A Robust and Highly Selective Catalytic System of Copper–Silica Nanocomposite and 1-Butanol in Fructose Hydrogenation to Mannitol

Article abstract of DOI:10.1002/cssc.202001323

We report for the first time the selective production of mannitol, a low-calorie sweetener and an important pharmaceutical ingredient, from fructose using Cu?SiO₂ nanocomposite as catalyst and 1-butanol as solvent. When compared with water and ethanol, a lower fructose solubility was achieved in 1-butanol, which caused a lower fructose conversion and higher mannitol selectivity by reducing formation of side products. Among various Cu-based catalysts in 1-butanol, Cu(80)?SiO₂ nanocomposite gave an unprecedented mannitol (83 %) and sorbitol (15 %) yield at 120 °C, 35 bar H₂, and 10 h reaction time. More importantly, this catalyst did not show any Cu leaching and its physicochemical properties were maintained after liquid-phase fructose hydrogenation whereas other Cu-based catalysts such as Cu(32)?Cr₂O and Cu(66)?ZnO did show significant leaching of Cu and Cr. Thus, Cu(80)?SiO₂ nanocomposite and 1-butanol are regarded as a robust and highly efficient catalytic system for the selective hydrogenation of fructose to mannitol. Also, density functional theory calculations supported that in addition to the stable initial structure of adsorbed fructose, the mannitol pathway was more thermodynamically favorable than the sorbitol pathway. Notably, the highly pure mannitol (99 %) could be recovered from the sorbitol-containing 1-butanol solution by simple filtration. Therefore, the present protocol is a novel and effective method to produce pure mannitol from fructose in both an environmental and an industrial context.

Full text of DOI:10.1002/cssc.202001323

Chemistry–Sustainability–Energy–Materials
Accepted Article
Title: A Robust and Highly Selective Catalytic System of Copper-Silica
Nanocomposite and 1-Butanol in Fructose Hydrogenation to
Mannitol
Authors: Pravin P. Upare, Young Kyu Hwang, Jin Chul Kim, Jeong
Hyeon Lee, Sang Kyu Kwak, and Dong Won Hwang
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: ChemSusChem 10.1002/cssc.202001323
Link to VoR: https://doi.org/10.1002/cssc.202001323
01/2020

10.1002/cssc.202001323
ChemSusChem
A Robust and Highly Selective Catalytic System of Copper-Silica
Nanocomposite and 1-Butanol in Fructose Hydrogenation to Mannitol
Pravin P. Upare[a], Young Kyu Hwang [a,b], Jin Chul Kim[c], Jeong Hyeon Lee[c], Sang Kyu Kwak[c]*,
Dong Won Hwang[a,b]
*
aGreen Carbon Catalysis Research Group, Korea Research Institute of Chemical Technology (KRICT), 141 Gajeongro,
Yuseoung, Daejeon 305–600, Republic of Korea
bDepartment of Advanced Materials and Chemical Engineering, University of Science and Technology (UST), 113
Gwahangno, Yuseong, Daejeon 305–333, Republic of Korea
cDepartment of Energy Engineering, School of Energy and Chemical Engineering, Ulsan National Institute of Science and
Technology (UNIST), 50 UNIST-gil, Ulju-gun, Ulsan 44919, Republic of Korea
We report for the first time the selective production of mannitol, a low calorie sweetener and an
important pharmaceutical ingredient, from fructose using Cu-SiO2 nanocomposite catalyst and 1-
butanol solvent. When compared with water and ethanol solvents, a lower fructose solubility was
achieved in 1-butanol, which caused a lower fructose conversion and higher mannitol selectivity
by reducing formation of side products. Among various Cu-based catalysts in 1-butanol, Cu(80)-
SiO2 nanocomposite gave an unprecedented mannitol (83%) and sorbitol (15%) yield at 120 °C,
35 bar H2, and 10 h reaction time. More importantly, this catalyst did not show any Cu leaching
and its physico-chemical properties were maintained after the liquid-phase fructose hydrogenation,
while other Cu-based catalysts such as Cu(32)-Cr2O and Cu(66)-ZnO did show a significant
leaching of Cu and Cr. Thus, Cu(80)-SiO2 nanocomposite and 1-butanol is regarded as a robust
and highly-efficient catalytic system for selective hydrogenation of fructose to mannitol. Also,
density functional theory (DFT) calculations supported that in addition to the stable initial structure
of adsorbed fructose, mannitol pathway was more thermodynamically favorable than sorbitol
pathway. Notably, the highly-pure mannitol of 99% could be recovered from the sorbitol-
containing 1-butanol solution by simple filtration. Therefore, the present protocol is a novel and
effective method to produce a pure mannitol from fructose in both an environmental and an
industrial context.
1. INTRODUCTION
Biomass, a renewable carbon source, offers great potential
among the available alternative resources for the production of
valuable chemicals and liquid fuels.1,2 In particular, carbohydrates
are the most abundant biomass and great attention has been paid to
chemical conversion into valuable chemicals and fuels.3-6
Mannitol is a kind of polyol which is mostly used as low calorie
sweetener in diabetic food and in chewing gums as it is poorly
absorbed from the intestines.7 It is also widely used in
pharmaceutical, cosmetic, surfactants and coating industries.8 The
global mannitol market size was valued at USD 209 million in 2015
and it is expected to reach USD 418 million by 2024.9
Mannitol is naturally occurring sugar and it can be extracted
form vegetables, fruits, manna, seaweed, and algae. However, the
extraction of mannitol from these plants is not economical.10
Currently, most of mannitol (around 50,000 tons/year) are
produced by the catalytic hydrogenation of fructose, which is more
abundant hexose in nature.11
Nickel-based catalysts such as Raney Ni are one of the most
common catalysts used for the aqueous phase hydrogenation of
fructose.12-14 However, nickel-based catalysts gives mannitol yield
of around 50% with a significant amount of sorbitol, an isomer of
mannitol. Some noble metal catalysts such as Ru and Pt were
reported to be more active than nickel-based catalysts, but these
catalysts are expensive and the selectivity to mannitol is still
limited to around 50%.15,16
In terms of mannitol selectivity, Cu-based catalysts are more
desirable than the noble and nickel-based catalysts, although the
catalytic activities were a little lower.17,18 Recently, Marchi et.al.
investigated various types of Cu-based catalysts for fructose
hydrogenation in ethanol-water (70:30) mixture as solvent.18 They
found that the metallic Cu nanoparticles dispersed on the surface of
SiO2 as a neutral support was the most active and selective catalyst,
which gave fructose conversion of 98% with mannitol selectivity
of 78% after 6 h at 100 C and 40 bar from the initial fructose
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001323
ChemSusChem
concentration of 1 wt.%. However, the catalyst stability such as Cu
leaching during the liquid-phase reaction is of concern.
solid was then dried at 120 oC in air for 12 h, calcined at 350 ºC in
air for 4 h.
Recently, we have developed a copper-silica nanocomposite
catalyst, which were very active, selective and robust for
hydrogenation of biomass derived carboxylic acid (C4-C6) and
their esters into desired products in vapor phase conditions.19,20 We
have also found that 1-butanol was very effective solvent in terms
of product selectivity as well as catalyst stability in liquid-phase
dehydration of fructose over Amberlyst-15 catalyst and
hydrogenation of 5-hydroxymethyl furfural (HMF) over copper-
silica nanocomposite and ruthenium-tin/zinc oxide catalyst.21-23 1-
butanol is also a sustainable solvent that can be produced from a
variety of waste biomass sources and it is widely used in production
of cosmetics and pharmaceuticals24.
The calcined catalysts were pressed into pellets, crushed, and
sieved (No. 20–40 mesh). Finally, the sieved catalysts were
reduced at 290 oC with a mixture of 5% H2 in N2 (50 ml/min) for 5
h before reaction.
2. 3 Catalyst Characterizations
The structure and crystallinity of catalyst samples were
determined by X-ray diffraction analysis (Rigaku, D/MAX IIIB or
Rigaku D/Max 2200 PC, Cu Kα radiation). The particle
morphology and particle size were analyzed with a transmission
electron microscope (TEM, TECNAI G2 RETROFIT). The
nitrogen sorption experiments were performed at -196 oC on a
volumetric sorption analyzer (Micromeritics TriStar 3020). The
surface area and pore volumes were evaluated using the Brunauer-
Emmett-Teller (BET) method for the N2 adsorption isotherm at -
o
o
196 C after outgassing under vacuum at 150 C for 12 h using
Micromeritics Tristar II 3000. H2-Temperature Programmed
Reduction (TPR) experiments were carried out in Micromeritics
model Pulse Chemisorb 2705 equipped with a thermal conductivity
detector (TCD) to monitor H2 consumption in the temperature
range of 100 oC to 800 oC in 5%H2/He atmosphere.
Thermogravimetric analysis (TGA) were carried out in a TGA2050
TA equipment from 25 to 600 °C at a heating rate of 5 °C/min under
air flow (100 ml/min).
Scheme 1. A reaction scheme of fructose hydrogenation to mannitol over
Cu-SiO2 nanocomposite in 1-butanol
To overcome the limitations of the prior art, we herein propose
a new process protocol for mannitol production from fructose based
on copper-silica nanocomposite catalyst and 1-butanol solvent
(Scheme 1). This protocol gave a unprecedented mannitol yield of
83% from high fructose concentration of 10 wt.% and interestingly,
the highly-pure mannitol could be recovered from the sorbitol-
containing 1-butanol solution by simple filtration.
2. 4 Fructose Hydrogenation
The fructose hydrogenation was carried out in a 100 ml
autoclave reactor. In a typical experiment, 5.0 g of fructose and 1.0
g of reduced catalyst were added to 45.0 g of 1-butanol in quartz
vessel and then it was loaded into the autoclave system. High
fructose concentration of 10 wt.% was not soluble in 1-butanol at
room temperature, but it was dissolved at temperatures above 85
C. The reactor was sealed, and purged with nitrogen for 5 min, and
finally pressurized with hydrogen. Fructose hydrogenation was
performed at 120 C, and 20 to 50 bar of H2 pressure for 1 to 10 h
using1000 rpm of stirring speed. After reaction, the reaction
mixture was cooled down to room temperature and mixed with an
equal amount of dimethylformamide (DMF) to dissolve the whole
reaction mixture, and then the catalyst was separated out by simple
filtration. Fructose conversion and product selectivity were
analyzed by HPLC (Youngin HPLC, Korea) equipped with a
refractive index detector and Comosil Sugar-D (4.6 ID x 250mm)
column using 80% acetonitrile in water as a mobile phase with a
flow rate of 0.5 ml/min.
2. EXPERIMENTAL SECTION
2.1 Materials
D-fructose (99%, Alfa Aeser) and 1-butanol (99.9%, Acros
Organic) were used for fructose hydrogenation without further
purification. Cu(NO3)2·3H2O (Wako), Zn (NO3)2.6H2O (Aldrich),
NaOH (AR grade, Samchun chemicals), and Ludox colloidal silica
(SM-30, Aldrich) were used for preparation of Cu(x)-SiO2 and Cu-
ZnO catalyst. D-mannitol (98%, Aldrich) and D-sorbitol (98%,
Aldrich) were used as standard chemicals for quantitative analysis
in high performance liquid chromatography. 2CuO·Cr2O3 (copper
chromite, Aldrich) was used as a hydrogenation catalyst after
reduction at 300 ºC for 2 h.
2. 5 Computational Details
To carry out theoretical DFT calculations on the Cu-SiO2
catalyst, we employed a slab model consisting of four atomic layers
of Cu with (111) surface of the crystal structure (Fig. S1a). Top two
layers of the model system were relaxed while bottom two layers
were fixed to represent the bulk phase during the optimization
calculation. The vacuum region of slab model was set to about 18
Å for all slab models to avoid self-interaction error. By checking
the distances between adsorbed hydrogens, the most stable
configuration of adsorbed hydrogen atoms on the surface of copper
slab model was found (Fig. S1b).
2.2 Catalyst Preparation
Cu(x)-SiO2 nanocomposites with different Cu loading (40 –
100 wt. %) were prepared by precipitation-deposition method. A
known amount of Cu(NO3)2⋅3ꢀH2O and colloidal silica were added
to distilled water. For precipitation, 0.2N NaOH and the above
solution was mixed at 0 oC while maintaining the suspension pH of
9.2 for homogeneous dispersion and control of nucleation rate. The
suspension was stirred for 12 h at 20 oC followed by stirring at 85
ºC for 5 h for nuclei growth. Next, the suspension was filtered and
washed with distilled water repeatedly until sodium could not be
detected in the filtrate. The solid was then dried at 110 ◦C in air for
12 h, calcined at 550 ºC in air for 6 h.
DFT calculations were conducted with Dmol3 program.25,26 We
used Perdew-Burke-Ernzerhof (PBE) exchange-correlation
functional27 and DNP 4.4 basis set with the DFT semi-core
pseudopotential (DSPP) core treatment. For the correction of
dispersion force, the Tkatchenko-Scheffler scheme was used.28 The
convergence criteria for energy, force, and displacement were set
to 1×10-5 Ha, 0.02 Ha/nm, and 0.0005 nm respectively. The
electrostatic contribution of solvent was described implicitly by
employing the Conductor-like Screening Model (COSMO) method
The Cu(66)-ZnO catalyst (Cu/Zn=2.0 in mole) was prepared by
co-precipitation method. A known amount of Cu(NO3)2·3H2O and
Zn (NO3)2.6H2O were dissolved in distilled water. For precipitation,
2M NaOH and the above solution was mixed at 65 oC while
maintaining the suspension pH of 6.5. The suspension was stirred
o
for 5 h at 65 C and then filtered and washed with distilled water
29
repeatedly until sodium could not be detected in the filtrate. The
with the dielectric constant of 1-butanol (ε = 17.7). As well as
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001323
ChemSusChem
reduction of Cu(x)-SiO2 nanocomposite catalysts at 290 C for 2h,
all CuO phase was transformed into Cu metallic state (Fig. 1B).
implicit solvation effect, a butanol molecule was added around the
adsorbed fructose molecule for the consideration of explicit
interaction with solvent. To calculate transition states in the
mechanism of fructose hydrogenation, we employed complete
single linear synchronous transit (LST) and quadratic synchronous
transit (QST) methods30,31, and the convergence criteria of the root
mean square force was set to be 0.02 Ha/nm.
A
B
3. RESULTS AND DISCUSSION
A series of Cu(x)-SiO2 nanocomposite catalysts (where x is the
Cu loading of 40–80 %) were prepared via a precipitation-
deposition method using colloidal silica (BET surface area,
345 m2/g; particle size, 7 nm) and Cu(NO3)2·3H2O. This method
successfully produced catalysts with high Cu loadings, while
maintaining high specific surface areas by blending metallic copper
nanoparticles with the silica nanoparticle matrix.
Figure 2. TEM Images of (A) fresh Cu(80)-SiO2 and (B) used Cu(80)-
SiO2
A
B
Table 1 summarizes the physicochemical properties and the
catalytic activities of the various Cu catalysts with different Cu
loading and support. The surface area of the Cu-SiO2
nanocomposite decreased from 347 to 145 m2/g upon increasing
the copper loading from 40 to 80 %. The average crystallite sizes
of the metallic Cu particles were calculated from line broadening
in the (111) plane at 2 of 43.3 using the Scherrer equation. More
specifically, the average crystallite size of the copper nanoparticles
was estimated to be ~16–20 nm, which remained relatively
constant for all Cu(x)-SiO2 catalyst samples with a little change of
pore size in a range of Cu loading from 40% to 60%. However,
when Cu loading was higher than 80%, the reduction in pore size
was more important due to the higher coverage of Cu on SiO2
nanoparticles. The copper particle size of Cu(80)-SiO2 based on
TEM image was in the range from ~9–23 nm and the average was
18 nm, which was consistent with the XRD result (Fig. 2A).
We initially examined the fructose hydrogenation using
Cu(50)-SiO2 nanocomposite at 25 bar and 120 C in various
solvents. This catalyst was selected due to its excellent activity,
selectivity, and robustness in the liquid-phase HMF hydrogenation.
