Catalysis and Stability Effect of Solvent Alcohol on the C6 Aldose Conversion toward Tetrose

Article abstract of DOI:10.1002/cctc.201900094

Conversions of biomass feedstock into various valuable chemicals are of great significance. As a typical route, retro-aldol condensation of monosaccharide greatly expands the variety of biomass-derived platform chemicals via a selective C?C splitting. Herein, we describe a solvent-catalysed strategy to high-selectively accumulate tetrose (four-carbon platform chemical) from C6 aldoses via the retro-aldol/aldol process. We find that alcohol solvents with Lewis acidity facilitate the C?C splitting process of hexose under the catalyst-free condition. The conversion is the fastest in methanol while it is the slowest in isopropanol. The product distribution is greatly influenced by the alcohols through shifting the equilibrium between tetrose and glycolaldehyde (GA). The addition of catalyst only accelerates the reaction rate, and does not change the product distribution. On the one hand, the acetalization of GA with methanol or ethanol shifts the equilibrium from tetrose toward GA, which results in a low yield of tetrose in methanol or ethanol solvent. On the other hand, tetrose can be well accumulated in isopropanol or n-butanol, and the yield of tetrose in isopropanol is higher than in n-butanol because tetrose can be well solvated and stabilized in it. This solvent-dependent reaction strategy provides a new possibility which contributes to the conversion of biomass feedback into valuable platform chemicals and accumulation of target products by utilizing the solvation effect.

Full text of DOI:10.1002/cctc.201900094

Accepted Article
Title: Catalysis and stability effect of solvent alcohol on the C6 aldose
conversion toward tetrose
Authors: wenrong Hou, Yueer Yan, Gang Li, Yulu Zhan, Lei Feng,
Ruohong Zhang, Zhenhua Li, Yahong Zhang, and Yi Tang
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.201900094
Link to VoR: http://dx.doi.org/10.1002/cctc.201900094
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201900094
FULL PAPER
Catalysis and stability effect of solvent alcohol on the C6 aldose
conversion toward tetrose
[
a]
[b]
[a]
[a]
[a]
[c]
[a]
Wenrong Hou , Yueer Yan , Gang Li , Yulu Zhan , Lei Feng , Ruohong Zhang , Zhenhua Li*
[
a]
[a]
Yahong Zhang* and Yi Tang
Abstract: Conversions of biomass feedstock into various valuable
chemicals are of great significance. As a typical route, retro-aldol
condensation of monosaccharide greatly expands the variety of
biomass-derived platform chemicals via a selective C-C splitting.
Herein, we describe a solvent-catalysed strategy to high-selectively
accumulate tetrose (four-carbon platform chemical) from C6 aldoses
via the retro-aldol/aldol process. We find that alcohol solvents with
Lewis acidity facilitate the C-C splitting process of hexose under the
catalyst-free condition. The conversion is the fastest in methanol
while it is the slowest in isopropanol. The product distribution is
greatly influenced by the alcohols through shifting the equilibrium
between tetrose and glycolaldehyde (GA). The addition of catalyst
only accelerates the reaction rate, and does not change the product
distribution. On the one hand, the acetalization of GA with methanol
or ethanol shifts the equilibrium from tetrose toward GA, which
results in a low yield of tetrose in methanol or ethanol solvent. On
the other hand, tetrose can be well accumulated in isopropanol or n-
butanol, and the yield of tetrose in isopropanol is higher than in n-
butanol because tetrose can be well solvated and stabilized in it.
This solvent-dependent reaction strategy provides a new possibility
which contributes to the conversion of biomass feedback into
valuable platform chemicals and accumulation of target products by
utilizing the solvation effect.
(C4) skeleton, such as α-hydroxy-γ-butyrolactone, methyl vinyl
glycolate and methyl-4-methoxy-2-hydroxybutanoate, possess
great potential as building blocks of green solvents,
pharmaceutical precursor and biodegradable polyester
plastics.[4] However, these C4 compounds are difficult to be
obtained from the direct conversion of hexoses, the most
abundant class of sugars with a carbon-six (C6) skeleton.[
They can only be produced in good yields when glycolaldehyde
(C2) or tetrose (C4) is used as substrate.[4d, 5] Obviously, tetrose
with a four-carbon bone, e.g. erythrose, is the root of most of
biomass-derived C4 products. Furthermore, erythrose can be
readily transformed into erythronic acid and erythritol by a simple
oxidation or reduction process. Erythritol can be used in
chemical, cosmetics, food and pharmaceutical industries due to
its very low caloric value and good sweetness. To enable the
generation of C4 compounds from C6 sugars, high-selective
methods for tailoring carbohydrate into C4 skeleton have to be
developed.
2c]
Introduction
Renewable biomass resource is of increasing interest in
academic and industrial regions because it is thought as the only
alternative carbon resource replacing fossil for the supplies of
chemicals and materials that human beings survive in the
future.[1] During the last three decades, extensive studies have
been performed on the conversion of sugars, the most abundant
part of biomass, towards various platform chemicals.[2]
Nowadays, a variety of platform chemicals have been obtained
by catalytic conversion of biomass-derived sugars.[3] Among
these platform chemicals, the compounds with a carbon-four
Scheme 1. Retro-aldol/Aldol condensation route of glucose/mannose as well
as the acetalization of produced GA with alcohol. The numbers shown under
the arrows are the changes of Gibbs free-energy (ΔG) of the two steps
computed by the M06-2x/6-31+G(d,p) method at 190 oC in methanol. The
small ΔG values of the two steps indicate that the retro-aldol process from
hexoses to ERO + GA, and that from ERO to GA are reversible.
