Role of the Strong Lewis Base Sites on Glucose Hydrogenolysis

Article abstract of DOI:10.1002/cctc.201800427

This work reports the individual role of strong Lewis base sites on catalytic conversion of glucose hydrogenolysis to acetol/lactic acid, including glucose isomerisation to fructose and pyruvaldehyde rearrangement/hydrogenation to acetol/lactic acid. La

Full text of DOI:10.1002/cctc.201800427

Accepted Article
Title: Role of the strong Lewis base sites on glucose hydrogenolysis
Authors: Parviz Yazdani, Bo Wang, Feng Gao, Sibudjing Kawi, and
Armando Borgna
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201800427
Link to VoR: http://dx.doi.org/10.1002/cctc.201800427
A Journal of
www.chemcatchem.org

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
Role of the strong Lewis base sites on glucose hydrogenolysis
P. Yazdani,[a,b] B. Wang,[a] F. Gao,[a] S. Kawi,* [b] A. Borgna* [a,b]
Abstract: This work reports the individual role of strong Lewis base
sites on catalytic conversion of glucose hydrogenolysis to
acetol/lactic acid, including glucose isomerisation to fructose and
pyruvaldehyde rearrangement/hydrogenation to acetol/lactic acid.
La2O3, Nd2O3, Sm2O3, and Pr6O11 were selected as representative
oxides consisting of different base sites. The basicity of these
lanthanide oxides was characterised by CO2-TPD and DRIFT
spectroscopy. It was found that lanthanide cation-oxygen pairs,
Brønsted OH group coordinated with one Lanthanide cation, and
lanthanide cation near lanthanide cation-oxygen pairs are the
dominant base sites as compared to the other base sites on La2O3,
Nd2O3, Sm2O3 and Pr6O11. In addition, the relative concentration of
the base sites is different over the examined lanthanide oxides.
These unique properties of lanthanide oxides are used to
understand the individual role of base sites on the reactions
mentioned above. Then, the catalytic results were correlated with the
the H2 atmosphere has also been reported in several studies
over heterogeneous catalysts. In this context, supported metal
catalysts, promoted by basic oxide have been developed for
converting glucose to acetol and propylene glycol [2a]. Catalysts
[8]
such as Cu-La2O3/Al2O3,[3] Ni–SnOx/Al2O3,[7] Ni/ZnO-CNT and
Ru/Ac+ZnO [9] were developed for this process. In addition, base
oxides, such as grafted SnIV onto a mesoporous silica MCM-41,
ZrO2, and CuCTAB/MgO have been reported to catalyse
glucose conversion to lactic acid under inert atmosphere.[2c]
In summary, these studies have shown that the key role of the
base sites is to convert first glucose to fructose, and then break
the carbon bond between the third and fourth carbon at fructose
to produce C3 chemicals.[3, 10] However, the individual role of the
different base sites is still not well understood. Ohyama et al.
reported that glucose isomerisation over alkaline earth metal
titanates depends on the base strength and concentration.[6a]
base properties of these lanthanide oxides. Based on this correlation, They have shown that the base strength of various titanates
the base site requirements for each reaction were identified. In this
regard, Brønsted OH group coordinated with one Lanthanide cation
over Nd2O3 is found to be the suitable base site for glucose
isomerisation. On the other hand, lactic acid is produced as the main
product in glucose hydrogenolysis over lanthanide cation near
lanthanide-oxygen pairs on Pr6O11. Finally, lanthanide-oxygen pairs
over La2O3 are suitable base sites for retro-aldol condensation of
fructose to C3 chemicals.
promoted by alkali and alkali earth metals are different. Similarly,
Deng et al. reported that Ni-SnOx/Al2O3 catalyses glucose
conversion to acetol under H2 atmosphere.[7] Basicity of
SnOx/Al2O3, CeOx/Al2O3, ZnOx/Al2O3 and AlOx/Al2O3 sample was
compared using CO2-TPD. It has been found that the
SnOx/Al2O3 sample with the higher concentration of strong base
sites is more selective to acetol compared to other oxides
examined in that study. Recently, glucose hydrogenolysis to
acetol and propylene glycol over Cu-La2O3/Al2O3 was reported.[3]
Catalytic studies and CO2-TPD characterisations revealed that
moderate and strong base sites play a vital role in this reaction.
The role of the different active components on the product
distribution of glucose hydrogenolysis discussed in this study. In
summary, copper was an appropriate active metal component
for the hydrogenation of the fragments formed after retro-aldol
condensation reactions, while lanthanum oxide was providing
the base sites for the glucose isomerization and retro-aldol
condensation reactions. Similar to other reports, strong base
sites played a vital role in the conversion of glucose to propylene
glycol. Hence, it is essential to understand the nature and
function of the strong base sites in the glucose conversion to
fructose and C3 chemicals, including lactic acid and
acetol/propylene glycol.
