Mechanistic Insight into a Sugar-Accelerated Tin-Catalyzed Cascade Synthesis of α-Hydroxy-γ-butyrolactone from Formaldehyde

Article abstract of DOI:10.1002/cssc.201500885

Applications of the formose reaction, which involves the formation of sugars from formaldehyde, have previously been confined to the selective synthesis of unprotected sugars. Herein, it is demonstrated that α-hydroxy-γ-butyrolactone (HBL), which is one of the most important intermediates in pharmaceutical syntheses, can be produced from paraformaldehyde. In the developed reaction system, homogeneous tin chloride exhibits high catalytic activity and the addition of mono- and disaccharides accelerates the formation of HBL. These observations suggest that the formose reaction may serve as a feasible pathway for the synthesis of important chemicals.

Full text of DOI:10.1002/cssc.201500885

DOI: 10.1002/cssc.201500885
Full Papers
Mechanistic Insight into a Sugar-Accelerated Tin-Catalyzed
Cascade Synthesis of a-Hydroxy-g-butyrolactone from
Formaldehyde
Sho Yamaguchi,* Takeaki Matsuo, Ken Motokura, Yasuharu Sakamoto, Akimitsu Miyaji, and
[a]
Toshihide Baba*
Applications of the formose reaction, which involves the for-
mation of sugars from formaldehyde, have previously been
confined to the selective synthesis of unprotected sugars.
Herein, it is demonstrated that a-hydroxy-g-butyrolactone
dehyde. In the developed reaction system, homogeneous tin
chloride exhibits high catalytic activity and the addition of
mono- and disaccharides accelerates the formation of HBL.
These observations suggest that the formose reaction may
serve as a feasible pathway for the synthesis of important
chemicals.
(
HBL), which is one of the most important intermediates in
pharmaceutical syntheses, can be produced from paraformal-
Introduction
C1 chemistry is based on an interconversion with C1 com-
pounds and a methodology for the production of important
chemicals containing more than two carbon members by
saccharide synthesis, have not been reported, to the best of
our knowledge. Strict control of the formose reaction, that is,
the coupling frequency of formaldehyde and the handling of
product mixtures, appears to be a considerable issue.
[1]
using syngas, methane, or methanol. Formaldehyde, which is
industrially produced by the oxidation of methanol, in particu-
lar, has high potential as an electrophile in intermolecular cou-
We previously reported the catalytic transformation of
a triose sugar (DHA) and formaldehyde into a-hydroxy-g-butyr-
[
2]
[3]
[8]
pling reactions or as a basic ingredient of polymer products.
olactone (HBL; 1); which is one of the most important inter-
[
9]
Therefore, if the use of formaldehyde leads to the production
of important chemicals, a new process for sustainable develop-
ment in the field of organic industrial chemistry will be ach-
ieved.
mediates in pharmaceutical syntheses. Among the examined
Lewis acid catalysts, tin chlorides exhibited the best per-
formance in the aforementioned coupling reaction. Herein, we
report an unprecedented cascade synthesis of 1 from formal-
dehyde catalyzed by tin chlorides, and the accelerated forma-
tion of 1 upon the addition of mono- and disaccharides. A de-
tailed screening of catalysts was investigated and, to clarify an
additional effect of glucose and the reaction mechanism, sev-
eral experiments were investigated. To consider the condition
of the active species, the influence of the abundance ratio of
The formose reaction involves the formation of sugars from
[
4]
formaldehyde. Butlerov proposed a mechanism for the reac-
[
4a]
tion, which consisted of the following steps: two formalde-
hyde molecules condensed to form glycolaldehyde (GA), which
reacted through an aldol reaction with another equivalent of
[5]
formaldehyde to afford glyceraldehyde (GLA). The aldose–
ketose isomerization of GLA forms 1,3-dihydroxyacetone
119
Sn/HCl on the production of 1 was investigated by Sn NMR
(
DHA), which can react with formaldehyde, GA, and GLA; sub-
spectroscopy measurements.
sequent isomerization results in the formation of tetrose, pen-
tose, and hexose, respectively. To date, successful examples of
the formose reaction, developed exclusively for the selective Results and Discussion
syntheses of unprotected sugars, have been carried out under
Screening of catalysts
[
6]
[7]
either a base or mineral. However, applications of the for-
mose reaction to important chemicals, with the exception of
Initially, the cascade synthesis of 1 from formaldehyde was car-
ried out in the presence of tin chloride (0.043 mmol) in 1,4-di-
[
10]
[
a] Dr. S. Yamaguchi, T. Matsuo, Dr. K. Motokura, Dr. Y. Sakamoto,
Dr. A. Miyaji, Prof. T. Baba
oxane (4.0 mL) at 1608C (Table 1). Solid paraformaldehyde
does not dissolve in any solvent, and therefore, we cannot cal-
culate the conversion of formaldehyde. Under these condi-
tions, one mole of 1 is produced from four moles of formalde-
hyde.