[
a]
Ms. Wenrong Hou, Dr. Gang Li, Ms. Yulu Zhan, Mr. Lei Feng, Prof.
Zhenhua Li, Prof. Yahong Zhang, Prof. Yi Tang
Department of Chemistry
Shanghai Key Laboratory of Molecular Catalysis and Innovative
Materials, Laboratory of Advanced Materials, Collaborative
Innovation Centre of Chemistry for Energy Materials (iChEM),
Fudan University
E-mail: zhangyh@fudan.edu.cn, zhenhuali@fudan.edu.cn
Dr. Yueer Yan
Institute for Preservation and Conservation of Chinese Ancient
Books, Fudan University Library, Fudan University
Ms. Ruohong Zhang
Retro-aldol condensation is an important process to enable
the selective C-C bond breaking of hexoses. For example,
fructose can be split into glyceraldehyde (GLA) and
dihydroxyacetone (DHA), two sugars with a carbon-three (C3)
skeleton, via a retro-aldol process.
[2c,
6]
However, in previous
studies, this process was always integrated with a tandem
reaction to convert the as-prepared C3 sugars into downstream
products with higher stability.[7] The retro-aldol process can also
convert C6 aldoses into tetrose and glycolaldehyde (GA), and
GA can be rebuilt into tetrose by an aldol process (Scheme 1).
Unfortunately, because of the negative ΔG from tetrose to GA
(Scheme 1), tetrose is extremely instable compared to GA, and
can hardly be accumulated in such process. It will be
[
[
b]
c]
Shanghai Institute of Quality Inspection and Technical Research
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900094
FULL PAPER
3
Table 1. Conversion of GLU and MAN in various alcohols with or without catalyst (MoO ).
Product yield (Selectiviey)[b]
Product yield (Selectiviey)[b]
(C mol%)
Carbon
balance
(mol%)
Carbon
balance
[c]
Convretro-aldo[a]
(mol%)
(C mol%)
Convretro-aldo[a]
(mol%)
Time
min)
[c]
Solvent
(
GA
Tetrose
GA
Substrate: MAN
28.3(40.5)
Tetrose
(mol%)
Substrate: GLU
15.5(45.2)
40
34.3
47.9
59.4
17.0
31.0
42.7
12.6
21.2
24.5
10.8
22.6
37.1
9.3(27.1)
6.8(14.2)
5.2(8.6)
90.5
79.8
70.1
99.4
88.0
80.8
99.8
99.0
98.3
98.6
92.7
85.4
69.8
88.3
92.6
17.1
34.4
45.6
16.7
27.1
43.4
25.5
36.2
47.5
12.2(17.5)
13.7(15.5)
12.9(13.9)
13.2(77.2)
19.9(57.8)
18.8(41.2)
14.4(86.3)
21.8(80.4)
34.2(78.8)
20.9(82.0)
26.0(71.8)
30.1(63.4)
70.8
63.4
62.7
98.9
90.7
82.0
99.5
98.2
96.3
99.4
96.3
91.1
MeOH
EtOH
IPA
80
60
20.9(43.6)
24.3(40.9)
1.9(11.2)
2.9(9.4)
38.0(43.0)
42.4(45.8)
2.7(15.8)
5.2(15.1)
8.9(19.5)
1.7(10.2)
3.6(13.3)
5.5(12.7)
4.0(12.3)
6.5(17.9)
8.5(14.3)
1
40
14.5(85.3)
16.0(51.6)
17.1(40.0)
11.2(88.9)
17.7(83.5)
20.1(82.0)
7.2(66.7)
11.6(51.3)
15.5(41.8)
80
160
6.4(15.0)
1.6(12.7)
2.5(11.8)
2.8(11.4)
2.1(19.4)
3.8(16.8)
6.9(18.6)
40
80
160
40
n-BA
80
160
Substrate: GLU (MoO
3
)
3
Substrate: MAN (MoO )
0
.03
50.2
51.3
53.9
75.6
83.0
87.0
38.8
55.1
58.2
31.7
62.8
87.7
15.5(30.9)
4.3(8.6)
4.3(8.4)
69.6
69.4
65.7
63.6
56.2
46.9
92.7
90.0
85.8
85.6
67.5
38.9
66.8
81.7
86.5
32.1
39.2
48.7
34.4
49.0
61.4
38.1
59.9
65.5
28.5(42.7)
37.3(45.7)
37.5(43.4)
4.3(13.4)
7.3(18.6)
10.2(20.9)
5.9(17.2)
12.3(25.1)
16.3(26.5)
3.2(8.4)
13.3(19.9)
15.0(18.4)
14.0(16.2)
26.4(82.3)
30.0(76.5)
33.4(68.6)
27.9(81.1)
34.2(69.8)
35.9(58.5)
18.9(49.6)
23.6(39.4)
23.5(35.9)
75.0
70.6
65.0
98.7
98.1
94.9
99.4
97.4
90.8
84.0
71.6
66.3
MeOH
EtOH
IPA
0.05
16.2(31.6)
16.3(30.2)
18.9(25.0)
19.1(23.0)
17.9(20.6)
3.9(10.1)
8.0(14.5)
10.2(17.5)
2.6(8.2)
0.07
3.2(5.6)
2
20.4(27.0)
20.1(24.2)
15.9(18.3)
27.6(71.1)
37.0(67.2)
33.8(58.1)
14.7(46.4)
22.1(35.2)
17.7(20.2)
3
4
5
15
2
5
5
n-BA
15
8.2(13.1)
8.9(10.5)
7.9(13.2)
8.3(12.0)
2
5
Reaction conditions: 100 mg of mannose was introduced into 5 g of alcohol solvent and reaction was conducted at 190 °C for desired time. [a] Convretro-aldol was
defined as consumed mannose for only getting through retro-aldol condensation, which calculated by Convretro-aldol = Conv.-YC6. [b] Yields of products were
calculated by carbon atom yield from original substrate. [c] Carbon balance = [(1-Conv.)+Y.C6+Y.C2+Y.C4].
