Versatile nickel-lanthanum(III) catalyst for direct conversion of cellulose to glycols

Article abstract of DOI:10.1021/cs501372m

Using cellulosic biomass to synthesize bulk quantities of high-value chemicals is of great interest for developing a sustainable biobased society. Especially, direct catalytic conversion of cellulose to glycols, important building blocks for polymers, remains a grand challenge. Herein, we report the development of a versatile binary nickel-lanthanum(III) catalyst for the conversion of cellulose to both ethylene glycol (EG) and propylene glycol (1,2-PG) in a yield of 63.7%, which is one of the best performances reported for this catalytic reaction. Especially, lanthanum(III) exhibited a high level of activity toward the degradation of cellulose (TON = 339) at a very low concentration (0.2 mmol/L). On the basis of density functional theory calculations and experimental analyses, we addressed a dual route for this catalytic mechanism: a major route involving the selective cracking of sugars into C₂ molecules and a minor route involving the hydrogenolysis of sugar alcohols. Lanthanum(III) catalyzes the cleavage of the C2-C3 bond in glucose via sequential epimerization and 2,3-hydride shift reactions to form glycolaldehyde, the precursor of EG.

Full text of DOI:10.1021/cs501372m

Research Article
pubs.acs.org/acscatalysis
Versatile Nickel−Lanthanum(III) Catalyst for Direct Conversion of
Cellulose to Glycols
Ruiyan Sun,†,‡ Tingting Wang,§ Mingyuan Zheng,*,† Weiqiao Deng,*,§ Jifeng Pang,† Aiqin Wang,†
Xiaodong Wang,† and Tao Zhang*,†
†State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China,
§State Key Laboratory of Molecular Reaction Dynamics, Dalian National Laboratory for Clean Energy, Dalian Institute of Chemical
Physics, Chinese Academy of Sciences, Dalian 116023, China
‡Graduate University of Chinese Academy of Sciences, Beijing 100049, China
S
* Supporting Information
ABSTRACT: Using cellulosic biomass to synthesize bulk quantities of high-value
chemicals is of great interest for developing a sustainable biobased society. Especially,
direct catalytic conversion of cellulose to glycols, important building blocks for
polymers, remains a grand challenge. Herein, we report the development of a
versatile binary nickel−lanthanum(III) catalyst for the conversion of cellulose to
both ethylene glycol (EG) and propylene glycol (1,2-PG) in a yield of 63.7%, which
is one of the best performances reported for this catalytic reaction. Especially,
lanthanum(III) exhibited a high level of activity toward the degradation of cellulose
(TON = 339) at a very low concentration (0.2 mmol/L). On the basis of density
functional theory calculations and experimental analyses, we addressed a dual route
for this catalytic mechanism: a major route involving the selective cracking of sugars
into C2 molecules and a minor route involving the hydrogenolysis of sugar alcohols.
Lanthanum(III) catalyzes the cleavage of the C2−C3 bond in glucose via sequential epimerization and 2,3-hydride shift reactions
to form glycolaldehyde, the precursor of EG.
KEYWORDS: cellulose, ethylene glycol, propylene glycol, nickel−lanthanum catalyst, dual route, theoretical calculation
conversion process was conducted in the presence of a series of
tungstenic catalysts, including Ni−W2C, WP, Ni−W, WO3, and
Ru/AC-H2WO4, the EG yields reached unprecedented levels of
49.8−69.5% (carbon yield).13,14,18−23 The use of both
heterogeneous tungstenic catalysts, such as tungsten carbides
and bimetallic catalysts,13,22 and homogeneous tungstenic
catalysts in combination with hydrogenation catalysts, such as
Raney Ni−H2WO4 and Ru/AC-ammonium metatungstate,24,25
has been reported to provide enhanced levels of performance
with regard to the conversion of cellulose. The tungstenic
species that dissolve in water during the reaction have been
found to be catalytically active toward the degradation of sugars
via a retro-aldol condensation to form glycolaldehyde, which is
a precursor of EG.15,19 However, once sugars have been
hydrogenated to the corresponding hexitols, they become so
chemically stable that they cannot be degraded to C2 molecules
by the tungstenic catalysts.15,22 With this in mind, an ideal
catalyst for the production of glycols from cellulose would
therefore need to perform two functions concurrently,
including the degradation of sugars and the degradation of
the resulting hexitols.
INTRODUCTION
■
The ever-diminishing reserves of fossil energy resources and
increasing levels of energy consumption will ultimately result in
global energy shortages and serious environmental problems.
The conversion of cellulose into chemicals and fuels could play
an important role in averting these problems, and research
toward this goal has grown considerably during the past
decade.1−3 Efficient reaction processes and catalysts for the
selective degradation of cellulose, however, have been difficult
to identify because of the robust crystalline structure and high
chemical stability of cellulose,4−6 and further work is therefore
needed to explore this area in greater detail. Among the variety
of chemical conversions reported for cellulose,1,7−11 its catalytic
transformation to ethylene glycol (EG) and 1,2-propylene
glycol (1,2-PG) by hydrolytic hydrogenation represents a
particularly attractive process because of its high atomic
economy, value-added products and large market demand.12−16
Glycols are widely used in the synthesis of a broad range of
materials, including polyester, antifreezing agents, and chemical
intermediates.12 For EG, in particular, the world capacity is
forecasted to expand from 28 million tons in 2011 to 36 million
by 2016.17
A breakthrough in the conversion of cellulose to EG and 1,2-
PG was achieved in 2008, with further developments being
reported in the years that followed. When the cellulose
Received: September 11, 2014
Revised: November 24, 2014
© XXXX American Chemical Society
874
DOI: 10.1021/cs501372m
ACS Catal. 2015, 5, 874−883

ACS Catalysis
Research Article
Besides tungstenic catalysts, several researchers have recently
explored the use of other catalysts for the conversion of
cellulose to glycols, including Ni/ZnO and Cu/Cr2O3.
Disappointingly, however, these catalysts provided relatively
low yields of EG (19.1−28.9%), with 1,2-PG being formed as
the main product (34.4−47.6%).26,27 Given that the market
capacity for EG is 1 order of magnitude higher than that of 1,2-
PG, catalysts with a high level of selectivity for EG would be
much more attractive for the conversion of cellulose to glycols.