Department of Environmental Chemistry and Engineering
Interdisciplinary Graduate School of Science and Engineering
Tokyo Institute of Technology, 4259-G1-14 Nagatsuta-cho
Midori-ku, Yokohama, Kanagawa 226-8502 (Japan)
Fax: (+81)045-924-5441
The reaction with HCl, which is a model Brønsted acid cata-
lyst, did not lead to the formation of 1, 2, or 3 (Table 1,
entry 1). This result indicates the importance of metal catalysts
to the reaction. When SnCl ·5H O was applied, a small amount
E-mail: syamaguchi@chemenv.titech.ac.jp
Homepage: http://www.chemenv.titech.ac.jp/lab/eng.html
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201500885.
4
2
ChemSusChem 2015, 8, 3661 – 3667
3661
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
chloride and iodide counterions, the amounts of 1 and 3
remain low, although that of 2 is similar, which confirms that
a chloride is the most suitable counteranion for this system.
However, the use of either nBu SnCl or Me Sn showed no im-
Table 1. Transformation of formaldehyde and glucose into 1 by using
Lewis acid catalysts.
[a]
3
4
provement on the amount of product, although they had a tet-
ravalent tin center (Table 1, entries 10 and 11). This result
shows that an absolute quantity of chloride counterion to a tin
center is important for the successful formation of 1. Having
established that chloride was the most suitable counterion in
this system, we next screened different metal chloride cata-
lysts, namely, AlCl , CrCl , FeCl , MgCl , and TiCl , and, conse-
3
3
2
2
4
quently, all were observed to have no activity in this reaction
(Table 1, entries 12–16). These results indicate that both tin
metal and a halogen counterion are critical elements for this
reaction system. Finally, when glucose was used and parafor-
maldehyde was not, product 1 was not obtained (Table 1,
entry 17).
[
b]
[c]
[c]
Entry
Catalyst
HCl
1 [mmol]
2 [mmol]
3 [mmol]
[
[
[
d]
e]
e]
1
2
3
4
5
6
7
8
9
1
<0.01
0.02
0.17
0.33
0.53
0.51
0.18
0.27
0.17
<0.01
<0.01
<0.01
0.026
0.15
0.21
0.032
0.19
<0.01
<0.01
<0.01
0.047
0.084
0.11
0.053
0.029
<0.01
0.031
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.048
SnCl
SnCl
SnCl
SnCl
4
2
4
2
·5H
·2H
·5H
·2H
2
2
2
2
O
O
O
O
Based on this experimental result, we added fresh glucose
and paraformaldehyde to the reaction mixture after an initial
run, and 1, 2, and 3 were afforded in almost the same amount
as that obtained in the first run. These results also support the
assertion that the tin catalyst continues to behave as a catalyst
without deactivation through multiple uses.
SnBr
SnBr
2
4
SnI
SnI
2
4
0.11
0
nBu
Me Sn
AlCl ·6H
CrCl ·6H
FeCl ·4H
MgCl ·6H
TiCl
SnCl
3
SnCl
0.05
0.093
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.11
1
1
4
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
1
1
1
1
1
1
2
3
4
5
6
7
3
2
O
3
2
O
2
2
O
2
2
O
Considerations on an additional effect of glucose
[f]
4
2
·2H
2
O
As shown in Table 1, the greatest amount of 1 was produced
when divalent tin chloride was used. In the presence of glu-
cose, the transformation of glucose into 1 occurs simultane-
ously owing to 1) the isomerization of glucose to fructose,
[
(
(
a] Reaction conditions: paraformaldehyde (6.25 mmol), glucose
0.625 mmol), 1,4-dioxane (4.0 mL), catalyst (0.043 mmol), naphthalene
20 mg), air, 3 h, 1608C. All conversions of glucose were >99%. [b] The
amount of 1 was determined by GC flame ionization detector (FID) analy-
sis. [c] The amounts of 2 and 3 were determined by H NMR spectroscopy
2
) retro-aldol reaction to DHA or GLA, and 3) an intermolecular
1
aldol reaction with formaldehyde. Therefore, product 1 could
be obtained from glucose and formaldehyde in this reaction
system. To determine whether the production of 1 was derived
from glucose or formaldehyde, an experiment was carried out
analysis. [d] 0.086 mmol. [e] Without glucose. [f] Without paraformalde-
hyde.