sequentially transformed into GA or other undesirable products
like tars.[8] Moreover, the coexistence of other side reaction such
as isomerization process among hexoses can further decrease
its selectivity and yield.[9] As a result, developing a proper
strategy to accumulate tetrose is prerequisite for C4 platform
chemicals. Recently, our group developed a borate-assistant
retro-aldol/aldol strategy to successfully accumulate erythrose
from C6 aldoses in methanol.[10] However, a further improvement
of tetrose yield as well as carbon atom utilization is still required.
Previous studies have demonstrated that the use of non-
aqueous polar solvents can greatly accelerate the reaction rate
and improve the product selectivity during the conversion of
biomass.[3b, 11] For example, the use of ionic liquid can lead to a
higher 5-hydroxymethylfurfural or sugar yield from cellulose than
in the aqueous phase.[11e-g] The use of γ-valerolactone as
solvent or cosolvent can significantly improve the reaction rate
and product selectivity of biomass conversion compared to its
conversion in the aqueous media.[
11h-k]
Solvents, in certain cases,
can even have a greater influence on the reaction selectivity
than catalysts.[11l] Herein, the solvent effects on the reaction
process and product distribution of the conversion of C6 sugars
have been investigated by studying the conversion of two C6
aldoses, glucose (GLU) and mannose (MAN), in non-aqueous
polar solvents, and to be specific, in four alcohols: methanol
(MeOH), ethanol (EtOH), isopropanol (IPA) and n-butanol (n-BA).
By doing so, a high-selective strategy to accumulate tetrose
from C6 aldoses via retro-aldol processes without extra catalyst
is proposed. A tetrose yield of up to 37% can be obtained in IPA.
Experimental and theoretical evidences show that the Lewis
acidity of the solvent alcohol is responsible for the retro-aldol
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900094
FULL PAPER
conversion rate and solvation effect is responsible for the high
tetrose yield in IPA.
In MeOH, the main product of GLU/MAN conversion is GA.
However, in the other three alcohols, tetrose is allowed to be the
major product, and the highest yield of tetrose is achieved in IPA
(
3
Table 1 and Figure S1). Moreover, when MoO catalyst is used,
a similar trend in product distribution is also observed. Although
catalyst accelerates the reaction rate, it does not change the
product distribution, i.e. tetrose yield is still the lowest in MeOH
and the highest in IPA. This result strongly suggests that the
product distribution is mainly determined by thermodynamics
rather than kinetics during the retro-aldol process of C6 aldoses.
The reaction routes of C6 aldose are summarized as follows. In
the first step, C6 aldose splits into ERO and GA (Scheme 1).
Subsequently, ERO breaks down into two GA molecules.
Through an aldol process, GA produced from the above two
steps can be reversibly reconverted into ERO. Therefore, the
product distribution is subtly controlled by the equilibrium
between ERO and GA. In MeOH system at 160 min, only a
small amount of tetrose could be detected, while the yields of
GA of GLU and MAN are 24.3% and 42.4%, respectively. The
results indicate that the equilibrium between ERO and GA favors
the generation of GA in MeOH. In contrast, the production of GA
is efficiently suppressed in IPA and the selectivity of tetrose
increases prominently. In IPA system, tetrose selectivity is
always higher than 80.0% while GA selectivity is just about
Results and Discussion
Conversion of GLU and MAN in four alcohols
Table 1 lists the conversions and yields of the retro-aldol
condensation of GLU/MAN in four alcohols without/with catalyst.
It should be noted that the yield of GA for the reaction in MeOH
is the total yield of GA and its acetalization products (see below).
It is interesting that the retro-aldol process of both GLU and
MAN could be carried out in four alcohols even without extra
catalyst. This is unexpected because previous studies on the
retro-aldol condensation of sugars have been all performed with
the catalysts.[6-7] Moreover, the conversion rate of GLU/MAN
greatly depends on the types of alcohols. For both substrates,
the conversion rate in MeOH is the highest while that in IPA is
the lowest, and those in EtOH and n-BA do not differ much
(
Table 1). Why can the retro-aldol process of GLU/MAN occur in
alcohol solvents and why do the conversion rates have such a
trend in the four solvents? Previous studies exhibited that the
retro-aldol process of sugar could be catalyzed by Lewis
acids.[7a, 9, 12] It is thus rational to conceive that the Lewis acidic
10.0% (Table 1), implying an effective accumulation of tetrose.