Furthermore, the reaction mechanism for the conversion of
cellulose to EG over these catalysts has not yet been elucidated
at the molecular level.
solution), cerium nitrate hexahydrate (Ce(NO3)3·6H2O), and
aluminum nitrate nonahydrate (Al(NO3)3·9H2O) were pur-
chased from Kermel Chemical Reagent Co. Lanthanum triflate
(La(CF3SO3)3) was purchased from Alfa Aesar. D2O was
purchased from Qingdao Tenglong Weibo Technology Co..
Lanthanum oxide (La2O3) was prepared by the calcination of
La(OH)3 at 1123 K in air for 5 h. Activated carbon (AC, SBET
=
1203 m2 g−1) was supplied by Beijing Guanghua-Jingke
Activated Carbon Co. 10% Ni/AC catalysts were prepared
according to the incipient wetness impregnation method using
an aqueous solution of Ni(NO3)2 as a precursor followed by
drying at room temperature overnight followed by a period of
heating at 393 K for 8 h. The resulting material was then
reduced under a stream of H2 at 523 K for 2 h. Prior to being
exposed to air, the catalysts were passivated under a stream of
1% O2 in N2 for at least 5 h. 10% Ni-0.5% Ir (Ru)/La2O3 was
prepared according to the wet impregnation method. In a
typical preparation, 0.8 g of nickel nitrate and 0.2 g of a 10%
aqueous solution of RuCl3 (H2IrCl6) were placed in a 250 mL
beaker containing 20 mL of deionized water, followed by 1.5 g
of La2O3 under continuous stirring. The resulting mixture was
then heated to 303 K to allow for the complete evaporation of
water. The resulting catalyst was then dried at 393 K for 8 h,
calcined at 773 K for 4 h under a stream of N2 flow, and lastly
reduced at 773 K for 5 h under a stream of H2.
Significant research efforts have been directed toward the use
of homogeneous catalysts for the conversion of biomass into
useful materials,28 as exemplified by the isomerization of
glucose to fructose by Al(III) and Cr(III),29,30 the conversion
of fructose to 5-hydroxymethyl furfural with Cr(III),29,31 and
the degradation of cellulose to lactic acid in the presence of
Pb(II).32 There are many unique advantages associated with
the use of homogeneous catalysts, including high reaction
activity and selectivity, enhanced accessibility to catalytically
active sites for bulky reactants such as the natural polymer of
cellulose, and the simplicity with which the reaction
mechanisms of these processes can be studied using computa-
tional methods.31,32 The use of homogeneous catalysis in
combination with heterogeneous catalysis could lead to
significant improvements in reaction efficiency,33 a reduction
in catalyst costs and allow for the exploration of novel reaction
processes.24,25 Herein, we disclose a versatile binary catalyst,
involving the combination of homogeneous lanthanum(III)
with a heterogeneous nickel catalyst, which produced EG and
1,2-PG in an overall yield of 63.7% (73.2% of which was EG)
from a cellulose conversion process. It is noteworthy that the
current Ni−La(III) catalyst exhibited a dual-route reaction
mode to produce EG. Furthermore, the concentration of
La(III) in the solution could be reduced to an extraordinarily
low level of 0.2 mmol/L, which is only 1/6 of the amount of
tungstenic catalyst required to give a similarly high yield of
EG,19 not to mention that the price of the La(III) catalyst is
just 1/3 of that of the tungstenic catalyst. In light of these
results, this newly developed Ni−La(III) catalyst represents a
cost-effective, practicable and affordable method for the
conversion of cellulose to EG. On the basis of theoretical
calculations and correlations with the experiment results, we
have also disclosed a mechanism to describe the way in which
the catalytic La(III) species facilitates the catalytic retro-aldol
condensation of glucose to allow for the formation of EG.
Characterization Methods. X-ray diffraction (XRD)
patterns were recorded on a PANalytical PW3040/60 X′ Pert
PRO diffractometer equipped with a Cu Kα radiation source (λ
= 0.15432 nm), operating at 40 kV and 40 mA. The metals ions
in solution after the reaction were determined using inductively
coupled plasma atomic emission spectrometry (ICP-AES) on
an IRIS intrepid II XSP instrument (Thermo Electron
Corporation). Total organic carbon (TOC) analysis was
conducted on a Vario EL III element analyzer (Elementar) to
determine the carbon content of the liquid products.
Transmission electron microscopy analysis was performed on
a JEM-2000EX (JEOL) microscope. The CO2-temperature-
programmed desorption (TPD) experiment was conducted on
a Micromeritics AutoChem II 2920 automated catalyst
characterization system. In detail, a La2O3 sample was heated
at 773 K for 1 h in an oxygen flow. CO2 was then dosed several
times at 323 K to allow for the adsorption of CO2 to reach
saturation. Physisorbed CO2 was then removed in a helium
stream at 323 K over 20 min. This stage was immediately
followed by the CO2-TPD experiment, which was conducted
under a helium flow (20 mL/min) from 323 to 1173 K at a
heating rate of 10 K/min, and the resulting signal was recorded
with a thermal conductivity detector (TCD). Solution NMR
analysis was performed on a Bruker Avance III 400 MHz
instrument (Bruker). All of the experiments were conducted in
D2O. Mass and MS/MS spectrometry measurements were
performed on an Agilent Techonlogies 6540 Ultra-High-
Definition (UHD) Accurate-Mass Q-TOF LC/MS system
(Agilent) equipped with an electrospray ionization source
operating in the positive ion mode. Electrospray ionization was
conducted at 3500 V, and a collision energy of 20 eV was used
for the collision-induced dissociation stage in the MS/MS
measurements. The MS and MS/MS measurement were
conducted over m/z range of 50 to 600.
