temperature
desorbed H2.
Brønsted acid sites and high concentration of surface Ni of Ni/
BEA. The phenol was absent in the products, which could be
attributed to the fast hydrogenolysis of phenol to benzene via
the surface Ni species, and the formed benzene was hydrogen-
ated to cyclohexane via the active H species. The phenol was
more easily activated by zeolite supported Ni catalyst compared
to anisole and toluene in HDO of a mixed feedstock.53
Based on above experiment results and research published
elsewhere, the proposed reaction pathways for the HDO of
anisole over BEA, MOR and natural zeolites supported Ni cata-
lysts are summarised in Scheme 1. The anisole is hydrogen-
ated to methoxycyclohexane (R1) via the active H species, fol-
lowed by the demethoxylation of methoxycyclohexane to cyclo-
hexene via the Brønsted acid sites (R2), which is the main reac-
tion pathway over bifunctional catalysts with Ni nanoparticles
and high H2 pressure.21,58 The Lewis acid sites have a negli-
gible activity in deoxygenation of methoxycyclohexane.13
In addition, the anisole can also undergo transalkylation
via acid sites to form methylanisole and phenol (R3), followed
by the hydrogenolysis of phenol (fast step, R5) and methyl-
anisole (slow step, R6) to benzene and toluene via the charge-
compensating Ni species which contain a low amount of active
H species at 230 °C. The alkylation reactions via the acid sites
was also observed in previous studies.8,20,57,59 The formed
benzene and toluene are hydrogenated to cyclohexene and
methylcyclohexene. Due to the high amount of surface Ni and
high concentration of strong Brønsted acid sites, transalkyla-
tion and hydrogenolysis (R3, R5) are significantly promoted
via the Ni/BEA catalyst, leading to a higher selectivity of
methylanisole compared to Ni/NZ and Ni/Escott (as shown in
Table 4). Benzene was observed to be the main product
without further hydrogenation over Ni-based catalysts in some
previous studies,20,59,60 which could be attributed to the
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 4673–4684 | 4681

View Article Online
Paper
Green Chemistry
Scheme 1 Proposed reaction pathways for HDO of anisole over Ni/Escott, Ni/NZ, Ni/BEA and Ni/MOR.
atmospheric pressure and higher reaction temperature extremely cheaper and abundant compared to synthetic zeo-
(300–450 °C) employed in their studies. High H2 pressure and lites. Therefore, the natural zeolites supported Ni catalysts
mild reaction temperature facilitate the full hydrogenation of have a great potential to be widely used in industrial hydro-
aromatic rings.21,52
treating reactors.
The formed cyclohexene is hydrogenated to cyclohexane (R9)
which is the primary final product. Besides, the cyclohexene
and methylcyclohexene can also react with anisole to form
cyclohexylanisole (R10) and methylcyclohexylanisole (R11) via
the acid sites.8 The formed cyclohexylanisole undergoes two
reaction pathways (R12 and 13: hydrogenation–demethoxylation
and R14 and 15: hydrogenolysis–hydrogenation) to 1,1′-bicyclo-
hexyl based on the Ni species contained in catalysts. The for-
mation of C–C coupling products facilitated by catalysts with
high acidity was also observed by other studies.21,61,62 The pres-
ence of protons is the main reason for the production of coup-
ling products (bicyclohexyl, cyclohexyl-anisole, etc.).53,61,63,64
To summarise, the natural zeolites supported Ni catalysts
exhibit desirable HDO and hydrogenation activities compared
to Ni/BEA and Ni/MOR catalysts. Moreover, natural zeolites are
4. Conclusions
Natural zeolites supported Ni catalysts were employed for the
HDO/hydrogenation reactions. Ni/Escott and Ni/NZ exhibit a
significantly higher hydrogenation activity in the hydrogen-
ation of toluene compared to Ni/BEA and Ni/MOR due to their
high concentration of active H species which were dissociative
adsorbed by bulk Ni particles and desorbed at a higher temp-
erature (210–220 °C). The strongly chemisorbed H species
promote the hydrogenation of toluene at the reaction tempera-
ture of 220 °C. While Ni/BEA and Ni/MOR exhibit a higher
metal dispersion, the smaller Ni species possess a stronger
metal–support interaction, leading to a higher concentration
4682 | Green Chem., 2021, 23, 4673–4684
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Paper
of weakly chemisorbed H species which were desorbed around 15 P. Yan, J. Mensah, A. Adesina, E. Kennedy and
150 °C. Also, the Ni/Escott and Ni/NZ exhibit a higher HDO M. Stockenhuber, Appl. Catal., B, 2020, 267, 118690.
activity compared to Ni/BEA and Ni/MOR in the HDO of 16 C. Zhao and J. A. Lercher, Angew. Chem., 2012, 124, 6037.
anisole due to their high hydrogenation activity. The high con- 17 X. Liu, W. An, C. H. Turner and D. E. Resasco, J. Catal.,
centration of strong Brønsted acid sites and large number of
surface Ni sites in Ni/BEA enhanced the transalkylation of 18 Y. Zheng, N. Zhao and J. Chen, Appl. Catal., B, 2019, 250,
anisole to phenol and hydrogenolysis of phenol to benzene, 280.
followed by the hydrogenation of benzene to cyclohexane, 19 H. Wang, M. Feng and B. Yang, Green Chem., 2017, 19,
2018, 359, 272.
leading to a relatively improved HDO activity compared to its
hydrogenation activity.
