Catalytic Transfer Hydrogenation of Furfural to Furfuryl Alcohol with Recyclable Al–Zr@Fe Mixed Oxides

Article abstract of DOI:10.1002/cctc.201701266

A series of magnetic, acid/base bifunctional Al–Zr@Fe₃O₄ catalysts were successfully prepared by a facile coprecipitation method and utilized in the catalytic transfer hydrogenation (CTH) of furfural to furfuryl alcohol with 2-propanol as hydrogen source. The physicochemical properties and morphologies of the as-prepared catalysts were characterized by various techniques, including XRD analysis, N₂ physisorption, vibrating sample magnetometry, thermal gravimetry analysis, X-ray fluorescence spectroscopy, NH₃/CO₂ temperature-programmed desorption, SEM, and TEM. The Al₇Zr₃@Fe₃O₄(1/1) catalyst with a Al³⁺/Zr⁴⁺/Fe₃O₄ molar ratio of 21:9:3 was found to exhibit a high furfuryl alcohol yield of 90.5 % in the CTH from furfural at 180 °C after 4 h with a comparatively low activation energy of 45.3 kJ mol^?1, as calculated from the Arrhenius equation. Moreover, leaching and recyclability tests confirmed Al₇Zr₃@Fe₃O₄(1/1) to function as a heterogeneous catalyst that could be reused for at least five consecutive reaction runs without significant loss of catalytic activity after simple recovery by an external magnet. Notably, the catalyst proved also efficient for hydrogenation of other biomass-derived furanic aldehydes.

Full text of DOI:10.1002/cctc.201701266

DOI: 10.1002/cctc.201701266
Full Papers
Catalytic Transfer Hydrogenation of Furfural to Furfuryl
Alcohol with Recyclable Al–Zr@Fe Mixed Oxides
[a, b]
[a]
[b]
[a]
Jian He,
Hu Li, Anders Riisager,* and Song Yang*
3
+
4+
A series of magnetic, acid/base bifunctional Al–Zr@Fe O cata-
Al /Zr /Fe O molar ratio of 21:9:3 was found to exhibit a
3 4
3
4
lysts were successfully prepared by a facile coprecipitation
method and utilized in the catalytic transfer hydrogenation
high furfuryl alcohol yield of 90.5% in the CTH from furfural at
1808C after 4 h with a comparatively low activation energy of
À1
(CTH) of furfural to furfuryl alcohol with 2-propanol as hydro-
45.3 kJmol , as calculated from the Arrhenius equation. More-
gen source. The physicochemical properties and morphologies
of the as-prepared catalysts were characterized by various
over, leaching and recyclability tests confirmed Al Zr @Fe O (1/
7
3
3
4
1) to function as a heterogeneous catalyst that could be
reused for at least five consecutive reaction runs without sig-
nificant loss of catalytic activity after simple recovery by an ex-
ternal magnet. Notably, the catalyst proved also efficient for
hydrogenation of other biomass-derived furanic aldehydes.
techniques, including XRD analysis, N physisorption, vibrating
2
sample magnetometry, thermal gravimetry analysis, X-ray fluo-
rescence spectroscopy, NH /CO temperature-programmed de-
3
2
sorption, SEM, and TEM. The Al Zr @Fe O (1/1) catalyst with a
7
3
3
4
Introduction
Catalytic upgrading of biomass-derived platform compounds
to fuels and value-added chemicals are attracting more and
more attention, because abundant biomass is a sustainable or-
ganic carbon resource and regarded as a promising alternative
to traditional fossil resources for producing fuels and chemi-
(FAOL), tetrahydrofurfuryl alcohol (THFAOL), 2-methylfuran (2-
[8]
MF), and 2-methyltetrahydrofuran (2-MTHF). As one of the
most important derivatives, FAOL has been widely applied as a
versatile raw material for the industrial synthesis of fine chemi-
cals including resins, lubricants, fibers, adhesives, vitamin C,
[
1–5]
[9,10]
cals.
Furfural (FF), as an important platform compound, is
and lysine.
Additionally, important furanic downstream
obtainable from hemicellulose (one of the main components
of lignocellulose) through an acid-catalyzed hydrolysis and de-
hydration process. Owing to the highly functionalized molecu-
lar structure of FF (i.e., CÀO, C=C, and C=O groups) it can
serve as an ideal feedstock for the synthesis of various biofuels
and value-added chemicals through different conversion reac-
tions, such as hydrogenation, oxidation, hydrogenolysis, aldoli-
products such as levulinic acid and g-valerolactone (GVL) can
[11]
be obtained through FAOL after additional conversion. Con-
sequently, the design of efficient catalytic systems for the syn-
thesis of FAOL from FF remains to be of high importance.
An efficient strategy for the production of FAOL from FF has
been developed involving molecular hydrogen as H donor in
the presence of supported zero-valent noble or nonnoble
[
6,7]
[12–16]
zation, and aldol condensation.
Particularly, hydrogenation
metal-based catalysts (e.g., Au, Pd, Ru, and Cu).
However,
or hydrogenolysis of FF has received immense attention be-
cause many important industrial chemicals and biofuel addi-
tives are obtainable directly from FF, such as furfuryl alcohol
the high preparation cost of H (usually prepared from fossil-
2
based resources) and metal catalysts (often reduced under H2)
as well as the flammable and explosive nature of the gas en-
courage exploration of alternative methods. On this basis, cata-
lytic transfer hydrogenation (CTH) utilizing alternative H
donors, such as formic acid and alcohols, has recently been
proposed as an attractive way to hydrogenate organic mole-
[a] J. He, Dr. H. Li, Prof. S. Yang
State Key Laboratory Breeding Base of
Green Pesticide & Agricultural Bioengineering,
Key Laboratory of Green Pesticide & Agricultural Bioengineering,
Ministry of Education, State-Local Joint Laboratory for
Comprehensive Utilization of Biomass,
Center for Research & Development of Fine Chemicals
Guizhou University
[17–19]
cules including biomass-derived compounds.