Introduction
Glucose is an intermediate platform chemical which can be
converted into a wide range of building blocks including sugar
alcohols, glycols, acids, and furanic compounds to produce
biobased chemicals.[1] Particularly, base-catalysed aqueous
conversion of glucose to acetol/propylene glycol has received
considerable attention.[2] This conversion involves a set of
parallel and series reactions including glucose isomerisation and
pyruvaldehyde
hydrogenation/rearrangement.[3]
Several
heterogeneous and homogeneous base catalysts were
4]
developed for these reactions.[2c,
Moreover, heterogeneous
base catalysts are preferred over homogeneous ones, mainly
due to alkali separation, neutralisation and corrosion issues.[5]
Furthermore, higher yields of products have been reported over
heterogeneous base catalyst compared to the homogeneous
ones in some studies.[6] For instance, it has been reported that
fructose yield upon glucose isomerisation reaches up to 40%
To achieve the aim mentioned above, a series of catalyst
systems with distinct base properties is necessary. In this study,
lanthanide oxides used as the base promoters. It is worth noting
that lanthanum oxide is known to be a suitable base promoter
[11]
3]
for glucose conversion to fructose
and C3 chemicals.[2a,
Moreover, it has been reported that base properties of
lanthanide oxides are attributed to lanthanide contraction.[12]
Sato et al. have shown that the base strength of the lanthanide
oxides decreased by decreasing the radius of the lanthanide
using cation-exchanged zeolites, mixed oxides,[6c] and
[6b]
metallosilicates
as heterogeneous catalysts. Besides,
catalytic hydrogenolysis of glucose to acetol/propylene glycol in
[a]
[b]
P. Yazdani, Dr. B. Wang, Dr. F. Gao, Dr. A. Armando
Heterogeneous Catalysis Division Institute of Chemical and
Engineering Sciences, A*STAR 1st Pesek Rd, 627833, Singapore E-
mail: Armando_borgna@ices.a-star.edu.sg
P. Yazdani, Prof. K. Sibudjing Department of Chemical and
Biomolecular Engineering National University of Singapore,
Singapore 21 Lower Kent Ridge Rd, Singapore 119260
Supporting information
document
this
is given in a separated
foTr his aartrictleicle is protected by copyright. All rights reserved.

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
cation. Inspired by these studies mentioned above, four different
lanthanide oxides with distinct strong Lewis base sites are
thoroughly characterised and applied in the catalytic conversion
of glucose. Indeed, two different active catalyst components,
namely Cu/Al2O3 and base sites, are physically mixed. Using this
method, the only catalyst variable will be base oxides. Then, the
differences in the nature and strength of the base sites over
La2O3, Nd2O3, Sm2O3, Pr6O11 can be used as an appropriate tool
to understand the role of the Lewis base sites on the catalytic
conversion of glucose, gaining more insights into the correlation
TPD signals seems to be more complicated. It has been shown
that thermal stability of bidentate and monodentate carbonates
could vary from 100 °C to 600 °C.[14] All lanthanide oxides used
in this study, possess two peaks in the temperature range
between 100 and 400 °C. CO2-TPD profile of the 5%Cu/Al2O3
also possesses two peaks in this temperature range. Besides,
CO2 desorption from La2O3 and Nd2O3 contain the third peak in
the temperature range of 450-650 °C. The observed CO2-TPD
profiles are in agreement with the ones presented by Sato et
al.[12] They have shown that rare metal oxides presented
between catalyst properties and the strength of Lewis base sites. different CO2-TPD profiles by calcination at 800 °C. Then, the
Eventually, this approach can be used to guide future catalyst
developments for glucose conversion.
CO2-TPD profiles of the oxides are de-convoluted, and the
amount of the desorbed CO2 is reported in Table 1. As it can be
seen, strong base sites dominate on the surface of La2O3. In
contrast, moderate base sites dominate on Nd2O3, Sm2O3, and
Pr6O11 and weak base sites dominate on 5%Cu/Al2O3. Also, low
concentration of week base sites is observed over all lanthanide
oxides used in this study. The surface area of the oxides is also
presented in Table 1. It is worth noting that the BET surface area
of Sm2O3 is 13.9 m2/g, which is almost three folds of that of the
other lanthanide oxides.
Results and Discussion
XRD characterisation
Figure 1 shows the XRD patterns of the lanthanide oxides,
showing sharp peaks and thereby confirming high crystallinity of
the oxides. Pure phases of hexagonal La2O3 (PDF 01-076-7398),
hexagonal Nd2O3 (PDF 00-041-1089), cubic Sm2O3 (PDF 03-
065-3183), and cubic Pr6O11 (00-042-1121) were obtained by
calcination of the corresponding lanthanide nitrate precursors.
Figure 2. CO2-TPD profile of the lanthanide oxides and 5%Cu/Al2O3
Figure 1. XRD spectra of lanthanide oxides.
Base properties and surface area
Figure 2 shows the CO2-TPD profiles of the lanthanide oxides.
Three different desorption peaks can be identified in the CO2-
TPD profiles. Di Cosimo et al. assigned these peaks to CO2
adsorbed on base sites over the surface of the oxides.[13]
According to this study, these peaks can be assigned to
hydroxyl groups (100-200 °C), metal–oxygen pairs (300-500 °C),
and oxygen anions (>500 °C) on the surface of the oxides.
However, the assignment of the surface base pairs to the CO2-
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
of polydentate carbonate. Polydentate carbonates are the most
stable carbonates formed on the surface of the La2O3 as these
species are stable up to 500 °C. Besides, monodentate
bicarbonate species reveals the presence of Brønsted OH group
coordinated with a Lanthanide cation (Ln3+-OH-). CO2 desorbs
over this base site by heating up to 500 °C. Finally, carboxylates
result from back bonding between d or f orbitals of La3+ ions and
Table 1. The concentration of the base sites and surface area of
the lanthanide oxides and 5%Cu/Al2O3.
Concentration of base sites,
SBET
(µmol/m2)
Lanthanides
(m2/g)
Weak
0.6
Moderate
1.0
Strong
2.1
the π* orbitals of
a C-O bond of CO2. However, these
carboxylates are not stable at the temperature higher than
150 °C. Linking the results of CO2-TPD and CO2-DRIFT is not
straight forward. It has been shown that different types of
monodentate and bidentate carbonates with different thermal
stabilities ranging from 100 °C to 600 °C could exist.[14]
Considering the previous discussion, the first peak of the CO2-
TPD profile may correspond to hydroxyl groups, carboxylates
and bridge carbonates. Moreover, the other two peaks at higher
temperature can mainly result from CO2 desorption from
La2O3
Nd2O3
3.9
2.9
1.4
1.6
1.6
0.5
6.8
6.2
4.0
0.1
1.2
Sm2O3
-
-
-
13.9
3.1
Pr6O11
5%Cu/Al2O3
90.5
bidentate
carbonates,
monodentate
bicarbonates
and
polydentate carbonates.