of 1 was obtained (Table 1, entry 2). The use of SnCl ·2H O af-
with [D ]paraformaldehyde was carried out. As shown in
2
2
2
forded 0.17 mmol of 1, which corresponded to a yield of 11 %,
based on the amount of formaldehyde (Table 1, entry 3). On
the other hand, the accelerating effect of the formose reaction
Scheme 1, 0.13 mmol of [D ]HBL (4) and 0.41 mmol of D -HBL
2
5
[11]
in the presence of sugar has been reported. This reaction
proceeded in an aqueous solution of Ca(OH) and was cata-
2
lyzed by complexes of calcium hydroxide and sugar. On the
basis of this report, we attempted to increase the amount of
1
produced by adding glucose. When SnCl ·5H O was applied,
4 2
0
.33 mmol (Table 1, entry 4) of 1 was obtained; surprisingly,
the use of SnCl ·2H O afforded 0.53 mmol of 1 (Table 1,
2
2
Scheme 1. Transformation of glucose and [D
2
]paraformaldehyde into
entry 5). These results showed that the addition of glucose
drastically increased the amount of 1. We subsequently investi-
gated the influence of the counteranion in the tin halides on
the formation of each product by conducting experiments
with anhydrous SnBr , SnBr , SnI , and SnI (Table 1, entries 6–
[
(
(
D ]HBL (4) and [D ]HBL (5). Reaction conditions: [D ]paraformaldehyde
6.25 mmol), glucose (0.625 mmol), 1,4-dioxane (4.0 mL), SnCl ·2H O
2 2
0.043 mmol), naphthalene (20 mg), air, 3 h, 1608C.
2
5
2
2
4
2
4
9
). The use of SnBr gave almost the same amounts of product
(5), the carbon selectivities of which were 9 and 27% yield, re-
spectively, were obtained and the total amount was almost
identical to the results shown in Table 1, entry 5. The amount
2
as those observed with SnCl ·2H O (Table 1, entry 6), whereas
when SnBr was employed the amount of 2 decreased dramat-
ically (Table 1, entry 7). SnI and SnI were both observed to
2
2
4
1
of 4 and 5 was determined by GC-FID and H NMR spectrosco-
2
4
have even lower activities and selectivities (Table 1, entries 8
and 9). When comparing similar catalysts that differ only in the
py analysis. In addition, GC-MS analysis confirmed the forma-
+
tion of 4 and 5. The m/z [M ] values of 4 and 5 are 104(1) and
ChemSusChem 2015, 8, 3661 – 3667
www.chemsuschem.org
3662
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
1
07(1), respectively (Figure S2 in the Supporting Information).
These results show that the product is a mixture of 4 and 5.
On the other hand, we could not observe unlabeled HBL (1),
which would arise from the coupling of two C2 fragments
from the retro-aldol of glucose (C2+C4 fragments), based on
1
results from H NMR spectroscopy analysis (Figure S3 in the
Supporting Information). This result is attributed to a previous
[
12]
report that the tin-catalyzed retro-aldol reaction of glucose
occurs mainly through resolution into C3+C3 fragments in-
stead of C2+C4 fragments.
On the other hand, an experiment with DHA and
[
D ]paraformaldehyde was also examined. As shown in
2
Table S1 in the Supporting Information, product 5 was not ob-
tained at all when either SnCl ·2H O or SnCl ·5H O was ap-
2
2
4
2
plied. This result and that shown in Figure S1 in the Support-
ing Information show that the combination of DHA and form-
aldehyde does not lead to the production of 1 based on a for-
mose reaction.
To determine the role of glucose, the influence of a specified
amount of glucose on the production of 1, 2, and 3 was inves-
tigated. As shown in Figure 1, we clarified that the total
Figure 2. a) Time profile measurement with glucose and paraformaldehyde.
Reaction conditions: glucose (0.625 mmol), paraformaldehyde (6.25 mmol),
2 2
SnCl ·2H O (0.043 mmol), 1,4-dioxane (4.0 mL), naphthalene (20 mg), air. All
conversions of glucose were >99%. The amount of 1 was determined by
1
GC-FID analysis. The amounts of 2 and 3 was determined by H NMR spec-
troscopy analysis. 1:
paraformaldehyde. Reaction conditions: glucose (0.625 mmol), paraformalde-
hyde (6.25 mmol), SnCl ·2H O (0.043 mmol), 1,4-dioxane (4.0 mL), naphtha-
&
, 2: ~, 3: *. b) Time profile measurement with only
Figure 1. Influence of the amount of glucose on the production of 1, 2, and
2
2
3
. Reaction conditions: glucose, paraformaldehyde (6.25 mmol), SnCl
2
·2H
2
O
lene (20 mg), air. All conversions of glucose were >99%. The amount of
(
0.043 mmol), 1,4-dioxane (4.0 mL), naphthalene (20 mg), air, 3 h, 1608C. All
1 was determined by GC-FID analysis. The amounts of 2 and 3 were deter-
1
mined by H NMR spectroscopy analysis. 1:
&
, 2: ~, 3: *.