Employing
abundant alcohols for the CTH process with the advantage of
high selectivity towards reduction of carbonyl groups to alco-
holic hydroxyl groups through the Meerwein-Ponndorf-Verley
Guiyang 550025 (P.R. China)
E-mail: yangsdqj@126.com
(
MPV) reduction seems more appealing than the use of corro-
[20,21]
sive formic acid.
Many Lewis acid-containing/base-contain-
[b] J. He, Prof. A. Riisager
Centre for Catalysis and Sustainable Chemistry, Department of Chemistry
Technical University of Denmark
DK-2800 Kgs. Lyngby (Denmark)
ing or acid/base bifunctional heterogeneous catalysts have
been developed for the highly efficient synthesis of GVL from
biomass-derived levulinic acid/levulinate ester by CTH using H
E-mail: ar@kemi.dtu.dk
[22–26]
donor alcohols
with the Lewis acid/base sites as crucial
Supporting information and the ORCID identification number(s) for the
author(s) of this article can be found under:
https://doi.org/10.1002/cctc.201701266.
[27]
sites for the MPV reduction. Very recently, several other cata-
[28]
lysts including Hf-Beta,
nitrogen-doped carbon-supported
ChemCatChem 2017, 9, 1 – 10
1
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[
29]
[30]
iron (Fe-L1/C-800), and Co-Ru/C have also been used as ef-
ficient solid catalysts for CTH of FF to FAOL with alcohols. Simi-
larly, we have shown that a mesoporous organotriphosphate–
zirconium nanohybrid (ZrPN) catalyst with Lewis acid/base
sites efficiently catalyzed the production of FAOL from FF in 2-
respectively, are compiled in Table 1. The surface areas of the
as-prepared catalysts slightly decreased with the increment of
Fe O content (Table 1, entries 1–4), whereas the mean pore
3
4
sizes of the materials were all around 4 nm. Furthermore, the
actual element compositions of Al, Zr, and Fe in the catalysts
were close to the controlled molar ratios as verified from X-ray
fluorescence (XRF) measurements (entries 2–4). Thermogravi-
metric analysis of the Al Zr @Fe O (1/1) catalyst (Figure S2) re-
[
31]
propanol through CTH. Despite these pioneering works, it
remains a much-sought-after goal to develop more efficient
and lower cost heterogeneous catalysts for FAOL production
by CTH.
7
3
3
4
vealed a total weight loss of 8.4% resulting from removal of
Low-cost, mixed-metal oxides are widely applied in catalysis
with renewables because of their tunable physicochemical
properties closely related to catalytic activity, for example, sur-
physical adsorbed water (4.0%; Tꢀ2008C) and remaining hy-
[23]
droxyl groups (ꢁ4.4%; Tꢂ3008C) for the material.
The magnetic properties of the as-prepared catalysts were
confirmed by measurement on a vibrating sample magneto-
meter (VSM). The results in Figure 1c show some superpara-
magnetic property of the AlZr@Fe catalysts attributed to the
absence of coactivity, and the saturation magnetization values
of Al Zr @Fe O (2/1), Al Zr @Fe O (1/1), and Al Zr @Fe O (1/2)
[32]
face area, pore size distribution, and acid/base character. For
instance, W–Zr mixed oxides can efficiently convert cellulose to
[
33]
lactic acid,
triose sugars over Sn/Al oxide,
fructose) be dehydrated into 5-hydroxymethylfurfural with Nb-
ethyl lactate can efficiently be obtained from
[
34]
and glucose can (through
7
3
3
4
7
3
3
4
7
3
3
4
[
35]
À1
doped WO3. In our previous work, Al–Zr mixed oxide was
confirmed to be an efficient catalyst for the production of GVL
were evaluated to be 6.8, 9.4, and 19.2 emug , respectively,
which increased with the increment of Fe O4 content (as
3
[
23]
from ethyl levulinate, and in continuation of our research on
CTH of biomass-derived chemicals, we expected also such cat-
alysts to be active in CTH of FF to FAOL. Solid materials with
magnetic properties are highly interesting catalyst supports
because they allow easy catalyst separation and recycling from
liquid reaction mixtures by applying an external magnetic
field. This renders the operation of such catalytic systems time-
saving and energy-efficient compared to traditional filtration
expected).
The acid/base properties of the as-prepared materials were
evaluated by NH /CO2 temperature-programmed desorption
3
(TPD) measurements as well (Figures 1d,e and Figures S3 and
S4). Considering that the materials were calcined at 3008C and
the apparent weight loss emerging at higher temperature (i.e.,
over 3008C, Figure S2), the highest desorption temperature for
NH or CO were set at 3008C during the TPD measurement
3
2
[
36,37]
and centrifugation operations.
In this regard, the design
(Figures 1d). As expected, the as-prepared catalysts possessed
both acid and base sites mainly originating from Al–Zr with
less contribution from the Fe O (Table 1, entries 1–5).
and preparation of a magnetic catalyst could be highly desira-
ble for the large-scale practical application of CTH of FF to
FAOL. Accordingly, we prepared herein a series of acid/base bi-
functional magnetic Al Zr @Fe O oxide catalysts by a facile co-
3
4
The microstructure of the Al Zr @Fe O (1/1) catalyst were
7
3
3
4
characterized by electron microscopy. As revealed from ele-
mental mapping by high-angle annular dark-field scanning
transmission electron microscopy (HAADF–STEM, Figure 2b)
the aluminum, zirconium, iron, and oxygen elements were con-
firmed to be homogeneously distributed in the catalyst matrix
consisting of individual particles with a size of approximately
10 nm (Figure 2c). In addition, it seemed as the Al–Zr mixed
oxides with amorphous nature encapsulated the magnetic
Fe O particles like a shell without the presence of larger bulk
7
3
3
4
precipitation method and explore their application in the CTH
of FF to FAOL through MPV reduction using 2-propanol as H
donor.