Formation of carbonate species on lanthanide oxides upon
CO2 adsorption
Upon exposure of the lanthanide oxides to CO2 at 30 °C, several
peaks appeared in the 2000-1000 cm-1 region of the FTIR
spectra. Then, the physically absorbed CO2 is desorbed in a flow
of He at room temperature, and the resulting spectra are
presented in Figures 3-5. Clearly, there are several carbonate
species formed on the surface of the lanthanide oxides and
5%Cu/Al2O3. The structures of the possible carbonate species
formed on the surface of the lanthanide oxides are presented in
Figure S1. The assignment to the FTIR signals was well
established by Sebastián et al.,[13] using previously reported
wavenumbers and width band splitting of the adsorbed
carbonate species on the base oxides, along with their thermal
stabilities to assign the FTIR peaks. The same approach was
[15]
used by Köck et al.
to identify the carbonate species formed
on the surface of Yttria-Stabilized ZrO2 upon CO2 adsorption.
Figure 3 shows the formation of different carbonate species on
the surface of the La2O3 and their thermal stability up to 500 °C.
The spectrum of the carbonates formed on the surface of the
La2O3 was analysed based on the methodology described above,
using the FTIR spectra presented in Figure 3 and previous
reports. As it can be seen in Figures 3-5, five distinct carbonates
can be identified on the surface of lanthanide oxides and
5%Cu/Al2O3. The presence of these carbonate species on the
surface of the lanthanide oxides and 5%Cu/Al2O3 at two different
temperatures is summarised in Table S1. The formation of the
Bridged and bidentate carbonates demonstrate the presence of
the Lewis acid-base pairs (Ln3+-O2-) on the surface of lanthanide
oxides.[13] Bidentate carbonate species persist up to 300 °C on
the IR spectra of La2O3, while these carbonate species are
removed under flowing He at 500 °C (Figure 3). Bridge
carbonates are easily removed by outgassing the sample up to
150 °C. On the other hand, CO2 adsorb on the Lewis acid sites
close to Lewis acid-base pairs (Ln3+ nearby Ln3+-O2-) in the form
Figure 3. Infrared spectra collected on La2O3 after desorption of formed
carbonates under flowing He upon heating to 500 °C. All peaks are assigned
to the corresponding carbonates.
It is clear that the relative concentrations of specific carbonates
on the surface of lanthanide oxides and 5%Cu/Al2O3 are different
(Figures
4 and 5). As seen in Figure 5, monodentate
bicarbonates dominate on the surface of Nd2O3, while the
concentration of polydentate carbonates is lower as compared to
the one on the other oxides. In addition, the relevant intensity of
bidentate carbonate is lower over Pr6O11 and Nd2O3 as
compared to the other carbonates in these oxides. On the other
hand, CO2 adsorbed on the surface of Pr6O11, Sm2O3, and
5%Cu/Al2O3 mostly in the form of polydentate carbonates
(Figure 4). In next section, the strength of the carbonates over
lanthanide oxides and 5%Cu/Al2O3 upon heating to 300 °C will
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
be investigated as the focus of this study is to understand the
individual effect of the strong Lewis base sites on glucose
hydrogenolysis.
bicarbonate, bidentate carbonate and polydentate carbonate,
which are shown in Table S1 as well. The integrated absorbance
profiles of the carbonates on the surface of Lanthanide oxides
and 5%Cu/Al2O3 after desorption at 300 °C are depicted in
Figure 8 to get a clear comparison of the base properties.
Polydentate carbonate is the only carbonate present on the
surface of Pr6O11. Furthermore, mainly polydentate carbonate is
observed on the surface of Sm2O3 and 5%Cu/Al2O3. In addition,
low concentrations of bidentate carbonates and monodentate
bicarbonates on the surface of the Sm2O3 also exist. Nd2O3
contains all three types of carbonates, with the dominance of
monodentate bicarbonate. For La2O3, the concentration of
bidentate carbonate is the highest among the four lanthanide
oxides investigated in this study. In summary, the difference in
the distribution of the strong base sites over lanthanide oxides in
this study is sufficiently significant to understand the role of the
different strong Lewis base sites in glucose hydrogenolysis.
More specifically, the individual function of Ln3+-O2-, Ln3+-OH-,
and Ln3+-(Ln3+-O2-) pairs over La2O3, Nd2O3, Sm2O3 and Pr6O11
will be correlated with the catalytic results of glucose
hydrogenolysis.
Figure 4. Infrared spectra collected on 5%Cu/Al2O3, Sm2O3 and La2O3 after
CO2 adsorption followed by purging under flowing He at 30 °C.
As shown by CO2-TPD and CO2-FTIR studies, the surface of the
lanthanide oxides contains weak and strong base sites.
Moreover, the type and amount of the strong base sites over the
series of lanthanide oxides are different. In following sections,
the role of strong Lewis base sites on the catalytic performance
is studied. Indeed, it should be noted that cooperative role of
weak base sites cannot be ruled out since they are present over
all lanthanide oxides.
Figure 5. Infrared spectra collected on Nd2O3 and Pr6O11 after CO2 adsorption
followed by purging under flowing He at 30 °C.
Strength of the carbonate species formed on the surface of
lanthanide oxides
Glucose hydrogenolysis to propylene glycol was reported to be
7]
more selective over moderate and strong base sites.[3,
In
addition, the yield of glucose isomerisation to fructose was also
shown to be higher over strong base sites.[6a] Hence, the
strength of the carbonates formed on the lanthanide oxides was
analysed by desorbing CO2 in flowing He at 300 °C to determine
the strength of the base sites on the surface of lanthanide oxides.
The IR spectra of the carbonates after desorption at 300 °C are
depicted in Figures 6 and 7. Similarly to the IR spectra at 30 °C,
the distribution of the carbonates over lanthanide oxides and
5%Cu/Al2O3 is different. However, only three carbonate species
are present on the surface of the oxides, including monodentate
Figure 6. Infrared spectra collected on Sm2O3 and La2O3 after desorption
under flowing He at 300 °C.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
base catalyst, following the procedure reported by Román et
al.[16] While glucose-D2 is converted to fructose-D1 when
glucose isomerisation is catalysed over Lewis acid sites,
glucose-D2 is converted to unlabelled fructose when the
reaction proceeds over base sites. In this case, unlabelled
glucose is produced by reverse isomerisation of the unlabelled
fructose. The 1H NMR spectra of glucose-D2 reacted over La2O3
is presented in Figure S2. 1H NMR spectra of glucose-D2 and
unlabelled glucose are also shown in this Figure. The 1H NMR
spectra of glucose-D2 reacted over La2O3 shows the presence of
regular glucose in the reaction solution confirmed by the
appearance of the peaks at δ=3.1. Hence, it is confirmed that
glucose isomerisation over lanthanide oxides is
a base
catalysed reaction. Moreover, some experiments were designed
to check whether the isomerisation reaction is heterogeneously
catalysed by lanthanide oxides. In these experiments, the
lanthanide oxides were separated from the liquid solution after
the catalytic study. Then, the obtained liquid solution is applied
in a subsequent isomerisation reaction at 100 °C and 5 bar of N2
for 30 min. No glucose was observed under this condition,
showing that the reaction is heterogeneously catalysed by the
lanthanide oxides.
Figure 7. Infrared spectra collected on 5%Cu/Al2O3, Nd2O3 and Pr6O11 after
desorption under flowing He at 300 °C.
As shown in Figure 9, 10% of glucose is converted to fructose
over lanthanide oxides at 100 °C after 30 min catalytic reaction
under 5 bar of N2 except for Pr6O11. Over Pr6O11, the rate of
glucose conversion is higher, reaching 10% only after 10 min
catalytic reaction at the same conditions. Glucose isomerisation
to fructose is more selective over Nd2O3 than over other
lanthanide oxides. Selectivity to fructose reaches 66.5% over
Nd2O3, while a selectivity of only 55.4%, 41.5%, and 43.4 is
achieved over La2O3, Sm2O3, and Pr6O11, respectively. With all
lanthanide oxides, near 7% mannitol and 4% lactic acid are
produced as side products along with fructose. It is worth noting
that total detected carbon species analysed by HPLC decreases
over
the
lanthanides
in
the
following
order,
Figure 8. Integrated absorbance signals of the carbonates on the surface of
the Lanthanide oxides and 5%Cu/Al2O3 (after desorption at 300 °C). A1:
Polydentate carbonates, A2: Bidentate carbonates, A3: Monodentate
bicarbonates, and A: A1 +A2 +A3
Nd2O3>La2O3>Sm2O3=Pr6O11. The carbon loss upon glucose
conversion in aqueous solution at high temperature is usually
ascribed to the formation of Humins and condensation
products,[3, 18] The first step in the formation of these undesired
products was reported to be dehydration reactions of glucose
and fructose along with hydrated HMF. Hence, it can be
concluded that the existence of Lewis acid sites near base pairs
promote the rate of the condensation reaction, resulting in a loss
of carbon. Considering the CO2-TPD profile of the lanthanide
oxides, the presence of the strong base sites is important to
prevent loss of carbon in glucose isomerisation. Interestingly,
Nd2O3 that contains less Ln3+-(Ln3+-O2-) sites but more Ln3+-OH-
base sites, yields more fructose compared to the other
lanthanides. So, it can be concluded that glucose isomerisation
to fructose is more selective over Brønsted OH group
coordinated with one Lanthanide cation (Ln3+-OH-) since these
sites are the dominating ones over Nd2O3. The correlation
between different strong Lewis base sites and glucose
Catalytic studies
The catalytic conversion of glucose isomerisation to fructose,
pyruvaldehyde hydrogenation/rearrangement to acetol/lactic
acid, and glucose hydrogenolysis to acetol/lactic acid were
studied, and results are summarised in Figures 9-12.
Glucose isomerisation to fructose has been reported to follow
two different mechanisms known as base catalysed, and Lewis
[12, 17]
acid catalysed isomerisation.[16] Sato et al.
analysed the
acid-base properties of lanthanide oxides using CO2-TPD and
NH3-TPD. Based on their analysis, these oxides only possess
base sites. To further assure that the glucose isomerisation over
these oxides follows proton transfer mechanism catalysed by
base sites, an NMR study has been carried out using La2O3 as a
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
isomerisation rate could be useful for developing new catalysts
with proper base sites to catalyse this reaction efficiently.
different lanthanide oxides. It is worth noting that detected
carbon species using La2O3 is lowest compared to other
lanthanide oxides either physically mixed with 5%Cu/Al2O3 or
not (Figures 10 and 11). As mentioned before, the carbon loss is
due to the Humins formation. Hence, this suggests that the
Humins yield is the highest when La2O3 is used as a base
promoter in pyruvaldehyde hydrogenation /rearrangement.
Figure 9. Glucose isomerisation to Fructose over lanthanide oxides. Reaction
condition: 5wt% glucose, 50 ml water, 1 gr catalyst, 100 0C, 5 bar of N2,
reaction time: 30 min, except for Pr6O11 (10 min)
In addition, pyruvaldehyde hydrogenation/rearrangement over
the lanthanide oxides physically mixed with Cu/Al2O3 at 200 °C
and 34 bar of H2 was conducted, and the results are presented
in Figure 10. Pyruvaldehyde is the intermediate of glucose
hydrogenolysis and can be converted to acetol over metallic
copper, and to lactic acid and Humins over acid-base sites
under the current reaction conditions.[3, 19] As it can be seen in
Figure 10, lactic acid is main product for all four base promoters
which are physically mixed with 5%Cu/Al2O3. La2O3 physically
mixed with Cu/Al2O3 produces the lowest yield of lactic acid at
40.4% compared to the other lanthanide oxides: 51.8% over
Nd2O3, and 57.0% for both Sm2O3 and Pr6O11. However, acetol
yield remains the same, 18.0%, for all catalysts in this reaction.
The yield of acetol increases to 27.0% using 5%Cu/Al2O3 without
base promoters. The reason is that acetol is the hydrogenation
product of pyruvaldehyde while lactic acid is the result of
pyruvaldehyde rearrangement. Interestingly, the lactic acid yield
is high as well over 5%Cu/Al2O3, 52.1%. This yield is even
higher than the yield over La2O3+5%Cu/Al2O3. This reaction
result raises the question of the role of lanthanide oxides in the
absence of 5%Cu/Al2O3. To answer this question,
pyruvaldehyde rearrangement /hydrogenation was performed at
the same condition over lanthanide oxides, and the results are
presented in Figure 11. Lactic acid yield is 58.1% over La2O3,
66.4% over Nd2O3 and 69.6% using Sm2O3 and Pr6O11.
Moreover, 13.0% of acetol is also produced even in the absence
of the 5%Cu/Al2O3. The same trend for the lactic acid yield is
Figure 10. Hydrogenation/re-arrangement of pyruvaldehyde over lanthanide
oxides physically mixed with 5%Cu/Al2O3, Reaction condition: 0.3wt%
pyruvaldehyde solution, 100ml water, 0.15 gr catalyst
(5%Cu/Al2O3+Lanthanide oxide), 200 °C, 34 bar of H2 gas, 10 min.
Figure 11. Hydrogenation/re-arrangement of pyruvaldehyde over lanthanide
oxides, Reaction condition: 0.3wt% pyruvaldehyde solution, 100ml water, 0.15
gr catalyst, 200 °C, 34 bar of H2 gas, 10 min.
observed
over
lanthanide
oxides
and
lanthanide
oxides+5%Cu/Al2O3. So, it can be concluded that the conversion
of pyruvaldehyde can be catalysed by lanthanide oxides and
5%Cu/Al2O3. In addition, the detected carbon species vary using
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
The function of Lewis acid-base pairs in condensation and
hydrolysis has been well explained in the literature.[20] It has
been shown that Lewis acid-base pairs catalyse nucleophilic
attack of water on the substrate. This reaction is catalysed by
the formation of metal cation-substrate bonds. Then, the newly
formed bond between the metal cation and the substrate
positions the substrate or polarises its electron distribution to
catalyse the reaction. This mechanism is accepted by several
authors for the conversion of pyruvaldehyde to lactic acid and
polyols to glycols over homogeneous or heterogeneous base
catalysts.[19-20, 21] Similarly, the same approach can be used to
understand the reaction mechanism of pyruvaldehyde
conversion over base oxides in this study. Based on the
concentration of lanthanide cations adjacent to lanthanide-
oxygen pairs. Recently, a similar correlation between catalytic
properties and pyruvaldehyde rearrangement to lactic acid was
proposed by Albuquerque et al.[19] These authors have shown
that pyruvaldehyde conversion to lactic acid proceeds through
activation of pyruvaldehyde molecule over coordinatively
unsaturated Zr4+ cations. Then water is dissociated over
Zr4+−O2− pairs, generating terminal OH groups to facilitate the
rearrangement reaction. Moreover, they showed that a minimum
strength of Lewis acidity is required to activate the
pyruvaldehyde molecule over acid sites. Therefore, the reaction
path I may not occur in the conversion of the pyruvaldehyde
over lanthanide oxides as these oxides do not possess Lewis
acidity.[12, 17] Thus another explanation for the different yields of
lactic acid over lanthanide oxides is required. It is shown that
La2O3 contains the higher concentration of lanthanide-oxygen
pairs. As discussed above, these base sites participate in both
lactic acid production and Humins formation. So, the reason for
different lactic acid yields over lanthanide oxides might be
mostly related to the formation of Humins over Ln3+-O2- pairs. In
summary, pyruvaldehyde is converted to lactic acid and acetol
over lanthanide oxides under the reaction conditions. However,
the presence of strong base sites known as Ln3+-O2- pairs over
La2O3 is decreasing the detected carbon species by 10-15 % via
Humins formation.
proposed reaction mechanisms for pyruvaldehyde conversion to
[21b, 22]
lactic acid
and the catalytic studies in this study, three
possible pathways for conversion of the pyruvaldehyde are
depicted in Scheme 1. Reaction path I proceed by adsorption of
pyruvaldehyde on lanthanide cations adjacent to lanthanide-
oxygen pairs through both carbonyl groups. Then a water
molecule is dissociated over lanthanide-oxygen pairs, and a
hydroxyl group is transferred to the activated pyruvaldehyde
molecule through a nucleophilic attack followed by a hydride
shift to produce lactic acid. On the other hand, the reaction path
II and III do not involve the activation of pyruvaldehyde over
lanthanide cations. Therefore, one may associate the higher
yield of lactic acid over Pr6O11 and Sm2O3 with the high
Scheme 1. Possible reaction pathways of pyruvaldehyde rearrangement to lactic acid and Humins over base sites
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
Finally, the role of the strong Lewis base sites on glucose
hydrogenolysis was investigated. Glucose hydrogenolysis
consists of cascade reactions which are catalysed over base
sites except for the hydrogenation of the glucose and the formed
fragments. These hydrogenation reactions are catalysed over
hydrogenolysis over both La2O3 and Nd2O3. Moreover, similar
results over these two oxides during glucose hydrogenolysis
shows that lanthanide-oxygen pairs over La2O3 and Brønsted
OH group coordinated with one Lanthanide cation over Nd2O3
are suitable base sites for this reaction. Interestingly, both
oxides have a third peak in their CO2-TPD profiles, indicating the
vital role of the strong base sites in this conversion in line with
previous reports.[3, 7]
[3]
5%Cu/Al2O3. Glucose hydrogenolysis over lanthanide oxides
physically mixed with Cu/Al2O3 was carried out at the reaction
conditions reported in Figure 12. Sorbitol and HMF are produced
with yields of 6.6% and 7.5% using 5%Cu/Al2O3. The yields of
these products are less than 1.5% when lanthanide oxides are
used as base promoters. It was reported in the literature that
aluminium is not an active element in glucose hydrogenolysis.[2a]
However, acetol and lactic acid are produced with a total yield of
14.4% using 5%Cu/Al2O3. This can be explained by the base
properties of 5%Cu/Al2O3 confirmed by CO2-DRIFT and TPD
measurements in this work. Indeed, Al3+-(Al3+-O2-) pairs and Al3+
-
O2- pairs are strong base sites. The lower yield of C3 chemicals
over 5%Cu/Al2O3 can be related to the lower concentration of
base sites over this catalyst; confirmed by CO2-TPD, or the
presence of acid sites. These two parameters cause the
increase in the sorbitol and HMF yield; respectively, decreasing
the yield of C3 chemicals.[3] As it can be seen in Figure 12, acetol
is the main product over La2O3, Nd2O3 and Sm2O3, while lactic
acid is the dominant product over Pr6O11. The yield of acetol
decreased from 40.1% to 36.5%, 29.5%, and 21.7% when La2O3,
Nd2O3, Sm2O3, and Pr6O11 are used as base promoters,
respectively. On the contrary, the yield of lactic acid increased
from 18.8% to 27.6% over these oxides. It was mentioned by
several authors that addition of base promotors to supported
metals cause an increase in the yield of lactic acid during
glucose hydrogenolysis.[2a, 3, 7, 23] Herein, we have shown that the
type of strong Lewis base sites is changing the yield of lactic
acid. The higher yield of lactic acid over Sm2O3 and Pr6O11 can
be explained considering the base properties of these oxides
discussed in the previous section. However, the yield of lactic
acid over these two oxides is not the same. This difference can
be explained by the rate of glucose isomerisation reaction over
these two oxides. It is worth noting that reaction rate of glucose
isomerisation over Pr6O11 is higher than other oxides. So,
pyruvaldehyde is produced in the reaction medium of glucose
hydrogenolysis at lower temperatures during the reactor heat-up
over this oxide. Then, lactic acid is produced in higher yield as
its formation is favoured at lower temperatures compared to
acetol. It is important to mention that Nd2O3 is more selective to
fructose during glucose isomerisation as compared to La2O3,
and Humins formation was higher over La2O3 Thus, one may
expect to produce more acetol upon glucose hydrogenolysis
over this oxide. However, the yield of acetol and lactic acid are
the same over these two oxides during glucose hydrogenolysis
as it can be seen in Figure 12. We can conclude that La2O3 is
more selective in the C3-C4 bond cleavage of fructose compared
to Nd2O3. It means that Ln3+-O2- base pairs are suitable base
sites for C3-C4 bond cleavage of fructose. This could explain why
similar catalytic results are obtained during glucose
Figure 12. Glucose hydrogenolysis over lanthanide oxides and 5%Cu/Al2O3,
Reaction condition: 0.3wt% glucose, 100 ml water, 0.15 gr catalyst
(5%Cu/Al2O3+Lanthanide oxide), 200 °C, 34 bar of H2 gas, 10 min, Others:
Fructose, Sorbitol, Arabitol, Erythritol, Pyruvaldehyde, Glycerol, Ethylene
glycol, 1,2-BDO, HMF.
Conclusions
In this study, the individual effect of strong base sites on glucose
isomerisation to fructose, pyruvaldehyde rearrangement
/hydrogenation, and glucose hydrogenolysis to acetol/lactic acid
was investigated. Four different lanthanide oxides with distinct
strong base sites were characterised and applied in the
mentioned reactions. The results show that different types of
base sites catalyse different reactions. More specifically, glucose
isomerisation to fructose is more selective over Brønsted OH
group coordinated with one Lanthanide cation (Ln3+-OH-). The
presence of lanthanide cations near lanthanide-oxygen pairs
catalyses glucose isomerisation along with condensation
reactions, resulting in Humins formation. Moreover, lanthanide-
oxygen pairs catalyse the unwanted side reactions, lowering
down the detected carbon species in pyruvaldehyde
hydrogenation/rearrangement. On the other hand, lanthanide
cation adjacent to lanthanide-oxygen base pairs are more
selective
towards
lactic
acid
during
pyruvaldehyde
hydrogenation/rearrangement. Finally, acetol is the primary
product of glucose hydrogenolysis when lanthanide-oxygen
pairs and Brønsted OH group coordinated with one Lanthanide
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
Catalyst characterisation
cation are used as base promoters, while mainly lactic acid is
produced over lanthanide cations near lanthanide-oxygen pairs
as the base promoter.
X-ray diffraction patterns (XRD) of the Lanthanide oxides were recorded
on a Bruker AXS D8 diffractometer using CuKα radiation employing tube
voltage and current of 40 kV and 40 mA, respectively. The Brunauer
Emmett Teller (BET)-surface area of the lanthanide oxides was analysed
using a Micromeritics ASAP 2420 system. The samples were pre-treated
at 200 °C for 2 hours under vacuum and then subjected to N2 physical
adsorption at liquid N2 temperature. CO2 temperature programmed
desorption (CO2-TPD) was used to determine the basicity of the lanthanide
oxides. Before analysis, the lanthanide oxides were pretreated in 30
ml/min flow of Ar at 500 °C for 1 hour with a heating rate of 10 °C/min.
Then the oxides were cooled down to 30 °C and saturated with CO2
using 30 ml/min flow of pure CO2 for 30 min. Subsequently, the oxides
were treated in a flowing N2 for 30 min at room temperature to remove
the physically adsorbed CO2. Finally, CO2-TPD was carried out by heating
up the TPD chamber to 800 °C in 30 ml/min flow of He with a heating rate of
10 °C/min. The desorbed CO2 was analysed using a TCD, and the TPD
profiles were recorded. For all base oxides, a blank CO2-TPD experiment
was carried out without adsorbing CO2. Then the resulting spectra were
subtracted from the CO2-TPD profiles of the base oxides. The CO2-TPD
profiles were deconvoluted, and the area under the peaks was calculated.
CO2-FTIR was carried out to evaluate the base sites and the base
strength of the lanthanide oxides. The spectroscopic analysis was carried
Experimental Section
Catalyst preparation
Nitrate precursors of Lanthanum, Praseodymium, Neodymium, and
Samarium were purchased from Sigma-Aldrich. The lanthanide oxides
were prepared by decomposing the nitrate precursors in the air at 800 °C
with a ramping rate of 2 °C/min for 2 hours. 5% Cu/Al2O3 was prepared
by wet impregnation. The appropriate amount of copper nitrate precursor
and gamma-alumina were added into a beaker containing 50 ml of water
and stirred overnight at room temperature. The resulting powder was
dried at 100 °C in a vacuum oven overnight. Then the dried catalyst was
calcined at 400 °C for 5 hours using a heating rate of 1°C /min under air
flow. The obtained catalyst was reduced in flowing H2 at 400 °C for 2
hours using heating rate of 2 °C/min. Then, the reduced catalyst was
passivated under
a flow of 1% O2 in N2 for 30 min at ambient
out on
a
Frontier MIR (PerkinElmer) equipped with
a
temperature before exposing to air. The prepared catalysts were stored
in a glove box with N2 atmosphere prior to catalytic studies.
mercury−cadmium−telluride (MCT) detector, following the procedure
described elsewhere.[24] In a typical measurement, the sample powder
was first placed in a high-pressure cell fitted with ZnSe windows. Before
analysis, all samples were pre-treated at 500 °C in flowing Ar (30 ml/min)
during 1 hour. Then the samples were cooled down to 30 °C. The spectra
of the samples after pre-treatment were taken and used as background.
After that, the pre-treated oxides were saturated in 30 ml/min flow of pure
CO2 for 30 min. Subsequently, physically adsorbed CO2 was removed
from the samples using a 10 ml/min flow of He for 30 min. The samples
were then heated up to different temperatures between 100 °C to 300 °C
in flowing He at 10 ml/min, and the spectra were recorded after cooling
down to room temperature with a resolution of 4 cm-1, recording 64 scans.
The IR bands are integrated after deconvolution of the IR signals using
Gaussian functions. NMR spectra were taken using a Bruker 400 MHz
NMR spectrometer. Deuterated glucose at the C2 position was used in
the isomerisation reaction to fructose. Then, the reaction solution was
separated from catalyst using a centrifuge. The obtained solution was
dried in a vacuum oven at 40 °C. The remaining powder was dissolved in
1 ml of D2O. Then, 1H NMR spectra of the glucose-D2 before and after
catalytic reactions were analysed to understand the mechanism of
glucose isomerisation to fructose, based on the procedure reported
elsewhere.[6a]
Catalytic activity
The hydrogenolysis of glucose and pyruvaldehyde hydrogenation/
rearrangement were carried out at 200 °C and 34 bar of H2 in a 300 ml
Teflon-lined Parr-reactor. 100 ml of 0.3wt% feed solution was used in
these experiments. 0.15 g of a physical mixture of 5%Cu/Al2O3 and
lanthanide oxide was used as the catalyst in these studies. The amount
of 5%Cu/Al2O3 was kept at 0.1 g for all reactions. Catalytic isomerisation
of 10wt% glucose to fructose was performed at 100 °C and 5 bar of N2.
50 ml of 10wt% glucose solution was used during isomerisation reaction,
using 1g of lanthanide oxides as a catalyst. First, high pressure H2/N2
was introduced in the reactor, and then the temperature was increased
up to the reaction temperature. The heating time from room temperature
to 200 °C and 100 °C were 35 min and 25 min, respectively. Time zero
was taken at this point. The resulting solution was collected after cooling
down the reactor to room temperature and analysed with HPLC. The
equations used to calculate conversion, selectivity and detected carbon
species are presented below. HPLC was performed using a 0.5 mM
aqueous solution of sulfuric acid as a mobile phase, at a flow rate of 0.5
ml/min. Aminex HPX-87H column was used as separation column
operating at 60 °C.
Acknowledgements
Converted carbon based mole of feed
Conversion (%)=
Selectivity (%)=
*100
*100
Carbon based mole of feed
This work was supported by Agency for Science, Technology
and Research (A*STAR) of Singapore, the National University of
Singapore, and the NEA (ETRP 1501 103). P. Yazdani is
grateful to A*STAR for the Singapore International Graduate
Award (SINGA) scholarship.
Carbon based mole of product
Converted carbon based mole of feed
n
Detected carbon species(%)= ꢀ Selectivity (%)
Keywords: Glucose isomerisation
•
Hydrogenolysis
•
k=0
Lanthanide oxides • Base properties • Propylene glycol
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
[1]
[2]
a) T. Werpy, G. Petersen, A. Aden, J. Bozell, J. Holladay,
J. White, A. Manheim, D. Eliot, L. Lasure, S. Jones,
Department of Energy Washington DC, 2004; b) E.