conversions of glucose were >99%. The amount of 1 was determined by
GC-FID analysis. The amounts of 2 and 3 were determined by H NMR spec-
1
troscopy analysis. 1:
&
(”: derived from four formaldehydes, *: derived from
glucose and formaldehyde), 2: ~, 3: *.
ly, but 2 and 3 were not formed at all (Figure 2b). These results
clearly show an accelerating effect in the initial stage of the re-
action.
amount of 1 increased in proportion to the amount of added
glucose and the product derived from four formaldehydes also
increased (Figure 1, dotted line). These results, and that shown
in Table 1, entry 3, indicate that the transformation of formal-
dehyde into 1 was accelerated in the presence of glucose.
We subsequently attempted to measure a time profile
Finally, we attempted to apply other sugars, such as mono-
and disaccharides, sugar alcohols, and amino sugars, instead of
glucose (Table 2). Initially, we applied fructose, galactose, and
mannose as isomers of glucose and 0.50, 0.52, and 0.40 mmol
of 1 were obtained, respectively (Table 2, entries 3–5). The use
of xylose as a representative of pentose afforded 0.45 mmol of
1 (Table 2, entry 6). We subsequently applied disaccharides cel-
lobiose and sucrose, and 0.32 and 0.54 mmol of 1 were ob-
tained, respectively (Table 2, entries 7 and 8). These results indi-
cate that sugar stereochemistry affects the amount of 1 pro-
(
Figure 2). When glucose and paraformaldehyde were applied
as substrates, the amount of each product increased with time
and the production of 1 stopped increasing after 6 h (Fig-
ure 2a). On the other hand, when only paraformaldehyde was
applied as a substrate, the amount of 1 increased proportional-
ChemSusChem 2015, 8, 3661 – 3667
www.chemsuschem.org
3663
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Table 2. Effect of the use of sugars or sugar alcohols instead of glucose
[a]
on the production of 1.
Entry
Accelerator
Amount formed [mmol]
[
b]
[c]
[c]
1
2
3
1
2
3
4
5
6
7
8
9
–
0.17
<0.010
0.15
0.18
0.17
0.12
0.15
0.039
0.17
<0.01
0.019
<0.010
0.084
0.090
0.11
0.084
0.070
0.051
0.10
glucose
fructose
mannose
galactose
xylose
cellobiose
sucrose
glucosamine
sorbitol
0.53
0.50
0.52
0.40
0.45
0.32
0.54
<0.01
0.16
<0.01
0.020
1
0
[
(
a] Reaction conditions: accelerators (mono- (0.625 mmol) or disaccharide
0.313 mmol)), paraformaldehyde (6.25 mmol), 1,4-dioxane (4.0 mL),
2 2
SnCl ·2H O (0.043 mmol), naphthalene (20 mg), air, 3 h, 1608C. All conver-
sions of accelerator were >99%. [b] The amount of 1 was determined by
1
GC-FID analysis. [c] The amounts of 2 and 3 were determined by H NMR
spectroscopy analysis.
Figure 3. Influence of the amount of formaldehyde on the production of 1,
2
, and 3. Reaction conditions: glucose (0.625 mmol), paraformaldehyde,
SnCl ·2H O (0.043 mmol), 1,4-dioxane (4.0 mL), naphthalene (20 mg), air, 3 h,
608C. All conversions of glucose were >99%. The amount of 1 was deter-
mined by GC-FID analysis. The amounts of 2 and 3 were determined by
2
2
1
1
H NMR spectroscopy analysis. 1:
, 2: ~, 3: *.