METHODS
■
Preparation of Catalysts. Lanthanum hydroxide (La-
(OH)3), nickel nitrate hexahydrate (Ni(NO3)2·6H2O), cobalt
nitrate hexahydrate (Co(NO3)2·6H2O), barium nitrate (Ba-
(NO3)2), D-(+)-glucose, and zirconium nitrate pentahydrate
(Zr(NO3)4·5H2O) were purchased from Sinopharm Chemical
Reagent Co. Ferric nitrate nonahydrate (Fe(NO3)3·9H2O),
lanthanum nitrate hexahydrate (La(NO3)3·6H2O), lanthanum
chloride heptahydrate (LaCl3·7H2O), lanthanum hexaboride
(LaB6), lanthanum acetate (La(OAc)3·nH2O), yttrium nitrate
hexahydrate (Y(NO3)3·6H2O), D-1-13C-glucose, and lead
nitrate (Pb(NO3)2) were purchased from Aladdin Chemical
Co. Zinc nitrate hexahydrate (Zn(NO3)2·6H2O), chromic
nitrate nonahydrate (Cr(NO3)3·9H2O), strontium nitrate
(Sr(NO3)2), manganous nitrate (Mn(NO3)2, 50% aqueous
Catalytic Reaction. The catalytic conversion of cellulose
(Merck, microcrystalline with an average particle size of about
90 μm and a specific surface area of 1.7 m2 g−1) was carried out
in a stainless-steel autoclave (Parr Instrument Company) at 5
875
DOI: 10.1021/cs501372m
ACS Catal. 2015, 5, 874−883

ACS Catalysis
Research Article
a
Table 1. Results of Cellulose Conversions with Binary Catalysts of Metal Compounds in Combination with 10%Ni/AC
c
yield (%)
b
d
d
entry
catalyst
conv. (%)
Ery
Gly
EG
1,2-PG
Man
1,2-BD
Sor
yieldGC (%)
yieldLC (%)
1
2
blank
87.8
85.4
87.2
94.0
85.2
85.1
89.1
91.6
90.4
86.6
90.4
94.8
92.6
95.3
94.3
95.0
100
1.9
2.4
2.2
3.0
2.2
1.9
2.0
2.0
1.0
2.5
2.2
1.6
1.2
2.1
1.4
1.3
1.1
1.2
2.2
1.4
0.7
2.5
3.3
3.3
5.9
3.1
3.2
3.5
4.1
2.0
4.2
3.5
4.8
4.7
5.7
5.0
3.7
1.6
4.2
5.5
5.7
22.5
9.5
9.8
9.5
6.2
6.3
5.8
5.7
6.3
5.7
5.6
5.8
2.8
5.5
5.3
1.9
1.0
2.0
1.0
1.1
0.8
1.0
1.7
1.2
0
5.0
4.2
3.8
4.7
4.5
4.4
4.3
4.3
4.5
4.7
3.5
3.3
4.3
4.6
1.3
3.1
0.9
1.3
3.6
3.2
3.9
16.6
12.8
11.9
5.9
12.2
11.1
11.8
9.9
5.4
6.4
7.9
0.7
0.6
1.6
1.1
3.3
1.4
0.4
3.3
0.6
0
2.0
3.3
3.6
6.0
3.4
4.6
3.5
5.0
2.5
6.2
5.1
7.6
10.1
8.5
5.0
2.7
2.0
2.6
3.5
5.1
7.4
74.2
76.2
75.4
80.8
74.5
74.7
75.8
77.6
73.8
75.4
74.5
82.4
71.2
78.6
84.6
81.5
71.0
73.5
84.3
82.2
82.2
Co(NO3)2
Fe(NO3)3
Zn(NO3)2
Al(NO3)3
Cr(NO3)3
Zr(NO3)4
Mn(NO3)2
Pb(NO3)2
Ba(NO3)2
Sr(NO3)2
Y(NO3)3
Ce(NO3)3
La(NO3)3
La(OAc)3
LaB6
15.6
14.1
16.7
15.4
16.2
15.5
16.6
17.2
18.0
17.1
29.5
30.2
32.6
33.9
32.0
20.5
22.3
36.0
38.4
25.1
3
9.5
4
12.3
8.8
5
6
9.3
7
9.6
8
11.7
12.7
10.7
9.4
9
10
11
12
13
14
15
16
17
18
19
20
21
11.5
15.6
13.8
13.1
15.2
10.0
8.8
LaCl3
La(OTf)3
La2O3
98.7
95.6
96.6
100
14.7
14.6
27.3
La(OH)3
e
La(OH)3
a
Reaction conditions: 0.15 g of 10% Ni/AC, 0.02 mmol metal cation, 0.25 g of cellulose, 25 mL of H2O, 518 K, 5 MPa of H2, 120 min, 800 rpm. Ery,
Gly, EG, 1,2-PG, Man, 1,2-BD and Sor are abbreviations for erythritol, glycerol, ethylene glycol, 1,2-propylene glycol, mannitol, 1,2-butanediol and
b
c
sorbitol, respectively. Cellulose conversion (wt %) was calculated by the change of cellulose weight before and after the reaction The yields of
polyols were calculated by using the equation: yield (C%) = (masses of carbon in the products)/(masses of carbon in cellulose charged into the
reactor) × 100%. YieldGC and YieldLC indicate the total carbon in the gas and liquid products, respectively, and were calculated on the basis of
carbon. 0.25 g of sorbitol was employed as substrate.
d
e
MPa of H2 pressure (measured at room temperature) and 518
K. For each reaction, 0.25 g of cellulose, a designated amount of
catalyst, and 25 mL of deionized water were placed in a 75 mL
autoclave, and the resulting mixture was stirred at 800 rpm. For
the recycling test, the catalysts were recovered after each run by
filtration and washed several times with deionized water before
being used in the next reaction together with the same amounts
of cellulose and water used in the first run. The liquid-phase
products were analyzed by HPLC on an Agilent 1200 system
equipped with a Shodex Sugar SC1011 column and a
differential refractive index detector (RID). HPLC analysis
was conducted at a column temperature of 318 K using water as
the mobile phase with a flow rate of 0.6 mL min−1. The samples
were injected with an injection volume of 5.0 μL. Gaseous
products were analyzed by GC on an Agilent 6890N system
equipped with a TDX-1 column and a TCD. Cellulose
conversions were determined by measuring the difference in
the cellulose weight before and after the reaction. The polyol
yields were calculated on the basis of the amount of carbon
according to the equation: yield (C%) = (mass of carbon in the
products)/(mass of carbon in the cellulose charged into the
reactor) × 100%. The carbon content of the cellulose fed into
the reactor was 4230 ppm.21 The yields of EG and 1,2-PG from
the cited literature reports have also been presented in the
carbon yields.
calculations using the PCM model with water as the solvent.
The harmonic frequencies at the equilibrium geometries were
calculated to confirm the first-order saddle points, local minima
on the potential energy surfaces, and the zero-point energy.
Correlations between the corresponding imaginary frequency
mode and intrinsic reaction coordinate calculations were
evaluated to confirm the transition states and the stable
structures.