solvents can trigger the retro-aldol conversion of GLU/MAN as
Furthermore, as shown in Scheme 1, GLU or MAN can be
converted into GA and ERO with a carbon yield ratio of 0.5 via a
retro-aldol process. ERO can not only split into GA by a
secondary retro-aldol process but also isomerize and epimerize
into ERU and THO. Instead, the produced GA can rebuild into
tetrose by aldol process. As shown in Table 1, the GA/tetrose
well. One of the quantity to characterize the Lewis acidity of a
solvent is the δ value,[13] which represents the difference
between the donor and acceptor numbers of electrons of the
solvent. The higher the δ value is and the lower the Lewis acidity
[13]
is. The δ values of the four solvents have a decreasing order
of IPA (25) > EtOH (-51) > n-BA (-78) > MeOH (-113).[13b] This
(
0
carbon yield) ratios in EtOH, IPA and n-BA are much lower than
.5 in the absence/presence of catalyst, further implying the
implies an increasing order of their Lewis acidities of MeOH > n-
BA > EtOH > IPA. The highest conversion rate in MeOH and the
lowest conversion rate in IPA correspond to the δ values of
MeOH and IPA (-113 and 25, respectively). The conversion
rates of GLU/MAN at 160 min in EtOH and n-BA are similar
because the δ values of EtOH and n-BA are close to each other
(
-51 and -78, respectively). The Lewis acidity and conversion
rate order can be further confirmed by the retro-aldol conversion
of fructose. The conversion rate of fructose in four alcohols
completely follows the order of MeOH > n-BA > EtOH > IPA
(
Table S1).
3
For comparison, MoO solid catalyst with Lewis acidity was
used as a heterogeneous catalyst to perform the retro-aldol
process of GLU/MAN in various alcohols.[12b] The results clearly
show that the addition of MoO
3
greatly accelerates the
conversion rate. For example, when GLU is used as substrate, a
o
tetrose yield of 37.0% can be reached at 190 C in IPA within
just 15 min. However, the acceleration effect of MoO
3
is different
in four solvents, and MoO accelerates the conversion of
3
GLU/MAN more deeply in MeOH and EtOH than in IPA and n-
BA. In MeOH, the conversion of both GLU and MAN can be over
5
0% within 2 s in the presence of MoO
3
. Obviously, except for
3
, other factors need to
the Lewis acidities of solvents and MoO
be considered.
Figure 1. MS results of reaction solution of GLU conversion in IPA (a) and
Besides the conversion rate, the type of solvent also displays
an influence on the product distribution of the retro-aldol process.
o
MeOH (b) at 190 C for 160 min
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900094
FULL PAPER
proceeding of aldol condensation of GA toward tetrose in these
solvents. Therefore, the equilibrium between ERO and GA
favors the generation of ERO in the three alcohols.
presence of catalyst, leading to a much faster conversion rate of
the aldoses in the two solvents than those in the other solvents
(Table 1).
The resultant reaction solutions of GLU in MeOH and IPA
were directly tested by mass spectrometry (MS). A signal at m/z
Table 2. Conversion and product distribution of GA in four alcohol solvents.
1
19, corresponding to the theoretical m/z value of tetrose under
negative mode, is indeed observed for the reaction in IPA
Figure 1a). Meanwhile, no GA signal can be detected. For the
Product yield
(C mol%)
Acetal
Carbon
[a]
Time
min)
Conv.
(mol %)
product
balance[
b]
Solvent
(
[c]
(
yield
Y.C4
Y.C6
(C mol %)
reaction in MeOH, no GA signal is detected either, but two
signals at m/z 105 and 119 can be clearly observed (Figure 1b).
Taking the result of HPLC (Figure S1a) that showed a very low
tetrose content into account, the signal at 119 cannot be
attributed to tetrose. More reasonably, the two m/z signals can
be assigned to the acetal products of GA with MeOH,
corresponding to glycolaldehyde dimethyl acetal (m=106) and
methanol-etherified glycolaldehyde dimethyl acetal (m=120),
respectively. The acetal products are further confirmed by GC
analysis of reaction solution (Figure S2a, b). Additionally, MS
(
C mol %)
4
0
5.5
2.7
1.0
0.2
2.4
2.4
1.4
3.0
6.3
7.2
2.5
5.1
8.7
2.3
4.9
5.2
97.9
99.9
96.4
98.9
90.0
77.1
98.7
92.1
91.0
95.5
90.9
75.8
82.1
MeOH
EtOH
IPA
80
60
94.4
1
2.2
0.1
98.9
40
80
41.2
49.9
60.9
28.2
47.4
55.9
32.5
50.6
59.0
37.1
33.6
26.3
24.4
34.4
38.3
25.7
36.6
29.6
3.6
10.5
1
60
20.5
40
80
-
-
-
-
-
-
o
analysis of ERO solution in MeOH heated at 190 C for 80 min
further validates this attribution: two predominant signals at m/z
1
60
105 and 119 are also observed (Figure S3a). These results
4
0
indicate that, in MeOH solvent, the produced GA during the
retro-aldol of GLU/ERO can further react with MeOH to form
acetal products, so as to shift the equilibrium between ERO and
GA toward GA. As a result, the yield of ERO is very low in
MeOH. On the contrary, the MS result of ERO solution in IPA
n-BA
80
60
1
Reaction conditions: 100 mg of GA, 5 g of alcohol, 190 oC. [a] Product yields
were calculated by carbon atom yield from original substrate. [b] Carbon
balance = [(1-Conv.)+Y.C6+Y.C4]. [c] Acetal product yields were determined by
GC.
o
heated at 190 C for 80 min shows the only signals of ERO and
its dimer (Figure S3b). This implies that the retro-aldol process
of ERO toward GA can be suppressed and the equilibrium
between ERO and GA favors ERO in IPA. This explains the high
yield of tetrose in IPA.
However, as showed in Table 2, GA can be converted into
tetrose with high selectivity in IPA and n-BA, and the yields of
tetrose are 38.3% and 29.9% at 160 min in IPA and n-BA,
severally. The MS (Figure S4c), HPLC (Figure S4d) and 1H-
NMR results (Figure S7) of its reaction solution in IPA show that
no acetal products of GA can be formed. Since no acetalization
process occurs to shift the equilibrium between tetrose and GA,
tetrose can be well accumulated in IPA. Moreover, it is worthy to
notice that a small amount of hexose could be detected (the
column Y.C6, Table 2) in all four solvents. This means that
although the retro-aldol splitting of hexose to tetrose + GA is
reversible, the equilibrium between hexose and tetrose + GA
favors more deeply to tetrose + GA in all four solvents. Hence,
the product distribution is mainly controlled by the equilibrium
between tetrose and GA. The equilibrium between tetrose and
GA favors tetrose more in IPA than in other solvents, and the
tetrose yield is therefore the highest in IPA.
Aldol condensation of GA in four alcohols
To further confirm the above assumption that the equilibrium
between the ERO and GA controls the product distribution, we
conducted the conversion of GA in the four alcohols under the
same reaction condition, and the results are listed in Table 2.
Without catalyst, the aldol process of GA toward tetrose and
hexose is also realized in all the four alcohol solvents. Moreover,
the product distribution is greatly affected by solvent. At 160 min,
the tetrose yield is just 0.1% in MeOH. Obviously, GA is inclined
to produce acetal products (> 90%, Table 2, Figure S2c), as
confirmed by the MS (Figure S4a) and HPLC (Figure S4b)
1
results of its reaction solution. Furthermore, the H-NMR results
of GA in MeOH before/after thermal treatment show that the
aldehyde H-1 resonated at 9.64 ppm disappears after thermal
treatment (Figure S5), further implying the acetal formation of
aldehyde in GA. Meanwhile, the acetal product is also detected
Theoretical study
So far, we can conclude that solvent alcohols with Lewis acidity
can directly perform the retro-aldol process of C6 sugars, and
different alcohols display different reaction activities. To further
understand why C6 sugars shows the highest conversion rate in
MeOH while the lowest one in IPA, we computed the free-
energy barriers of the retro-aldol process of GLU in all the four
solvents using the M06-2x/6-31+G(d,p) method with the implicit
solvation model. The free-energy barriers of the retro-aldol
(
Table 2 and Figure S6) in EtOH, although its yield is low and
the major product is still tetrose at 160 min. The decreasing of
the tetrose yield with time shows that the acetalization reaction
of GA in EtOH is slower than the aldol process of GA to tetrose.
Therefore, a part of tetrose initially produced from the fast aldol
process of GA is slowly converted back to GA through the retro-
aldol process with the acetalization of GA. Also, the acetalization
of GA in MeOH/EtOH promotes the conversion of aldose at the
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900094
FULL PAPER
Table 3. Molecular configurations of ERO, eight-hydrogen-bonded alcohol molecule and ERO-eight-hydrogen-bonded alcohol molecule and the solvation free
energies of ERO at 298.15 K in IPA and n-BA.
[
a]
Erythrose(gas) + (ROH) (SMD) → erythrose ∙ (ROH) (SMD)
8
8
Solvation free energy
kJ/mol]
ERO
(ROH)
8
ERO(ROH)
8
[
-
95.5
-
87.0
[
a] The “gas” and “SMD” in the parenthesis indicate that their energies were computed in gas phase and in solution using the SMD solvation model,
respectively.
o
process from GLU to ERO + GA at 190 C are 179.6，179.5,
80.3, 179.2 kJ/mol in the MeOH, EtOH, IPA and n-BA,
rate of hexose in EtOH is slightly faster than that in n-BA
although they possess similar free-energies barriers of the retro-
1
respectively. The free-energy barrier in IPA is indeed the highest, aldol process (179.5 vs 179.2 kJ/mol) and δ values (-51 vs -78).
implying the retro-aldol process is slowest in IPA. However,
those values for the other three solvents are very close to each
Also, because of the absence of acetalation of tritose in
MeOH/EtOH[15], the conversion rate of fructose in four alcohols
other and cannot explain why the reaction in MeOH is the fastest. completely depends on their δ values, i.e., MeOH > n-BA >
We speculate that the solvation model used may cause the error
of the free-energy barrier. However, the results do confirm the
experimental observation that the reaction rates do not differ
much in the four solvents. For example, for GLU substrate, the
highest conversion rate (in MeOH) is at most three times of the
lowest one (in n-BA) in Table 1, which only corresponds to a
free-energy barrier difference of just 4 kJ/mol, qualitatively in
well agreement with the closeness of the four free-energies
barriers computed by the M06-2x method. Moreover, we note
that the mean absolute error of the M06-2x method in computing
energy barriers is beyond 4 kJ/mol.[14] Therefore, it is reasonable
that such small differences in reaction rates may not be
quantitatively reproduced by the M06-2x method. Herein, we
assign the faster conversion of hexoses in MeOH than in EtOH
and n-BA to the fast acetalation of GA in it, which greatly
accelerates the conversion rate of GLU/MAN/ERO toward GA in
it. Thanks to the slow acetalation of GA in EtOH, the conversion
EtOH > IPA (Table S1).
Moreover, the experimental results clearly indicate that the
acetalization of GA in MeOH and EtOH is the major factor
influencing their product distribution because it shifts the
equilibrium between tetrose and GA toward GA, leading to a low
tetrose yield in the two solvents. However, it cannot explain why
the tetrose yield is higher in IPA. To explain this, a plausible
hypothesis is proposed. ERO is solvated better in IPA so that
the equilibrium between ERO and GA favors ERO more deeply
in IPA. Furthermore, it is worth reminding that a tetrose molecule
has three hydroxyl groups and one carbonyl group, thus the
hydrogen bonds between them and surrounding alcohol solvent
molecules have to be taken into consideration. Therefore, only
using an implicit solvation model which treats the solvents as
polarizable continuum to compute the solvation energy of
erythrose may not be reliable. We computed the solvation free-
energies of ERO in the IPA and n-BA solvents with a more
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900094
FULL PAPER
accurate explicit-implicit model as described in the
computational method section (the molecular configurations of
the models are presented in Table 3). The computed solvation
free-energy of ERO is now much more negative in IPA than in n-
BA (-95.5 vs. -87.0 kJ/mol, Table 3). The results indicate that
ERO is solvated much better in IPA than in n-BA. Therefore, the
equilibrium between ERO and GA favors ERO more deeply in
IPA than in n-BA.
100 mg of substrate and 5 g of alcohol were added into the
reaction tube, and then the glass tube containing the reactants
was put into a microwave instrument with a magnetic stirrer. The
reaction temperature was controlled at 190 °C for 1 s – 160 min.
Upon the end of the reaction, the tube was cooled to room
temperature by compressed air as fast as it can be. After that
the composition of mixture was sampled and then analyzed by
high-performance liquid chromatography (HPLC, Schimadzu).
Product analysis
Conclusions
Conversions of the substrates and yields of the products were
determined by the Schimadzu HPLC (RI detector) equipped with
a Shodex SC1011 (8.0 mm × 300 mm) sugar column maintained
at 80 °C, with deionized water as the mobile phase with a flow
rate of 1.0 mL/min. Meanwhile, THO was determined by Aminex
The conversion of hexoses (GLU and MAN) in four alcohol
solvents (MeOH, EtOH, IPA and n-BA) has been systematically
studied. We find that the retro-aldol condensation of hexoses
proceeds in the four alcohols without extra catalyst, and the
alcohol solvent with higher Lewis acidity displays faster
conversion rate. Moreover, the main product is GA in MeOH
whereas tetrose in the other three alcohols. A tetrose yield up to
HPX-87H (7.8 mm × 300 mm), with H SO4 solution of 8.0
2
mmol/L as the mobile phase with a flow rate of 1.0 mL/min.
Quantities of the substances were determined based on their
own calibration curves by using the external standard method.
The yield of tetrose was determined by the total yields of ERO,
ERU and THO. Mass spectrometry (Bruker McriOTOF11) under
negative ions model and 1H nuclear magnetic resonance (1H
NMR, Brucker AVANCE III HD, 400 MHz) were used to confirm
the products and the reaction route.
37% could be obtained in IPA. The addition of catalyst can only
accelerate the reaction rate but it cannot change the product
distribution. Experimental evidences and theoretical calculations
suggest that the retro-aldol process of the hexose and tetrose in
the four solvents are reversible and the product distribution is
subtly controlled by the equilibrium between tetrose and GA. In
MeOH and EtOH solvents, the acetalization of GA shifts the
equilibrium between tetrose and GA toward GA, resulting in a
low tetrose yield. In addition, higher tetrose yields in IPA and n-
BA solvents are achieved because of no acetalization of GA.
Moreover, tetrose can be better solvated in IPA than in n-BA so
that the equilibrium between tetrose and GA favors tetrose,
which brings a higher tetrose yield in IPA than that in n-BA. This
solvent alcohol-stabilized reaction strategy provides a new
glimpse on the accumulation of the instable products via a
solvation effect. Moreover, this catalyst-free reaction system
also favors their further conversion in a cascade process. Future
works will focus on the one-pot conversion of tetrose in alcohol
towards various C4 skeleton compounds, such as subsequent
transfer hydrogenation, hydrogenation and oxidation reactions.
Computational details
To investigate the solvation effect on the reaction, first principle
density functional theory (DFT) calculations were performed with
the M06-2X/6-31+G(d,p) method[14, 16] and the IEFPCM solvation
model[17] with radii and non-electrostatic terms for Truhlar and
coworkers’ SMD solvation model.[18] To investigate the reaction
mechanism in the four solvents, the solvation free energies of
the species involved in the reaction were computed with an
implicit solvation model in which the interaction between the
solute molecule and solvent is treated as the interaction with
polarizable continuum solvent media. Furthermore, to accurately
compute the solvation energy of ERO in the alcohol solvents
where the hydrogen bonding interaction between ERO and the
surrounding alcohol molecules cannot be well described by an
implicit solvation model, we used an explicit + implicit solvation
model. To be specific, taking the structural characters of ERO
and solvents into account, the solvation free-energy of ERO in
an alcohol solvent was computed as the free-energy change at
Experimental Section
Materials
298.15 K for the following reaction:
Mannose (MAN, 99%), glucose (GLU, 99.5%), fructose (FRU,
ERO(gas) + (ROH) (SMD) → ERO ∙ (ROH) (SMD)
8
8
9
(
9%), glycolaldehyde dimer (GA, 95%), 1,3-dihydroxyacetone
DHA), glyceraldehyde (GLA), erythrose (ERO, 75%) and
erythrulose (ERU, 85%) were purchased from Sigma-Aldrich.