&
duced. Finally, the use of an amino sugar gave no product 1,
whereas when we utilized a sugar alcohol the amount of 1 re-
mained unchanged (Table 2, entries 9 and 10). This result
shows that amino sugars and sugar alcohols do not accelerate
the formation of 1, whereas mono- and disaccharides, in par-
ticular, have a distinct effect on the reaction.
aldol condensation with another formaldehyde provided the
C3 sugar (Scheme 2, steps 1 and 2). Tin can coordinate with
enolate 6 due to its strong Lewis acid character, so activated
complex 6 is formed and readily undergoes the aldol reaction
with formaldehyde to afford 7 (Scheme 2, step 3). If this reac-
tion proceeds based on this route, compounds 1, 2, and 3 are
generated at a constant rate, as reported previously for DHA
Considerations of the reaction mechanism
To consider the reaction mechanism, we investigated the influ-
ence of a specific amount of paraformaldehyde on the produc-
tion of 1, 2, and 3 (Figure 3). The
[
9c]
and formaldehyde.
amount of 1 and 3 increased
commensurately and that of 2
remained unchanged with the
amount of formaldehyde used.
This result shows that
2 is
mainly composed of elements
derived from glucose.
The reaction pathway in
Scheme 2 is discussed in detail
in the next sections and is sup-
ported by the presentation of
experimental findings based on
the results shown in Figures 2
and 3. There are two possible
routes: I: the production of C3
sugar, DHA, or GLA through the
condensation of three formalde-
hydes, or II: an intermolecular
aldol reaction of two GAs.
Route I
Two formaldehydes were con-
densed to generate GA and then Scheme 2. Formation of 1 from four formaldehydes through the aldol reaction.
ChemSusChem 2015, 8, 3661 – 3667
www.chemsuschem.org
3664
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Route II
(Figure 4b). On the other hand, the trend for the production of
2
and 3 was similar to that seen for the formation of 1. These
Another possible mechanistic route to 1 is based on an aldol
reaction of two GAs. After aldol coupling and tautomerization,
the aldehyde appears prone to elimination of the hydroxyl
group in the b position of the unsaturated carbon (Scheme 2,
steps 4 and 5). Therefore, dehydration of tetrose 8 at the C3
position occurs to provide 7. If this reaction proceeds based
on this route, compounds 1 and 3 are generated at a constant
rate, whereas little of 2 is obtained.
Time profile measurements showed that 2 and 3 were not
afforded when using only formaldehyde as a substrate. In addi-
tion, when the amount of formaldehyde increased, the
amount of 2 remained unchanged. Based on these experimen-
tal results, this transformation of formaldehyde in the presence
of glucose mainly proceeds through route II (Scheme 2).
results indicate that the production of 1 and 3 from formalde-
hyde may be inhibited by an increase in HCl and the produc-
tion of 2 generated from glucose through a retro-aldol reac-
tion is also inhibited. However, according to the results shown
in Table 1, entries 10 and 11, the reaction did not proceed in
the absence of halide anions. These results and that shown in
Figure 4 and Figure S4 in the Supporting Information indicate
that a tetravalent tin cation in the reaction mixture interacts
with glucose and coordinates with a chloride anion.
In addition, as shown in Table 1, entries 2–5, the amount of
1 obtained was different in the presence or absence of glu-
cose, leading us to hypothesize that the addition of glucose
could change the conditions of the active species. Therefore,
119
we performed Sn NMR spectroscopy studies to compare the
Catalytic behavior of tin chlorides
Based on the results shown in Table 1, entries 2–5, it seems
that the interaction between the Lewis and Brønsted acidity
derived from the tin chloride salts in the reaction medium is
crucial to the overall rate of this cascade reaction. On the
other hand, the transformation of glucose and formaldehyde
in the presence of HCl yields no 1, 2, or 3. Although the impor-
tance of Lewis acidic tin is clear, we need to confirm the differ-
ence in activities between tin(II) and tin(IV). In addition, the
balance between Lewis and Brønsted acidity is key to under-
standing this cascade reaction, and directing it to higher pro-
duction and selectivities.
The results depicted in Table 1 showed that the reaction was
different in the presence of SnCl ·2H O and SnCl ·5H O. There-
2
2
4
2
fore, we initially attempted to investigate the influence of
a specific amount of tin catalyst on the production of 1. The
amount of each product obtained when either SnCl ·2H O or
2
2
SnCl ·5H O was utilized is shown in Figure S4a and S4b in the
4
2
Supporting Information, respectively. When 0.043 mmol of
SnCl ·2H O was used, the greatest production of 1 was ob-
2
2
tained; on the other hand, only 0.020 mmol of SnCl ·5H O, less
4
2
than half the catalytic amount, was required to obtain the
highest amount of 1. This result indicates that divalent tin may
be oxidized in air to generate tetravalent tin. On the other
hand, the use of SnCl ·2H O provided a greater amount of
2
2
each product than the use of SnCl ·5H O. Based on this result,
4
2
we speculated that HCl was obtained in greater amounts with
SnCl ·5H O than with SnCl ·2H O and, consequently, unpre-
4
2
2
2
dictable side reactions could be induced.
