Cu/C-catalyzed Hydrogenolysis of Sorbitol to Glycols–On the Influence of Particle Size and Base

Article abstract of DOI:10.1002/cctc.201900299

Herein, a series of Cu catalysts supported on activated carbon with increasing particle size was tested for the conversion of sorbitol to glycols including 1,2-propanediol and ethylene glycol in the presence of Ca(OH)₂. Activity measurements reveal a distinct correlation between TOF and particle size, which increases until 14 nm (N₂O chemisorption based) after which a plateau is reached. However, variation of the particle size has no apparent effect on the selectivity for C₃/C₃ and C₂/C₄ cleavage. In contrast, variation of the Ca(OH)₂ equivalence has a profound impact on selectivity. The ratio of C₃/(C₂+C₄) products reaches an optimum at 0.45 equiv. of Ca(OH)₂. Notably, the time course of the reaction demonstrates that the by-products lactic acid and glycerol are in-situ converted to the desired product 1,2-propanediol, contributing to the unprecedented combined glycol selectivity of 84.5 % at 513 K, 5 MPa H₂ and 0.3 equiv. of Ca(OH)₂. Furthermore, recycling tests reveal a rapid deactivation which is attributed to aggregation of Cu observed in TEM.

Full text of DOI:10.1002/cctc.201900299

Accepted Article
Title: Cu/C-catalyzed hydrogenolysis of sorbitol to glycols–on the
influence of particle size and base
Authors: Xinde Wang, Anna Katherina Beine, Peter Hausoul, and
Regina Palkovits
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.201900299
Link to VoR: http://dx.doi.org/10.1002/cctc.201900299
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201900299
FULL PAPER
Cu/C-catalyzed hydrogenolysis of sorbitol to glycols–on the
influence of particle size and base
Xinde Wang, Anna Katharina Beine, Peter J.C. Hausoul and Regina Palkovits*
Abstract: Herein, a series of Cu catalysts supported on activated
carbon with increasing particle size was tested for the conversion of
sorbitol to glycols including 1,2-propanediol and ethylene glycol in
the presence of Ca(OH)
correlation between TOF and particle size, which increases until 14
nm (N O chemisorption based) after which a plateau is reached.
However, variation of the particle size has no apparent effect on the
selectivity for C /C and C /C cleavage. In contrast, variation of the
Ca(OH) equivalent has a profound impact on selectivity. The ratio of
/(C +C ) products reaches an optimum at 0.45 eq. of Ca(OH)
2
. Activity measurements reveal a distinct
Scheme 1. Metal-catalyzed hydrogenolysis of sorbitol under basic conditions
to short chain polyols and acids.
2
3
3
2
4
2
decarbonylation and deoxygenation, leading to degradation and
formation of gaseous products.[10] Banu et al. investigated the
conversion of sorbitol over Ru-, Pt- and Ni-NaY catalysts in the
presence of base.[ Ni-NaY showed the best results with over
C
3
2
4
2
.
Notably, the time course of the reaction demonstrates that the by-
products lactic acid and glycerol are in-situ converted to the desired
product 1,2-propanediol, contributing to the unprecedented
11]
6
2 3
0% conversion and 70% selectivity to glycols. Ni/Al O was
combined glycol selectivity of 84.5% at 513 K, 5 MPa H
2
and 0.3 eq.
investigated by Ye et al..[ It was found that the addition of Ce
significantly improved the activity. However, also here the
formation of considerable amounts of gaseous products (e.g.
12]
of Ca(OH) . Furthermore, recycling tests reveal a rapid deactivation
2
which is attributed to the aggregation of Cu observed in TEM.
CH
4
and CO
2
)
is
a
disadvantage. Therefore, strong
hydrogenolysis of C-C and C-O bonds and poor hydrothermal
stability make Ni an unsuitable alternative.[13]
Introduction
Montassier et al. first investigated the conversion of polyols (i.e.
sorbitol, xylitol, erythritol and glycerol) over a series of carbon
supported catalysts.[14] Ru, Ni, Rh and Ir catalyzed both C-C and
C-O cleavage leading to the formation of methane, whereas
over Cu mainly C-O cleavage was favored. This suggests that
Cu possesses a high potential to obtain glycols with high
selectivity. Recently, Jin et al. developed a multi-functional
Biomass is a cheap and abundant carbon source, which is
playing an increasingly important role in the utilization of
alternative resources.[1] Among the various biomass-based
platform chemicals, sorbitol, which can be obtained from the
hydrogenation of glucose, has attracted considerable attention.[2]
The hydrogenolysis of sorbitol under alkaline conditions results
in the formation of glycerol (GLY),[3] lactic acid (LA), 1,2-
[4]
Cu/CaO-Al O3 catalyst over which the combined selectivity to
2
propanediol (1,2-PDO),[5] ethylene glycol (EG)[6] as well as a
1,2-PDO and EG reached 61.5% with full sorbitol conversion at
503 K after 6h.[15]
A synergistic two-phase mechanism, where
the CaCuAlO phase presents the active site for dehydrogenation
minor amount glycolic acid (GA)[7] (Scheme 1). It is generally
accepted that the major reaction pathways under these
conditions are dehydrogenation of the polyol, base-catalyzed
and retro aldolization, and spinel CuAl
2
O
4
species catalyze
hydrogenation, was proposed. Ca2 was considered as crucial
for the cleavage of C-C bonds to smaller products (e.g. 1,2-PDO,
LA and EG). However, the deeper understanding of the
+
retro-aldolization to form
C
2
and
C
3
products and
rehydrogenation. The resulting products have been widely
applied in industry for renewable plastics, paints, antifreeze,
cosmetics and pharmaceuticals.[8]
The employed catalyst systems for polyol hydrogenolysis can be
divided into noble metal based (e.g. Ru, Pd, Pt, Rh) and non-
noble metal based (e.g. Ni, Cu, W). Guo et al. observed
excellent performance for Ru supported on carbon nanotubes
2+
respective roles of Cu and Ca on the reaction is still missing.
Furthermore, the low intrinsic activity (TOF) of Cu/CaO-Al is
2
O
3
a considerable disadvantage as compared to noble metal
catalysts. Huang et al. revealed that the activity of Cu catalyst in
the conversion of xylitol to glycols is dependent on particle
size.[ But contrary results regarding the optimum particle size
16]
CNTs) in the conversion of sorbitol.[9] Full conversion and 60%
selectivity to glycols was obtained over the WO -promoted
Ru/CNTs catalyst in the presence of Ca(OH) at 478 K. A kinetic
(
of Cu supported on SiO
2
were proposed in different papers.[17]
x
Therefore, more and straightforward techniques should be used
to better determine the particle size. In addition, the highest
selectivity to glycols in former papers is low (61.5%). Thus, a
study on improving the activity of Cu catalyst meanwhile
accelerating the selectivity to glycols is highly important.
2
study by our group on sorbitol and xylitol hydrogenolysis under
neutral/acidic conditions, revealed that Ru/C also catalyzes
Here we present a study concerning the influence of Cu particle
[
a]
X. Wang, A.K. Beine, Dr. P.J.C. Hausoul, Prof. Dr. R. Palkovits
Institut fꢀr Technische und Makromolekulare Chemie
RWTH Aachen University
Worringerweg 2, 52074 Aachen, Germany
Email: palkovits@itmc.rwth-aachen.de
size and equivalents of Ca(OH)
the hydrogenolysis of sorbitol to glycols including 1,2-PDO and
EG. Other than SiO used as support in former papers, activated
2
on the activity and selectivity of
2
carbon due to its low strong metal-support interactions (SMSI)
with metal is employed to support Cu in this work.[ The effects
18]
Supporting information for this article is given via a link at the end
of the document.
of metal loading, reaction conditions, equivalence of Ca(OH)
and recycling are investigated.
2
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900299
FULL PAPER
Table 1. Physical properties of Cu catalysts with different loadings.
Particle size
Particle size
(XRD)
(nm)
Particle size
(TEM)
BET area [
a]
Pore volume
[a]
Dispersion [b]
(%)
Cu area
(m /g)
[b]
Catalyst
C
(N
2
O) [
b]
[c]
[d]
2
3
2
(m /g)
(cm /g)
(nm)
(nm)
1803
1698
1606
1288
1112
742
0.89
0.84
0.79
0.64
0.54
0.38
-
-
-
-
-
5%Cu/C
17.1
13.5
11.5
7.1
115.9
91.5
78.1
48.3
28.6
5.8
7.4
8.7
4.9
3.8 (1.2) [e]
n.d.
10%Cu/C
15%Cu/C
25%Cu/C
40%Cu/C
8.0
9.3
9.5 (3.0)
16.4 (5.4)
25.0 (4.7)
14.0
23.7
18.3
26.0
4.2
[
2
a] Measured by BET. [b] Measured by N O Chemisorption. [c] Calculated from XRD. [d] Calculated from TEM. [e] Standard deviation.
ranges with increasing loading. Moreover, the presented data of
particle sizes determined by three different methods are in line
with each other. The minor discrepancy indicates that the results
of particle size are highly reliable.
Results and Discussion
A series of Cu/C catalysts with different loadings (i.e. 5 - 40
w.t %) was prepared using wet impregnation. The results of N
physisorption, N O-chemisorption (CS) and XRD of the reduced
2
2
catalysts are summarized in Table 1. Notably, with increasing
2
loading the BET area decreases from 1803 m /g for activated
2
carbon to 742 m /g for 40 w.t % Cu/C (see Figure S1 for
isotherms). At the same time, the pore volume decreases from
3
0
.89 to 0.38 cm /g. The significant decrease of both BET surface
area and pore volume with increasing metal loading suggests an
incremental filling of the micropores by Cu. N O-CS is a well-
known method for the evaluation of the dispersion of Cu
2
catalysts. In short, first the Cu(NO
reduced using H
3
)
2
loaded materials are
0
2
. The resulting accessible Cu surface is
oxidized to Cu
2
O by N
2
O, which in the final step is reduced back
0
to Cu , giving the total accessible surface. Based on the
obtained data the dispersion, specific Cu surface and crystallite
size are calculated (Table 1).[19]
0
The data confirm that with increasing loading the dispersion
decreases from 17.1% to 4.2%, the specific Cu surface area
2
declines from 115.9 to 28.6 m /g and the particle size increases
from an average of 5.8 nm to 23.7 nm. Thus a clear trend is
observed between loading and mean particle size. This is further
confirmed by XRD and TEM analysis of the reduced catalysts.