RESULTS AND DISCUSSION
■
Screening of Catalysts for Cellulose Conversion. A
variety of metal salts were initially screened in combination
with a 10% Ni/AC catalyst for the cellulose conversion process.
As shown in Table 1, compared with the blank experiment
involving the use of 10% Ni/AC in isolation (Table 1, entry 1),
the addition of different transition metal salts led to significant
changes in yields of EG and 1,2-PG. The addition of La(III)
salts afforded the highest levels of selectivity of all of the salts
tested toward the production of EG and 1,2-PG. Hexitols
(22.8% yield) were observed as the major polyol products in
the blank experiment, with EG and 1,2-PG being formed in
yields of 9.5 (10.8% EG selectivity) and 9.8%, respectively, after
2 h at 518 K (Table 1, entry 1). Pleasingly, however, the
addition of La(III) led to a significant increase in the EG yield
to 32.6%, which was accompanied by a 3-fold increase in the
EG selectivity (i.e., 34.2% vs 10.8%). The 1,2-PG yield also
increased to 13.8% (Table 1, entries 1 and 14). The addition of
Ce(III) and Y(III) also led to an increase in EG selectivity,
which could be attributed to similarities in the chemical
properties of the lanthanide elements or elements belonging to
group 3 of the periodic table (Table 1, entries 12 and 13). In
Computation Methods. All calculations were performed
using the Gaussian 09 software package based on the DFT
method. The standard 6-31G** basis set was used for the
geometry optimizations, frequency calculations, IRC, and
single-point energy calculations (the LanL2DZ basis set was
employed for La). The solvent effects were considered in the
876
DOI: 10.1021/cs501372m
ACS Catal. 2015, 5, 874−883

ACS Catalysis
Research Article
contrast, none of the other metal salts tested in the current
study lead to similar changes in the selectivity of the reaction
toward the formation of EG (Table 1, entries 2−11).
PG yields remained close to 38 and 15%, respectively, as the
mass ratio of cellulose/La(OH)3 increased from 2.5 to 100.
Surprisingly, the use of only 1 mg of La(OH)3 (0.2 mmol/L
La(III)) in the reaction (i.e., a cellulose/La(OH)3 mass ratio is
250) still afforded high levels of conversion with EG and 1,2-
PG yields of up to 30.9 and 14.4%, respectively. The TON for
the formation of EG and 1,2-PG at this concentration of La(III)
was 339. The ability of such an extraordinarily low
concentration of La(III) to realize this level of conversion is
of remarkable significance for the practical and low cost
application of an homogeneous catalyst. Furthermore, the
La(III) ions exhibited catalytic activity toward the cellulose
conversion, as evidenced by the higher level of cellulose
conversion when the reaction was conducted in the presence of
increased concentration of La(III).
The effects of the reaction time and temperature on the
cellulose conversion are shown in Figure 2a,b. The yields of EG
and 1,2-PG increased consistently from the beginning of
reaction and eventually reached a plateau around 39 and 13%,
respectively. Compared with the Ru/AC-H2WO4 catalyst,
which realized the complete conversion of cellulose in 30
min,19 the binary Ni−La(III) catalyst required a longer reaction
time to allow for the complete conversion of cellulose. The
requirement for a longer reaction time in this case could be
related to weaker acidity of Ni−La(III), which would lead to a
lower rate of cellulose hydrolysis. Extending the reaction time
to 190 min did not lead to a change in the product distribution,
which demonstrated that the polyol products were stable to the
Ni−La(III) catalyst under the current conditions. The cellulose
conversion and the yields of EG and 1,2-PG increased with
increasing reaction temperature up to 518 K. At higher
temperatures, the hydrogenolysis of the polyols was enhanced
to form more gases at the expense of EG and 1,2-PG
production.25
The nature of the anion of La(III) salt had a significant
impact on the EG yields (Table 1, entries 14−20). For example,
nitrate ions were superior to chloride and triflate ions, most
likely because the nitrate anions would have decomposed under
the reaction conditions. La(OH)3 and La2O3 provided the best
performances of all of the La(III) salts tested in the current
study in terms of their selectivity for EG (Table 1, entries 19
and 20), and La(OH)3 and La2O3 were therefore used as the
La(III) source in all of the following reactions. The use of these
compounds would also eliminate the possibility of the anion of
the La(III) salt having an adverse impact on the reaction
because these compounds can dissolve in hot-compressed water
to release La(III) ions in situ, in a similar manner to a tungsten
acid catalyst that we studied previously.19
Catalytic Performance of Ni−La(III) for Cellulose
Conversion. The effect of the mass ratio of cellulose to
La(OH)3 in the reaction is shown in Figure 1. The EG and 1,2-
The stabilities of the different Ni−La(III) catalysts were
examined in the recycling experiments, and the results are
shown in Figure 3. In the case of 10% Ni/AC-La(OH)3 (Figure
3a), the EG yield decreased remarkably from an initial value of
38.0 to 12.3% after three consecutive runs. The TEM images
(Figure 4a,b) and XRD patterns (Figures S1a,b) of the catalysts
revealed that the nickel particles on the 10% Ni/AC were
significantly sintered after the reaction, which could be
attributed to the Ostwald ripening of the Ni particles following
their exposure to hot water and a H2 atmosphere. It is known
Figure 1. Effect of the mass ratio of cellulose to La(OH)3 on the
product yields in the cellulose conversion. Reaction conditions: 518 K,
110 min, 5 MPa of H2, 800 rpm, 0.15 g of 10% Ni/AC, 0.25 g of
cellulose and 25 mL of H2O.
Figure 2. Results of cellulose conversion with the binary catalyst of La(OH)3 and 10%Ni/AC depending on (a) reaction time, reaction temperature:
518 K; (b) reaction temperature, reaction time: 110 min; Reaction conditions: 0.15 g of 10% Ni/AC, 0.1 g of La(OH)3, 0.25 g of cellulose, 25 mL of
H2O, 5 MPa of H2, 800 rpm.
877
DOI: 10.1021/cs501372m
ACS Catal. 2015, 5, 874−883

ACS Catalysis
Research Article
Figure 3. Recycling stabilities of different catalysts in cellulose conversion. (a) 10% Ni/AC-La(OH)3, reaction conditions: 0.15 g of 10% Ni/AC, 0.1
g of La(OH)3, 110 min; (b) 10% Ni-0.5% Ir/La2O3, reaction conditions: 0.15 g of 10% Ni-0.5% Ir/La2O3, 150 min; (c) 10% Ni-0.5% Ru/La2O3, the
same reaction conditions as (b). (0.25 g of cellulose, 25 mL of H2O, 518 K, 5 MPa of H2, 800 rpm).