We subsequently attempted to investigate the influence of
the HCl/Sn ratio on the production of 1, 2, and 3. The amount
of each product with different HCl/Sn ratios, when using either
SnCl ·2H O or SnCl ·5H O, are shown in Figure 4a and b, re-
2
2
4
2
spectively. To vary the Lewis and Brønsted acid contents, exter-
Figure 4. Influence of the ratio of HCl/Sn on the production of 1, 2, and 3
nal HCl was added. As shown in Figure 4a, when SnCl was
with SnCl
(
(
2
·2H
0.625 mmol), paraformaldehyde (6.25 mmol), SnCl
0.043 mmol), 4m HCl in 1,4-dioxane, 1,4-dioxane (4.0 mL), air, 3 h, 1608C. All
2
O (a) and SnCl
4
·5H
2
O (b). Reaction conditions: glucose
2
2
·2H O/SnCl ·5H
2
4
2
O
used, the highest amount of 1 was observed with no extra
equivalents of HCl and that of 1 decreased with increasing
amounts of HCl. Furthermore, when SnCl4 was used, the
amount of 1 also decreased with increasing amounts of HCl
conversions of glucose were >99%. The amount of 1 was determined by
GC-FID analysis. The amounts of 2 and 3 was determined by H NMR spec-
troscopy analysis. 1:
, 2: ~, 3: *.
1
&
ChemSusChem 2015, 8, 3661 – 3667
www.chemsuschem.org
3665
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
conditions of tin metal in different reaction environments
Figure 5).
Figure 5a and b shows the Sn NMR spectra of SnCl ·2H O
cation coordinated with chloride anions could be an active
119
(
species. Finally, Sn NMR spectroscopy measurements provid-
ed indications of changing conditions around the tin metal
center, particularly with how the tin metal interacted with glu-
cose. The details of this reaction, especially with regard to in-
teractions between tin metal and glucose, are being
investigated in our laboratory and will be reported in
due course.
119
2
2
and SnCl ·5H O, respectively, in CD CN. The singlets at d=
4
2
3
À244 (^) and À639 ppm (
&
) are observed and may be derived
On the other hand, almost all carbon atoms
coming from glucose as an accelerator were trans-
formed into some byproducts, such as 2, 3, or an un-
expected one, through side reactions. From the sus-
tainability point of view, we need to consider the ef-
fective utilization of glucose, which is representative
of mono- and disaccharides, and therefore, the devel-
opment of catalysts for the conversion of glucose
into HBL is a considerable issue to overcome.
Experimental Section
1
19
Figure 5. Sn NMR spectra in CD
c) SnCl ·2H O (0.171 mmol) and glucose (0.625 mmol) after heating to 808C for 3 h, and
d) SnCl ·5H O (0.171 mmol) and glucose (0.625 mmol) after heating to 808C for 3 h.
3 2 2 4 2
CN (1.0 mL) of a) SnCl ·2H O at RT, b) SnCl ·5H O at RT,
2
2
General
4
2
Materials were purchased from Wako Pure Chemicals,
Tokyo Kasei Co., Kanto Kagaku Co., and Aldrich and
were used without further purification. A Shimadzu
from di- and tetravalent tin metals, respectively. After the solu-
+
GC14B instrument equipped with a 30 m Ultra ALLOY -1 column
tion of SnCl ·2H O was heated to 808C and stirred for 3 h, the
2
2
was used for GC-FID analysis. The temperature program was as fol-
signal at d=À244 ppm (^) derived from SnCl ·2H O was com-
À1
2
2
lows: 1) 323 K for 2 min, 2) a linear ramp of 3 Kmin to 373 K,
À1
pletely absent and the signal at d=À639 ppm (
&
), which was
3) 373 K for 5 min, 4) a linear ramp of 10 Kmin to 553 K, and
also seen in the spectrum of SnCl ·5H O, was exclusively ob-
served (Figure 5). We subsequently examined the Sn NMR
5) 553 K for 20 min. A Shimadzu QP2010 plus instrument equipped
with a DB-1 column was used for GC-MS analysis. Products were
confirmed by the comparison of their GC retention time, mass
4
2
119
spectrum of SnCl ·2H O in the presence of glucose (Figure 5c).
2
2
1
spectrum, and H NMR spectrum with those of authentic samples.