22 When comparing the different solvent systems, we observed that
a higher mannitol selectivity was achieved in 1-butanol compared
with those in both water and ethanol (Fig. 3 and Fig. S2).
(111)
(200)
Cu(60)-SiO2
Cu(80)-SiO2
(220)
used Cu(80)-SiO2
fresh Cu(80)-SiO2
fresh Cu(60)-SiO2
fresh Cu(50)-SiO2
CuO
fresh Cu(40)-SiO2
100
150
200
250
300
350
400
450
500
20
30
40
50
60
70
80
Reduction Temperature (C)
2/ degree
Figure 1. Physicochemical properties of Cu-based materials. (A) H2-TPR
profile of the Cu(60)-SiO2, Cu(80)-SiO2 and CuO calcined at 550 C for 6h;
(B) XRD patterns of the Cu-SiO2 catalysts with different Cu loading
TPR profile of the calcined Cu(60)-SiO2, Cu(80)-SiO2 and CuO
showed only one hydrogen uptake peak (Fig. 1A). But, the
reduction temperature centered at 240 C of calcined Cu(60)-SiO2
and Cu(80)-SiO2 was much lower than that of CuO centered at 297
C. These results indicated that Cu-SiO2 nanocomposite even with
high copper loading, have properties different to bulk CuO, with a
high reducibility of the supported CuO owing to the copper oxide
nanoparticles dispersed on the silica nanoparticles.19,20 After
Table 1. Catalytic activity and catalyst stability of fructose hydrogenation over various Cu catalystsa
Pore
Crystallite
size
(nm)
carbon
balance
(%)
Product Yield (%)
Mannitol Sorbitol
SBET
Fructose
Conv. (%)
Metal Leaching
(wt. %)
Catalyst
volume
(m2/g)
(cm3/g)
Cu(40)-SiO2
Cu(50)-SiO2
Cu(60)-SiO2
Cu(80)-SiO2
Cu(100)-SiO2
Cu(32)-Cr2O3
Cu(66)/ZnO
347
226
198
145
106
39
0.34
0.36
0.37
0.27
0.18
0.13
0.24
20
16
18
18
24
27
6
96
75
77
81
83
86
51
74
14
17
16
15
9
93
94
97
98
95
54
93
0.00
100
100
100
100
100
100
0.00
0.00
0.00
Cu (6.4)
Cu (3.3), Cr (2.9)
Cu (0.1)
3
52
19
a Reaction conditions: 120 °C, 35 bar H2, catalyst 1.0 g, fructose 5.0 g, 1-butanol 45 g, stirring 1000 rpm, and reaction time 10 h
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001323
ChemSusChem
In the presence of water, fructose conversion reached 100%
after 6 h, but mannitol and sorbitol yield was only 43% and 18%,
respectively, which indicated that a significant amount of the
fructose degradation and/or condensation product was produced as
a byproduct. It was noticeable that Cu (0.1 wt. %) was significantly
leached out from Cu(50)-SiO2 nanocomposite catalyst during
fructose hydrogenation in water, which might result from the
fructose degradation products. When ethanol was used as the
reaction medium for fructose hydrogenation, fructose conversion
decreased slightly to 97% under the same reaction condition, but
mannitol and sorbitol yield increased to 51% and 18%, respectively.
Figure 4. Effect of reaction parameter on the activity of Cu(x)/SiO2. (A) the
effect of reaction pressure in fructose hydrogenation over Cu(50)-SiO2 at
120 °C and 10 h. C = conversion, Y = yield.; (B) the effect of
fructose/catalyst ratio in fructose hydrogenation over Cu(80)-SiO2 at 120 °C,
35 bar H2 and 10 h. C = conversion, Y = yield.
The catalytic activities of the Cu-SiO2 with different Cu loading
were then compared at 35 bar H2 pressure, 120 C, and 10 h reaction
time (Table 1). Upon increasing Cu loading from 40 to 80%, the
mannitol yield increased from 75% to 83%, while the fructose
conversion reached 100%, except Cu(40)-SiO2 (96%). The
unsupported Cu metal (Cu(100)-SiO2) showed the highest mannitol
yield of 86% at 100% fructose conversion, but a noticeable copper
leaching (6.4 wt. %) was observed in the reaction mixture.
Interestingly, Cu leaching was not observed after fructose
hydrogenation in 1-butanol for all Cu-SiO2 nanocomposite catalyst
with Cu loading from 40 to 80%.
C(fructose)
Y(mannitol)
Y(sorbitol)
100
80
60
40
20
0
We also investigated the catalytic activity of Cu(66)-ZnO
(Cu/Zn molar ratio of 2:1) and Cu(32)-Cr2O3 (the reduced form of
2CuO·Cr2O3) for comparison with Cu-SiO2 nanocomposite
catalysts. Cu(32)-Cr2O3 showed much lower mannitol yield of 51%
and sorbitol yield of 3% at 100% fructose conversion than Cu-SiO2.
Interestingly, the carbon balance (excluding oligomeric side
products) over Cu(32)-Cr2O3 was only 54%, which was much
lower than those over Cu-SiO2 (93 to 98%). It was noticeable that
Cu (3.3 wt. %) and Cr (2.9 wt. %) was significantly leached out
from Cu(32)-Cr2O3 catalyst during the reaction. These results
might result from the strong adsorption of fructose on the acidic
surface of Cr2O3,33 where the strongly-adsorbed fructose can be
transformed into the oligomeric products such as humin by
condensation reaction.21
Cu(66)-ZnO gave a slightly lower mannitol yield of 74 % and
higher sorbitol yield of 19% than Cu-SiO2 under equivalent
reaction condition, where a fructose conversion of 100 % was
achieved. However, Cu leaching of 0.1 wt. % was observed after
hydrogenation reaction over this catalyst even in 1-butanol.
Therefore, only Cu-SiO2 nanocomposite was found to be both
catalytically active as well as stable in the liquid-phase fructose
hydrogenation using 1-butanol solvent.
water (6 h)
ethanol (6 h)
1-butanol (6 h) 1-butanol (10 h)
Figure 3. Catalytic activity of the fructose hydrogenation in different
solvents. Reaction conditions: Cu(50)-SiO2, 1.0 g; fructose, 5 g; solvent, 45
g; temperature, 120 °C; H2 pressure, 25 bar; reaction time, 6 h and 10 h. C
= conversion, Y = yield.
In 1-butanol solvent, fructose conversion decreased to 79%
after 6 h, where mannitol and sorbitol yield was 53% and 10%,
respectively. This result indicates that the mannitol selectivity is
higher in butanol than those in water and ethanol solvent, although
the reaction rate is about 20% lower. Interestingly, fructose
conversion reached 98% after 10 h, where mannitol and sorbitol
yield was 70% and 17%, respectively. The higher solubility of
hydrogen in 1-butanol could in principle give an improved reaction
rate32, but 1-butanol had a negative impact on the hydrogenation
rate (fructose conversion). Instead, the mannitol selectivity in 1-
butanol was higher than those in ethanol and water. It seems that
the low fructose solubility affected negatively the hydrogenation
reaction rate, which might result in the selective hydrogenation of
fructose into mannitol without formation of side products.
The catalytic activity of Cu(50)-SiO2 was highly dependent on
the hydrogen pressure (Fig. 4A). As the reaction pressure increased
from 20 to 25 bar, the fructose conversion and mannitol yield
increased significantly from 74% to 98% and from 53% to 70%,
respectively. As the reaction pressure increased from 25 to 50 bar,
the mannitol yield increased almost linearly with the pressure from
70% to 82%, while the fructose conversion and sorbitol yield was
saturated as 100% and 17%, respectively. These results indicate
that a significant amount of fructose is strongly chemisorbed on the
metallic copper surface and this strongly chemisorbed fructose is
selectively hydrogenated to mannitol as hydrogen pressure
increases.
Next, the effect of the fructose/catalyst ratio in 1-butanol on the
catalytic activity of Cu(80)-SiO2 was investigated at 120 C and
35 bar of H2 pressure (Fig. 4B). At a low fructose/catalyst ratio (0.1
to 1.0), fructose was converted completely to mannitol and sorbitol
(carbon balance > 98%). However, as the fructose/catalyst ratio
further increased to 1.5 and 2.0, the fructose conversion was
reduced significantly to 77% and 59%, while the enantioselectivity
of hydrogenation (ratio of mannitol/sorbitol) was not significantly
affected. Since hydrogenation of fructose was catalysed by Cu
metallic site, the decrease in fructose conversion with increasing
fructose/catalyst ratio was ascribed to the lack of Cu metallic site.
120
C(fructose)
Y(mannitol)
Y(sorbitol)
100
80
60
40
20
0
B
A
100
80
60
40
20
0
100
80
60
40
20
0
C(fructose)
Y(mannitol)
Y(sorbitol)
C(fructose)
Y(mannitol)
Y(sorbitol)
1
2
3
4
5
cycle
0
0.5
1
1.5
2
20
25
30
35
40
45
50
fructose/catalyst ratio (in wt.)
reaction pressure (bar)
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001323
ChemSusChem
Figure 5. Catalytic activity of Cu(80)-SiO2 in the fructose hydrogenation.
Reaction conditions: 120 °C, 35 bar H2, 1.0 g catalyst, 5.0 g fructose, 45 g
1-butanol, 10 h. C = conversion, Y = yield.
butanol solvent via density functional theory (DFT) calculations.
We considered dissociative adsorption of hydrogen molecule on
the Cu (111) surface and sought the configuration of adsorbed
fructose on the surface. When an adsorbed hydrogen atom is
detached and bound to the C=O bond of fructose, the product
becomes mannitol or sorbitol by the difference of relative position
of hydrogen atoms (Fig. 7). Thus, it is important to check the
chirality of hydrogenated product, which could be changed
depending on the adsorbed orientation of fructose.
First, solvent molecule (i.e., 1-butanol) stably interacted
through hydrogen bonding with carbonyl group of adsorbed
fructose, where the hydrogenation by the adsorbed hydrogen took
place (Fig. S6). Next, between the copper surface and fructose,
three types of interactions were mainly observed, i.e., interactions
between Cu‧‧‧HO-C, Cu‧‧‧H-C, and Cu-H‧‧‧O=C. For the
comparison of the adsorbed configurations, we identified which
and how many interactions were included for the atom pairs that
closely interacted within 3 Å (Fig. 8A and B).
In a heterogeneous catalytic process, catalyst stability is of
particular importance. We therefore investigated the stability of the
Cu(40)-SiO2 and Cu(80)-SiO2 catalyst at 120 °C and 35 bar over
10 h. After the reaction, the catalyst was recovered by filtration and
washing with water, and dried at 120 °C in oven. The recovered
Cu(40)-SiO2 without reduction showed a lower fructose conversion
of 89% and mannitol yield of 70% than fresh catalyst (Table S1).
The decrease of this catalytic activity was mainly due to formation
of Cu+ phase during recovery process, as confirmed by XRD
analysis (Fig. S3). The recovered catalyst was then reduced at
290 °C prior to commencing the following cycle. As shown in Fig.
5, the recycled Cu(80)-SiO2 catalyst exhibited a consistent
performance over five cycles, which indicated that this partially
Cu+ phase was reduced to Cu metal completely. In addition, no
leaching of the active sites was observed during fructose
hydrogenation in the liquid phase, as confirmed by ICP analysis of
the product mixture. In addition, when the catalyst was filtered out
after 1 h reaction, the fructose conversion and polyol yield was not
changed, which indicates that the fructose hydrogenation was
catalysed only by the heterogeneous Cu (80)-SiO2, not by the
leached Cu metallic component (Fig. S4). Catalyst stability was
also confirmed by examination of the used Cu(80)-SiO2 catalyst by
XRD (Fig. 1B) and TEM (Fig. 2B) after the 5th cycle reaction,
where the agglomeration of Cu particles of Cu(80)-SiO2 was
negligible after reaction. TGA results showed that a loss of 15 wt. %
was observed for the used catalyst after 5th cycle reaction and
drying (curve B in Fig. S5), which could result from the fructose
strongly chemisorbed on the surface. However, deposition of
carbonaceous species on the Cu(80)-SiO2 surface was negligible
after reduction (curve C in Fig. S5).
Sorbitol
Mannitol
formation
formation
Fructose
Figure 7. Orientation-dependency of the hydrogenation reaction of
adsorbed fructose on the Cu-SiO2 surface.
After fructose hydrogenation over Cu(80)-SiO2 catalyst in 1-
butanol, the polyol mixture solution including the catalyst was
cooled to 25 °C (Fig. 6B), at which temperature mannitol
crystallized selectively owing to its negligible solubility, whereas
most sorbitol was still solubilized in 1-butanol. The crystallized
mannitol and catalyst were separated from the sorbitol/1-butanol
mixture by filtration. A solid sample was mixed with water to
dissolve the mannitol crystal and then the catalyst was filtered out
from the aqueous mannitol solution. The purity of the mannitol
sample was identified to be over 99% by HPLC analysis (Fig. 6C).
As far as we know, this is the first report on the production of pure
mannitol from fructose by catalytic hydrogenation and simple
quenching to room temperature without complex separation
technology.
At the adsorbed configuration where the formation of mannitol
was preferred (ISad,m), there were four Cu‧‧‧HO-C, two Cu‧‧‧H-C,
and one Cu-H‧‧‧O=C interactions, while at the fructose
configuration of forming sorbitol (ISad,s), there were one Cu‧‧‧HO-
C, two Cu‧‧‧H-C, and one Cu-H‧‧‧O=C interactions. Due to large
coulombic interaction energies in Cu‧‧‧HO-C interaction compared
to that in Cu‧‧‧H-C interaction, configuration of ISad,m was
thermodynamically favored. Also, in terms of the binding energy
(∆EB.E.), ∆EB.E. of ISad,m was -44.18 kcal/mol, whereas ∆EB.E. of
ISad,s was -34.19 kcal/mol. From this result, selective formation of
mannitol was expected to be done by the preferred orientation of
adsorbed fructose on the copper surface.
Mannitol
(C)
Sorbitol
(B)
Fructose
(A)
10
15
20
25
retention time (min)
Figure 6. HPLC chromatograms of (A) fructose sample; (B) after fructose
hydrogenation; (C) after quenching and filtration
To investigate the selective formation of mannitol by
hydrogenation of fructose using Cu-SiO2 catalyst, we compared the
formation mechanism of mannitol to that of sorbitol under 1-
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001323
ChemSusChem
that of 1-butanol solvent (i.e., ∆∆EB.E.‘s of 5.36, 7.93, and 9.47 for
water, ethanol and butanol conditions, respectively), indicating the
reduced selectivity on mannitol formation at water and ethanol as
shown in the experimental observation.
10
ISad,m
ISad,s
40
DDEB.E.
8
6
4
2
0
35
30
5
0
water
ethanol
1-butanol
Figure 8. Magnified view of the optimized configurations of fructose at the
(A) ISad,m, and (B) ISad,s states. (C) Reaction energy profiles of two fructose
hydrogenation reactions. In the (A) and (B), closely interacting atom pairs
and their interaction types are indicated by dashed lines with red for
Cu‧‧‧HO-C, black for Cu‧‧‧H-C, and blue for Cu-H‧‧‧O=C interactions,
respectively. For the clear view, top layer of Cu surface and adsorbed
hydrogen are shown by space-filling model, while the fructose molecule is
shown by ball-and-stick model. Copper and adsorbed hydrogen are colored
in orange and sky blue, respectively. For fructose molecule, carbon,
hydrogen, and oxygen atoms are colored in gray, white, red, respectively.
Figure 9. Binding energies of mannitol and sorbitol (∆EB.E.) and differences
of binding energies (∆∆EB.E.) in each solvent.
4. CONCLUSION
We developed a promising protocol for the production of
mannitol crystal from fructose using Cu(80)-SiO2 nanocomposite
and 1-butanol solvent for the first time. More importantly, this
catalyst did not show any Cu leaching and its physico-chemical
properties were maintained after the liquid-phase hydrogenation. In
addition, the highly-pure mannitol crystal could be recovered from
the sorbitol-containing 1-butanol solution by simple filtration.
Therefore, the present protocol is a novel and effective method to
produce a pure mannitol from fructose in both an environmental
and an industrial context.
We further investigated the selective formation by tracking
hydrogenation mechanisms for the formations of mannitol and
sorbitol on Cu surface, respectively (Fig. 8C and Fig. S7). By
considering the previous mechanistic study on the hydrogenation
of fructose34, a hydrogen atom adsorbed on the copper surface was
transferred to the carbon of the C=O bond in fructose at the first
step (i.e., from IS to IM). The activation energy (Ea) and heat of
reaction (∆E) for the first hydrogen transfer in mannitol formation
were estimated to be 7.89 and 6.36 kcal/mol, respectively, while in
the sorbitol formation reaction, the energies of 10.39 and 7.20
kcal/mol were required for Ea and ∆E, respectively. After the
hydrogen transfer step, at the IM to FS step, another hydrogen atom
was attached to the oxygen in C(H)-O- of intermediate structure. In
this reaction step, each reaction route for mannitol and sorbitol
formation required marginal amount of Ea and exhibited large
exothermic heat of ∆E (i.e., Ea and ∆E for mannitol pathway were
5.47 and -30.29 kcal/mol, respectively, and 3.24 and -25.97
kcal/mol for sorbitol pathway, respectively). Thus, comparing the
∆EB.E., Ea and ∆E obtained from theoretical calculations, in addition
to the stable initial structure of adsorbed fructose at ISad,m, mannitol
pathway was more thermodynamically favorable than sorbitol
pathway. Theoretically, mannitol was expected to be formed
selectively on the Cu-SiO2 catalyst, which was consistent with
experimental observation.
We conducted further calculations under the other solvent
conditions. In this calculation, we considered the implicit solvent
condition (i.e., COSMO) for estimating the difference between
binding energies (∆∆EB.E.) of ISad,m and ISad,s under each solvent
(Fig. R1). Here, ∆∆EB.E represents the selective formation of
mannitol over sorbitol. The binding energies of ISad,m under water
and ethanol solvent conditions were slightly smaller than that
estimated in 1-butanol (i.e., ∆EB.E.‘s of -37.87, -38.56, and -38.85
for water, ethanol and 1-butanol conditions, respectively). On the
contrary, the binding energies of ISad,s under water and ethanol
were larger than that estimated in 1-butanol (i.e., ∆EB.E.‘s of -32.51,
-30.63, and -29.38 for water, ethanol and 1-butanol conditions,
respectively). Thus, the differences of binding energies between
ISad,m and ISad,s under water and ethanol solvents were smaller than
ACKNOWLEDGEMENTS
This work was supported by the R&D Program of the
Institutional Research Program of the KRICT (SI2011-10) and
Korea Research Fellowship Program funded by the National
Research Foundation of Korea [2017H1D3A1A02053077]. S.K.K.
acknowledges
the
financial
support
from
NRF-
2014R1A5A1009799.
REFERENCES
(1) Geilen, F. M. A.; Engendahl, B.; Harwardt, A.; Marquardt, W.;
Klankermayer, J.; Leitner, W. Selective and Flexible Transformation
of Biomass‐Derived Platform Chemicals by
a Multifunctional
Catalytic System. Angew. Chem. Int. Ed. 2010, 49, 5510-5514.
(2) Huber, G. W.; Iborra, S.; Corma, A. Synthesis of Transportation Fuels
from Biomass:ꢀ Chemistry, Catalysts, and Engineering. Chem. Rev.
2006, 106, 4044-4098.
(3) Corma, A.; Iborra, S.; Velty, A. Chemical Routes for the Transformation
of Biomass into Chemicals. Chem. Rev. 2007, 107, 2411-2502.
(4) Kamm, B. Production of platform chemicals and synthesis gas from
biomass. Angew. Chem., Int. Ed. 2007, 46, 5056–5058.
(5) Bozell, J. J.; Petersen, G. R. Technology development for the
production of biobased products from biorefinery carbohydrates—the
US Department of Energy’s “Top 10” revisited. Green Chem. 2010, 12,
539-554.
(6) Alonso, D. M.; Bond, J. Q.; Dumesic, J. A. Catalytic conversion of
biomass to biofuels. Green Chem. 2010, 12, 1493-1513.
(7) Makkee, M.; Kieboom, A. P. G.; Bekkum, H. Production methods of D-
mannitol. Starch 1985, 37, 136-141.
(8) Dai, Y.; Meng, Q.; Mu, W.; Zhang, T. Recent advances in the
applications and biotechnological production of mannitol. J. Func.
Foods 2017, 36, 404-409.
(9) Alternative Sweeteners – A Global Market Overview, April 2017.
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001323
ChemSusChem
(10) Bhatt, S. M.; Mohan, A.; Srivastava, S. K. Challenges in Enzymatic
Route of Mannitol Production. ISRN Biotechnol. 2013, 914187, 1-13.
(11) Saha B. C.; Racine, F. M. Biotechnological production of mannitol and
its applications. Appl. Microbiol. Biotechnol. 2011, 89, 879-891.
(12) Zhang, J.; Xu, S.; Wu, S.; Liu, Y. Hydrogenation of fructose over
magnetic catalyst derived from hydrotalcite precursor. Chem. Eng. Sci.
2013, 99. 171-176.
(13) Toukoniitty, B.; Kuusisto, J.; Mikkola, J. P.; Salmi, T.; Murzin, D. Y.
Effect of ultrasound on catalytic hydrogenation of D-fructose to D-
mannitol. Ind. Eng. Chem. Res. 2005, 44, 9370-9375.
(14) Liaw, B. J.; Chen, C. H.; Chen, Y. Z. Hydrogenation of fructose over
amorphous nano-catalysts of CoNiB and polymer-stabilized CoNiB.
Chem. Eng. J. 2010, 157, 140-145.
(15) Heinen, A. W.; Peters, J. A.; Bekkum, V. Combined Hydrolysis and
Hydrogenation of Inulin Catalyzed by Bifunctional Ru/C. Carbohydr.
Res. 2001, 330, 381-390.
(16) Crezee, E.; Hoffer, W.; Berger, J.; Makkee, M. Three-phase
Hydrogenation of D-glucose over A carbon Supported Ruthenium
Catalyst—Mass Transfer and Kinetics. Appl. Catal. A 2003, 251, 1-17.
(17) Kuusisto, J.; Mikkola, J.-P.; Casal, P. P´.; Karhu, H.; V¨ayrynen, J.;
Salmi, T. Kinetics of the catalytic hydrogenation of d-fructose over a
CuO-ZnO catalyst. Chem. Eng. J. 2005, 115, 93-102.
(18) Zelin, J.; Meyer, C. I; Regenhardt, S.A.; Sebastianc, V.; Garetto, T.F. ;
Marchi, A. J. Selective liquid-phase hydrogenation of fructose to d-
mannitol over copper-supported metallic nanoparticles. Chem. Eng. J.
2017, 319, 48-56.
(19) Upare, P. P.; Hwang, Y. K.; Lee, J. M.; Hwang, D. W.; Chang, J. S.
Chemical Conversions of Biomass-Derived Platform Chemicals over
Copper-Silica Nanocomposite Catalysts. ChemSusChem, 2015, 8,
2345-2357.
(20) Upare, P. P.; Lee, J.-M.; Hwang, Y. K.; Hwang, D. W.; Lee, J.-H.;
Halligudi, S. B.; Hwang, J. S.; Chang, J. S. Direct Hydrocyclization of
Biomass‐Derived Levulinic Acid to 2‐Methyltetrahydrofuran over
Nanocomposite Copper/Silica Catalysts. ChemSusChem 2011, 4, 1749
-1752.
(21) Upare, P. P.; Hwang, D. W.; Hwang, Y. K.; Lee, U-H.; Hong, D.-Y.;
Chang, J.-S. An integrated process for the production of 2,5-
dimethylfuran from fructose. Green Chem. 2015, 17, 3310-3313.
(22) Upare, P. P.; Hwang, Y. K.; Hwang, D. W. An integrated process for
the production of 2,5-dihydroxymethylfuran and its polymer from
fructose. Green Chem. 2018, 20, 879-885.
(23) Upare, P. P.; Chamas, A.; Lee, J. H.; Kim, J. C.; Kwak, S. K.; Hwang,
Y. K.; Hwang, D. W. Highly Efficient Hydrotalcite/1-Butanol Catalytic
System for the Production of the High-Yield Fructose Crystal from
Glucose. ACS Catal. 2020, 10, 1388-1396.
(24) Garcia V.; Pakkila, J.; Ojamo, H.; Muurinena, E.; Keiski, R. L.
Challenges in biobutanol production: How to improve the efficiency?
Renewable Sustainable Energy Rev., 2011, 5, 964-980.
(25 Delley, B. An All‐electron Numerical Method for Solving the Local
Density Functional for Polyatomic Molecules. J. Chem. Phys. 1990, 92,
508-517.
(26) Delley, B. From Molecules to Solids with the DMol3 Approach. J.
Chem. Phys. 2000, 113, 7756-7764.
(27) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient
Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865-3868.
(28) Tkatchenko, A.; Scheffler, M. Accurate Molecular Van Der Waals
Interactions from Ground-State Electron Density and Free-Atom
Reference Data. Phys. Rev. Lett. 2009, 102, 073005.
(29) Klamt, A.; Schüürmann, G. COSMO: A new approach to dielectric
screening in solvents with explicit expressions for the screening energy
and its gradient. J. Chem. Soc., Perkin Trans. 1993, 2, 799-805.
(30) Bell, S.; Crighton, J. S. Locating transition states. J. Chem. Phys. 1984,
80, 2464-2475.
(31) Halgren, T. A.; Lipscomb, W. N. The synchronous-transit method for
determining reaction pathways and locating molecular transition states.
Chem. Phys. Lett. 1977, 49, 225-232.
(32) M. S. Wainwright, T. Ahn, D. L. Trimm, N. W. Cant, Solubility of
hydrogen in alcohols and esters. J. Chem. Eng. Data 1987, 32, 22-24.
(33) Cheng, W.-H. Development of Methanol Decomposition Catalysts for
Production of H2 and CO. Acc. Chem. Res. 1999, 32, 685-691.
(34) Castoldi, M. C. M.; Câmara, L. D. T.; Monteiro, R. S.; Constantino, A.
M.; Camacho, L.; Carneiro, J. W. M.; Aranda, D. A. G. Experimental
and theoretical studies on glucose hydrogenation to produce sorbitol.
React. Kinet. Catal. Lett. 2007, 91, 341-352.
This article is protected by copyright. All rights reserved.