XRD patterns of the reduced catalysts are shown in Figure S2.
Metallic Cu exhibits diffraction peaks at 2θ = 43.3° (Cu(111)),
5
0.4° (Cu(200)) and 74.1° (Cu(220)). Characteristic reflexes of
Cu(I) are missing, confirming that Cu has been fully reduced by
after 4 h at 623 K. The Cu(111) reflex was used to determine
H
2
the particle size using the Debye-Scherrer equation (Table 1).
The data emphasize that the particle size increases with
increasing loading from 4.9 nm for 5 w.t % to 26 nm for 40 w.t %
is in well agreement with data obtained by chemisorption. Also
TEM analysis of the reduced catalysts confirms a clear trend
between the loading and the particle size. Representative TEM
images and Cu particle distributions are illustrated in Figure 1.
As can be seen, the particle size distribution migrates to higher
Figure 1. Particle distributions of Cu catalysts and corresponding electron
microscopy images. Cu
n n
: Cu average particle size. σ : standard deviation of
n
Cu . The scale in TEM is equal to 100 nm.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900299
FULL PAPER
Table 2. Performance of Cu catalysts in the conversion of sorbitol. [a]
Selectivity
Conversion
Catalyst
3 2 4
C /(C +C )
(%)
1.2-PDO
EG
GLY
LA
GA
Others [c]
C
- [b]
-
-
-
-
-
-
-
5
%Cu/C
17.2
30.8
35.9
42.3
20.6
39.3
39.1
36.3
34.4
41.0
10.4
12.3
12.2
12.3
10.5
10.2
14.4
18.9
25.0
3.1
23.4
21.9
19.4
16.3
26.5
11.2
7.0
6.1
6.3
8.9
5.3
5.2
7.2
7.5
10.2
2.7
3.1
3.0
3.2
2.5
10%Cu/C
15%Cu/C
25%Cu/C
40%Cu/C
[
2 2 2 4 6
a] Conditions: 20 mL H O, 0.66 g sorbitol, 10 mg Cu, 0.2g Ca(OH) , 433 K, 2 h, 5 MPa H , 750 rpm. [b] No conversion. [c] Including C -C
1
hexitols, pentitols, triols, diols and trace amount of C chemicals.
that mass transfer limitation due to gas diffusion can also be
neglected. Table 2 summarizes the obtained results. Activated
carbon was also tested and gave no conversion, confirming it is
inert in the reaction. 5%Cu/C enabled 17.2% conversion with
3
2
9.3% selectivity to 1,2-PDO, 10.4% to EG, 10.2% to GLY,
3.4% to LA and 11.2% to GA. With increasing Cu loading, the
conversion increases to an optimum of 42.3% over 25%Cu/C
followed by a decrease to 20.6% over 40%Cu/C. The ratio
between C and (C +C ) products (C /(C +C )) remains roughly
3 2 4 3 2 4
constant, indicating that the particle size has little effect on the
selectivity of the reaction. However, there is significant variation
in the production of 1,2-PDO, LA and GLY. The combined
selectivity of 1,2-PDO and LA decreases from 62.7% over
5%Cu/C to 50.7% over 25%Cu/C, whereas the selectivity to
GLY increases from 10.2% to 25.0%. This suggests that a
stronger hydrogenation catalyst favors the production of GLY
while the total production of C products remains constant.
3
In order to identify the role of the Cu particle size on sorbitol
conversion, the turnover frequency (TOF) was determined in
relation to the average Cu particle size obtained by N
2
O CS
(
Figure 3). The reported TOF is based on the accessible copper
surface and represents the TOF per hour/Cu atom. The data
reveals a clear trend. Increasing the particle size from 5.8 nm to
Figure 2. High resolution TEM image of 25%Cu/C.
1
4 nm leads to a significant increase of the TOF from 10.4 to
-1
54.8 h . From 14 nm onwards the TOF remains constant. The
A high resolution TEM of 25%Cu/C is presented in Figure 2.
Large particles with extended lattice planes are observed. The
distance between the observed lattice fringes is 2.08 Å, which
corresponds to the (111) interplanar spacing of metallic Cu
crystals. The hydrogenolysis of sorbitol was performed in the
observed activity of nanoparticles can be the result of several
factors such as strong metal support interactions (SMSI),
cooperative effects at the metal-support interface and the
proportions of different active surface sites.[18a, 20] Given the low
SMSI of carbon supports, we consider that the size-dependent
2 2
presence of Ca(OH) at 433 K for 2 h under 5 MPa H . The
dehydrogenation performance of Cu is due to the changes of the
amount of Cu in the reaction was kept constant. Due to the
microporous nature of activated carbon, it is assumed that the
Cu particles are located mainly on the external surface. As a
result the effect of intra-particle mass transfer on the reaction
rate can be neglected. In addition, we observed that the reactor
pressure does not change significantly during the reaction. The
Cu surface rather than the interaction with carbon.[
18c]
In any
case the apparent TOF is the average of various different TOFs
due to crystallographic features such as Miller-index planes,
steps, kinks, edges, corners and their respective proportions in
the accessible surface.[ It is considered that with increasing
particle size the proportion of the more stable Cu(111) surface
grows. In line with this the above results suggest that this is also
the catalytically most active surface. Furthermore, the data
21]
influence of initial H
2
pressure on conversion and products
distribution is shown in Figure S3 and shows no significant
influence within a range of 3 to 7 MPa. Therefore we propose
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900299
FULL PAPER
Figure 3. TOF of sorbitol conversion as a function of Cu average particle size.
(Conditions: TOF determined at < 35% conv., 1-2 h, 433 K, 5 MPa H
2
, 750
rpm. (Particle size determined by N O CS)).
2
Figure 4. Products evolution over 25%Cu/C. (Conditions: 0.18 M sorbitol, 0.08
g Cu, 0.2g Ca(OH) , 20 ml H O, 513 K, 5 MPa H , 750rpm, full conversion.)
2 2 2
confirm that beyond 14 nm the specific surface activity of Cu is
independent on particle size. This suggests that for larger
particles the proportion of densely packed surfaces, e.g.
Cu(111), is high regardless of particle size. Thus, since the
specific surface area decreases with increasing particle size it
can be postulated that 14 nm is the ideal particle size in case of
the Cu-catalyzed hydrogenolysis of sorbitol.
Nevertheless, the hydrogenolysis of polyols is the result of a
2
synergy between Cu-catalyzed dehydrogenation and Ca(OH) -
promoted isomerization and retro aldolization. Therefore, not
only reaction temperature and time but also the equivalence of
Ca(OH) was investigated. The influence of temperature on the
2
products distribution was studied using 4h reaction in the
temperature range 433–513 K (Figure S4). In all cases sorbitol
was fully converted at the end of the reaction. The combined
selectivity to glycols (EG and 1,2-PDO) increases with higher
temperature. Notably, the selectivity towards acids (i.e. LA and
GA) significantly decreases at higher temperature. Thus, the
data emphasize that higher reaction temperature favours the
production of the desired glycols. The evolution of the product
distribution over reaction time was monitored at 513 K (Figure 4).
Time zero was set at the time when the reactor first reached the
reaction temperature. During the heating phase (approx. 30 min)
sorbitol was already fully converted. The selectivity to 1,2-PDO
increases while those of LA and GLY decrease, suggesting that
GLY and LA are converted to 1,2-PDO during the reaction. In
contrast, EG and 1,2-butanediol (1,2-BDO) remain constant. At
particularly the
3 2 4
C /(C +C ) ratio provides insight into the
cleavage position of the retro-aldol reaction and the identity of
the preceding carbohydrate intermediate.[22b, 23] For example,
retro-aldol cleavage of an aldose yields one C
2
and one C
4
fragment, whereas cleavage of a 2-ketose yields two C
3
fragments (Scheme 2). Figure 5 shows the product distribution
and the carbon-based ratio of C products (i.e., 1.2-PDO, LA and
GLY) and C +C products (i.e., EG, GA, BDOs and BTO) as a
function of Ca(OH) equivalents w.r.t. sorbitol. The C /(C +C
ratio increases from 1.9 at 0.15 eq. Ca(OH) to a plateau of
approx. 6 at 0.45 eq. This suggests that with increasing amount
of Ca(OH) , the relative amount of ketose in the reaction
increases leading to higher amount of C products. In principle,
metal-catalyzed dehydrogenation of hexitols can yield 1-aldose,
-ketose and 3-ketose intermediates. Dehydrogenation at the 1-
3
2
4
2
3
2
4
)
2
2
3
2
pos. is generally more facile than at the 2-pos. due to a lower
activation barrier.[ It is proposed that the increasing C
ratio with increasing amount of Ca(OH)
15]
3
2 4
/(C +C )
2
is the result of base-
24h, full conversion and a selectivity to glycols (EG and 1,2-
PDO) of up to 84.5% was achieved. To the best of our
knowledge, this is the highest selectivity reported for EG and
,2-PDO so far.[ The conversion of GLY to 1,2-PDO over Cu
15]
1
catalysts is well known, whereas the conversion of LA to 1,2-
PDO over Cu has rarely been reported.[8,
22]
A reaction
performed with LA, under otherwise similar conditions, gave
0% conv. with 1,2-PDO as the sole product. This is the first
2
report on the conversion of lactic acid over Cu under aqueous
reaction conditions. Therefore, it is proposed that the
unprecedented selectivity to glycols is the result of an in-situ
conversion of the side products LA and GLY to the desired 1,2-
PDO.
The impact of Ca(OH)
2
as a promoter on the product distribution
was identified by varying the amount added in the reaction. Here,
Scheme 2. Conversion pathways of hexitols. The main pathway to produce
.2-PDO is highlighted in red.
1
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900299
FULL PAPER
3 2 4
Figure 5. Product distribution and selectivity ratio of C and C +C products
versus the amount of Ca(OH)
2 3
. (C products include 1,2-PDO, GLY and LA
and C +C products include GA, EG, Butanediols). Reaction conditions: 0.18
2
4
M sorbitol, 0.08 g Cu (25%Cu/C), 20 ml H
full conversion.
2
O, 4 h, 513 K, 5 MPa H
2
, 750 rpm,
Figure 7. Particle distributions of used and recalcined Cu catalysts and
corresponding electron microscopy images. Cu
n n
: Cu average particle size. σ :
standard deviation of Cu . The scale in TEM is equal to 100 nm. Used
n
represents 5-time used catalyst. Recalcined catalyst is obtained from
calcination of 5-time used catalysts.
catalyzed isomerization (i.e. via the Lobry-de-Bruyn-van-
Ekenstein reaction). Due to the low production of C and/or C
1 5
[
24]
chemicals, the route involving the formation of 3-ketose appears
not to be preferred. Besides this, the selectivity for LA formation
also increases, partly at the expense of 1,2-PDO. It should be
Cu steadily decreases over the runs. ICP-OES analysis of the
filtered solutions reveals that leaching (< 0.05%) is negligible in
each run (Table S1). A TEM image and particle size distribution
of the catalyst 5 times used is displayed in Figure 7. Cu particles
with increasing size w.r.t the pristine catalyst are observed.
Moreover the average particle size has shifted from 16.5 to 42.6
nm and now also significant amounts of particles large than 60
nm are observed. Therefore, the deactivation of Cu catalyst is
attributed to sintering during the hydrothermal reaction.
noted that the employed amount of Ca(OH)
in all cases, thus all reactions are performed in saturated
Ca(OH) solution with a fraction of solid Ca(OH)
.[25]
In fact, the amount of Ca(OH) is sufficient (i.e., even when
2
exceeds solubility
2
2
2
taking the formation of lactic acid into account) to remain
saturated throughout the whole experiment. Given the strong
dependence of the product distribution on the total amount of
Ca(OH) , this suggests that the solid fraction also participates in
2
the reaction. Compared to the invariant selectivity ratio observed
for the Cu particle size data, these results reveal that the total
Liu et al. reported that aggregated Cu particles supported on
SiO
2
can be redispersed by a thermal treatment (re-calcination)
with recovery of the catalytic activity.[ In line with this we have
17]
attempted reactivation of the aggregated catalyst by first
amount of Ca(OH)
product distribution.
The stability of the 25%Cu/C catalyst was tested in
2
plays a much more important role on the
washing with diluted H
3
PO
4
soln., followed by calcination under
at 623K. XRD patterns in
N
2
at 823 K and reduction under H
2
5
recalcined catalyst (Figure S2) show no peaks other than Cu
confirming the removal of all Ca salts. A TEM image of the
treated catalyst (Figure 6) shows no improvement in the particle
size distribution and also the activity was not recovered.
Presumably the poor stabilization of atomic copper on active
carbon prevents a redispersion at high temperature.
subsequent runs at 513 K, 5 MPa. The obtained results are
shown in Figure 6. The data reveal that both the conversion of
sorbitol and the selectivity to glycols decrease significantly.
Moreover, the selectivity to LA increases at the expense of 1,2-
PDO. These results indicate that the hydrogenation activity of
Conclusions
Cu catalysts supported on activated carbon with different particle
sizes were thoroughly investigated by XRD, N O chemisorption
2
and TEM. All methods confirm a clear trend that particle size is
increased with higher loading. Moreover, the results of particle
size determined by different methods are in line with each other.
Cu catalysts with different particle size were used for the
conversion of sorbitol to produce glycols in the presence of
2
Ca(OH) . The TOF reveals a distinct relation between the
intrinsic activity and particle size. It is proposed that Cu particles
of 14 nm present an optimum in terms of activity. The particle
size dependence is attributed to the fraction of Cu(111) lattice
plane of the accessible surface. However, Cu particle size has
Figure 6. Conversion and products distribution of catalyst recycling
experiments. Reaction conditions: 0.18 M sorbitol, 0.08 g Cu (25%Cu/C), 20
ml H O, 4 h, 513 K, 5 MPa H , 750 rpm, full conversion.
2 2
no significant effect on the ratio of C
3
/(C
2
+C
4
) products which is
/(C +C ) increases
controlled by the eq. Ca(OH) . The ratio of C
2
3
2
4
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900299
FULL PAPER
from 1.9 at 0.15 eq. Ca(OH)
2
to a plateau of approx. 6 at 0.45 eq. by ultrasonication. Drops of the suspension were applied onto a nickle
grid-supported transparent carbon foil and dried in air. The particle size
2
Therefore, the respective roles of Cu and Ca(OH) on sorbitol
conversion to glycols are identified: (1) conversion of sorbitol is
controlled by the Cu particle size while (2) the selectivity to
was determined by software ImageJ.
Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES).
ICP-OES was performed on an ICP Spectroflame D by spectro.
3 2 4 2
C /(C +C ) products is affected by the equivalence of Ca(OH) .
Notably, the time course indicates in-situ conversion of the side
product lactic acid to 1,2-PDO, leading to full conversion and an
unprecedented selectivity of 84.5% to glycols (EG and 1,2-PDO)
2
N O Chemisorption (N
2
O CS). The specific Cu surface areas, dispersion
and crystallite size were measured by N
Pulsar TPR/TPD automated chemisorption analyzer.[19a-c] The
chemisorption process consists of three sequential steps:
2
O chemisorption on a ChemBET
at 513 K, 5 MPa H
2
and 0.3 eq. Ca(OH)
2
. Recycling tests show
2
N O
that the Cu/C-catalyst was deactivated due to the aggregation of
particles verified by TEM.
Thus it is demonstrated that Cu/C is a promising catalyst for
sorbitol hydrogenolysis that can achieve excellent selectivity to
glycols. Nevertheless, high temperatures and high amounts of
base are needed and the propensity for copper to sinter is still a
challenge that needs to be addressed. Therefore it would be of
importance to design new catalysts that prevent the aggregation
of Cu nanoparticles.
CuO + H → Cu
0
+ H O
H consumption= A1
(Eq. 2)
(Eq. 3)
2
2
2
0
2Cu + N
2
O → Cu
2
O + N
2
0
Cu O + H
2
2
→ 2Cu + H
2
O
H
2
consumption= A2
(Eq. 4)
First, CuO is reduced to Cu0 (Eq. 2). In this step, 5 vol.% H
in Ar (100
2
mL/min) was used as a reducing agent (H
temperature raised from r.t. to 623 K (10 K/min). Then Cu is oxidized to
2
consumption A1), and the
0
Cu
was cooled to 333 K in He (100 mL/min) and purged with He for 30 min.
Then, pure N O (100 mL/min) was introduced to the catalyst at 333 K for
h. Subsequently, the catalyst was purged with He (100 mL/min) for 1 h
to remove the residual N O. Eq. 4 represents the reduction of Cu
surface species. In this step, 5 vol.% H /Ar (100 mL/min) was used as the
reducing agent (H consumption A2) and the temperature raised from
33 K to 623 K with a heating rate of 10 K/min. The dispersion (D) of Cu
2 2
O by N O (Eq. 3). This step was initiated after the reduced catalyst
Experimental Section
2
All chemicals were used as received. Cu(NO
3
)
2
·3H
98%,) 1,2-propanediol (99%,) ethylene glycol (99.8%) glycolic acid
99%,) 1,2-butanediol (98%,), 1,2,3-butanetriol, 1,2,4-butanetriol (≥ 90%,)
2 2
O (≥ 98%), Ca(OH)
1
(
(
2
2
O
2
and sorbitol (≥ 98%) were obtained from Sigma Aldrich. Xylitol (100%)
and glycerol (100%) were obtained from SUPELCO. Activated carbon
supra was obtained from ACROS Organics. Lactic acid (90%) was
obtained from Fluka.
2
3
[19d]
was calculated using Eq. 5:
2
퐴2
퐷 =
× ꢀ00%
(Eq. 5)
퐴1
Copper catalysts supported on activated carbon with different loadings
were prepared via
amounts of Cu(NO
mL) with stirring for 1 h in a flask followed by addition of activated carbon
1 g). The resulting slurry was stirred at around 500 rpm for at least 2 h at
room temperature in order to obtain a well-dispersed suspension. The
slurry was dried using a rotary evaporator (323 K, 7 kPa, 160 rpm) and
resulting materials were additionally dried in a drying oven at 353 K for
a
wetness impregnation method. The required
Specific area of metallic copper was calculated from the amount of H
2
3
)
2
·3H O was dissolved into deionized water (DI, 10
2
consumption (A2) with 1.4 × 1019 copper atoms/m . The accessible
surface area (SCu) and particle size of Cu (d) were calculated using Eq. 6
and Eq. 7: [19d]
2
(
2
∙퐴2∙푁푎푣
1353 퐴2
퐴1
푆 =
푑 =
≈
= 676.ꢃ 퐷
(Eq. 6)
(Eq. 7)
퐴1∙푀퐶푢∙1.4 푥 1ꢁꢂ9
2
overnight. The catalysts were activated in H flow at 623 K for 4 h. The
pristine catalysts were stored under Ar atmosphere before activity tests.
ꢄ
1
≈
S∙ρ퐶푢
ꢅ
With Nav = 6.02 × 1023 (mol ), MCu = 63.546 (g/mol), ρCu = 8.92 (g/cm3).
-1
Brunauer−Emmett−Teller (BET) Measurement. The textual properties of
the samples were determined by
N
2
adsorption at 77 K using
QUADRASORB SI automated surface area and pore size analyzer.
Samples were prepared in vacuum at 473 K for overnight. The pore-size
distribution was estimated by the Barrett-Joyner-Halenda (BJH) method
from the desorption branches of the adsorption isotherms. Surface areas
were calculated using the BET method.
Hydrogenolysis of sorbitol was carried out in a 50 mL autoclave with
Teflon insert. In a typical run, aqueous sorbitol solution (20 g, 5 w.t%),
Ca(OH)
2
(0.2 g) and Cu catalyst were charged into the autoclave. The
. The
reactor was purged five times and pressurized to 5 MPa with H
2
sealed autoclave was heated to 513 K with 750 rpm stirring. Samples
were taken periodically.
X-ray diffraction (XRD). The wide-angle XRD patterns were recorded on
a Bruker D8 Advance X-ray diffractometer using nickel-filtered Cu Kα
radiation (휆 = 0.15406 nm) with a scanning angle (2θ) range of 20–80°, a
scanning speed of 4°/min with a voltage of 40 kV, and a current of 20 mA.
The crystallite sizes (d) of Cu in the samples were calculated by the
Scherrer equation (Eq. 1):
The gas-phase was collected using a gasbag and analyzed by GC
performed on a Agilent HP 6890 instrument with a Shin Carbon ST
micropacked GC column (308-373 K, TCD, He). Liquid-phase products
were analyzed by HPLC performed on a Shimadzu system (Rezex ROA
Organic Acid H+ (8%) column by Phenomenex, eluent: 0.05 M H
SO ).
2 4
The sorbitol conversion (Eq. 8) and selectivity (Eq. 9) were calculated on
the basis of carbon. The TOF was calculated using (Eq. 10).
푘λ
푑 =
(Eq. 1)
훽푐표푠휃
푚표푙ꢋ 표푓 ꢌꢋꢍ푐푡ꢍꢎ푡 푐표ꢎ푠ꢏ푚ꢋꢐ
ꢆꢇ푛ꢈ푒푟ꢉꢊꢇ푛 =
푆푒ꢑ푒ꢒꢓꢊꢈꢊꢓ푦 =
푇푂퐹 (ꢖ−1) =
× ꢀ00%
(Eq. 8)
(Eq. 9)
(Eq. 10)
Where k is the Debye–Scherrer constant (0.90), θ is the diffraction angle
푚표푙ꢋ 표푓 ꢌꢋꢍ푐푡ꢍꢎ푡 푐ℎꢍꢌ푔ꢋꢐ
and β is the full width at half maximum. The average crystallite size of
푚표푙ꢋ 표푓 푝ꢌ표ꢐꢏ푐푡 × ꢔ ꢍ푡표푚푠 ꢕꢎ 푝ꢌ표ꢐꢏ푐푡
0
Cu was estimated from Cu(111) (2θ = 43.3°).
× ꢀ00%
푚표푙ꢋ 표푓 ꢌꢋꢍ푐푡ꢍꢎ푡 푐표ꢎ푠ꢏ푚ꢋꢐ × C atoms in reactant
푚표푙ꢋ 표푓 ꢌꢋꢍ푐푡ꢍꢎ푡 푐표ꢎ푠ꢏ푚ꢋꢐ
Transmission electron microscopy (TEM). Images were obtained using a
Philips TECNAI G2 F20 system electron microscope at 100 kV equipped
with a field emission gun. The sample powder was dispersed in ethanol
× ꢀ00%
푚표푙ꢋ 표푓 푠ꢏꢌ푓ꢍ푐ꢋ 푐표푝푝ꢋꢌ ꢍ푡표푚푠 × 푡ꢕ푚ꢋ (ℎ)
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900299
FULL PAPER
[
[
[
11]
12]
13]
M. Banu, P. Venuvanalingam, R. Shanmugam, B. Viswanathan, S.
Sivasanker, Top. Catal. 2012, 55, 897-907.
L. M. Ye, X. P. Duan, H. Q. Lin, Y. Z. Yuan, Catal. Today 2012,
Acknowledgements
183, 65-71.
The authors kindly appreciate Heike Fickers-Boltz, Elke Biener
and Hannelore Eschmann for the GC analyses, Noah Avraham
and Simon Petring for HPLC analyses, Karl Vaeßen for XRD
measurements, Heike Bergstein for ICP-OES analyses and
Shun Yu for experimental support. X. Wang acknowledges a
scholarship from the Chinese Scholarship Council.
a) B. Chen, B. Zhang, Y. B. Zhang, X. G. Yang, ChemCatChem
2016, 8, 1929-1936; b) M. A. Dasari, P. P. Kiatsimkul, W. R.
Sutterlin, G. J. Suppes, Appl. Catal. A: Gen. 2005, 281, 225-231;
c) T. van Haasterecht, C. C. I. Ludding, K. P. de Jong, J. H. Bitter,
J. Energy Chem. 2013, 22, 257-269.
C. Montassier, D. Giraud, J. Barbier, J. P. Boitiaux, Bull. Soc. Chim.
Fr. 1989, 148-155.
X. Jin, J. Shen, W. J. Yan, M. Zhao, P. S. Thapa, B. Subramaniam,
R. V. Chaudhari, ACS Catal. 2015, 5, 6545-6558.
[14]
[
[
15]
16]
Z. W. Huang, J. Chen, Y. Q. Jia, H. L. Liu, C. G. Xia, H. C. Liu,
Appl. Catal. B: Environ. 2014, 147, 377-386.
Keywords: Sorbitol • 1,2-propanediol • ethylene glycol • copper
particle • Ca(OH)
2
• hydrogenolysis
[17]
H. L. Liu, Z. W. Huang, C. G. Xia, Y. Q. Jia, J. Chen, H. C. Liu,
ChemCatChem 2014, 6, 2918-2928.
[
18]
a) S. J. Tauster, S. C. Fung, R. T. K. Baker, J. A. Horsley, Science
[
1]
a) G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044-
1981, 211, 1121-1125; b) S. J. Tauster, Acc. Chem. Res. 1987, 20,
4098; b) M. Besson, P. Gallezot, C. Pinel, Chem. Rev. 2014, 114,
389-394; c) T. Daio, A. Staykov, L. M. Guo, J. F. Liu, M. Tanaka, S.
1827-1870; c) I. Delidovich, P. J. C. Hausoul, L. Deng, R.
M. Lyth, K. Sasaki, Sci. Rep. 2015, 5, 13126.
a) O. Hinrichsen, T. Genger, M. Muhler, Chem. Ing. Tech. 2000,
Pfutzenreuter, M. Rose, R. Palkovits, Chem. Rev. 2016, 116,
540-1599; d) P. Gallezot, Chem. Soc. Rev. 2012, 41, 1538-1558.
a) A. Corma, S. Iborra, A. Velty, Chem. Rev. 2007, 107, 2411-
[
19]
1
72, 94-98; b) G. C. Chinchen, C. M. Hay, H. D. Vandervell, K. C.
[
2]
Waugh, J. Catal. 1987, 103, 79-86; c) J. L. Gong, H. R. Yue, Y. J.
Zhao, S. Zhao, L. Zhao, J. Lv, S. P. Wang, X. B. Ma, J. Am. Chem.
Soc. 2012, 134, 13922-13925; d) C. J. G. Vandergrift, A. F. H.
Wielers, B. P. J. Joghi, J. Vanbeijnum, M. Deboer, M.
Versluijshelder, J. W. Geus, J. Catal. 1991, 131, 178-189.
a) F. B. M. Vanzon, S. D. Maloney, B. C. Gates, D. C.
Koningsberger, J. Am. Chem. Soc. 1993, 115, 10317-10326; b) D.
C. Koningsberger, B. C. Gates, Catal. Lett. 1992, 14, 271-277; c) B.
Skarman, L. R. Wallenberg, P. O. Larsson, A. Andersson, J. O.
Bovin, S. N. Jacobsen, U. Helmersson, J. Catal. 1999, 181, 6-15.
Z. P. Liu, P. Hu, J. Am. Chem. Soc. 2003, 125, 1958-1967.
a) R. D. Cortright, M. Sanchez-Castillo, J. A. Dumesic, Appl. Catal.
B: Environ. 2002, 39, 353-359; b) K. Tajvidi, P. J. C. Hausoul, R.
Palkovits, ChemSusChem 2014, 7, 1311-1317.
2
502; b) X. G. Zhang, K. Wilson, A. F. Lee, Chem. Rev. 2016, 116,
12328-12368; c) B. H. Wang, L. Zeng, W. Huang, F. S. Melkonyan,
W. C. Sheets, L. F. Chi, M. J. Bedzyk, T. J. Marks, A. Facchetti, J.
Am. Chem. Soc. 2016, 138, 7067-7074.
Y. G. Zheng, X. L. Chen, Y. C. Shen, Chem. Rev. 2010, 110,
[
[
3]
4]
[
20]
1807-1807.
S. Iqbal, S. A. Kondrat, D. R. Jones, D. C. Schoenmakers, J. K.
Edwards, L. Lu, B. R. Yeo, P. P. Wells, E. K. Gibson, D. J. Morgan,
C. J. Kiely, G. J. Hutchings, ACS Catal. 2015, 5, 5047-5059.
Y. Liu, C. Luo, H. C. Liu, Angew. Chem. Int. Ed. 2012, 51, 3249-
[
[
[
[
[
[
5]
[
[
21]
22]
3253.
6]
N. Ji, T. Zhang, M. Zheng, A. Wang, H. Wang, X. Wang, J. G.
Chen, Angew. Chem. Int. Ed. 2008, 47, 8510-8513.
R. Y. Sun, M. Y. Zheng, J. F. Pang, X. Liu, J. H. Wang, X. L. Pan,
A. Q. Wang, X. D. Wang, T. Zhang, ACS Catal. 2016, 6, 191-201.
A. M. Ruppert, K. Weinberg, R. Palkovits, Angew. Chem. Int. Ed.
7]
[
[
23]
24]
P. J. C. Hausoul, L. Negahdar, K. Schute, R. Palkovits,
ChemSusChem 2015, 8, 3323-3330.
8]
a) I. Delidovich, R. Palkovits, ChemSusChem 2016, 9, 547-561; b)
A. K. Beine, A. J. D. Kruger, J. Artz, C. Weidenthaler, C. Glotzbach,
P. J. C. Hausoul, R. Palkovits, Green Chem. 2018, 20, 1316-1322.
B. Kutus, A. Gacsi, A. Pallagi, I. Palinko, G. Peintler, P. Sipos,
RSC Adv. 2016, 6, 45231-45240.
2012, 51, 2564-2601.
9]
X. C. Guo, J. Guan, B. Li, X. C. Wang, X. D. Mu, H. Z. Liu, Sci.
Rep. 2015, 5, 16451.
P. J. C. Hausoul, A. K. Beine, L. Neghadar, R. Palkovits, Catal. Sci.
Technol. 2017, 7, 56-63.
[
25]
10]
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900299
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Sorbitol conversion to glycols is a
synergistic reaction catalyzed by Cu
Xinde Wang, Anna Katharina Beine,
Peter J.C. Hausoul, Regina Palkovits*
2
and Ca(OH) . The activity of Cu
depends on particle size with a
maximum TOF reached at 14 nm.
Regarding the decreased specific
surface area for bigger particle, the
catalyst with 14 nm Cu particle
diameter shows highest activity.
However, particle size does not
Page No. – Page No.
Copper-catalysed hydrogenolysis of
sorbitol to glycols–on the influence of
particle size and base
influence the ratio of C
products which is controlled by the
equivalence of Ca(OH) charged.
3 2 4
/(C +C )
2
This article is protected by copyright. All rights reserved.