Soghrati, C. Choong, C. K. Poh, S. Kawi, A. Borgna,
ChemCatChem 2017, 9, 1402-1408; c) S. Van de Vyver, J.
Geboers, P. A. Jacobs, B. F. Sels, ChemCatChem 2011, 3,
82-94.
a) M. Zheng, J. Pang, R. Sun, A. Wang, T. Zhang, ACS
Catalysis 2017, 7, 1939-1954; b) H. Li, S. Yang, S.
Saravanamurugan, A. Riisager, ACS Catalysis 2017, 7,
3010-3029; c) W. Deng, Q. Zhang, Y. Wang, Journal of
Energy Chemistry 2015, 24, 595-607.
[12]
[13]
[14]
[15]
S. Sato, R. Takahashi, M. Kobune, H. Gotoh, Applied
Catalysis A: General 2009, 356, 57-63.
S. E. Collins, M. A. Baltanás, A. L. Bonivardi, The Journal
of Physical Chemistry B 2006, 110, 5498-5507.
Y. Yanagisawa, K. Takaoka, S. Yamabe, T. Ito, The
Journal of Physical Chemistry 1995, 99, 3704-3710.
E.-M. Köck, M. Kogler, T. Bielz, B. Klötzer, S. Penner,
The Journal of Physical Chemistry C 2013, 117, 17666-
17673.
Y. Román‐Leshkov, M. Moliner, J. A. Labinger, M. E.
Davis, Angewandte Chemie International Edition 2010, 49,
8954-8957.
S. Sato, R. Takahashi, T. Sodesawa, A. Igarashi, H.
Inoue, Applied Catalysis A: General 2007, 328, 109-116.
H. Rasmussen, H. R. Sørensen, A. S. Meyer,
Carbohydrate research 2014, 385, 45-57.
E. M. Albuquerque, L. E. Borges, M. A. Fraga, C. Sievers,
ChemCatChem 2017, 9, 2675-2683.
a) M. Bicker, S. Endres, L. Ott, H. Vogel, Journal of
Molecular Catalysis A: Chemical 2005, 239, 151-157; b) J.
E. Coleman, Current opinion in chemical biology 1998, 2,
222-234.
a) X. Jin, J. Shen, W. Yan, M. Zhao, P. S. Thapa, B.
Subramaniam, R. V. Chaudhari, ACS Catalysis 2015, 5,
6545-6558; b) L. Li, F. Shen, R. L. Smith, X. Qi, Green
Chemistry 2017, 19, 76-81; c) L. Yang, J. Su, S. Carl, J. G.
Lynam, X. Yang, H. Lin, Applied Catalysis B:
Environmental 2015, 162, 149-157.
[16]
[3]
[4]
P. Yazdani, B. Wang, Y. Du, A. Borgna, Catalysis
Science & Technology 2017, 7, 4680-4690.
M. Schlaf, Z. Zhang, Reaction pathways and mechanisms
in thermocatalytic biomass conversion I: Cellulose
[17]
[18]
[19]
[20]
structure,
heterogeneous catalysts, Springer 2015.
M. Xian, Sustainable Production of Bulk Chemicals:
Integration of Bio ,Chemo Resources and Processes,
depolymerization
and
conversion
by
[5]
[6]
‐
‐
Springer Dordrecht, Netherlands, 2015.
a) J. Ohyama, Y. Zhang, J. Ito, A. Satsuma,
ChemCatChem 2017, 9, 2864-2868; b) S. Lima, A. S. Dias,
Z. Lin, P. Brandão, P. Ferreira, M. Pillinger, J. Rocha, V.
Calvino-Casilda, A. A. Valente, Applied Catalysis A:
General 2008, 339, 21-27; c) C. Moreau, R. Durand, A.
Roux, D. Tichit, Applied Catalysis A: General 2000, 193,
257-264.
[21]
[7]
[8]
T. Deng, H. Liu, Journal of Molecular Catalysis A:
Chemical 2014, 388, 66-73.
C. van der Wijst, X. Duan, I. Skeie Liland, J. C. Walmsley,
J. Zhu, A. Wang, T. Zhang, D. Chen, ChemCatChem 2015,
7, 2991-2999.
Y. Hirano, K. Sagata, Y. Kita, Applied Catalysis A:
General 2015, 502, 1-7.
E. Girard, D. Delcroix, A. Cabiac, Catalysis Science &
Technology 2016, 6, 5534-5542.
[22]
[23]
P. i. Mäki-Arvela, I. L. Simakova, T. Salmi, D. Y. Murzin,
Chemical reviews 2013, 114, 1909-1971.
a) R. Sun, M. Zheng, J. Pang, X. Liu, J. Wang, X. Pan, A.
Wang, X. Wang, T. Zhang, ACS Catalysis 2015, 6, 191-
201; b) R. Figueiredo, H. Andrade, J. Fierro, Brazilian
Journal of Chemical Engineering 1998, 15.
[9]
[24]
C. Choong, Z. Zhong, L. Huang, A. Borgna, L. Hong, L.
Chen, J. Lin, ACS Catalysis 2014, 4, 2359-2363.
[10]
[11]
N. Essayem, R. L. De Souza, F. Rataboul, C. Feche, D.
Cardoso, D. Fabiano, in U.S. Patent Application, Vol.
14/241,213, 2015.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800427
ChemCatChem
FULL PAPER
Yazdani, B. Wang,
F Gao, S. Kawi,* A.
Borgna*
Strong Lewis base pairs
play unique roles in
glucose hydrogenolysis
to
glycol. Strong base sites
over La and Nd
acetol/propylene
Page No. – Page
No.
2O3
2O3
are identified as the
proper base sites for
this conversion.
Individual role of
strong Lewis base
sites on glucose
hydrogenolysis to
Acetol/Propylene
glycol
This article is protected by copyright. All rights reserved.