As shown in Figure 5c, the main signal at d=À614 ppm (”) is
observed. In addition, in case of the combination of
SnCl ·5H O and glucose, the main signal at d=À598 ppm (~)
1
H NMR spectra were recorded in CDCl3 or [D ]DMSO with an
6
119
AVANCE 400 spectrometer operated at 400 MHz. The Sn NMR
spectra were recorded with an AVANCE 400 spectrometer operated
at 186.46 MHz. Chemical shift values were determined in relation
4
2
is different from that observed with SnCl ·2H O (Figure 5d).
2
2
These results indicate that tin metal interacts with glucose, al-
though we could not identify what species these signals were
derived from.
to an external reference of SnMe (d=0 ppm).
4
Typical procedure for the catalytic reaction (Table 1, entry 5)
A 50 mL autoclave with a Teflon liner was charged with glucose
Conclusions
(
113 mg), paraformaldehyde (188 mg), SnCl ·2H O (9.7 mg), naph-
2 2
In this study, we considered the specific catalytic activity of ho-
mogeneous tin chloride for converting formaldehyde into 1 in
the presence of glucose. Initially, we screened a number of
Lewis acid catalysts and tin halides. Tin chlorides, in particular,
had the highest activity in this reaction system. Based on an
thalene (20 mg, internal standard), and 1,4-dioxane (4.0 mL) and
was pressurized with air. The autoclave was heated to the preset
temperature of 1608C. After being stirred at the same temperature
for 3 h, the autoclave was cooled and the reaction mixture was an-
alyzed by GC-FID and H NMR spectroscopy. All yields were calcu-
lated on a carbon basis, that is, two moles of 1 could be formed
1
experiment with [D ]paraformaldehyde and a study on the
2
from one mole of glucose. Carbohydrate conversion was deter-
amount of glucose used, we discovered that the addition of
glucose led to an acceleration of the conversion of formalde-
hyde into 1. Although the addition of mono- and disaccharides
had an effect similar to that of glucose, the use of amino
sugars and sugar alcohols did not show an accelerated effect.
In addition, a time profile measurement and a study on the
amount of paraformaldehyde used suggested that this reac-
tion might not proceed via C3 sugar as an intermediate. On
the other hand, we clarified that HCl generated from the chlo-
ride anion inhibited this reaction system and a tetravalent tin
1
mined by H NMR spectroscopy in [D ]DMSO or D O.
6
2
119
Typical Sn NMR spectroscopy measurement (Figure 5c)
The reaction was performed in a 20 mL Schlenk flask. SnCl ·2H O
2
2
(
38.6 mg) was combined with glucose (113 mg) and deuterated
acetonitrile (1.00 mL, Kanto Kagaku Co.) at room temperature.
A magnetic stirrer bar was then added and the flask was placed in
an oil bath. After being stirred at 808C for 3 h in air, the mixture
was cooled to room temperature, and was then inserted directly
ChemSusChem 2015, 8, 3661 – 3667
www.chemsuschem.org
3666
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
into an NMR tube. According to this procedure, the experiments
shown in Figure 5 were conducted.
Res. 1984, 134, C4–C6; d) A. W. Schwartz, R. M. de Graaf, J. Mol. Evol.
1993, 36, 101–106; e) R. Krishnamurthy, S. Pitsch, G. Arrhenius, Orig.
Life Evol. Biosph. 1999, 29, 139–152; f) S. Pitsch, R. Krishnamurthy, G. Ar-
rhenius, Helv. Chim. Acta 2000, 83, 2398–2411; g) L. E. Orgel, Proc. Natl.
Acad. Sci. USA 2000, 97, 12503–12507; h) A. Ricardo, M. A. Carrigan,
A. N. Olcott, S. A. Benner, Science 2004, 303, 196; i) G. Springsteen, G. F.
Joyce, J. Am. Chem. Soc. 2004, 126, 9578–9583; j) A. N. Simonov, O. P.
Pestunova, L. G. Matvienko, V. N. Snytnikov, O. A. Snytnikova, Yu. P. Tsen-
talovich, V. N. Parmon, Adv. Space Res. 2007, 40, 1634–1640; k) J. B.
Lambert, S. A. Gurusamy-Thangavelu, K. Ma, Science 2010, 327, 984–
Keywords: carbohydrates · domino reactions · homogeneous
catalysis · sustainable chemistry · lactones
[
[
1] a) J. L. G. Fierro, Catal. Lett. 1993, 22, 67–91; b) G. K. Surya Prakash, G. A.
Olah, Isr. J. Chem. 2014, 54, 1412.
2] For a one-carbon homologation reaction, for example, an asymmetric
hydroxymethylation, see: a) S. Ishikawa, T. Hamada, K. Manabe, S. Ko-
bayashi, J. Am. Chem. Soc. 2004, 126, 12236–12337; b) H. Torii, M. Naka-
dai, K. Ishihara, S. Saito, H. Yamamoto, Angew. Chem. Int. Ed. 2004, 43,
986.