1668.
20 X. Xu, E. Jiang, Z. Li and Y. Sun, Fuel, 2018, 221, 440.
21 D. P. Gamliel, B. P. Baillie, E. Augustine, J. Hall,
G. M. Bollas and J. A. Valla, Microporous Mesoporous Mater.,
2018, 261, 18.
22 S. Xu, A. Zheng, Y. Wei, J. Chen, J. Li, Y. Chu, M. Zhang,
Q. Wang, Y. Zhou, J. Wang, F. Deng and Z. Liu, Angew.
Chem., 2013, 125, 11778.
Conflicts of interest
There are no conflicts to declare.
23 G. Bonilla, I. Diaz, M. Tsapatsis, H. K. Jeong, Y. Lee and
D. G. Vlachos, Chem. Mater., 2004, 16, 5697.
24 A. S. M. Junaid, C. Street, W. Wang, M. M. Rahman, W. An,
W. C. McCaffrey and S. M. Kuznicki, Fuel, 2012, 94, 457.
25 R. H. Ibrasheva and K. A. Zhubanov, Stud. Surf. Sci. Catal.,
2000, 130, 2447.
26 J. E. Sánchez-Velandia, J. F. Gelves, L. Dorkis,
M. A. Márquez and A. L. Villa, Microporous Mesoporous
Mater., 2019, 287, 114.
Acknowledgements
Authors thanks the Zeolite Australia Limited for providing the
natural zeolites. Funding from the Australian Research
Council is gratefully acknowledged. Authors thank the
Electron Microscope and X-ray Unit of the University of
Newcastle for the use of their facilities.
27 O. Akpolat, G. Gunduz, F. Ozkan and Nu. Besün, Appl.
Catal., A, 2004, 265, 11.
References
1 C. Liu, H. Wang, A. M. Karim, J. Sun and Y. Wang, Chem. 28 Q. Han, M. U. Rehman, J. Wang, A. Rykov, O. Y. Gutierrez,
Soc. Rev., 2014, 43, 7594.
2 Y. Mathieu, L. Sauvanaud, L. Humphreys, W. Rowlands,
Y. Zhao, S. Wang, X. Ma and J. A. Lercher, Appl. Catal., B,
2019, 253, 348.
T. Maschmeyer and A. Corma, ChemCatChem, 2017, 9, 29 H. Fang, J. Zheng, X. Luo, J. Du, A. Roldan, S. Leoni and
1574. Y. Yuan, Appl. Catal., A, 2017, 529, 20.
3 D. D. Laskar and B. Yang, Biofuels, Bioprod. Biorefin., 2013, 30 S. Velu and S. K. Gangwal, Solid State Ionics, 2006, 177, 803.
7, 602.
31 A. Borodzinski and M. Bonarowska, Langmuir, 1997, 13,
4 D. R. Vardon, B. K. Sharma, J. Scott, G. Yu, Z. C. Wang,
5613.
L. Schideman, Y. H. Zhang and T. J. Strathmann, Bioresour. 32 M. L. M. Bonati, R. W. Joyner and M. Stockenhuber, Catal.
Technol., 2011, 102, 8295. Today, 2003, 81, 653.
5 H. Wang, H. Ruan, H. Pei, H. Wang, X. Chen, M. P. Tucker, 33 P. A. Redhead, Vacuum, 1962, 12, 203.
J. R. Cort and B. Yang, Green Chem., 2015, 17, 5131.
6 P. Yan, G. Bryant, M. M. J. Li, J. Mensah, E. Kennedy and
M. Stockenhuber, Microporous Mesoporous Mater., 2020,
309, 110561–110574.
7 E. Furimsky, Appl. Catal., A, 2000, 199(2), 147.
8 P. Yan, M. M. J. Li, A. A. Adesina, G. Zhao, A. Setiawan,
E. M. Kennedy and M. Stockenhuber, Catal. Sci. Technol.,
2020, 10, 810.
9 W. Song, Y. Liu, E. Barath, C. Zhao and J. A. Lercher, Green
Chem., 2015, 17, 1204.
10 Z. Yang, W. Chen, J. Zhang, Z. Yang, N. Zhang, C. Zhong
and B. H. Chen, J. Catal., 2018, 363, 52.
11 M. S. Zanuttini, C. D. Lago, M. S. Gross, M. A. Peralta and
C. A. Querini, Ind. Eng. Chem. Res., 2017, 56, 6419.
12 L. Nie and D. E. Resasco, J. Catal., 2014, 317, 22.
13 P. Yan, J. Mensah, M. Drewery, E. Kennedy, T. Maschmeyer
and M. Stockenhuber, Appl. Catal., B, 2021, 281, 119470.
14 L. Zhou and A. Lawal, Catal. Sci. Technol., 2016, 6, 1442.
34 K. Gora-Marek, K. Brylewska, K. A. Tarach, M. Rutkowska,
M. Jablonsa, M. Choi and L. Chmielarz, Appl. Catal., B,
2015, 179, 589.
35 O. Mazaheri and R. J. Kalbasi, RSC Adv., 2015, 5, 34398.
36 M. O. de Souza, F. M. T. Mendes, R. F. de Souza and
J. H. Z. dos Santos, Microporous Mesoporous Mater., 2004,
69, 217.
37 J. Valyon, G. Onyestyak and L. V. C. Rees, J. Phys. Chem. B,
1998, 102, 8994.
38 F. Yin, A. L. Blumenfeld, V. Gruver and J. J. Fripiat, J. Phys.
Chem. B, 1997, 101, 1824.
39 A. Martinez, M. A. Arribas, P. Concepcion and S. Moussa,
Appl. Catal., A, 2013, 467, 509.
40 C. Minchev, S. A. Zubkov, V. Valtchev, V. Minkov,
N. Micheva and V. Kanazirev, Appl. Catal., A, 1994, 119,
195.
41 J. van de Loosdrecht, A. M. van der Kraan, A. J. van Dillen
and J. W. Geus, J. Catal., 1997, 170, 217.
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 4673–4684 | 4683

View Article Online
Paper
Green Chemistry
42 A. J. Maia, B. Louis, Y. L. Lam and M. M. Pereira, J. Catal., 54 J. F. Denayer, W. Souverjins, P. A. Jacobs, J. A. Martens and
2010, 269, 103. G. V. Baron, J. Phys. Chem. B, 1998, 102, 458.
43 M. Che, Z. Cheng and C. Louis, J. Am. Chem. Soc., 1995, 55 J. Jae, G. A. Tompsett, A. J. Foster, K. D. Hammond,
117, 2008.
S. M. Auerbach, R. F. Lobo and G. W. Huber, J. Catal., 2011,
279, 257.
44 S. T. Ceyer, Acc. Chem. Res., 2001, 34, 737.
45 S. Ewald, S. Standl and O. Hinrichsen, Appl. Catal., A, 2018, 56 H. Wang and M. Frenklach, Combust. Flame, 1994, 96, 163.
549, 93.
57 X. Zhu, L. L. Lobban, R. G. Mallinson and D. E. Resasco,
46 L. Znak and J. Zielinski, Langmuir, 2006, 22, 8758.
J. Catal., 2011, 281, 21.
47 N. A. Eberhardt and H. Guan, Chem. Rev., 2016, 116, 58 Y. Yang, C. Ochoa-Hernandez, V. A. de La Pena O’Shea,
8373.
P. Pizarro, J. M. Coronado and D. P. Serrano, Appl. Catal., B,
2014, 145, 91.
59 S. Pichaikaran and P. Arumugam, Green Chem., 2016, 18,
2888.
60 W. S. Lee, Z. Wang, W. Zheng, D. G. Vlachos and A. Bhan,
Catal. Sci. Technol., 2014, 4, 2340.
61 L. Chen, J. Xin, H. Dong, D. Yan, X. Lu and S. Zhang, Green
Chem., 2016, 18, 2341.
48 S. Smeds, T. Salmi, L. P. Lindfors and O. Krause, Appl.
Catal., A, 1996, 144, 177.
49 J. T. Miller, B. L. Meyers, M. K. Barr, F. S. Modica and
D. C. Koningsberger, J. Catal., 1996, 159, 41.
50 M. W. Schreiber, D. Rodriguez-Nino, O. Y. Gutiérrez and
J. A. Lercher, Catal. Sci. Technol., 2016, 6, 7976.
51 W. Song, C. Zhao and J. A. Lercher, Chem. – Eur. J., 2013,
19, 9833.
62 E. Furimsky and F. E. Massoth, Catal. Today, 1999, 52, 381.
52 C. Tu, J. Chen, W. Li, H. Wang, K. Deng, V. A. Vinokurov 63 D. Y. Hong, S. J. Miller, P. K. Agrawal and C. W. Jones,
and W. Huang, Sustainable Energy Fuels, 2019, 3, 3462. Chem. Commun., 2010, 46, 1038.
53 P. Yan, M. Drewery, J. Mensah, J. C. Makie, E. Kennedy and 64 R. K. M. R. Kallury, T. T. Tidwell, D. G. B. Boocock and
M. Stockenhuber, Top. Catal., 2020, 63, 778. D. H. L. Chow, Can. J. Chem., 1984, 62, 2540.
4684 | Green Chem., 2021, 23, 4673–4684
This journal is © The Royal Society of Chemistry 2021