Figure 4. TEM images of fresh and used catalysts: (a) fresh and (b) used 10% Ni/AC, (c) fresh and (d) used 10% Ni-0.5% Ir/La2O3.
that La2O3 can stabilize transition metals when used as a
catalyst support or additive.34−38 The stability of a nickel
catalyst can also be improved by the introduction of a small
amount of a noble metal.39,40 With these things in mind, we
designed a supported bimetallic catalyst consisting of 10% Ni-
0.5% Ir/La2O3 that would simultaneously provide hydro-
878
DOI: 10.1021/cs501372m
ACS Catal. 2015, 5, 874−883

ACS Catalysis
Research Article
genation sites and La(III) ions for the reaction. As shown in
Figure 3b, the bimetallic catalyst exhibited much greater
stability than the previous catalyst over nine cycles. A plot of
the polyols yields observed during the recycling experiments
revealed that the overall yield of EG and 1,2-PG reached the
highest value of 63.7% (73.2% of which was EG) during the
sixth run. A similar phenomenon was also observed for the 10%
Ni-0.5% Ru/La2O3 catalyst (Figure 3c). The concentration of
La(III) ions in solution after the reaction was found to be in the
range of 100−200 ppm for runs 1−7 (Table S1) meanwhile
sintering of nickel particles was not notable, as shown in Figure
4c,d, which demonstrated that La2O3 could be used as a
suitable support for stabilizing the nickel particles at the same
time as releasing La(III) ions in situ for the reaction.
Reaction Pathways. In terms of the reaction mechanisms
involved in the conversion of cellulose to EG and 1,2-PG, one
possible pathway involving the degradation of hexitols has been
studied extensively.41−46 To determine whether this mechanism
was occurring in our reaction, we used sorbitol as substrate in a
conditional experiment (Table 1, entry 21). Following a
reaction time of 2 h, the sorbitol had been completely degraded
to give glycerol, EG and 1,2-PG in yields of 22.5, 25.1 and
27.3%, respectively. This result demonstrated that the Ni−
La(III) catalyst was catalytically active toward the conversion of
hexitols. This result was therefore consistent with that for the
conversion of cellulose in the presence of Ni−La(III) (Table 1,
entries 1 and 14−20), where the yields of EG and 1,2-PG
increased at the expense of those of the hexitols. The results of
a CO2-desorption experiment on La2O3 showed a significant
CO2 desorption peak at 953 K (Figure S2), which indicated
that the La(III) salts were becoming basic when they were
hydrolyzed in hot water during the reaction.47 The hexitols
could therefore be degraded by the Ni−La(III) catalyst via the
same route that occurs in the presence of basic and
hydrogenation catalysts. Catalytic behavior of this type from
an La(III)-based catalyst is quite different from that of the
tungstenic catalysts, which exhibit no activity toward the
degradation of hexitols.15,22
It is noteworthy, however, that the EG:1,2-PG:Gly product
ratio of the sorbitol conversion was 1.1:1.2:1, which is quite
different from that of the cellulose conversion, which was
6.7:2.6:1 (Table 1, entries 20 and 21). The Ni−La(III) catalyst
demonstrated preferential selectivity for the production of EG
in the cellulose conversion process. In this respect, the La(III)-
based catalyst behaved like the tungstenic catalyst, which shows
an especially high level of selectivity toward the production of
EG.
The reaction intermediate glycolaldehyde (GA) was
monitored throughout the cellulose conversion process to
develop a deeper understanding of the role of La(III) in the
degradation of cellulose to EG. When the reaction was
conducted in the presence of La(III) and the absence of the
Ni catalyst at temperatures above 518 K, the yields of GA were
much greater than the yield obtained in the blank experiment
(Figure 5). This catalytic behavior of La(III) was very similar to
that of the tungstenic catalyst.19 However, the maximal GA
yield achieved by La(III) catalysis was slightly lower than that
of the tungstenic catalyst. The starting temperatures for the
formation of GA with the La(III) catalysts were also 20 K
higher than those of the tungstenic catalysts.19 This result could
therefore account for the lower EG yields observed for the Ni−
La(III) catalysts in the cellulose conversion process compared
with the tungstenic catalyst.
Figure 5. Glycolaldehyde formation in the cellulose conversion over
different La(III) compounds without hydrogenation catalyst at 518 K
and 5 MPa of H2 for 30 min.
On the basis of the experimental results described above and
our understanding of the mechanisms involved in similar
processes,15 we have proposed a tentative dual-route reaction
mode mechanism for the current conversion, which is shown in
Scheme 1. During the conversion of cellulose to EG, cellulose
would be initially hydrolyzed to glucose and oligosaccharides by
the H+ ions released by the hot water.11 The resulting sugars
would then be converted to EG via two different routes. The
first and major route (i) would involve the La(III)-catalyzed
degradation of the sugars via a retro-aldol condensation to form
GA, which would be hydrogenated over the Ni sites to form
EG. The second and minor route (ii) would involve the initial
hydrogenation of the sugars to form hexitols, which would
undergo a hydrogenolysis reaction to form EG and 1,2-PG.
Reaction Mechanism. Given that extensive studies have
been conducted toward developing a deeper understanding of
the catalytic degradation of hexitols in the presence of
hydrogenation catalysts under basic conditions, we decided to
focus exclusively on the mechanism of the proposed dominant
route (i). In a similar manner to Pb(II)−OH,32 Cr(III)−OH,31
and Sn(IV)−OH,48,49 La(III) can be hydrolyzed to La(III)−
OH in aqueous solution, and this species would have acted as
the active species under our current reaction conditions.50−53
Ultra-High-Definition (UHD) Accurate-Mass Q-TOF LC/MS
analysis allowed us to confirm that La(III)−OH was present in
the aqueous solutions of La(III) used in the current study
(Figure S3). On the basis of this result, La(III)−OH was used
to conduct a theoretical calculation to develop a deeper
understanding of the way in which La(III) catalyzes the
degradation of glucose to GA. The results of this calculation are
shown in Figure 6 and Tables S2−22, and a detailed
mechanistic explanation to Figure 6 is shown in Scheme S1
as well. In view of these results, it was possible to divide the
reaction into four stages, including (i) hydrogen transfer
following the formation of a complex between La(III)−OH and
glucose; (ii) C2−C3 bond cleavage via sequential glucose
epimerization and 2,3-hydride shift reactions; (iii) loss of GA;
and (iv) cracking of the erythrose into two molecules of GA.
The initial step would be a hydrogen transfer reaction from
C2-OH to La(III)−OH (stage (i). This step would occur quite
readily because of its low activation energy (ΔG‡), which was
determined to be as low as 0.61 kcal/mol. The length of the
C2−O bond would therefore decrease as a consequence of the
stronger interactions, which would promote the cleavage of the
C2−C3 bond (stage (ii). The cleavage of the C2−C3 bond
879
DOI: 10.1021/cs501372m
ACS Catal. 2015, 5, 874−883

ACS Catalysis
Research Article
Scheme 1. Proposed Reaction Pathways for the Conversion of Cellulose to Ethylene Glycol and Propylene Glycol with the Ni−
La(III) Catalyst
Figure 6. Theoretical calculation for the La(III)−OH-catalyzed conversion of glucose to glycolaldehyde. Calculated Gibbs energy profiles (in kcal/
mol) and optimized structures of the species involved in the catalytic process, including intermediates and transition states.
would comprise two steps, including (a) epimerization of
glucose with the reversion of C1 and C2 positions and (b) the
cleavage of the C2−C3 bond via a 2,3-hydride shift. The ΔG‡
for the first step was determined to be 23.54 kcal/mol, which is
the highest activation energy of the four stages. The calculated
apparent activation energy of 18.97 kcal/mol (i.e., the
difference in the Gibbs energies between I and IV) was very
close to the experimental value of 20.0 kcal/mol (Figure S4).
This result therefore highlighted the validity of the proposed
reaction model. Following the epimerization of glucose, the
cleavage of the C2−C3 bond would proceed more readily
(ΔG‡ = 17.63 kcal/mol).
GA would be readily formed during stage (iii) via a hydrogen
transfer reaction (ΔG‡ = 2.03 kcal/mol). The remaining
erythrose would undergo further cracking into two molecules of
GA during stage (iv). In contrast to glucose, the C2−C3 bond
of erythrose (i.e., the C4−C5 bond of glucose) would break
directly following the formation of a stable structure (X)
through the torsion of the C−C bonds, and this cleavage
process would occur via a hydrogen transfer process in the
absence of epimerization. The highest ΔG‡ for this stage was
determined to be 20.38 kcal/mol, which is smaller than the
value for stage (ii). Stage (iv) would therefore proceed much
more readily than stage (ii). This is also consistent with the
observed production of much larger quantities of GA during
the reaction compared with the erythrose intermediates.
To further confirm the catalytic role of La(III)−OH in the
cleavage of the C2−C3 bond of glucose, we compared the
energy barriers for the reactions conducted both with and
without La(III)−OH. As shown in Figure 7, the ΔG‡ of the
reaction conducted in the absence of La(III)−OH was 32.26
kcal/mol, which is notably higher than that of the reaction
conducted in the presence of La(III)−OH (23.54 kcal/mol).
This result therefore demonstrates that La(III)−OH catalyzes
the degradation of glucose to form GA.
Experimental evidence in support of the catalytic function of
La(III) in the cleavage of the C2−C3 bond of glucose was
provided by the analysis of the reaction solution by tandem
mass (MS/MS) spectrometry. The mass (MS) and MS/MS
results revealed that the La(III)−glucose complex and the
fragments of glucose degradation (i.e., GA and erythrose) were
formed via the cleavage of the C2−C3 bond in glucose (Figure
8). In contrast, the analysis of the reaction solution without
La(III) revealed no GA or erythrose. Instead, these results
revealed the presence of a Na(I)−glucose species as well as
several unknown fragments (Figure 9). These results therefore
confirmed that glucose was activated by the La(III) species
toward selective catalytic degradation.
880
DOI: 10.1021/cs501372m
ACS Catal. 2015, 5, 874−883

ACS Catalysis
Research Article
HPLC chromatograph (Figure S5) and could be attributed to
the isomerization of glucose to fructose, which would be
catalyzed by the basic La(III). The computational results
showed that the activation energy for the isomerization of
glucose isomerization to fructose (18.52 kcal/mol) was lower
than that for the epimerization of glucose to mannose (23.54
kcal/mol), which would explain why more fructose was
produced by this reaction than mannose (Figure S6). However,
the theoretical calculation also showed that the activation
energy for the degradation of fructose via the cleavage of its
C3−C4 bond was 27.86 kcal/mol, which is higher than that
calculated for the degradation of glucose via the cleavage of its
C2−C3 bond (23.56 kcal/mol) (Figure S7). This result
therefore indicated that the presence of La(III) was advanta-
geous for the degradation of glucose over fructose in terms of
the kinetics of the process. This could explain why EG was
observed as the dominant product over 1,2-PG under the
current reaction conditions.
Figure 7. Calculated Gibbs energy profiles (in kcal/mol) and
optimized structures of the C2−C3 bond cleavage of glucose with
and without La(III)−OH, including intermediates and transition
states.
CONCLUSIONS
■
In summary, we have developed a versatile binary Ni−La(III)
catalyst for the conversion of cellulose to EG and 1,2-PG, which
were obtained in a combined yield of 63.7% (with EG
accounting for 73.2%). This new Ni−La(III) catalyst is efficient
for the degradation of cellulose to EG and 1,2-PG even at the
very low La(III) concentration of 0.2 mmol/L. This Ni−La(III)
catalyst presents two routes for the production of EG.
According to the main route, sugars derived from the hydrolysis
of cellulose would be directly degraded to GA with high
selectivity by the La(III) catalyst. The GA would then be
hydrogenated over the Ni sites to form EG. For the minor
route, the sugars would initially undergo a hydrogenation
reaction to form hexitols, which would be catalytically degraded
to form EG and 1,2-PG by basic La(III) and metallic nickel.
Using theoretical calculations, we have clearly shown that
glucose undergoes epimerization and 2,3-hydride shift steps in
the presence of the La(III) catalyst to form GA, and the results
of these computational experiments were in good agreement
with the experimental findings.