[
8] For the utilization of HBL as a starting material in the field of pharma-
ceutical syntheses, for example, see: a) S. Shiury, J. Partridge, M. Usko-
kovic, J. Org. Chem. 1988, 53, 1040–1046; b) D. Green, J. Kiddle, C.
Thompson, Tetrahedron 1995, 51, 2865–2874; c) D. Schinzer, A. Bauer, J.
Schieber, Chem. Eur. J. 1999, 5, 2492–2500; d) J. Mulzer, A. Mantoulidis,
E. çhler, J. Org. Chem. 2000, 65, 7456–7467; e) J. White, P. Hrnciar, J.
Org. Chem. 2000, 65, 9129–9142; f) P. Phukan, S. Sasmal, M. Maier, Eur.
J. Org. Chem. 2003, 1733–1740.
1
983–1986; Angew. Chem. 2004, 116, 2017–2020; c) S. Kobayashi, T.
Ogino, H. Shimizu, S. Ishikawa, T. Hamada, K. Manabe, Org. Lett. 2005, 7,
729–4731; d) M. Kokubo, C. Ogawa, S. Kobayashi, Angew. Chem. Int.
4
Ed. 2008, 47, 6909–6911; Angew. Chem. 2008, 120, 7015–7017.
3] For utilization as an ingredient for polymer products, for example, Bake-
lite and urea formaldehyde resin, see: a) B. Tollens, Ber. Dtsch. Chem.
Ges. 1884, 17, 653–659; b) B. Tollens, Ber. Dtsch. Chem. Ges. 1896, 29,
[
[
9] a) T. Kagayama, S. Sakaguchi, Y. Ishii, Tetrahedron Lett. 2005, 46, 3687–
3689; b) S. Yamaguchi, K. Motokura, Y. Sakamoto, A. Miyaji, T. Baba,
2
2
751–2752; c) C. Goldschmidt, Ber. Dtsch. Chem. Ges. 1896, 29, 2438–
439; d) C. Goldschmidt, Chem. Ztg. 1897, 21, 460; e) W. Hesse, Industri-
Chem. Commun. 2014, 50, 4600–4602; c) S. Yamaguchi, T. Matsuo, K.
Motokura, Y. Sakamoto, A. Miyaji, T. Baba, ChemSusChem 2015, 8, 853–
al Polymers Handbook (Ed.: E. S. Wilks), Wiley-VCH, Weinheim, 2001,
pp 1129–1151.
860; d) S. Van de Vyver, C. Odermatt, K. Romero, T. Prasomsri, Y. Romµn-
Leshkov, ACS Catal. 2015, 5, 972–977.
[
[
[
4] a) A. C. R. Butlerov, Acad. Sci. 1861, 53, 145–147; b) O. Loew, J. Prakt.
Chem. 1886, 34, 51–55.
5] a) R. Breslow, Tetrahedron Lett. 1959, 1, 22–26; b) R. F. Socha, A. H.
Weiss, M. M. Sakharov, J. Catal. 1981, 67, 207–217.
6] For reactions under basic conditions, see: a) R. Breslow, J. Am. Chem.
Soc. 1958, 80, 3719–3726; b) T. Matsumoto, S. Inoue, J. Chem. Soc.
Perkin Trans. 1 1982, 9, 1975–1979.
[
[
10] We also investigated the optimal reaction temperature (Figure S1 in the
Supporting Information).
11] a) T. Sakai, M. Ishizaki, K. Kobayashi, M. Goto, Bull. Chem. Soc. Jpn. 1981,
5
4, 2083–2088; b) T. Sakai, M. Ishizaki, M. Goto, Bull. Chem. Soc. Jpn.
982, 55, 2409–2414.
12] M. S. Holm, S. Saravanamurugan, E. Taarning, Science 2010, 328, 602–
05.
1
[
6
[
7] For reactions with mineral, see: a) A. G. Cairns-Smith, P. Ingram, G. L.
Walker, Theor. Biol. 1972, 35, 601–604; b) Y. Shigemasa, S. I. Akagi, E.
Waki, R. Nakashima, J. Catal. 1981, 69, 58–68; c) Y. Shigemasa, A. Matsu-
da, M. Ueda, R. Nakashima, K. Takeda, M. Suzuki, S. Saito, Carbohydr.
Received: June 30, 2015
Published online on October 6, 2015
ChemSusChem 2015, 8, 3661 – 3667
www.chemsuschem.org
3667
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim