CeO2 promoting allyl alcohol synthesis from glycerol direct conversion over MoFe/CeO2 oxide catalysts: morphology and particle sizes dependent

Article abstract of DOI:10.1007/s11164-018-3694-4

MoFe-N, MoFe/c–CeO₂, MoFe/p₁–CeO₂, and MoFe/p₂–CeO₂ (where N, c, and p stand for non-supported, nanocube, and nanoparticle) oxide catalysts were designed for gas-glycerol direct catalytic conversion into allyl alcohol. The catalysts also were characterized by XRD, TEM, BET, H₂-TPR, and NH₃-TPD. Mo–Fe oxides were highly dispersed on the surface of c-CeO₂ and p-CeO₂ supports, different with the MoFe-N consist of crystalline Fe₂(MoO₄)₃ and Fe₂O₃ crystalline phase. The support effect and special natural property of CeO₂ significantly improve the allyl alcohol selectivity from gas-glycerol over MoFe/CeO₂. The p-CeO₂ with low particle sizes and crystalline degree was superior to high-crystalline nanocube c-CeO₂ to promote its interaction with the MoFe oxide active components, and improve the surface acid site concentration and reducibility of MoFe/CeO₂ as well as catalytic activity and stability for allyl alcohol synthesis from gas-glycerol without any extra hydrogen donors. Over the MoFe/p₂–CeO₂, the glycerol conversion reached 97.1%, and the selectivity of allyl alcohol, enthanal, propanoic acid, and acrylic acid were 23.3%, 8.6%, 12.6%, and 7.8%, respectively, yielding allyl alcohol of 22.6%.

Full text of DOI:10.1007/s11164-018-3694-4

Research on Chemical Intermediates
https://doi.org/10.1007/s11164-018-3694-4
CeO promoting allyl alcohol synthesis from glycerol direct
2
conversion over MoFe/CeO oxide catalysts: morphology
2
and particle sizes dependent
Hai Lan1,2 · Jia Zeng · Biao Zhang · Yi Jiang
3
2
2
Received: 1 October 2018 / Accepted: 28 November 2018
©
Springer Nature B.V. 2018
Abstract
MoFe-N, MoFe/c–CeO , MoFe/p –CeO , and MoFe/p –CeO (where N, c, and p
2
1
2
2
2
stand for non-supported, nanocube, and nanoparticle) oxide catalysts were designed
for gas-glycerol direct catalytic conversion into allyl alcohol. The catalysts also were
characterized by XRD, TEM, BET, H -TPR, and NH -TPD. Mo–Fe oxides were
2
3
highly dispersed on the surface of c-CeO and p-CeO supports, diꢀerent with the
2
2
MoFe-N consist of crystalline Fe (MoO ) and Fe O crystalline phase. The support
2
4 3
2
3
eꢀect and special natural property of CeO signiꢁcantly improve the allyl alcohol
2
selectivity from gas-glycerol over MoFe/CeO . The p-CeO with low particle sizes
2
2
and crystalline degree was superior to high-crystalline nanocube c-CeO to promote
2
its interaction with the MoFe oxide active components, and improve the surface acid
site concentration and reducibility of MoFe/CeO as well as catalytic activity and
2
stability for allyl alcohol synthesis from gas-glycerol without any extra hydrogen
donors. Over the MoFe/p –CeO , the glycerol conversion reached 97.1%, and the
2
2
selectivity of allyl alcohol, ethanol, propanoic acid, and acrylic acid were 23.3%,
8
.6%, 12.6%, and 7.8%, respectively, yielding allyl alcohol of 22.6%.
Keywords Mo–Fe oxides · Allyl alcohol · Glycerol · CeO support eꢀect · Biomass
2
resource
*
Jia Zeng
LH602388274@126.com
1
2
3
Department of Applied Chemistry, School of Materials Science and Engineering, Chongqing
JiaoTong University, Chongqing 400074, People’s Republic of China
Chengdu Institute of Organic Chemistry, Chinese Academy of Science, Chengdu 610041,
People’s Republic of China
Chongqing Academy of Environmental Science, Chongqing 401147, People’s Republic of China
Vol.:(012
1
3456789)
3

H. Lan et al.
Introduction
Nowadays, depletion of fossil fuel resources and the global warming problem
have promoted the development of renewable energies (biomass and solar energy,
etc.) all over the world, [1–3]. Typically, biodiesel, as a successful utilization of
biomass, has been produced on a large scale during the past decade, and the pro-
duction has continually increased [4, 5]. The most widely adopted biodiesel pro-
duction method is transesteriꢁcation of triglycerides obtained from vegetable oils
and animal fats; however, glycerol is produced as a main byproduct at roughly a
tenth the mass of the biodiesel [6, 7]. The chemocatalytic technologies to convert
glycerol into high-value chemicals receive great attention as an eꢀective way to
deal with glycerol, which has created an international research hotspot in recent
years [8, 9]. Glycerol comprises a versatile starting material for the preparation of
various compounds possessing a C3 backbone, such as 1,2- or 1,3-propanediol,
acrolein, dihydroxyacetone, and so on [10].
Inspiringly, a few works have focused on allyl alcohol, which is used as a
large-scale industrial chemical intermediate, promoting additions, substitutions,
decomposition, oxidation, re-arrangement, and polymerisation reactions based on
this commodity to produce resins, paints, coatings, silane coupling agents, and
polymer crosslinking agents [11]. The basic motivation is that the current indus-
trial processes of allyl alcohol production are fossil fuel-based routes, involving
hydrolysis of allyl chloride (from propylene), Meerwein–Pondorf reduction of
acrolein with isopropanol, and hydrolysis of allyl acetate (from propylene react-
ing with acetic acid) [12, 13]. Synthesis of allyl alcohol from glycerol shows
potential signiꢁcance on energy and environmental aspects.
Recently, reported chemocatalytic conversion of glycerol into allyl alcohol
involved two eꢀective ways. One is in the presence of sacriꢁcial additives (extra
hydrogen donors), such as formic acid-assisted liquid-glycerol DODH (deoxyde-
hydration), H -sacriꢁced liquid-glycerol DODH over a heterogeneous ReO -Au/
2
x
CeO catalyst, and H -sacriꢁced gas-glycerol dehydration–hydrogenation over
2
2
the bifunctional Ag/ZSM-5 catalysts, which, however, showed obvious drawbacks
of requiring a special solvent, high pressure, prolonged reaction time, low conver-
sion or low selectivity [14–16]. The other is gas-glycerol direct conversion into
allyl alcohol without any hydrogen donors, ꢁrst reported by Liu et al., that about
2
4% of allyl alcohol yielded from glycerol (Con. 99%) over the FeO catalysts
x
[
17]. Attracted by advantages of low raw-material cost, high-eꢂciency simple
reaction, and mild reaction condition et al., Fe-based oxide catalysts were further
developed for gas-glycerol direct conversion into allyl alcohol, such as Fe/Al O ,
2
3
γ-Al O /(Fe O ) , ZrO –FeO , and H-ZSM5/Fe [18–20]. However, the selec-
2
3
2
3 0.16
2
x
tivity of allyl alcohol was 10–15%, which can be obviously improved to above
2
0% over the catalysts after adding base metals (K, Rb etc.) to reduce surface
acidic property, such as ZrO (7%)-FeO /Rb, K/ZrO -FeO , γ-Alumina/Fe/Rb, and
2
x
2
x
H-ZSM5/Fe/Rb [20–22]. The moderate-acid sites and redox sites both are key
active centers for this reaction, which are important for preparing new catalysts
with high activity, selectivity and stability [17–24].
1
3

CeO promoting allyl alcohol synthesis from glycerol direct…
2
Following previous studies that bimetal MoFe oxides having higher catalytic per-
formance than single Mo or Fe oxides for gas-glycerol conversion to allyl alcohol
and the CeO without high speciꢁc surface area and porous structure also showing
2
relatively excellent support eꢀect comparing with alkaline MgO, acidic ZrO , and
2
neutral CNTs,, this work aims to prepared MoFeO /CeO bimetal oxides catalysts
x
2
with diꢀerent nano-morphologies and particle sizes for gas-glycerol direct catalytic
conversion to allyl alcohol [23, 24]. The catalysts design was inspired by CeO has
2
redox sites and can catalyze the glycerol conversion to produce intermediate-radical
hydrogen ([H·]), and ꢁnally obtaining methanol, which is diꢀerent with the SiO
2
and Al O supports and may beneꢁt allyl alcohol production [25]. At the same time,
2
3
physicochemical property and catalytic performance of CeO (or metal/CeO cata-
2
2
lysts) are always dependent on the morphology and nanoparticle sizes [26–29]. The
prepared MoFe/CeO oxide catalysts are also characterized by using X-ray diꢀrac-
2
tion (XRD), Brunauer–Emmett–Teller (BET), transmission electron microscopy
(
TEM), NH -temperature programmed desorption (NH -TPD), and H -temperature
3
3
2
programmed reduction (H -TPR) analyses.
2
Experimental
Catalyst preparation
The nanocube c-CeO was prepared by hydrothermal methods followed the Mai
2
et al.’s recipe [24]. Typically, 1.96 g Ce(NO ) ·6H O and 16.88 g NaOH was dis-
3
3
2
solved in 40 mL and 30 mL deionized water, respectively. The NaOH solution was
added dropwise into the Ce(NO ) solution under stirring at room temperature. The
3
3
mixed solution was adequately stirred for an additional 30 min at room temperature
and then transferred into a 100-mLTeꢃon bottle. The Teꢃon bottle was tightly sealed
and hydrothermally treated in a stainless-steel autoclave at 180 °C for 24 h. After
cooling, the obtained precipitate was collected, washed with distilled water, dried at
1
00 °C for 24 h, and ꢁnally calcined at 500 °C for 4 h in air.
The nanoparticles CeO was prepared by precipitation methods, using dilute
2
ammonia water and NaOH solution (0.5 mol/L) as the precipitates, respectively.
3
.0 g Ce(NO ) ·6H O were dissolved in 100 mL deionized water. The NaOH solu-
3 3 2
tion (0.5 mol/L) was then added dropwise in the Ce(NO ) solution under stirring,
3
3
until the pH of the solution reached 9.0. The obtained precipitate was thoroughly
+
washed with distilled water in ordered to remove the undesired Na ions, and sub-
sequently dried overnight in an oven at 100 °C for 24 h. The dried sample was cal-
cined at 500 °C for 4 h, and the obtained CeO sample is denoted as p -CeO . The
2
1
2
sample, prepared by following the same process using the dilute ammonia water as
the precipitate, denoted as p -CeO .
2
2
MoFe/CeO catalysts were prepared by simple wet impregnation. Brieꢃy,
2
a certain amount of (NH ) Mo O ·4H O and Fe(NO ) ·9H O (n /n =0.3;
4
6
7
24
2
3 2
2
Mo Fe
ꢁ
nal Mo–Fe oxide loading of 5 wt%) were dissolved in distilled water, followed
by adding several drops of HNO (67%). Then, a corresponding amount of CeO
3
2
(
c-CeO , p -CeO or p -CeO ) was added to the solution and stirring for 2 h at room
2
1
2
2
2
1
3

H. Lan et al.
temperature. After staying for 24 h, the solution was spin-ꢃash evaporated at 80 °C
until the solvent being completely removed. Through drying overnight, the sample
was calcinated at 500 °C for 3 h. Finally, the catalyst was obtained and designated as
MoFe/c–CeO , MoFe/p –CeO , and MoFe/p –CeO , respectively. The unsupported
2
1
2
2
2
MoFe composite oxides catalyst (denoted as MoFe-N) was prepared following the
same process without adding CeO .
2
Catalyst characterization
XRD characterization
X-ray diꢀraction (XRD) patterns were recorded on a Fangyuan DX-1000 powder
X-ray diꢀractometer (China) with Cu Kα radiation at a tube voltage of 40 kV, and
the data of 2θ were collected from 10° to 80° range with 0.02° of the step size at the
−1
rate of 5° min . Low-angle XRD patterns were recorded on a Rigaku D/Max–2500/
PC diꢀractometer with a rotating anode using Ni ꢁltered Cu–Kα (as radiation source
(
λ=0.15418 nm) radiation at 40 kV of a tube voltage and 200 mA of a tube current.
The data were corrected from 0.5° to 5° with 0.02° increment and recording time of
00 s at each increment.
1
BET characterization
Adsorption and desorption isotherms were collected on Autosorb-6 at −196 °C.
Prior to the measurement, all samples were degassed at 200 °C until a stable vac-
uum of ca. 5 m Torr was reached. The speciꢁc surface area was assessed using the
Brunauer–Emmett–Teller (BET) method from adsorption data in a relative pressure
range from 0.06 to 0.10.
TEM analysis
2
Transmission electron microscopy (TEM) was performed using a Tecnai G Spirit
microscope operating with an acc3eleration voltage of 120 kV. For the TEM meas-
urement, the samples were prepared by ultrasonication in ethanol, evaporating a
drop of the resultant suspension onto a carbon–coated copper grid.
H2‑TPR studies
H temperature programmed reduction (H –TPR) was conducted using a conven-
2
2
tional apparatus equipped with a TCD detector. 25 mg of sample was placed in a
quartz tube (4.0 mm ID), and pretreated at 250 °C for 30 min in a pure N ꢃow of
2
−
1
3
0 mL min . After the temperature cooled down, H –TPR was performed by heat-
2
−1
ing the samples at 10 °C min from 50 to 700 °C in a 5% H –N mixture ꢃowing at
2
2
−1
3
0 ml·min .
1
3

CeO promoting allyl alcohol synthesis from glycerol direct…
2
NH ‑TPD studies
3
NH temperature programmed desorption (NH –TPD) was conducted using a con-
3
3
ventional apparatus equipped with a TCD detector. 50 mg of sample was placed in a
quartz tube (4.0 mm ID), and pretreated with gas He (30 mL/min) at 300 °C for 1 h;
as the temperature cooled down to 50 °C, NH was chemisorbed until equilibrium
3
−
1
was reached. Then TPD was performed by heating the samples at 10 °C min from
−1
5
0 to 600 °C in gas He ꢃowing at 30 mL min .
Catalytic activity measurement
Catalytic tests of all catalysts were carried out in a continuous ꢃow ꢁxed–bed quartz
tubular reactor (i.d. 8 mm) under atmospheric pressure. The quartz tube reactor con-
taining 200 mg of catalysts mixed with 200 mg of quartz (both sieved, 380–700 μm)
was placed inside a tubular furnace. The carrier gas (N ) was passed downward
2
through the reactor containing the catalyst bed, while electronic mass ꢃow controller
−
1
(
D07–7A/ZM, China) was used to control the ꢃow rate at 10 mL min . After the
temperature of the reactor had reached 340 °C, a 35 wt% glycerol aqueous solution
−1
was pumped in at a ꢃow rate of 1 mL h using a syringe pump (LSP01-1A, china).
The products were collected in every 2 h using ethanol cooled by an ice-water trap,
and the obtained product yields showed the average of each 2 h. The collected liq-
uid samples were analyzed with a HP 5890 Series II gas chromatograph equipped
with a HP-INNOWAX (19091 N-113, 30 m*0.32 mm*0.25 mm) and an FID detec-
tor. 1,2-propanediol was selected as the internal standard for the quantiꢁcation of
glycerol and products. In addition, all samples were also checked without standard
and no 1,2-propanediol was detected in all cases. The product selectivity was calcu-
lated with mol-based method, and the products yields were calculated based on the
amount of glycerol fed to the reactor.
Results and discussion
XRD characterization
Figure 1 presents the XRD patterns of the prepared MoFe-N and MoFe/CeO
2
catalysts, in which, MoFe-N showed a series of typical XRD peaks attributed to
Fe (MoO ) (PDF#97-010-0606) and Fe O (PDF#00-033-0664). Diꢀerently,
2
4 3
2
3
for MoFe/CeO catalysts, it can be only observed similar prominent XRD peaks
2
occurred at 2θ=28.56°, 33.12°, 47.46°, 56.32°, 59.14°, 69.44°, 76.66°, and 79.14°,
which are associated with the ꢃuorite-like CeO crystalline phase [26–30]. The
2
absence of Mo and/or Fe oxide XRD peaks for MoFe/CeO catalysts conꢁrmed
2
that Mo and/or Fe oxides were highly-dispersed on the surface of CeO support or
2
under the detection limit of XRD resulting from the low mass loading of Mo and
Fe oxides, which may enhance the interaction between the Mo–Fe oxides and CeO
2
1
3

H. Lan et al.
MoFe / c-CeO2
MoFe / p -CeO
1
2
MoFe / p -CeO
2
2
MoFe-N
Fe (MoO ) PDF # 97-010-0606
2
4 3
Fe O
PDF # 00-033-0664
2
3
2
0
25 30 35 40 45 50 55 60 65 70 75 80
ο
2θ /
Fig. 1 The wide-angle XRD patterns of the MoFe-N and MoFe/CeO catalysts
2
[
27–29]. Moreover, it can be observed the CeO XRD peaks of MoFe/p –CeO and
2 1 2
MoFe/p –CeO are much lower in intensity than that of the MoFe/c–CeO . The low
2
2
2
degree crystalline of p-CeO (supports) usually existed many crystalline defects and
2
exhibit diꢀerent physico-chemical property comparing well to crystalline c-CeO , so
2
as to aꢀect the surface acidity and reducibility of MoFe/CeO [29, 30].
2
TEM characterization
The morphology and particle sizes of the MoFe oxides supported on the CeO pre-
2
pared by diꢀerent methods were examined by TEM, and the micrographs of the
obtained nano-oxide samples were shown in Fig. 2.
The MoFe/c–CeO , of which the CeO prepared through hydrothermal methods,
2
2
exhibited nanocube shapes with uniform particle sizes (about 25 nm) and regular
shapes, in accord with the previous literatures [26–30]. In contrast, the CeO pre-
2
pared by precipitation methods gave the MoFe/p –CeO and MoFe/p –CeO with
1
2
2
2
irregular particle shapes. Diꢀerently, the particle sizes MoFe/p –CeO were dis-
1
2
tributed in the range of 11–23 nm, lager than that of MoFe/p –CeO distributed
2
2
in diameter range of 7–17 nm, which can be attributed to the diꢀerent precipitants
−
of NaOH and NH ·H O with diꢀerent OH release rates. The diꢀerent shapes and
3
2
2
particle sizes of CeO well revealed CeO prepared by hydrothermal and precipi-
2
tation methods having diꢀerent degree of crystallinity (XRD results in Fig. 1). At
the same time, the speciꢁc surface areas of the MoFe/c–CeO , MoFe/p –CeO and
2
1
2
MoFe/p –CeO were also characterized and calculated through the BET equation
2
2
2
(
as shown in Table 1), which were 23.2, 20.5, and 41.8 m /g, respectively. The mor-
phology, particle sizes, and speciꢁc surface area of the CeO -based catalysts were
2
always key factors aꢀecting adsorption and activation of reactants for gas–solid het-
erogeneous catalytic reaction [27–30].
1
3

CeO promoting allyl alcohol synthesis from glycerol direct…
2
Fig. 2 The TEM images of the MoFe/c–CeO (a), MoFe/p –CeO (b), and MoFe/p –CeO (c) oxide cat-
2
1
2
2
2
alysts
Table 1 The physico-chemical
2
Sample
SBET (m /g) Acid site concen- Acid site
property data of MoFe-N and
tration (mmol/g) density (µmol/
MoFe/CeO catalysts
2
2
m )
MoFe-N
MoFe/c–CeO2
6.27
23.2
20.5
0.134
0.148
0.234
0.339
0.214
0.064
0.114
0.081
MoFe/p –CeO
1
2
MoFe/p –CeO
41.8
2
2
H2‑TPR study
The H -TPR test was conducted on the prepared MoFe-N and MoFe/CeO cat-
2
2
alysts, and the H -TPR proꢁles are shown in Fig. 3. It can be observed from
2
1
3

H. Lan et al.
5
08
6
37
6
MoFe / c-CeO2
MoFe / p -CeO
457
68
1
2
421
MoFe / p -CeO
2
2
5
51
MoFe-N
1
00 150 200 250 300 350 400 450 500 550 600 650 700
o
Temperature / C
Fig. 3 The H -TPR spectrum for the prepared CeO catalysts
2
2
Fig. 3, that the H -TPR curve of MoFe-N deviated from the baselines (as shown
2
by the arrow) at approximately 470 °C, and formed two multiple H consump-
2
tion peaks with the lower-temperature one centered at about 550 °C. However,
The MoFe/CeO shifted the H -TPR peaks to lower reaction temperatures. The
2
2
MoFe/c–CeO showed the H -TPR curve with two peaks centered at 508 °C and
2
2
6
37 °C, respectively. The H -TPR curve of MoFe/p –CeO have two H consump-
2 1 2 2
tion peaks centered at 457 °C and 668 °C, respectively. For MoFe/p –CeO , the
2
2
H reducing reaction started at 260 °C, and formed an obvious peak centered at
2
about 420 °C in the H -TPR curve.
2
H -TPR study is an important technology for metal oxides catalysts to infer the
2
dispersion and reducibility of reducible species from the temperatures, intensities
and areas of the TPR hydrogen consumption peaks [31, 32]. It should be pointed
out that Fe O has reduction temperature greater than 400 °C following a two-step
2
3
reduction process: Fe O →Fe O →FeO (lower than 800 °C), and the composite
2
3
3
4
Fe–Mo oxides such as Fe (MoO ) and FeMoO are more stable and diꢂcult to be
2
4 3
4
reduced towards individual Fe oxides [33–36]. Thus, according to the XRD results,
the H -TPR peak of MoFe-N at 550 °C is mainly attributed to the reduction of Fe O
2
2
3
species [33, 34]. For the MoFe/CeO catalysts, the shift of the H -TPR peaks to low
2
2
reduction temperatures was beneꢁt from the CeO and Mo–Fe oxides being well dis-
2
persed on the surface of CeO . But, it is not easy to clearly attribute the H reduction
2
2
peaks to the concise metal oxides (CeO or Fe O ) because the CeO has a relatively
2
2
3
2
high reduction temperature the same as Fe O [29, 30]. For this, the MoFe/p –CeO
2
3
1
2
and MoFe/p –CeO have lower reduction temperatures than MoFe/c–CeO , which
2
2
2
can be ascribed to the p -CeO and p -CeO naturally having lower reduction tem-
1
2
1
2
peratures than the c-CeO because of the smaller particles and lower crystallinity of
2
p-CeO (conꢁrmed by XRD and TEM). Moreover, the p -CeO and p -CeO with
2
1
2
1
2
small particles and low crystallinity can also promote the reduction of the surface
1
3

CeO promoting allyl alcohol synthesis from glycerol direct…
2
loading Fe and Mo oxides, so that the MoFe/p –CeO and MoFe/p –CeO have low
1
2
2
2
reduction temperatures, especially for the MoFe/p –CeO [28–30].
2
2
In terms of the reduction temperature results in Fig. 3, the prepared catalysts can
be ranked as MoFe/p –CeO >MoFe/p –CeO >MoFe/c–CeO >MoFe-N, follow-
2
2
1
2
2
ing the reducibility from high to low. The diꢀerent reducibility of MoFe/CeO oxide
2
catalysts demonstrated that the catalysts provided redox sites with diꢀerent catalytic
oxidability at a certain reaction temperature for glycerol converting to ethylene gly-
col radical or glycerol (or intermediate products) being oxidized to CO , which may
x
have obvious eꢀect on the products distribution of the glycerol catalytic conversion
[
17, 22, 24].
NH ‑TPD
3
The surface acid sites of catalysts seem to be the initial active centers for the glyc-
erol molecules in gas–solid catalytic conversion reaction [17–22]. To investigate
the surface acidity and acid strength, NH -temperature programmed desorption
3
(
NH -TPD) experiments were carried out on the MoFe-N and MoFe/CeO cata-
3
2
lysts, and the results were shown in Fig. 4. As shown in Fig. 4, the NH -TPD curves
3
of the MoFe-N, deviate from the baseline (as shown by the arrow) at about 83 °C,
and formed a prominent NH -TPD peak centered at about 117 °C. The MoFe/CeO
3
2
catalysts shifted the initial NH desorption temperatures to above 100 °C, with the
3
curves forming multiple NH -TPD peaks in the range of 125–260 °C. The MoFe/
3
p –CeO has NH -TPD peaks with relatively higher desorption temperature than
2
2
3
the MoFe/p –CeO and MoFe/c–CeO . The occurrence of clear NH -TPD peaks
1
2
2
3
in the range of 100–300 °C indicates the surface of the MoFe-N and MoFe/CeO
2
catalysts only hold moderate (weak) acidic property, though the acidity of these
catalysts having diꢀerent strength following the order: MoFe/p –CeO >MoFe/
2
2
MoFe / c-CeO2
MoFe / p -CeO2
1
MoFe / p -CeO2
2
MoFe-N
50
100
150
200
250
300
350
400
o
Temperature / C
Fig. 4 The NH -TPD curves of the MoFe-N and MoFe/CeO oxide catalysts
3
2
1
3

H. Lan et al.
p –CeO –MoFe/c–CeO >MoFe-N [21–24]. The acid sites of the MoFe-N were
1
2
2
provided by the Fe O and Fe (MoO ) in terms of the XRD result, while the sur-
2
3
2
4 3
face acid sites of the MoFe/CeO catalyst samples also come from additional CeO
2
2
in addition to the MoFe oxides, because the CeO naturally has weak acidity.
2
At the same time, the amounts of surface acid sites of the MoFe-N and MoFe/
CeO catalysts can be inferred from the corresponding NH -TPD peak area of the
2
3
samples [23, 24]. The acid site density and acid site concentration of the sam-
ples in terms of the NH -TPD (combining with catalyst mass and speciꢁc surface
3
area from BET) results were calculated and showed in Table 1. The correspond-
ing data showed the MoFe-N and MoFe/c–CeO have surface acid site concentra-
2
tion of 0.134 mmol/g and 0.148 mmol/g, respectively. For the MoFe/p –CeO and
1
2
MoFe/p –CeO catalysts, the surface acid site concentration obviously increased to
2
2
0
.234 mmol/g and 0.339 mmol/g, respectively. Based on this reference data, the cat-
alysts, can be ranked as MoFe/p –CeO >MoFe/p –CeO –MoFe/c–CeO >MoFe-
2
2
1
2
2
N, with the same order as the surface acidic strength of the studying catalyst sam-
ples. This phenomenon suggests the CeO supports can enhance the surface acid site
2
concentration and acidic strength of the supported MoFe oxide catalysts to a certain
extent, but which was obviously inꢃuenced by the crystallinity degree and particle
size of CeO . Combining with the XRD, TEM, and BET results, the CeO with low-
2
2
degree crystallinity and small particles can be a suitable metal-oxide support to has
strong interaction with the surface MoFe oxides and to improve the surface acid site
concentration and acidic strength of the MoFe/CeO catalyst system. The distinct
2
surface acid strength and acid site concentration may lead to diꢀerent catalytic activ-
ities of MoFe-N and MoFe/CeO catalysts for the gas-glycerol catalytic conversion,
2
because of the surface acid sites are the reactive sites for glycerol conversion, but
also can cause coke formation [21, 22, 27].
Catalytic activity
Table 2 presents the results that the investigated catalytic performance of the MoFe-
N and MoFe/CeO catalysts for gas-glycerol direct conversion. As shown in Table 2,
2
the catalytic conversion of glycerol over the MoFe-N, MoFe/c–CeO , MoFe/
2
p –CeO , and MoFe/p –CeO was 71.3%, 71.7%, 81.0%, and 97.1%, respectively.
1
2
2
2
The selectivity of allyl alcohol over MoFe-N was 14.6%, which increased to above
2
2% over the MoFe/CeO catalysts. Prominently, the MoFe/p –CeO has the high
2 2 2
DODH catalytic activity with the highest allyl alcohol yield of 22.6%. Additionally,
the investigated products from glycerol catalytic conversion over the MoFe-N and
MoFe/CeO catalysts included propanol, acrolein, ethanol, acetol, acetone, propa-
2
noic acid, and acrylic acid, but with low selectivity under 10%. The propanoic acid
and acrylic acid selectivity over the MoFe/p –CeO reached 12.6 and 7.8%, respec-
2
2
tively, which were obviously higher than that over the other MoFe-based oxides cat-
alysts. The CeO supported MoFe oxides catalysts (MoFe/p –CeO ) seems superior
2
2
2
to the previously reported ZrO -FeO , γ-Al O (Fe O ) , γ-Alumina/Fe/Rb, and
2
x
2
3/
2
3 0.16
H-ZSM-5/Fe/Rb catalysts with allyl alcohol selectivity less than 20% from glycerol
direct conversion through gas–solid catalytic reaction [18–22].
1
3

CeO promoting allyl alcohol synthesis from glycerol direct…
2
1
3

H. Lan et al.
It can be known from the results in Table 2, the allyl alcohol selectivity increased
about 10% over the prepared MoFe/CeO catalysts in comparison to the MoFe-N,
2
suggesting the CeO support can obviously improve the catalytic activity of MoFe
2
bimetal oxides for allyl alcohol synthesis from gas-glycerol. As is known, the ally
alcohol produced from glycerol direct conversion needs a co-function of moderate
acid centers and non-acid centers (redox centers), through dehydration and follow-
up hydrogen transfer (hydrogen reduction) process [17, 21, 24]. According to the
H -TPR results, the MoFe/CeO samples have higher reducibility (oxidability) than
2
2
the MoFe-N resulting from the increasing speciꢁc surface area and strong interac-
tion between MoFe oxides and the CeO support. Importantly, the CeO revealed
2
2
to have high catalytic performance for carbon–carbon (C–C) bonds of glycerol to
produce hydrogen radical and ꢁnally obtained methanol [25]. Here, the hydrogen
radical was needed for allyl alcohol synthesis, and the hydrogen donor came from
the glycerol molecules as the previous veriꢁed [17–24]. The enhanced catalytic
activity of MoFe/CeO catalysts for allyl alcohol selectivity can be attributed to the
2
enhanced oxidability of MoFe/CeO and natural special property of CeO facili-
2
2
tating the dehydration and C–C bonds cleavage of glycerol over the MoFe/CeO
2
catalysts. In summary, the support eꢀect and special natural property of CeO sig-
2
niꢁcantly improve the allyl alcohol selectivity of MoFe/CeO catalysts from gas-
2
glycerol direct conversion.
Based on the glycerol conversion and allyl alcohol yields from high to low,
followed the order of MoFe/p –CeO > MoFe/p –CeO > MoFe/c–CeO > MoFe-
2
2
1
2
2
N, that is, the CeO as the support signiꢁcantly improves the catalytic activity of
2
MoFe bimetal oxides for gas-glycerol direct conversion to allyl alcohol without
any extra hydrogen donors. According to the results in Figs. 4 and 5, the glycerol
conversion over the MoFe-N and MoFe/CeO keep a positive linear relation with
2
the strength and concentration and surface moderate acid sites (Figs. 4 and 5).
1
00
100
9
8
7
6
5
4
3
0
0
0
0
0
0
0
9
5
0
5
0
5
0
5
0
9
8
8
7
7
6
6
20
10
0.10
0.15
0.20
0.25
0.30
0.35
-
1
Surface acid site concentration / mmol
•
g
Fig. 5 The relationship of glycerol conversion and allyl alcohol selectivity with the surface acid site con-
centration of the MoFe-N and MoFe/CeO oxide catalysts
2
1
3

CeO promoting allyl alcohol synthesis from glycerol direct…
2
Thus, the increment of the glycerol conversion can be attributed to the increasing
surface acid site concentration (surface acid amounts) and surface acidic strength
of the MoFe/CeO catalysts, due to the glycerol conversion starting at the activa-
2
tion and dehydration of the hydroxyl groups over the surface acid centers of cata-
lysts [21–24]. Moreover, the results also tell that the morphology, particle sizes
and crystalline of CeO have important inꢃuence on the catalytic performance, in
2
terms of the XRD and TEM results. The CeO with smaller particle sizes and low
2
crystalline degree was superior to nanocube CeO with high crystalline degree
2
to promote its interaction with the MoFe oxide active components and improve
the surface acid site concentration and reducibility of MoFe/CeO (H -TPR and
2
2
NH -TPD results in Figs. 3 and 4). Thus, the MoFe/p –CeO showed higher cata-
3
2
2
lytic activity than the MoFe/p –CeO and MoFe/c–CeO for allyl alcohol yielding
1
2
2
from gas-glycerol conversion. In addition, it need to be explained that the higher
propanoic acid and acrylic acid selectivity over the MoFe/p –CeO resulted from
2
2
its higher oxidability [28–30].
The existence of various byproducts of acrolein, acetol, acetaldehyde, propi-
onaldehyde, and acetone is a typical phenomenon for allyl alcohol synthesis from
gas-glycerol direct conversion because the 1,2 or 1,3 dehydration and cleavage of
carbon–carbon bonds of glycerol can undergo over the acidic-redox MoFe-based
oxide catalysts [21–25]. According to the previous studies, acrolein and acetol
were produced from the 1,2 or 1,3 dehydration of glycerol over the acid centers,
respectively, and the subsequent reaction of these molecules over the acid and/or
redox sites produced acetaldehyde, propionaldehyde, acetone, propanoic acid and
acrylic acid [21–24].
Catalytic stability
Glycerol can easily undergo condensation on the acid oxide catalysts. The products
allyl alcohol, acrolein, and acetol are more active than the reactant glycerol over
acid catalysts because of the coexistence of carbon–carbon double bonds and car-
bon–oxygen double bonds or a hydroxyl group in the molecule. These can all form a
mass of carbonaceous deposits continuously increase on the catalysts, leading to sig-
niꢁcant deactivation observed over the reported γ-alumina/Fe and K [5] /ZrO -FeO
2
x
catalysts with the time-on-stream of 3 h [20–28]. For this, the catalytic stability of
the MoFe/p –CeO and MoFe/c–CeO catalyst was investigated, and the correspond-
2
2
2
ing results are shown in Fig. 6a and b. As Fig. 6a presented, for MoFe/p –CeO ,
2
2
the catalytic conversion of glycerol gradually reduced in average rate of 5%/h in the
range of time-on-stream 2–10 h. But, the products including allyl alcohol acrolein,
ethanol, acetol, and propionaldehyde keep a good constant electivity in 10 h. For
MoFe/c–CeO , the glycerol conversion badly decreased in the investigated reaction
2
time of 8 h, and the allyl alcohol selectivity also obviously decreased. In compari-
son, the MoFe/p –CeO showed higher catalytic stability than the MoFe/c–CeO ,
2
2
2
probably beneꢁting from the CeO with small nano particle sizes and low degree of
2
crystalline improving the carbon resistance of MoFe/p –CeO .
2
2
1
3

H. Lan et al.
(
a)
3
5
0
5
0
5
0
5
0
110
Allyl alcohol
Ethanal
Acrolein
Acetol
1
9
8
7
6
5
4
00
0
Propanal
Acetone
3
2
2
1
1
Propanoic acid
Acrylic acid
Allyl alcohol yield
0
0
0
0
0
30
2
1
0
0
0
2
4
6
8
10
Time-on-stream / h
(
b)
4
3
3
2
2
1
1
0
5
0
5
0
5
0
5
0
100
90
Allyl alcohol
Acrolein
Acetol
Ethanal
Propanal
Acetone
Propanoic acid
Acrylic acid
Allyl alcohol yield
8
7
6
0
0
0
50
4
3
2
1
0
0
0
0
0
2
4
6
8
Time-on-stream / h
Fig. 6 The glycerol conversion and products distribution as a function of time-on-stream over the MoFe/
p –CeO (a) and MoFe/c–CeO (b) catalysts
2
2
2
Conclusions
Nanocube and nano-particle MoFe/CeO2 catalysts were successfully prepared
through simple wet impregnation methods, using nanocube and nanoparticle CeO
2
as the supports. Diꢀerent from the MoFe-N (without supports) consisting of crys-
talline Fe (MoO ) and Fe O crystalline phase, Mo–Fe oxides were well-dispersed
2
4 3
2
3
on CeO for the MoFe/CeO (c-CeO , p -CeO , and p -CeO ) catalysts. The CeO
2
2
2
1
2
2
2
2
eꢀectively improve the reducibility and surface moderate acid sites (acidic strength)
of MoFe/CeO , and nanoparticle CeO (especially for p -CeO ) with low particle
2
2
2
2
sizes and crystalline degree was superior to high-crystalline nanocube c-CeO . The
2
1
3

CeO promoting allyl alcohol synthesis from glycerol direct…
2
allyl alcohol selectivity from glycerol direct conversion obviously increased over
the MoFe/CeO in comparison with that over the MoFe-N, which beneꢁted from
2
the support eꢀect and special natural property of CeO . The catalysts, based on
2
the catalytic performance for glycerol conversion to allyl alcohol can be ranked as
MoFe/p –CeO >MoFe/p –CeO >MoFe/c–CeO >MoFe-N, showing positive
2
2
1
2
2
relationship with the surface acid concentration and reducibility of catalysts. Over
the MoFe/p –CeO , allyl alcohol of 22.6% yielded from the glycerol (conversion of
2
2
9
7.1%), along with acrolein, ethanol, propanoic acid, and acrylic acid produced with
the selectivity of 6.9%, 8.6%, 12.6%, and 7.8%, respectively. The MoFe/p –CeO
2
2
showed better catalytic stability than the MoFe/c–CeO , but the glycerol conversion
2
gradually decreased as the reaction continued.
References
1
2
.
.
D.E. Bloom, Science 333, 562 (2011)
C.T. Wu, K.M.K. Yu, F.L. Liao, N. Young, P. Nellist, A. Dent, A. Kroner, S.C.E. Tsang, Nat. Com-
mun. 3, 1050 (2012)
3
4
5
6
.
.
.
.
B. Obama, Science 355, 126 (2017)
C.H. Zhou, J.N. Beltramini, Y.X. Fan, G.Q. Lu, Chem. Soc. Rev. 37, 527 (2008)
F.X. Yang, M.A. Hanna, R.C. Sun, Biotechnol. Biofuels 5, 13 (2012)
P. Cintas, S. Tagliapietra, E. Calcio Gaudino, G. Palmisano, G. Cravotto, Green Chem. 16 1056
(
2014)
7
8
9
.
.
.
R.A. Sheldon, Green Chem. 16, 950 (2014)
Y. Nakagawa, M. Tamura, K. Tomishige, Res. Chem. Intermed. 44, 3879 (2018)
H.T. Nguyen, G.S. Kamali Kannangara, Chem. Soc. Rev. 42, 9454 (2013)
1
1
0. D.L. Sun, Y. Yamada, S. Sato, W. Ueda, Appl. Catal. B: Environ. 193, 75 (2016)
1. NPCS, B.o. Consultants, Engineers, Industrial Alcohol Technology Handbook, Asia Paciꢁc Busi-
ness Press Inc. (2010)
1
1
1
1
2. K. Weissermel, H.J. Arpe, Industrial organic chemistry, 4th edition, 312 (1994)
3. J.G. Speight, Chemical and Process Design Handbook, McGraw Hill, (2002)
4. E. Arceo, P. Marsden, R.G. Bergman, J.A. Ellman, Chem. Commun. 203, 3357 (2009)
5. S. Tazawa, N. Ota, M. Tamura, Y. Nakagawa, K. Okumura, K. Tomishige, ACS Catal. 6, 6393
(
2016)
1
1
6. G.M. Lari, Z.P. Chen, C. Mondelli, J. Pérez-Ramírez, ChemCatChem. 9, 2195 (2017)
7. Y. Liu, H. Tüysüz, C.J. Jia, M. Schwickardi, R. Rinaldi, A.H. Lu, W. Schmidt, F. Schüth, Chem.
Commun. 46, 1238 (2010)
1
8. T. Yoshikawa, T. Tago, A. Nakamura, A. Konaka, M. Mukaida, T. Masuda, Res. Chem. Intermed.
3
7, 1247 (2011)
1
2
9. L. Harvey, G. Sánchez, E.M. Kennedy, M. Stockenhuber, Asia-Pac. J. Chem. Eng. 10, 598 (2015)
0. G. Sánchez, B.Z. Dlugogorski, E.M. Kennedy, M. Stockenhuber, Appl. Catal. A: Gen. 509, 130
(
2016)
2
2
1. A. Konak, T. Tag, T. Yoshikaw, A. Nakamur, T. Masud, Appl. Catal. B: Environ. 146, 267 (2014)
2. G. Sánchez, J. Friggieri, C. Keast, M. Drewery, B.Z. Dlugogorski, E. Kennedy, M. Stockenhuber,
Appl. Catal. B: Environ. 152–153, 117 (2014)
2
2
2
3. H. Lan, X. Xiao, S.L. Yuan, B. Zhang, G.L. Zhou, Y. Jiang, Acta Phys. Chim. Sin. 33, 2301 (2017)
4. H. Lan, X. Xiao, S.L. Yuan, B. Zhang, G.L. Zhou, Y. Jiang, Catal. Lett. 147, 2187 (2017)
5. M.H. Haider, N.F. Dummer, D.W. Knight, R.L. Jenkins, M. Howard, J. Moulijn, S.H. Taylor, G.J.
Hutchings, Nature Chem. 7, 1028 (2015)
2
6. H.X. Mai, L.D. Sun, Y.W. Zhang, R. Si, W. Feng, H.P. Zhang, H.C. Liu, C.H. Yan, Nanorods, and
Nanocubes. J. Phys. Chem. B 109, 24380 (2005)
2
2
2
7. Y.M. Liu, L.F. Luo, Y.X. Gao, W.X. Huang, Appl. Catal. B Environ. 197, 214 (2016)
8. S.J. Chang, M. Li, Q. Hua, L.J. Zhang, Y.S. Ma, B.J. Ye, W.X. Huang, J. Catal. 293, 195 (2012)
9. Y.J. Lee, G.H. He, A.J. Akey, R. Si, I.P. Herman, J. Am. Chem. Soc. 133, 12952 (2011)
1
3

H. Lan et al.
3
0. G.L. Zhou, B.G. Gui, H.M. Xie, F. Yang, Y. Chen, S.M. Chen, X.X. Zheng, J. Ind. Eng. Chem. 20,
1
60 (2014)
3
3
3
3
3
3
1. G.L. Zhou, H. Lan, T.T. Gao, H.M. Xie, Chem. Eng. J. 246, 53 (2014)
2. H. Lan, G.L. Zhou, C.J. Luo, Y.R. Yu, G.Z. Zhang, Int. J. Chem. React. Eng. 14, 757 (2016)
3. R.M.M. Abbaslou, A. Tavassoli, J. Soltan, A.K. Dalai, Appl. Catal. A: Gen. 367, 47 (2009)
4. Z.B. Lei, S.Y. Bai, L.Q. Dang, H.A. Xia, S.B. Liu, Micropor. Mesopor. Mater. 123, 306 (2009)
5. Y.S. Li, Y. Chen, L. Li, J.L. Gu, W.R. Zhao, L. Li, J.L. Shi, Appl. Catal. A: Gen 366, 57 (2009)
6. S.D. Qina, C.H. Zhang, J. Xu, B.S. Wu, H.W. Xiang, Y.W. Li, J. Mol. Catal. A: Chem. 304, 128
(
2009)
1
3