The results of these computational studies confirmed that
La(III) catalyzed the cleavage of the C2−C3 bond of glucose
through the epimerization of glucose to mannose followed by a
2,3-hydride shift. The epimerization of glucose was confirmed
by 13C nuclear magnetic resonance (NMR) experiments using
isotopically labeled glucose at C1 (D-1-13C-glucose) as the
substrate. 13C NMR analysis of the reaction of D-1-13C-glucose
with La(III) following 1 h in deuterium oxide (D2O) at 413 K
revealed a new set of resonance signals at δC = 70.7 and 71.3
ppm (Figure 10), which were assigned to the chemical shifts of
the α- and β-pyranose units of D-2-13C-mannose, respectively.
These results indicated that D-2-13C-mannose was formed via
the rearrangement of the carbon skeleton through exchanges at
the C1 and C2 positions.54,55 This result therefore provided
further evidence in support of the theoretical calculations.
Furthermore, a large amount of D-1-13C-fructose (as evidenced
by two new resonances at δC = 62.7 and 63.9 ppm
corresponding to the α- and β-pyranose units of D-1-13C-
fructose, respectively) was also observed in the 13C NMR
spectra (Figure 10).55 This observation was consistent with the
In contrast to the tungstenic catalysts that have been studied
extensively in glucose conversion processes, the current Ni−
La(III) catalyst exhibited several attractive features for the
Figure 8. MS and MS/MS spectra of glucose in the presence of La(III). (a) MS spectrum. (b) MS/MS spectrum of the [La(C6H12O6)-H]2+ species
with m/z of 158.9797, which was selected as target ion to conduct MS/MS experiment.
881
DOI: 10.1021/cs501372m
ACS Catal. 2015, 5, 874−883

ACS Catalysis
Research Article
Figure 9. MS and MS/MS spectra of glucose in the absence of La(III). (a) MS spectrum. (b) MS/MS spectrum of the [Na(C6H12O6)]+ species with
m/z of 203.0529, which was selected as target ion to conduct MS/MS experiment.
curve and apparent activation energy of retro-aldol
condensation of glucose with La(III), HPLC chromato-
graph of the D-(+)-glucose solution, and detailed
computational results (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
*E-mail: myzheng@dicp.ac.cn. Fax: (+86) 411-84691570.
*E-mail: dengwq@dicp.ac.cn.
*E-mail: taozhang@dicp.ac.cn.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the National Nature Science
Foundation of China (21376239, 21306191, and 21176235)
Figure 10. 13C NMR spectra of the reactant solution in the absence
and National Basic Research Program (973 program
(a) and presence (b) of La(III) (9.2 mmol/L) employing D-1-13C-
2009CB226102). The authors thank Dr. Fan Yang for her
glucose (50.5 mmol/L) as the substrate in D2O at 413 K for 1 h.
kind help in Q-TOF LC/MS and NMR analysis.
conversion of cellulose to EG and 1,2-PG, especially in terms of
its practical application. For instance, La(III) can work
effectively in this process at very low concentrations, and the
price of the La(III) catalyst is only one-third of the tungstenic
catalyst. The La(III) was only mildly basic, making it noncaustic
toward the reactor equipment and nickel catalysts. The Ni−
La(III) catalyst produced EG and 1,2-PG via two routes,
effectively ensuring that all of the C6 molecules would be
degraded into small C2 and C3 polyols. Although the stability of
supported Ni−La(III) catalysts still needs further improving,
such as by combining hydrothermally stable skeletal Ni
catalysts with La(III), such an effective Ni−La(III) catalyst
represents a promising candidate and inspiring strategy for the
development of efficient catalysts for the conversion of cellulose
to EG and 1,2-PG.
REFERENCES
■
(1) Yabushita, M.; Kobayashi, H.; Fukuoka, A. Appl. Catal., B 2014,
145, 1−9.
(2) Delidovich, I.; Leonhard, K.; Palkovits, R. Energy Environ. Sci.
2014, 7, 2803−2830.
(3) Climent, M. J.; Corma, A.; Iborra, S. Green Chem. 2014, 16, 516−
547.
(4) Klemm, D.; Heublein, B.; Fink, H.-P.; Bohn, A. Angew. Chem., Int.
Ed. 2005, 44, 3358−3393.
(5) Rinaldi, R.; Schuth, F. ChemSusChem 2009, 2, 1096−1107.
̈
(6) Vilcocq, L.; Castilho, P. C.; Carvalheiro, F.; Duarte, L. C.
ChemSusChem 2014, 7, 1010−1019.
(7) Climent, M. J.; Corma, A.; Iborra, S. Green Chem. 2011, 13, 520−
540.
(8) Geboers, J. A.; Van de Vyver, S.; Ooms, R.; Op de Beeck, B.;
Jacobs, P. A.; Sels, B. F. Catal. Sci. Technol. 2011, 1, 714−726.
(9) Ruppert, A. M.; Weinberg, K.; Palkovits, R. Angew. Chem., Int. Ed.
2012, 51, 2564−2601.
ASSOCIATED CONTENT
* Supporting Information
The following file is available free of charge on the ACS
■
(10) Fukuoka, A.; Dhepe, P. L. Angew. Chem., Int. Ed. 2006, 45,
5161−5163.
S
(11) Luo, C.; Wang, S.; Liu, H. Angew. Chem., Int. Ed. 2007, 46,
7636−7639.
Publications website at DOI: 10.1021/cs501372m.
Characterization of catalysts and reaction solution (ICP-
AES, XRD, Q-TOF LC/MS and CO2-TPD), Arrhenius
(12) Yue, H.; Zhao, Y.; Ma, X.; Gong, J. Chem. Soc. Rev. 2012, 41,
4089−4089.
882
DOI: 10.1021/cs501372m
ACS Catal. 2015, 5, 874−883

ACS Catalysis
Research Article
(13) Ji, N.; Zhang, T.; Zheng, M.; Wang, A.; Wang, H.; Wang, X.;
Chen, J. Angew. Chem., Int. Ed. 2008, 47, 8510−8513.
(14) Liu, Y.; Luo, C.; Liu, H. Angew. Chem., Int. Ed. 2012, 51, 3249−
3253.
(15) Wang, A.; Zhang, T. Acc. Chem. Res. 2013, 46, 1377−1386.
(16) Zheng, M.; Pang, J.; Wang, A.; Zhang, T. Chin. J. Catal. 2014,
35, 602−613.
(17) Overview of World Fiber Production. http://dnfi.org/wp-
content/uploads/2013/08/Portugal-Paper.pdf.
(18) Ji, N.; Zhang, T.; Zheng, M.; Wang, A.; Wang, H.; Wang, X.;
Shu, Y.; Stottlemyer, A. L.; Chen, J. G. Catal. Today 2009, 147, 77−85.
(19) Tai, Z.; Zhang, J.; Wang, A.; Zheng, M.; Zhang, T. Chem.
Commun. 2012, 48, 7052−7054.
(20) Zhang, Y.; Wang, A.; Zhang, T. Chem. Commun. 2010, 46, 862−
864.
(21) Ji, N.; Zheng, M.; Wang, A.; Zhang, T.; Chen, J. G.
(48) Roman-Leshkov, Y.; Moliner, M.; Labinger, J. A.; Davis, M. E.
Angew. Chem., Int. Ed. 2010, 49, 8954−8957.
(49) Assary, R. S.; Curtiss, L. A. J. Phys. Chem. A 2011, 115, 8754−
8760.
(50) Roman
1580.
́
-Leshkov, Y.; Davis, M. E. ACS Catal. 2011, 1, 1566−
(51) Bush, M. F.; Saykally, R. J.; Williams, E. R. J. Am. Chem. Soc.
2008, 130, 9122−9128.
(52) Kobayashi, S.; Nagayama, S.; Busujima, T. J. Am. Chem. Soc.
1998, 120, 8287−8288.
(53) Kobayashi, S. Pure Appl. Chem. 2007, 79, 235−245.
(54) Hayes, M. L.; Pennings, N. J.; Serianni, A. S.; Barker, R. J. Am.
Chem. Soc. 1982, 104, 6764−6769.
(55) Gunther, W. R.; Wang, Y.; Ji, Y.; Michaelis, V. K.; Hunt, S. T.;
́
Griffin, R. G.; Roman-Leshkov, Y. Nat. Commun. 2012, 3, 1109.
ChemSusChem 2012, 5, 939−944.
(22) Zheng, M.; Wang, A.; Ji, N.; Pang, J.; Wang, X.; Zhang, T.
ChemSusChem 2010, 3, 63−66.
(23) Zhao, G.; Zheng, M.; Wang, A.; Zhang, T. Chin. J. Catal. 2010,
31, 928−932.
(24) Zhao, G.; Zheng, M.; Zhang, J.; Wang, A.; Zhang, T. Ind. Eng.
Chem. Res. 2013, 52, 9566−9572.
(25) Tai, Z.; Zhang, J.; Wang, A.; Pang, J.; Zheng, M.; Zhang, T.
ChemSusChem 2013, 6, 652−658.
(26) Wang, X.; Meng, L.; Wu, F.; Jiang, Y.; Wang, L.; Mu, X. Green
Chem. 2012, 14, 758−765.
(27) Xiao, Z.; Jin, S.; Pang, M.; Liang, C. Green Chem. 2013, 15,
891−895.
(28) Deuss, P. J.; Barta, K.; de Vries, J. G. Catal. Sci. Technol. 2014, 4,
1174−1196.
(29) Zhao, H.; Holladay, J. E.; Brown, H.; Zhang, Z. C. Science 2007,
316, 1597−1600.
(30) Choudhary, V.; Pinar, A. B.; Lobo, R. F.; Vlachos, D. G.;
Sandler, S. I. ChemSusChem 2013, 6, 2369−2376.
(31) Choudhary, V.; Mushrif, S. H.; Ho, C.; Anderko, A.; Nikolakis,
V.; Marinkovic, N. S.; Frenkel, A. I.; Sandler, S. I.; Vlachos, D. G. J.
Am. Chem. Soc. 2013, 135, 3997−4006.
(32) Wang, Y.; Deng, W.; Wang, B.; Zhang, Q.; Wan, X.; Tang, Z.;
Wang, Y.; Zhu, C.; Cao, Z.; Wang, G.; Wan, H. Nat. Commun. 2013, 4
10.1038/ncomms3141.
(33) Geboers, J.; Van de Vyver, S.; Carpentier, K.; de Blochouse, K.;
Jacobs, P.; Sels, B. Chem. Commun. 2010, 46, 3577−3579.
(34) Zheng, X.; Lin, H.; Zheng, J.; Duan, X.; Yuan, Y. ACS Catal.
2013, 2738−2749.
(35) Thyssen, V. V.; Maia, T. A.; Assaf, E. M. Fuel 2013, 105, 358−
363.
(36) Zhi, G.; Guo, X.; Wang, Y.; Jin, G.; Guo, X. Catal. Commun.
2011, 16, 56−59.
(37) Kam, R.; Selomulya, C.; Amal, R.; Scott, J. J. Catal. 2010, 273,
73−81.
(38) Huang, Z.; Wang, X.; Wang, Z.; Zou, X.; Ding, W.; Lu, X. RSC
Adv. 2014, 4, 14829−14832.
(39) Pang, J.; Wang, A.; Zheng, M.; Zhang, Y.; Huang, Y.; Chen, X.;
Zhang, T. Green Chem. 2012, 14, 614−617.
(40) Liang, G.; He, L.; Arai, M.; Zhao, F. ChemSusChem 2014, 7,
1415−1421.
(41) Tajvidi, K.; Hausoul, P. J. C.; Palkovits, R. ChemSusChem 2014,
7, 1311−1317.
(42) Liang, G.; He, L.; Cheng, H.; Li, W.; Li, X.; Zhang, C.; Yu, Y.;
Zhao, F. J. Catal. 2014, 309, 468−476.
(43) Huang, Z.; Chen, J.; Jia, Y.; Liu, H.; Xia, C.; Liu, H. Appl. Catal.,
B 2014, 147, 377−386.
(44) Sun, J. Y.; Liu, H. C. Green Chem. 2011, 13, 135−142.
(45) Saxena, U.; Dwivedi, N.; Vidyarthi, S. R. Ind. Eng. Chem. Res.
2005, 44, 1466−1473.
(46) Clark, I. T. Ind. Eng. Chem. 1958, 50, 1125−1126.
(47) Chambon, F.; Rataboul, F.; Pinel, C.; Cabiac, A.; Guillon, E.;
Essayem, N. ChemSusChem 2013, 6, 500−507.
883
DOI: 10.1021/cs501372m
ACS Catal. 2015, 5, 874−883