Results and Discussion
Characterization of catalysts
3
4
The XRD patterns of the as-prepared catalysts are depicted in
Figure 1a. The Al Zr exhibited an amorphous nature with no
Al–Zr oxide particles, as presented in Figure 2d. Furthermore,
the crystal faces of Fe O were clearly observed in the high-res-
7
3
3
4
clear diffraction peaks detected consistent with our previous
olution (HR) TEM image (Figure 2e), both observations in good
consistency with the results from XRD analysis (Figure 1a).
[
23]
report. With regard to Fe O , five diffraction peaks at 30.3,
3
4
35.6, 43.3, 57.2, and 62.98 were clearly observed and assigned
to the (220), (311), (400), (511), and (440) crystalline planes
Catalyst screening for CTH of furfural to furfuryl alcohol
[
38]
of Fe O ,
respectively. In the case of the as-prepared
3
4
Al Zr @Fe O catalysts, the Al–Zr mixed oxides remained amor-
The catalytic results of CTH of FF to FAOL in 2-propanol at low
FF conversion (i.e., 1208C, 0.5 h) obtained with the as-prepared
Al Zr @Fe O catalysts are shown in Table 2. As expected,
7
3
3
4
phous in nature, and no additional diffraction peaks emerged
indicating that no other crystal phases were generated.
7
3
3
4
The N adsorption–desorption isotherms of the as-prepared
almost no FF conversion and FAOL product was detected if
the reaction was conducted without any catalyst or in the
presence of Fe O only (Table 2, entries 1 and 6). In contrast,
2
catalysts are displayed in Figure 1b; all are type-IV isotherms
with typical H3 hysteresis loop. In addition, all catalysts exhibit-
ed a relative narrow pore distribution (Figure S1 in the Sup-
porting Information) revealing that the catalysts possessed ir-
regular mesoporous structures. Data on surface areas and
mean pore sizes obtained from BET and BJH methods,
3
4
relative high FAOL yield (7.5%) and formation rate
À1
À1
(125.0 mmolg min ) was found for the Al Zr3 catalyst
7
(entry 2), indicating that the Al–Zr mixed oxides provided the
main activity sites. Specifically, Al Zr @Fe O (1/2) through Al Zr
3
7
3
3
4
7
&
ChemCatChem 2017, 9, 1 – 10
www.chemcatchem.org
2
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 1. (a) XRD patterns of the as-prepared Al
7
Zr
3
, Fe
3
O
4
and Al
7
Zr
3
@Fe
3
O
4
catalysts, (b) N
2
adsorption–desorption isotherms of the as-prepared Al
7
Zr @Fe O
3 3 4
catalysts, (c) magnetization curves of the Al
catalyst.
7
Zr
3
@Fe
3
O
4
catalysts, (d) NH
3
-TPD profile of Al
7
Zr
3
@Fe
3
O
4
(1/1) catalyst, and (e) CO
2
-TPD profile of Al
7
Zr
3
@Fe
3
O
4
(1/1)
Table 1. Physicochemical properties of the as-prepared catalysts
[
a]
[b]
[c]
Entry
Catalyst
BET surface area
Mean pore size
[nm]
Acid/base amount
Al/Zr/Fe (mole ratio)
Controlled
3
À1
À1
[d]
[cm g
]
[mmolg
]
Measured
1
2
3
4
5
Al
Al
Al
Al
7
7
7
7
Zr
Zr
Zr
Zr
3
3
3
3
199.8
193.9
186.3
175.5
–
3.90
3.88
4.10
4.32
–
1.10/0.13
0.70/0.12
0.68/0.11
0.61/0.08
0.2/0.03
–
–
@Fe
@Fe
@Fe
3
3
3
O
O
O
4
(2/1)
4
(1/1)
4
(1/2)
4.7:2:1
2.3:1:1
1.2:0.5:1
–
4.9:1.9:1
2.5:0.9:1
1.4:0.4:1
–
3 4
Fe O
[
a] BET surface areas determined by N
2
adsorption–desorption measurements. [b] Calculated by the Barrett–Joyner–Halenda method. [c] Evaluated by NH
3
/
CO
2
-TPD method. [d] Determined by XRF analysis.
ChemCatChem 2017, 9, 1 – 10
www.chemcatchem.org
3
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Figure 2. (a) HAADF–STEM image of Al
Al Zr @Fe (1/1).
7
Zr
3
@Fe
3
O
4
(1/1), (b) EDX elemental mappings of the Al
7
Zr
3
@Fe
3
O
4
(1/1) particle shown in (a), and (c)–(e) TEM images of
7
3
3 4
O
a
Table 2. CTH of FF to FAOL over different catalysts
À1
À1 [b]
Entry
Catalyst
none
FF conv. [%]
FAOL yield [%]
FAOL select. [%]
FAOL formation rate [mmolg min
]
1
2
3
4
5
6
1.4
21.9
16.8
7.8
7.6
1.9
0
0
0
Al
Al
Al
Al
7
Zr
7
Zr
7
Zr
7
Zr
3
3
3
3
7.5
6.3
5.2
4.2
0.9
34.2
37.5
66.7
55.3
47.4
125.0
105.0
86.7
70.0
15.0
@Fe
@Fe
@Fe
3
3
3
O
O
O
4
(2/1)
4
(1/1)
4
(1/2)
Fe
3
O
4
[a] Reaction conditions: FF (0.1922 g, 2 mmol), catalyst (0.04 g), 2-propanol (10 mL), 1208C, 0.5 h. [b] Calculated as: mole of FAOL per 30 min per mass of
catalyst.
exhibited a gradual increase in the FAOL formation rate rather
than a linear relationship as a function of the acid or base
amount (Figure S6), implying that specific acid/base sites in
the catalysts played a key role in the CTH of FF through MPV
Al Zr @Fe O (1/1) (Table 2, entry 3), possibly owing to its ap-
7 3 3 4
À1
propriate amounts of acid/base sites (0.68/0.11 mmolg ) and
2
À1
relatively high specific surface area (186.3 m g ). As the
Al Zr @Fe O (1/1) catalyst also had an acceptable saturation
7
3
3
4
[
23]
À1
reduction.
However, only 34.2% FAOL selectivity was ob-
magnetization (9.4 emug ), this catalyst was selected for the
following studies.
served over the Al Zr catalyst, and much acetalization product
7
3
of FF with 2-propanol (i.e., 2-(diisopropoxymethyl)furan,
DIPMF) formed in the reaction mixture as demonstrated by
GC–MS analysis (Figure S5). This is because the acetalization of
Effect of reaction temperature and time
[
39]
FF with alcohol is facilitated by acid sites, which the Al Zr3
The influence of reaction temperature and time on the CTH of
FF to FAOL over Al Zr @Fe O (1/1) catalyst was investigated in
7
catalyst contained to the largest degree of the prepared cata-
7
3
3
4
lysts (Table 1, entry 1). Interestingly, the introduction of Fe O4
detail and the results are compiled in Figure 3. Clearly, the
temperature played an important role in the reaction. For ex-
ample, only 15% FF conversion with 12.6% FAOL yield was ob-
tained if the reaction was conducted at 1208C for 1 h, whereas
65.2% yield of FAOL at 69.8% FF conversion could be attained
if the reaction temperature was increased to 1808C within a re-
action time of 1 h. Moreover, it was established that relatively
lower FAOL selectivity was obtained at low temperature (i.e.,
120 and 1408C) whereas high temperature (i.e., 160 and
1808C) improved the FAOL selectivity (Figure S7). This outcome
can be elucidated as resulting from the competitive acetaliza-
tion of FF with 2-propanol forming DIPMF, which at low
3
as catalyst support not only rendered the catalysts magnetic,
but it apparently also resulted in catalysts with improved
FAOL selectivity (Table 2, entries 2–5). However, despite
Al Zr @Fe O (1/2) exhibited the highest saturation magnetiza-
7
3
3
4
À1
tion (19.2 emug ) among the as-prepared Al Zr @Fe O cata-
7
3
3
4
lysts it gave the lowest FAOL yield (4.2%) and FAOL formation
À1
À1
rate (70.0 mmolg min , entry 5), which probably resulted
from its relatively low amount of acid/base sites (0.61/
À1
0
.08 mmolg ) and relatively small specific surface area
2
À1
(
175.5 m g ). In contrast, a higher FAOL yield (5.2%) and a rel-
atively high FAOL selectivity (66.7%) were achieved with
&
ChemCatChem 2017, 9, 1 – 10
www.chemcatchem.org
4
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 3. Synthesis of FAOL from FF by CTH over Al
7
Zr
3
@Fe
3
O
4
(1/1) at (a) 1208C, (b) 1408C, (c) 1608C, and (d) 1808C from 0.5 to 5 h. Reaction conditions: FF
(0.1922 g, 2 mmol), catalyst (0.04 g), 2-propanol (10 mL).
reaction temperature will be more pronounced than at higher
temperature because the reaction is reversible. Additionally,
the reaction time also had a large influence on the reaction.
Prolonging the reaction time to 4 h at 1808C gave nearly full
conversion of FF (99.1%) and 90.5% yield of FAOL. Neverthe-
less, the FAOL yield decreased slightly to 88.9% after extend-
ing the reaction time to 5 h at 1808C, which resulted from the
etherification of the generated FAOL with 2-propanol forming
displayed from GC–MS spectra (Figure S8). Based on the above
observations the possible reaction pathways shown in
Scheme 1 can be inferred for FAOL and byproducts formation
from FF (IPF, FHB, FB) in the catalytic system.
To confirm the heterogeneous catalytic nature of
Al Zr @Fe O (1/1) during the reaction, a filtration experiment
7
3
3
4
was performed. In this case, the solid catalyst was removed
from the reaction solution after 1 h with the help of an exter-
nal magnet, and the solution was subsequently allowed to
react for another 4 h under identical reaction condition. The re-
sults in Figure 4b show that no further FAOL formation oc-
curred in the absence of Al Zr @Fe O (1/1). Moreover, the fil-
2
-isopropoxyfuran (IPF). Based on these observations, the opti-
mal reaction temperature and time for the reaction were set as
808C and 4 h, respectively.
1
7
3
3
4
trate was subsequently subjected to inductively coupled
plasma optical emission spectroscopy (ICP–OES), and only low
concentrations of Al (ꢁ2 ppm), Zr (ꢁ2 ppm), and Fe (<
0.2 ppm) were measured, which did not contribute to FAOL
formation (Figure 4b). This further implies that the reaction
proceeded by heterogeneous catalysis.
Effect of catalyst dosage and leaching experiments
The influence of Al Zr @Fe O (1/1) catalyst dosage on the
7
3
3
4
FAOL yield of the reaction at 1808C and 4 h was also studied.
As given in Figure 4a, the blank experiment resulted in 23.5%
FF conversion with only 3.6% yield of FAOL, while the main
product was the acetalization product DIPMF (verified by GC-
MS analysis), implying that the CTH process hardly took place
without any catalyst. In contrast, the FF conversion and FAOL
yield increased with the increment of the amount of catalyst
Kinetic studies
To obtain more insight into the CTH of FF to FAOL over the
Al Zr @Fe O (1/1) catalyst, its kinetics was investigated at four
(i.e., with more active sites available) reaching maximum FAOL
7
3
3
4
yield (90.5%) with 0.04 g Al Zr @Fe O (1/1). Further increase in
different reaction temperatures (i.e., 120, 140, 160, and 1808C)
assuming the FF conversion to be a first-order reaction. As
compiled in Figure 5a, reaction rate constants (k) were evaluat-
ed at each temperature from the corresponding slope of the
7
3
3
4
catalyst amount (i.e., 0.06 g) lead to lower FAOL yield, which is
primarily attributed to the occurrence of more side reaction,
wherein the amounts of byproducts such as IPF, 4-(furan-2-
yl)but-3-en-2-one (FB), and 4-(furan-2-yl)-4-hydroxybutan-2-one
plots, and the apparent activation energy (E ) determined to
a
À1
(FHB, Scheme 1) were enhanced in the reaction mixture as
be 45.3 kJmol
from the corresponding Arrhenius plot
ChemCatChem 2017, 9, 1 – 10
www.chemcatchem.org
5
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Figure 4. (a) Effect of Al
FAOL from FF. Reaction conditions: FF (0.1922 g, 2 mmol), 2-propanol
10 mL), 1808C, 4 h. (b) FAOL yield profiles of the reaction solution with
Al Zr @Fe (1/1) catalyst or without Al Zr @Fe (1/1) catalyst (separated
by a magnet after 1 h). Reaction conditions: FF (0.1922 g, 2 mmol), 2-propa-
nol (10 mL), Al Zr @Fe (1/1) (0.04 g), 1808C.
7 3 3 4
Zr @Fe O (1/1) catalyst dosage on the production of
Figure 5. (a) Kinetic profiles of the FF to FAOL conversion by the
Al Zr @Fe O (1/1) catalyst (X, FF conversion) and (b) Arrhenius plot of forma-
7 3 3 4
tion of FAOL from FF.
(
7
3
3
O
4
7
3
3 4
O
7
3
3 4
O
H as H donor (Table S2). This further confirms the excellent
2
catalytic property of the Al Zr @Fe O (1/1) catalyst.
7
3
3
4
Catalyst recyclability
The reusability of the Al Zr @Fe O (1/1) catalyst in the CTH of
7
3
3
4
FF to FAOL was evaluated under optimized reaction conditions
i.e., 1808C and 4 h). In the reusability tests, the
Al Zr @Fe O (1/1) catalyst was separated by an external
(
7
3
3
4
magnet after each run, washed with ethanol and acetone
twice (2ꢁ5 mL), dried at 808C for 2 h, and then directly used
for the next run. The results in Figure 6 demonstrate that the
catalytic activity of the catalyst decreased gradually through
five consecutive reaction runs resulting in 95.9% FF conversion
and 83.8% FAOL yield in the fifth reaction run.
The XRD pattern of the used catalyst after five reaction runs
exhibited no clear crystallographic changes compared to the
fresh catalyst (Figure 7a), and both catalyst mesoporosity (Fig-
ure 7b) as well as acid/base properties (Figure S9) also re-
mained largely unchanged. Additionally, ICP–OES analysis dem-
onstrated that the filtrate only contained approximately 2 ppm
of Al, 3 ppm of Zr, and <0.2 ppm of Fe (detection limit) after
reaction for 4 h at 1808C, indicating a very low degree of
leaching of activity species into the reaction mixture. In con-
trast, the surface area of the used catalyst declined 15% (to
Scheme 1. Possible reaction pathways for the formation of (furfuryl alcohol)
FAOL and byproducts from FF.
presented in Figure 5b. This Ea value is slightly lower than
those of other reported heterogeneous catalysts for the pro-
duction of FAOL from FF with CTH using alcohols as H donors,
and more comparable to those of metal-based catalysts using
&
ChemCatChem 2017, 9, 1 – 10
www.chemcatchem.org
6
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 6. Recyclability of Al
Reaction conditions: FF (0.1922 g, 2 mmol), 2-propanol (10 mL),
Al Zr @Fe (1/1) (0.04 g), 1808C, 4 h.
7 3 3 4
Zr @Fe O (1/1) in the synthesis of FAOL from FF.
7
3
3 4
O
2
À1
1
59.2 m g ). In combination with a larger weight loss in ther-
mal gravimetry analysis (TGA, Figure 7c), this suggests that
carbon residues deposited on the catalyst during reaction,
which could also explain the slight decline observed in FAOL
yield during the reuse.
Substrate scope
Biomass-derived furanic aldehydes besides FF, such as 5-hy-
droxymethylfurfural (HMF), 5-methylfurfural, and 2,5-diformyl-
furan, are also recognized as important platform chemicals
and, therefore, relevant substrates for CTH. Accordingly, the
feasibility of the Al Zr @Fe O (1/1) catalyst for CTH with these
7
3
3
4
substrates was also examined with 2-propanol as H donor, and
the results are summarized in Table 3. Clearly, the developed
Al Zr @Fe O (1/1) catalyst exhibited also good catalytic per-
7
3
3
4
formance for the alternative furanic aldehydes resulting in
more than 70% yield of the corresponding products along
with a considerable amount of the corresponding etherifica-
tion products (not quantified). This clearly suggests the
Al Zr @Fe O (1/1) catalyst to be a highly versatile catalyst for
7
3
3
4
CTH of biomass-derived furanic aldehydes in general.
Figure 7. (a) XRD patterns, (b) N
c) TG curves of fresh Al Zr @Fe
respectively, after five reaction runs.
2
adsorption–desorption isotherms and
O (1/1) and used Al Zr @Fe O (1/1) catalysts,
3 4 7 3 3 4
(
7
3
Conclusions
Low-cost, magnetic acid/base bifunctional Al Zr @Fe O cata-
7
3
3
4
lysts were prepared by a facile coprecipitation method, physi-
cochemical characterized, and demonstrated to be efficient
catalysts for the catalytic transfer hydrogenation (CTH) of furfu-
ral (FF) to furfuryl alcohol (FAOL) by using 2-propanol as H-
donor. Under optimized reaction conditions (1808C, 4 h) 99.1%
conversion of FF with 90.5% yield of FAOL was achieved in the
presence of Al Zr @Fe O (1/1). The apparent activation energy
be reused several times under the applied reaction conditions
without significant decline in catalytic activity. Additionally, the
Al Zr @Fe O (1/1) catalyst was found to exhibit good feasibility
7
3
3
4
for CTH of alternative biomass-derived furanic aldehydes. It is
envisaged that the developed low-cost, efficient, durable, and
easily-separated catalyst holds a promising potential for CTH
of biomass-derived molecules.
7
3
3
4
of the CTH of FF to FAOL over Al Zr @Fe O (1/1) was found to
7
3
3
4
À1
be 45.3 kJmol , which was lower than activation energies of
comparable catalyst systems previously reported. Filtration ex-
periments confirmed that the CTH reaction proceeded in a het-
erogeneous manner allowing the Al Zr @Fe O (1/1) catalyst to
7
3
3
4
ChemCatChem 2017, 9, 1 – 10
www.chemcatchem.org
7
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
based on the above procedure using 42/18/3 mmol and
[a]
Table 3. CTH of alternative biomass-derived aldehydes over Al
7
Zr
3
@Fe
3
O
4
.
3+
4+
2
1/9/6 mmol of Al /Zr /Fe O , respectively. Also, refer-
3 4
ences of pure Al Zr (Al–Zr mixed oxide) and Fe O were
Entry Substrate Product Temp. Time Conv. Yield
7
3
3
4
prepared by the described coprecipitation method.
[
8C]
[h]
4
[%]
[%]
[
b]
1
2
3
180
180
180
93.6 80.5
82.7 71.0
Catalyst characterization
Powder X-ray diffraction (XRD) measurements were con-
ducted on a Huber G670 diffractometer with CuKa radia-
tion at room temperature in the 2q interval 5–808. Ele-
ment mapping were obtained by energy-dispersive X-
ray analysis from a Quanta 200 ESEM FEG operated at
4
6
[
c]
70.6 /
+
>99
[
d]
23.2
1
0 kV. TEM images were provided using a FEI Tecnai mi-
[
a] Reaction conditions: Substrate (2 mmol), 2-propanol (10 mL), Al
alyst (0.08 g). [b] 0.06 g Al Zr @Fe (1/1) catalyst. [c] Yield of 2,5-furandimethanol.
d] Yield of HMF.
7
Zr
3
@Fe
3
O
4
(1/1) cat-
croscope operated at 200 kV. N2 physisorption experi-
ments were performed at À1968C with a Micromeritics
ASAP 2020 instrument. Before the N2 physisorption
measurements, all samples were degassed in vacuum at
7
3
3 4
O
[
2
008C for 4 h. The specific surface areas of as-prepared
Experimental Section
materials were measured by the BET method, and mean
pore sizes were evaluated by the Barrett-Joyner-Halenda (BJH)
method. XRF tests were performed using a PANalytical Epsilon3-XL
analyzer with standard addition method. Firstly, the sample was
dissolved in acidic solution containing HNO and H SO , then dilut-
Materials
Fe(NO ) ·9H O (ꢂ98%) and Al(NO ) ·9H O (98%) were purchased
3
3
2
3 3
2
from Riedel–de Haen, FeCl ·4H O (ꢂ99%) from Merck, and
3
2
4
2
2
ed into 10 mL and hereof 0.5 mL solution dispersed in 0.5 g TiO2.
ZrOCl ·8H O (ꢂ99%) and furfural (FF, ꢂ99%) from Fluka. Ammoni-
2
2
After fully mixing, the mixed sample was dried at 1008C and then
um hydroxide solution (25–28% NH basis), poly(ethylene glycol)
3
directly measured. The magnetization data were acquired in hyste-
resis mode at 178C by using a Quantum-Design MPMS-XL SQUID
(
(
9
Maverage =2000), FAOL (98%), 5-methylfurfural (99%), 2-propanol
99.5%), ethanol (99.5%), acetone (ꢂ99%), and naphthalene (>
magnetometer. TGA was fulfilled under a dynamic air atmosphere
9%, internal standard) were procured from Sigma–Aldrich. 5-hy-
À1
(
30 mLmin ) using a METTLER TOLEDO thermal analyzer in the
droxymethylfurfural (HMF, ꢂ95%), 2,5-furandicarboxaldehyde (ꢂ
5%), 5-methylfuran-2-methanol (ꢂ95%), and 2,5-furandimethanol
ꢂ95%) were provided by Bepharm Ltd.
temperature interval 25–6008C with a constant heating rate of
9
(
À1
1
08Cmin . NH /CO -TPD were performed on a Micromeritics Au-
3
2
toChem II 2920 apparatus. The sample (50 mg) was firstly degassed
À1
at 3008C for 1 h under helium flow (25 mLmin ), cooled down to
5
08C followed by purging with NH /CO –He mixture gas flow
Catalyst preparation
3 2
À1
(15 mLmin ) for 1 h. After flushing with high-purity He for addi-
The Fe O catalyst support particles were synthesized by a facile
3
4
tional 1 h to remove the remaining and weakly adsorbed NH or
CO , NH /CO TPD profiles were subsequently recorded by increas-
ing the temperature from 50 to 3008C with a heating rate of
3
[38]
coprecipitation method.