Threose (THO, 60%) was purchased from J&K. MoO , methanol
MeOH), ethanol (EtOH), and n-butanol (n-BA) were purchased
from Sinopharm Chemical Reagent Co., Ltd. Isopropanol (IPA,
9.9%) were purchased from Macklin.
where the “gas” and “SMD” in the parenthesis indicate that the
energy of that species is computed in gas phase and in the
solution using the implicit SMD solvation model, respectively.
3
8 8
The (ROH) and ERO∙(ROH) represent an eight hydrogen-
(
bonded alcohol molecular cluster and a hydrogen-bonded
complex formed between ERO and eight alcohol molecules,
respectively. Default geometry convergence criterion of
Gaussian 09 were used. Harmonic vibrational frequency
analyses were performed on all the species at the M06-2X/6-
9
Experimental procedure
3
1+G(d,p) level to verify if the optimized geometry was a
Retro-aldol condensation of C6 sugars and aldol condensation
of GA were both carried out in a thick-wall pressure glass tube.
genuine minima or a transition state and to compute zero-point
vibrational energies and temperature dependent thermodynamic
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900094
FULL PAPER
quantities. All calculations were performed with the Gaussian 09
[7]
a) M. S. Holm, S. Saravanamurugan, E. Taarning, Science 2010,
quantum software package.[19]
328, 602-605; bJ. Zhang, X. Liu, M. Sun, X. Ma, Y. Han, ACS
Catal. 2012, 2, 1698-1702; c) Z. Tang, W. Deng, Y. Wang, E. Zhu,
X. Wan, Q. Zhang, Y. Wang, ChemSusChem 2014, 7, 1557-1567.
a) A. A. Peterson, F. Vogel, R. P. Lachance, M. Fröling, J. M. J.
Antal, J. W. Tester, Energy Environ. Sci. 2008, 1, 32-65; b) B. M.
Kabyemela, T. Adschiri, R. M. Malaluan, K. Arai, Ind. Eng. Chem.
Res. 1997, 36, 1552-1558; c) M. Sasaki, B. Kabyemela, R.
Malaluan, S. Hirose, N. Takeda, T. Adschiri, K. Arai, J. Supercrit.
Fluid. 1998, 13, 261-268.
[
8]
Acknowledgements
This work was supported by the 973 Program (No.
2
018YFA0209402), NSFC (Nos. 21673045, 21573044,
1433002, 21805042) and State Administration for Market
2
[
[
[
9]
S. Lin, X. Guo, K. Qin, L. Feng, Y. Zhang, Y. Tang,
ChemCatChem 2017, 9, 4179-4184.
Regulation Research Program (No. 2017QK084).
10] Y. Yan, L. Feng, G. Li, S. Lin, Z. Sun, Y. Zhang, Y. Tang, ACS
Catalysis 2017, 7, 4473-4478.
Keywords: biomass • tetrose • acetalation • solvent-catalysed •
11] a) C. B. Rasrendra, J. N. M. Soetedjo, I. G. B. N. Makertihartha, S.
Adisasmito, H. J. Heeres, Top. Catal. 2012, 55, 543-549; b) Y.
Wang, H. Wang, H. Lin, Y. Zheng, J. Zhao, A. Pelletier, K. Li,
Biomass Bioenergy 2013, 59, 158-167; c) M. A. Mellmer, D.
Martin Alonso, J. S. Luterbacher, J. M. R. Gallo, J. A. Dumesic,
Green Chem. 2014, 16, 4659-4662; d) P. Weerachanchai, S. K.
Kwak, J.-M. Lee, Bioresour. Technol. 2014, 170, 160-166; e) Y.
Román-Leshkov, J. N. Chheda, J. A. Dumesic, Science 2006, 312,
solvation energy
[
1] a) G. Berndes, M. Hoogwijk, R. van den Broek, Biomass
Bioenergy 2003, 25, 1-28; b) A. J. Ragauskas, C. K. Williams, B.
H. Davison, G. Britovsek, J. Cairney, C. A. Eckert, W. J. Frederick,
J. P. Hallett, D. J. Leak, C. L. Liotta, J. R. Mielenz, R. Murphy, R.
Templer, T. Tschaplinski, Science 2006, 311, 484-489; c) N.
Mosier, C. Wyman, B. Dale, R. Elander, Y. Y. Lee, M. Holtzapple,
M. Ladisch, Bioresour. Technol. 2005, 96, 673-686.
1
933-1937; f) E. I. Gürbüz, S. G. Wettstein, J. A. Dumesic,
ChemSusChem 2012, 5, 383-387; g) Y. J. Pagán-Torres, T. Wang,
J. M. R. Gallo, B. H. Shanks, J. A. Dumesic, ACS Catal. 2012, 2,
[
[
2]
a) A. Osatiashtiani, A. F. Lee, D. R. Brown, J. A. Melero, G.
Morales, K. Wilson, Catal. Sci. Technol. 2014, 4, 333-342; b) I. T.
Horváth, H. Mehdi, V. Fábos, L. Boda, L. T. Mika, Green Chem.
9
30-934; h) L. Qi, I. T. Horváth, ACS Catal. 2012, 2, 2247-2249; i)
D. M. Alonso, J. M. R. Gallo, M. A. Mellmer, S. G. Wettstein, J. A.
Dumesic, Catal. Sci. Technol. 2013, 3, 927-931; j) E. I. Gürbüz, J.
M. R. Gallo, D. M. Alonso, S. G. Wettstein, W. Y. Lim, J. A.
Dumesic, Angew. Chem. Int. Ed. 2013, 52, 1270-1274; k) L. Qi, Y.
F. Mui, S. W. Lo, M. Y. Lui, G. R. Akien, I. T. Horváth, ACS Catal.
2
008, 10, 238-242; c) M. Sasaki, K. Goto, K. Tajima, T. Adschiri, K.
Arai, Green Chem. 2002, 4, 285-287.
3]
a) A. Corma, S. Iborra, A. Velty, Chem. Rev. 2007, 107, 2411-
2
502; b) H. Zhao, J. E. Holladay, H. Brown, Z. C. Zhang, Science
007 316, 1597-1600; c) P. L. Dhepe, A. Fukuoka,
1, 969-975; d) R. Rinaldi, F. Schüth,
2
,
2
2
014, 4, 1470-1477; l) L. Shuai, J. Luterbacher, ChemSusChem
016, 9, 133-155.
ChemSusChem 2008
,
Energy Environ. Sci. 2009, 2, 610-626; e) D. M. Alonso, S. G.
Wettstein, J. A. Dumesic, Chem. Soc. Rev. 2012, 41, 8075-8098; f)
A. Wang, T. Zhang, Acc. Chem. Res. 2013, 46, 1377-1386.
[
[
[
12] a) Y. Wang, W. Deng, B. Wang, Q. Zhang, X. Wan, Z. Tang, Y.
Wang, C. Zhu, Z. Cao, G. Wang, H. Wan, Nat. Commun. 2013, 4,
2
1
141; b) M. Orazov, M. E. Davis, P. Natl. Acad. Sci. 2015, 112,
[
4]
a) S. Yamaguchi, K. Motokura, Y. Sakamoto, A. Miyaji, T. Baba,
Chem. Commun. 2014, 50, 4600-4602; b) S. Van de Vyver, Y.
Román-Leshkov, Angew. Chem. Int. Ed. 2015, 54, 12554-12561;
c) A. Sølvhøj, E. Taarning, R. Madsen, Green Chem. 2016, 18,
1777-11782.
13] a) V. Gutmann, The donor-Acceptor Approach to Molecular
Interactions, Plenum press, New York, 1978; b) Y. Macurs, Chem.
Soc. Rev. 1993, 22, 409-416; c) M. Chatterjee, T. Ishizaka, H.
Kawanami, Green Chem. 2014, 16, 4734-4739.
5
448-5455; d) M. Dusselier, P. Van Wouwe, F. de Clippel, J.
Dijkmans, D. W. Gammon, B. F. Sels, ChemCatChem 2013, 5,
69-575.
14] Y. Zhao, D. G. Truhlar, J. Chem. Phys. 2006, 124, 224105.
5
[
15] L. Feng, G. Li, Y. Yan, W. Hou, Y. Zhang, Y. Tang, RSC Adv. 2018, 8,
0163-30170.
[
[
5]
6]
a) R. De Clercq, M. Dusselier, C. Christiaens, J. Dijkmans, R. I.
Iacobescu, Y. Pontikes, B. F. Sels, ACS Catal. 2015, 5, 5803-
3
[
16] R. Krishnan, J. S. Binkley, R. Seeger, J. A. Pople, J. Chem. Phys.
1980, 72, 650-654.
5
811; b) M. Dusselier, P. Van Wouwe, S. De Smet, R. De Clercq,
L. Verbelen, P. Van Puyvelde, F. E. Du Prez, B. F. Sels, ACS
Catal. 2013, 3, 1786-1800.
[
[
17] M. J. F. G. Scalmani, J. Chem. Phys. 2010, 132, 110-114.
18] A. V. Marenich, C. J. Cramer, D. G. Truhlar, J. Phys. Chem. B
a) T. M. Aida, K. Tajima, M. Watanabe, Y. Saito, K. Kuroda, T.
Nonaka, H. Hattori, R. L. Smith, K. Arai, J. Supercrit. Fluid. 2007
2, 110-119; b) S. Matsuoka, H. Kawamoto, S. Saka, J. Anal. Appl.
Pyrolysis 2012, 93, 24-32.
,
2
009, 113, 6378-6396.
19] R. A. Gaussian 09, Frisch, M. J. et al. Gaussian, Inc., Wallingford
CT, 2009
4
[
.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900094
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Methanol
Conv.retro-aldol 59.4%
n-Butyl alcohol
37.1%
Ethanol
42.7%
Isopropanol Decreasing
24.5%
Lewis acidity
A catalyst-free strategy was
proposed to accumulate tetrose
from C6 aldoses where the
Lewis acidic alcohol solvent
performs the reaction and
tetrose can be effectively
accumulated due to the
stabilization effect of the
alcohol solvent.
Wenrong Hou, Yueer Yan, Gang
Li, Yulu Zhan, Lei Feng,
Ruohong Zhang, Zhenhua Li*
Yahong Zhang* and Yi Tang
Selectivity /%
Solvation energy between
ERO and alcohol( kJ/mol )
-87.0
-95.5
90
Page No. – Page No.
60
30
0
C4
C2
Catalysis and stability effect of
solvent alcohol on the C6
aldose conversion toward
tetrose
Methanol
n-Butyl alcohol
Ethanol
Isopropanol
This article is protected by copyright. All rights reserved.