Briefly, Fe(NO ) ·9H O (6 mmol) and
3 3 2
2
3
2
FeCl ·4H O (3 mmol) were dissolved in deionized water (50 mL)
2
2
under nitrogen at room temperature, and then poly(ethylene
glycol) (1 g) was added to the solution to prevent the magnetic
particles from aggregating during the following generation pro-
À1
1
08Cmin under pure He flow and maintaining this temperature
for additionally 1 h. ICP–OES analysis was completed on a Perki-
nElmer Optima 8000 equipment on filtrates after catalyst removing
by an external magnet to examine the leaching of metal species
upon catalyst reuse.
[
39]
cess. Afterwards, the resulting solution was stirred at 858C (oil
bath) for 15 min, and ammonium hydroxide solution was dropwise
added under vigorous stirring until pH 9 was reached (black slurry)
followed by continued stirring for another 30 min at 858C. Subse-
quently, the black Fe O particles were collected by the use of a
3
4
CTH and product analysis
permanent external magnet, washed to neutral pH with deionized
water, and left in deionized water (150 mL) until they were used
for catalyst preparation.
CTH of FF was performed in a 50 mL stainless-steel autoclave con-
taining a magnetic stirring bar. Typically, FF (0.1922 g, 2 mmol), 2-
propanol (10 mL), and catalyst (0.04 g) were charged into the reac-
tor, which was then sealed and heated to a designed temperature
for an intended reaction time. Time zero of the experiments was
considered when the reaction temperature reached the target tem-
perature occurring at 120, 140, 160 and 1808C after 20, 22, 26 and
A series of Al Zr @Fe O catalysts with different ratios of Al Zr (Al/
7
3
3
4
7
3
Zr molar ratio=7:3) and Fe O were prepared through a coprecipi-
tation method at room temperature. Firstly, the as-synthesized
3
4
[23]
3
+
4+
Fe O particles (3 mmol) were dispersed into aqueous Al /Zr so-
3
4
lution (200 mL; 7.88 g Al(NO ) ·9H O, 21 mmol; 2.90 g ZrOCl ·8H O,
3
3
2
2
2
3
0 min, respectively. After completion of the reaction, the reactor
9
mmol) with ultrasonication for 30 min. Then, the pH value of the
was rapidly cooled to room temperature in a cold water bath.
suspension was adjusted to 9 by slow addition of ammonium hy-
droxide solution followed by stirring for another 5 h. Afterwards,
the precipitate was collected by using a permanent external
magnet, washed thoroughly with deionized water until neutral pH
following dried at 808C overnight. Finally, the dried precipitate was
calcined at 3008C under nitrogen atmosphere for 4 h resulting in
the catalyst henceforth referred to as Al Zr @Fe O (1/1). Analogous
Identification of liquid products in the reaction mixture was ach-
ieved by GC–MS (Agilent 6850-5975C) equipped with HP-5MS ca-
pillary column (30.0 mꢁ250 mmꢁ0.25 mm). The reactant and prod-
uct samples were quantitatively analyzed on the basis of standard
curves with commercial samples using naphthalene as internal
standard on a GC (Agilent 6890n) equipped with FID detector and
HP-5MS capillary column (30.0 mꢁ250 mmꢁ0.25 mm). The
7
3
3
4
Al Zr @Fe O (2/1) and Al Zr @Fe O (1/2) catalysts were prepared
7
3
3
4
7
3
3
4
&
ChemCatChem 2017, 9, 1 – 10
www.chemcatchem.org
8
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
temperature program used in the analysis was 608C for 1 min fol-
lowed by heating to 2308C with a heating ramp of 108Cmin and
hold for 4 min.
[15] S. Sitthisa, T. Sooknoi, Y. Ma, P. B. Balbuena, D. E. Resasco, J. Catal. 2011,
À1
2
77, 1–13.
16] R. V. Sharma, U. Das, R. Sammynaiken, A. K. Dalai, Appl. Catal. A 2013,
54, 127–136.
[
4
[
[
17] M. Grasemann, G. Laurenczy, Energy Environ. Sci. 2012, 5, 8171–8181.
18] R. Bigler, R. Huber, M. Stçckli, A. Mezzetti, ACS Catal. 2016, 6, 6455–
Acknowledgements
6
464.
[
[
[
19] H. Li, Z. Fang, S. Yang, ACS Sustainable Chem. Eng. 2016, 4, 236–246.
20] M. Chia, J. A. Dumesic, Chem. Commun. 2011, 47, 12233–12235.
21] L. Bui, H. Luo, W. R. Gunther, Romꢂn-Y. Leshkov, Angew. Chem. Int. Ed.
J.H. thank the Chinese State Scholarship Fund (No.
201606670008) for financial support of the work, and acknowl-
2
013, 52, 8022–8025; Angew. Chem. 2013, 125, 8180–8183.
edge the Department of Chemistry, Technical University of Den-
mark for support.
[
[
[
[
[
[
22] X. Tang, H. Chen, L. Hu, W. Hao, Y. Sun, X. Zeng, L. Lin, S. Liu, Appl.
Catal. B 2014, 147, 827–834.
23] J. He, H. Li, Y. Lu, Y. Liu, Z. Wu, D. Hu, S. Yang, Appl. Catal. A 2016, 510,
1
1–19.
24] J. Song, B. Zhou, H. Zhou, L. Wu, Q. Meng, Z. Liu, B. Han, Angew. Chem.
015, 127, 9531–9535.
Conflict of interest
2
25] J. He, H. Li, Y. Liu, W. Zhao, T. Yang, W. Xue, S. Yang, J. Ind. Eng. Chem.
2016, 43, 133–141.
26] A. H. Valekar, K. Cho, S. K. Chitale, D. Hong, G. Cha, U. Lee, D. W. Hwang,
C. Serre, J, Chang, U. K. Hwang, Green Chem. 2016, 18, 4542–4552.
27] J. R. Ruiz, C. Jimꢃnez-Sanchidriꢂn, Curr. Org. Chem. 2007, 11, 1113–1125.
The authors declare no conflict of interest.
Keywords: aldehydes · alcohols · biomass · hydrogenation ·
magnetic properties
[28] M. Koehle, R. F. Lobo, Catal. Sci. Technol. 2016, 6, 3018–3026.
[
[
[
29] J. Li, J. Liu, H. Zhou, Y. Fu, ChemSusChem 2016, 9, 1339–1347.
30] Z. Gao, L. Yang, G. Fan, F. Li, ChemCatChem 2016, 8, 3769–3779.
31] H. Li, J. He, A. Riisager, S. Saravanamurugan, B. Song, S. Yang, ACS Catal.
[
1] D. M. Alonso, S. G. Wettstein, J. A. Dumesic, Chem. Soc. Rev. 2012, 41,
8075–8098.
2
016, 6, 7722–7727.
32] J. Guo, S. Zhu, Y. Cen, Z. Qin, J. Wang, W. Fan, Appl. Catal. B 2017, 200,
11–619.
[
[
2] M. J. Gilkey, B. Xu, ACS Catal. 2016, 6, 1420–1436.
3] H. Li, Z. Fang, R. L. Smith, Jr., S. Yang, Prog. Energy Combust. Sci. 2016,
[
[
6
5
5, 98–194.
4] H. Li, S. Saravanamurugan, S. Yang, A. Riisager, Green Chem. 2016, 18,
26–734.
5] S. Saravanamurugan, A. Riisager, E. Taarning, S. Meier, Chem. Commun.
016, 52, 12773–12776.
6] R. Mariscal, Maireles-P. Torres, M. Ojeda, I. Sꢂdaba, M. L. Granados,
Energy Environ. Sci. 2016, 9, 1144–1189.
7] H. Li, S. Yang, A. Riisager, A. Pandey, R. S. Sangwan, S. Saravanamurugan,
R. Luque, Green Chem. 2016, 18, 5701–5735.
33] F. Chambon, F. Rataboul, C. Pinel, A. Cabiac, E. Guillon, N. Essayem,
Appl. Catal. B 2011, 105, 171–181.
34] E. Pighin, V. K. Dꢅez, J. I. D. Cosimo, Appl. Catal. A 2016, 517, 151–160.
35] C. Yue, G. Li, E. A. Pidko, J. J. Wiesfeld, M. Rigutto, E. J. M. Hensen, Chem-
SusChem 2016, 9, 2421–2429.
[
[
[
[
[
[
7
[
[
2
[
36] C. Opris, B. Cojocaru, N. Gheorghe, M. Tudorache, S. M. Coman, V. I. Par-
vulescu, B. Duraki, F. Krumeich, J. A. van Bokhoven, J. Catal. 2016, 339,
209–227.
[
[
[
37] H. H. T. Vu, T. S. Atabaev, N. D. Nguyen, Y. Hwang, H. Kim, J. Sol-Gel Sci.
Technol. 2014, 71, 391–395.
38] M. Wang, G. Fang, P. Liu, D. Zhou, C. Ma, D. Zhang, J. Zhan, Appl. Catal.
B 2016, 188, 113–122.
39] J. M. Rubio-Caballero, S. Saravanamurugan, P. Maireles-Torres, A. Riisag-
er, Catal. Today 2014, 234, 233–236.
8] M. Audemar, C. Ciotonea, K. D. O. Vigier, S. Royer, A. Ungureanu, B.
Dragoi, E. Dumitriu, F. Jꢃrꢄme, ChemSusChem 2015, 8, 1885–1891.
9] A. Corma, S. Iborra, A. Velty, Chem. Rev. 2007, 107, 2411–2502.
[
[
10] F. H. Isikgor, C. R. Becer, Polym. Chem. 2015, 6, 4497–4559.
11] D. M. Alonso, S. G. Wettstein, J. A. Dumesic, Green Chem. 2013, 15, 584–
95.
12] M. Li, Y. Hao, Cꢂrdenas-F. Lizana, M. A. Keane, Catal. Commun. 2015, 69,
19–122.
5
[
Manuscript received: August 1, 2017
1
Revised manuscript received: September 6, 2017
[13] S. T. Thompson, H. H. Lamb, ACS Catal. 2016, 6, 7438–7447.
[
14] Q. Yuan, D. Zhang, L. van Haandel, F. Ye, T. Xue, J. Mol. Catal. A 2015,
Accepted manuscript online: && &&, 0000
Version of record online: && &&, 0000
406, 58–64.
ChemCatChem 2017, 9, 1 – 10
www.chemcatchem.org
9
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

FULL PAPERS
J. He, H. Li, A. Riisager,* S. Yang*
Magnetic hydrogenation: The bifunc-
tional Al–Zr@Fe O catalyst shows excel-
3
4
&& – &&
lent catalytic performance in the catalyt-
ic transfer hydrogenation of furfural and
other biomass-derived furanic aldehydes
using alcohol as hydrogen donor and
offers facile reuse in consecutive reac-
tions after its easy recovery by an exter-
nal magnet.
Catalytic Transfer Hydrogenation of
Furfural to Furfuryl Alcohol with
Recyclable Al–Zr@Fe Mixed Oxides
&
ChemCatChem 2017, 9, 1 – 10
www.chemcatchem.org
10
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!