Influence of the Interaction between a Tin Catalyst and an Accelerator on the Formose-Inspired Synthesis of α-Hydroxy-γ-butyrolactone

Article abstract of DOI:10.1002/cctc.201501377

In this study, we focused on the tin-catalyzed transformation of formaldehyde into α-hydroxy-γ-butyrolactone (HBL) in the presence of an α-hydroxy carbonyl compound as the accelerator. The screening of various accelerators aided in clarifying the structural prerequisites of the accelerator for the formose-inspired synthesis of HBL. To investigate the influence of the interactions between the tin metal and the accelerator on the catalytic activity, we performed a deuterium-exchange experiment with α-hydroxyacetophenone followed by in situ ¹¹⁹Sn NMR spectroscopy and X-ray absorption fine structure measurements. On the basis of the experimental results, we proposed a reaction mechanism to obtain HBL.

Full text of DOI:10.1002/cctc.201501377

DOI: 10.1002/cctc.201501377
Full Papers
Influence of the Interaction between a Tin Catalyst and an
Accelerator on the Formose-Inspired Synthesis of
a-Hydroxy-g-butyrolactone
Sho Yamaguchi,* Takeaki Matsuo, Ken Motokura, Akimitsu Miyaji, and Toshihide Baba*[a]
In this study, we focused on the tin-catalyzed transformation
of formaldehyde into a-hydroxy-g-butyrolactone (HBL) in the
presence of an a-hydroxy carbonyl compound as the accelera-
tor. The screening of various accelerators aided in clarifying the
structural prerequisites of the accelerator for the formose-in-
spired synthesis of HBL. To investigate the influence of the in-
teractions between the tin metal and the accelerator on the
catalytic activity, we performed a deuterium-exchange experi-
ment with a-hydroxyacetophenone followed by in situ
119Sn NMR spectroscopy and X-ray absorption fine structure
measurements. On the basis of the experimental results, we
proposed a reaction mechanism to obtain HBL.
Introduction
Formose is a portmanteau of formaldehyde and aldose.[1] The
formose reaction involves the formation of sugars from formal-
dehyde through aldol condensation in the presence of a cata-
lyst system comprising a base and a divalent metal, such as
calcium.[2] Sakai et al. reported that the presence of sugars has
an accelerating effect on the formose reaction.[3] The formose
reaction proceeds in aqueous Ca(OH)2 solution, catalyzed by
a calcium hydroxide–sugar complex. However, to date, applica-
tions of the formose reaction have been restricted to the selec-
tive synthesis of unprotected sugars.
alyst interacts with the added sugar, remains unclear. Hence, in
this study, we investigated the influence of the interaction be-
tween the tin catalyst and the accelerator on the formose-in-
spired synthesis of HBL. Screening of various accelerators was
performed to identify the structural prerequisites for the accel-
erating effect. A deuterium-exchange experiment based on
1H NMR spectroscopy analysis by using a model compound fol-
lowed by in situ 119Sn NMR spectroscopy and X-ray absorption
fine structure (XAFS) measurements provided information on
the interaction between the tin metal and the accelerator. Fur-
thermore, we proposed a route to HBL on the basis of the ex-
perimental results.
We previously reported a formose-inspired cascade synthesis
of a-hydroxy-g-butyrolactone[4,5] (HBL), which is one of the
most important intermediates in pharmaceutical synthesis,
from formaldehyde.[6] In the developed reaction system, homo-
geneous tin chloride exhibits high catalytic activity, and the ad-
dition of mono- and disaccharides accelerates the formation of
HBL (Scheme 1). However, the actual accelerating effect of
sugars on the reaction, that is, the manner in which the tin cat-
Results and Discussion
Initially, the cascade synthesis of HBL from formaldehyde was
performed in the presence of SnCl2·2H2O (0.043 mmol) in 1,4-
dioxane (4.0 mL) at 1608C, and 0.17 mmol of HBL was ob-
tained (Table 1, entry 1). Solid paraformaldehyde does not dis-
solve in any solvent, and therefore, we were unable to calcu-
late the conversion of formaldehyde. To increase the yield of
HBL, various accelerators were screened. We previously clari-
fied that the use of mono- and disaccharides, with the excep-
tion of sugar alcohols and amino sugars, accelerates the syn-
thesis of HBL from formaldehyde.[6] In the present study, upon
using glucose as the accelerator for the reaction, 0.53 mmol of
HBL was obtained, of which 0.41 mmol was confirmed to be
derived from formaldehyde on the basis of the result obtained
Scheme 1. Formose-inspired synthesis in the presence of an accelerator with
I) Ca under basic conditions and II) Sn under Lewis acidic conditions.
[a] Dr. S. Yamaguchi, T. Matsuo, Dr. K. Motokura, Dr. A. Miyaji, Prof. Dr. T. Baba
Department of Environmental Chemistry and Engineering
Interdisciplinary Graduate School of Science and Engineering
Tokyo Institute of Technology
by using [D2]paraformaldehyde (Table 1, entry 2). The produced
[7]
amounts of other byproducts, such as lactic acid (LA)
and
vinyl glycolic acid (VG),[8] also increased significantly. In addi-
tion, we purified the product mixture by column chromatogra-
phy and obtained pure HBL in 22% yield. Upon using fructose,
an isomer of glucose, an increase in the amount of HBL pro-
duced was also observed (Table 1, entry 3). On the basis of
these results, we hypothesized that a combination of adjacent
4259-G1-14 Nagatsuta-cho
Midori-ku, Yokohama, Kanagawa 226-8502 (Japan)
E-mail: syamaguchi@chemenv.titech.ac.jp
tbaba@chemenv.titech.ac.jp
Supporting Information for this article can be found under http://
dx.doi.org/10.1002/cctc.201501377.
ChemCatChem 2016, 8, 1386 – 1391
1386
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Table 1. Screening of accelerators possessing an a-hydroxy carbonyl moiety.[a]
Entry
Accelerator
HBL
[mmol][b]
Yield of
HBL [%][c]
LA
[mmol][b]
VG
[mmol][b]
1
2
3
4
5
6
7
8
–
0.17
11
26
26
25
26
22
4
25
4
5
<0.01
0.15
0.18
0.12
0.14
0.06
<0.01
0.07
<0.01
<0.01
–
<0.01
0.08
0.09
0.06
0.07
0.03
<0.01
0.04
<0.01
<0.01
–
glucose
fructose
0.53 (0.41)[d]
0.50 (0.40)[d]
0.52 (0.39)[d]
0.53 (0.40)[d]
0.34
1,3-dihydroxyacetone
glyceraldehyde
hydroxyacetone
acetoin
a-hydroxyacetophenone
2-hydroxypropiophenone
benzoin
0.06
0.39
0.06
0.07
9
10
11
a-hydroxy-g-butyrolactone
0.625 (0.00)[d]
0
[a] Reaction conditions: accelerator (0.625 mmol), paraformaldehyde (6.25 mmol), 1,4-dioxane (4.0 mL), SnCl2·2H2O (0.043 mmol), naphthalene
(0.156 mmol), air atmosphere, 3 h, 1608C. [b] Amount produced was determined by 1H NMR spectroscopy. [c] One mole of HBL is produced from four
moles of formaldehyde. [d] Amount derived from formaldehyde.
carbonyl and hydroxyl groups, that is, an a-hydroxy carbonyl
moiety, could be essential to achieve an accelerating effect.
Therefore, we examined various accelerators possessing the a-
hydroxy carbonyl moiety. Triose sugars, such as 1,3-dihydroxy
acetone (possessing a ketone and a terminal hydroxy group)
and glyceraldehyde (possessing an aldehyde and a nonterminal
hydroxy group), were evaluated, and they provided 0.52 and
0.53 mmol of HBL, respectively (Table 1, entries 4 and 5). As
a ketone and an aldehyde easily isomerize by 1,2-hydride shift,
we can regard these substrates as equivalent. Then, the reac-
tion was performed with hydroxyacetone (HA; possessing a ter-
minal a-hydroxy carbonyl moiety), and it afforded 0.34 mmol
of HBL (Table 1, entry 6). However, the yield of HBL decreased
drastically upon using acetoin (possessing a nonterminal a-hy-
droxy carbonyl moiety) (Table 1, entry 7). These results indicate
that the cascade conversion of formaldehyde into HBL may be
accelerated by the use of compounds possessing a terminal a-
hydroxy carbonyl moiety. a-Hydroxyacetophenone (HAP),
which has a terminal a-hydroxy carbonyl moiety, afforded HBL
in reasonable yield (i.e., 0.39 mmol), whereas the yield of HBL
decreased upon using 2-hydroxypropiophenone and benzoin,
both of which have a nonterminal a-hydroxy carbonyl moiety
(Table 1, entries 8–10). On the basis of the results correspond-
ing to entries 2–10 in Table 1, we should consider that tin cata-
lysts have some interaction with the terminal a-hydroxy car-
bonyl moiety of the accelerator, that is, a structural prerequi-
site for HBL production. Finally, we checked whether the reac-
tion product itself could act as an accelerator (Table 1,
entry 11), but the results were negative (no HBL formation
from formaldehyde occurred).
HAP on the amount of each product formed in the reaction
(Figure S1, Supporting Information). Upon increasing the con-
centration of HAP, the total amount of HBL, as well as the
amounts of other products such as LA and VG, increased. In
addition, we performed an experiment by using HAP and
[D2]paraformaldehyde, and then only [D5]HBL was obtained.
This result also shows that the obtained HBL is derived from
formaldehyde (Scheme S1). In this reaction system, all conver-
sions of HAP were >99%, and then HAP was mostly converted
into phenylglyoxal (PG), which could be produced by oxidation
of HAP. As the amount of HAP used increased, the amount of
PG produced also increased. However, a more detailed discus-
sion as to why PG is given in large amounts is required.
To clarify the accelerating effect of HAP, we measured the
time profile of the reaction (Figure S2). Upon using HAP and
paraformaldehyde as substrates, the amount of HBL increased
drastically in the initial stages of the reaction. In addition, the
majority of HAP was converted into PG after a very short reac-
tion time. However, HBL production slowed down if the reac-
tion was performed in the absence of HAP. The results in Figur-
es S1 and S2 clearly demonstrate the accelerating effect of
HAP on the formose-inspired cascade synthesis of HBL.
We subsequently investigated the influence of the valence
of tin on the production of HBL (Table 2). Upon using
SnCl4·5H2O and SnCl2·2H2O for the reaction in the absence of
HAP, 0.02 and 0.17 mmol of HBL were obtained, respectively
(Table 2, entries 1 and 3). In the presence of HAP, SnCl4·5H2O
afforded 0.26 mmol of HBL and SnCl2·2H2O afforded
0.39 mmol of HBL (Table 2, entries 2 and 4), which revealed the
lower catalytic activity of SnCl4·5H2O. However, on the basis of
these results alone we cannot explain the difference in the cat-
alytic activities of SnCl4·5H2O and SnCl2·2H2O.
Next, we evaluated the accelerating effect for the formose-
inspired synthesis of HBL by using HAP as a model com-
pound.[9] HAP is a suitable accelerator for analysis of the inter-
action with tin catalysts because its solubility in 1,4-dioxane is
higher than that of sugars. To understand the reaction in
detail, we investigated the influence of the concentration of
To confirm an interaction between the tin catalyst and an ac-
celerator possessing an a-hydroxy carbonyl moiety, we initially
performed a deuterium-exchange experiment (Figures 1 and
1
2). Figure 1a shows the H NMR spectrum of HAP in CD3CN.
ChemCatChem 2016, 8, 1386 – 1391
www.chemcatchem.org
1387
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
complex SnCl4[H2O]2.[10] In addition, a single reso-
^
nance at d=À244 ppm ( ) is observed in the
119Sn NMR spectrum of SnCl2·2H2O in CD3CN (Fig-
ure 3b). Upon heating SnCl2·2H2O to 808C, the reso-
^
nance at d=À244 ppm ( ) disappeared and a single
resonance at d=À639 ppm ( ), attributable to
*
SnCl4[H2O]2, appeared (Figure 3c). This result shows
that divalent tin is easily oxidized to tetravalent tin in
CD3CN solution. HAP was subsequently added to the
SnCl2·2H2O solution (1:1 ratio). After this mixture was
heated to 408C for 0.5 h, a single resonance at d=
*
À611 ppm ( ), which, presumably, considering the
chemical shift, is derived from tetravalent tin, was ex-
clusively observed (Figure 3d). If we used an excess
amount of HAP as the accelerator, the same result as
that shown in Figure 3d was given. On the other
hand, if SnCl4·5H2O was used in the presence of HAP,
Figure 1. 1H NMR spectra in CD3CN: a) HAP at RT, b) a mixture of HAP and SnCl2·2H2O
0.5 h after the addition of D2O, and c) a mixture of HAP and SnCl2·2H2O 3.5 h after the
addition of D2O.
a resonance at d=À639 ppm ( ), corresponding to
*
the resonances in Figure 3a,c, was observed (Fig-
ure 3e). On the basis of these results, we speculated
that, in the presence of HAP, divalent tin could be
Table 2. Transformation of formaldehyde and HAP into HBL by using
a tin catalyst.[a]
Entry Tin catalyst HAP
HBL
Yield of
LA
VG
[mmol] [mmol][b] HBL [%][c] [mmol][b] [mmol][b]
1
2
3
4
SnCl4·5H2O
SnCl4·5H2O 0.625
SnCl2·2H2O
SnCl2·2H2O 0.625
0
0.02
0.26
0.17
0.39
1
17
11
25
<0.01
0.04
<0.01
0.07
<0.01
0.04
<0.01
0.04
0
[a] Reaction
conditions:
paraformaldehyde
(6.25 mmol),
HAP
(0.625 mmol), 1,4-dioxane (4.0 mL), tin catalyst (0.043 mmol), naphthalene
(0.156 mmol), air atmosphere, 3 h, 1608C. [b] Amount of HBL was deter-
mined by 1H NMR spectroscopy. [c] One mole of HBL is produced from
four moles of formaldehyde.
Figure 2. Percent deuterium exchange for the skeletal CÀH bond of HAP.
^
^
.
SnCl2·2H2O: , SnCl4·5H2O:
HAP and SnCl2·2H2O were mixed and heated to 408C
for 0.5 h, and then, D2O was added to the solution
over 0.5 h (Figure 1b) and 3.5 h (Figure 1c). A com-
parison of the spectra in Figure 1b,c with the spec-
trum in Figure 1a reveals that the signals correspond-
ing to the two a-protons at approximately d=5 ppm
gradually diminish. Figure 2 shows the rate of
proton–deuterium exchange. Upon using SnCl2·2H2O,
the rate of proton–deuterium exchange increased
drastically relative to that if SnCl4·5H2O was used.
These results show that the tin catalyst directly inter-
acts with the a-hydroxy carbonyl moiety of HAP,
which thereby accelerates HBL production.
We subsequently performed in situ 119Sn NMR spec-
troscopy measurements (Figure 3). Figure 3a shows
the 119Sn NMR spectrum of SnCl4·5H2O in CD3CN. The
Figure 3. 119Sn NMR spectra in CD3CN: a) SnCl4·5H2O at RT, b) SnCl2·2H2O at RT,
c) SnCl2·2H2O after heating to 808C for 1.5 h, d) a mixture of HAP and SnCl2·2H2O after
heating to 408C for 0.5 h, and e) a mixture of HAP and SnCl4·5H2O after heating to 408C
single resonance at d=À639 ppm ( ) is thought to
*
be derived from the hexacoordinate tetravalent tin for 0.5 h.
ChemCatChem 2016, 8, 1386 – 1391
www.chemcatchem.org
1388
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
converted into a tetravalent tin complex that is different from
SnCl4[H2O]2.
oxygen atom bonds, was longer than the SnÀO bond of SnÀOH
or SnO2. On the basis of these results, we propose that, during
the reaction, an oxygen atom derived from glucose coordinates
to a divalent tin species to form six-coordinate SnIV, which coor-
dinates with four oxygen atoms and two chloride atoms.
Next, Sn K-edge XAFS spectroscopy was performed. In the
Sn K-edge X-ray absorption near-edge structure (XANES) spec-
trum, the edge energy of SnCl2·2H2O and glucose in CH3CN
was similar to that of SnCl4·5H2O in CH3CN but different from
that of SnCl2·2H2O in CH3CN. This similarity suggested the for-
mation of a tetravalent tin complex in the solution of
SnCl2·2H2O and glucose. Figure 4 shows the Fourier transform
(FT) of the k3-weighted Sn K-edge extended X-ray absorption
fine structure (EXAFS) data. In the case of SnO2, SnCl2·2H2O in
CH3CN, and SnCl4·5H2O in CH3CN, signals corresponding to
SnÀO and SnÀCl bonds are observed at approximately r=1.6
and 2.0 Š, respectively (Figure 4a–c). Furthermore, signals cor-
responding to both the SnÀO and SnÀCl bonds are clearly ob-
served for SnCl2·2H2O and glucose in CH3CN (Figure 4d). This
result qualitatively shows that the addition of glucose increas-
es the coordination number of the oxygen atoms.
To summarize, the results in Table 1 and the deuterium-ex-
change experiment show that the tin catalysts directly interact
with the a-hydroxy carbonyl moieties. Furthermore, the results
of the in situ 119Sn NMR spectroscopy and XAFS measurements
show that hexacoordinate SnIV formed by mixing divalent tin
with an accelerator is coordinated to four oxygen atoms and
two chloride atoms. Therefore, we propose a route to HBL
from four molecules of formaldehyde, as shown in
Scheme 2a.[14] Tetravalent tin complex
2 generated from
SnCl2·2H2O and accelerator 1 is converted into 3, which has an
ene–diol ligand, under heating conditions. Another aldol con-
densation between 3 and formaldehyde, followed by 1,2-hy-
dride shift for CÀC bond growth, provides 5 and 7. Then,
glyoxal and glycolaldehyde (GA) or glyceraldehyde (GLA) are
given by retro-aldol reaction of a ligand of 5 or 7, respectively.
Accelerator 1 (HAP) is oxidized to glyoxal (PG) and, alternative-
ly, the tetravalent tin complex is reduced to a divalent one.
That is, one accelerator is needed to provide one HBL in this
reaction system. The observed experimental results of Table 1
and Figure S1 support this proposal, because GA[8] is converted
into HBL and VG, and GLA[4b,c] with formaldehyde is converted
into HBL, VG, and LA. Furthermore, Sakai et al. previously re-
ported a selective synthesis of unprotected sugars by using
calcium and a-hydroxy carbonyl compounds on the basis of
a comparable cascade of hydride and alkyl shifts[15] for a similar
compound.[3] As is clear from this report, our proposed mecha-
nism (Scheme 2a and Scheme S2) with Sn catalysts is similar to
that of the previous work with Ca catalysts, and therefore, the
Ca-catalyzed formose reaction also supports our proposal route.
On the basis of the above proposed mechanism, we can un-
derstand that using substrates possessing nonterminal a-hy-
droxy carbonyl moieties, such as acetoin, 2-hydroxypropiophe-
none, and benzoin, does not lead to an accelerating effect, be-
cause the nonterminal a-hydroxy carbonyl moiety exhausts
the hydrogen attached to the a-carbon atom by an aldol reac-
tion with formaldehyde and then terminates the reaction
(Table 1, entries 7, 9, and 10). In case of the absence of an ac-
celerator, complex 8 could be generated from SnCl2·2H2O and
two molecules of formaldehyde (Scheme 2b). Considering the
result of Table 2, low amounts of HBL and other byproducts
are produced, because the condensation between two mole-
cules of formaldehyde (acyloin condensation) is slower than
the aldol reaction between formaldehyde and the accelerator,
as reported previously.[16] Furthermore, in case of the absence
of a formaldehyde and an accelerator, divalent tin is easily con-
verted into tetravalent tin 10 on the basis of the results of the
119Sn NMR spectroscopy study, and the disproportional reaction
may provide an oxidized tin complex (Scheme 2c). On the
other hand, divalent tin has higher catalytic activity than tetra-
valent tin. For cases in which tetravalent tin is used, ligand ex-
change between 1 and the chloride atoms of complex 10, hex-
acoordinate tetravalent tin complex SnCl4[H2O]2, is absolutely
Figure 4. FT of k3-weighted Sn K-edge EXAFS for a) SnO2, b) SnCl2·2H2O in
CH3CN, c) SnCl4·5H2O in CH3CN, and d) SnCl2·2H2O and glucose in CH3CN.
The signals in the r=1.0–2.54 Š range in Figure 4d were
well fitted by using the SnÀO and SnÀCl shell parameters.
Curve-fitting analysis of the solution of SnCl2·2H2O and glucose
in CH3CN suggested that four oxygen atoms and two chloride
atoms were coordinated to a tetravalent tin species, with dis-
tances of 2.13 and 2.39 Š, respectively.[11] The distance between
the Sn atom and the nearest chloride atoms was essentially
consistent with the value determined by X-ray crystallography
for the SnÀCl bond (r=2.41 Š) in SnCl4·5H2O.[10a] However, the
distance between the Sn atoms and the nearest oxygen atoms
differed slightly from the value determined by X-ray crystallog-
raphy for the SnÀO bond (r=2.06 Š) in SnO2.[12] X-ray structural
analysis of the complexes of SnCl4 and a-alkoxy carbonyl com-
pounds has previously been performed.[13] The SnÀalkoxide
bond, which includes SnÀcarbonyl oxygen atom and SnÀether
ChemCatChem 2016, 8, 1386 – 1391
www.chemcatchem.org
1389
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
teracts with HAP. In situ 119Sn NMR spectroscopy and
X-ray absorption fine structure measurements indi-
cated that a tetravalent tin complex is formed by co-
ordination with four oxygen atoms and two chloride
atoms. On the basis of these results, we proposed
a reaction mechanism to obtain HBL through an
aldol reaction and subsequent 1,2-hydride shift with
a complex of the tin catalyst and an accelerator.
In Scheme 2a and Scheme S2, the so-called “accel-
erator” may be nothing but a reagent that can lead,
in the presence of the Sn catalyst, to a more favora-
ble route towards the formation of the desired prod-
uct than if formaldehyde is used alone. However, we
define an “accelerator” as a reagent that brings
a change to the velocity of the desired chemical reac-
tion route. In this paper, the addition of substrates
possessing an a-hydroxy carbonyl moiety clearly ac-
celerated the desired chemical reaction route to-
wards the formation of HBL. Therefore, we believe
that substrates possessing an a-hydroxy carbonyl
moiety can be called “accelerators”. On the other
hand, this “accelerator” is not catalytic and we need
an excess amount of the accelerator equal to the
amount of HBL generated. Therefore, we should aim
for a catalytic accelerator system in the future.
This study demonstrated that the use of formalde-
hyde alone is sufficient for the production of impor-
tant chemicals, and thus, a new process for sustaina-
ble development in the field of organic industrial
chemistry was established. Further mechanistic work
and optimization of the catalytic conditions could
lead to the design of heterogeneous catalysts for the
title reaction.
Scheme 2. a) Proposed route to HBL from four molecules of formaldehyde with a com-
plex of SnCl2·2H2O and an accelerator. b) In the absence of an accelerator. c) In the ab-
sence of a formaldehyde and an accelerator. d) A complex of SnCl4·5H2O and an acceler-
ator.
Experimental Section
Catalytic tests
In a typical procedure (Table 1, entry 5), a 50 mL Teflon-
lined autoclave was charged with HAP (0.625 mmol), par-
aformaldehyde (6.25 mmol), SnCl2·2H2O (9.7 mg), naph-
slow, although coordination of 1 proceeds smoothly if divalent
tin is used (Scheme 2d). The difference in the exchange veloci-
ties, as shown in Figures 1 and 2, supports this proposal.
thalene (0.156 mmol, internal standard), and 1,4-dioxane (4.0 mL)
and then pressurized with air. The autoclave was heated to
a preset temperature of 1608C. The mixture was stirred at the
same temperature for 3 h. Subsequently, the autoclave was cooled,
1
and the mixture was analyzed by GC–MS and H NMR spectrosco-
Conclusions
py.
In this study, we clarified the interactions between the tin cata-
lyst and an accelerator that showed catalytic activity in the for-
mose-inspired synthesis of a-hydroxy-g-butyrolactone (HBL).
We proposed a route to HBL from four molecules of formalde-
hyde with a complex of the tin catalyst and an accelerator.
Screening experiments revealed that compounds possessing
a terminal a-hydroxy carbonyl moiety have an accelerating
1H NMR spectroscopy measurements
Ina typical procedure for percent deuterium exchange (Figures 1
and 2), the Sn catalyst (0.171 mmol) was combined with HAP
(0.171 mmol) in CD3CN (1.0 mL) at room temperature in a flask,
which was placed in an oil bath. After stirring with a magnetic stir
bar at 408C for 0.5 h under ambient air, the mixture was cooled to
room temperature. Deuterium oxide (100 mg) was added to the
mixture, which was then transferred directly to an NMR tube.
1
effect on the formose-inspired synthesis of HBL. H NMR spec-
troscopy measurements with a-hydroxyacetophenone (HAP) as
the model compound showed that the tin catalyst directly in-
ChemCatChem 2016, 8, 1386 – 1391
www.chemcatchem.org
1390
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
119Sn NMR spectroscopy measurements
Krishnamurthy, G. Arrhenius, Helv. Chim. Acta 2000, 83, 2398–2411;
i) L. E. Orgel, Proc. Natl. Acad. Sci. USA 2000, 97, 12503–12507; j) A. Ri-
cardo, M. A. Carrigan, A. N. Olcott, S. A. Benner, Science 2004, 303, 196;
k) G. Springsteen, G. F. Joyce, J. Am. Chem. Soc. 2004, 126, 9578–9583;
l) A. N. Simonov, O. P. Pestunova, L. G. Matvienko, V. N. Snytnikov, O. A.
Snytnikova, Yu. P. Tsentalovich, V. N. Parmon, Adv. Space Res. 2007, 40,
1634–1640; m) J. B. Lambert, S. A. Gurusamy-Thangavelu, K. Ma, Science
2010, 327, 984–986.
In a typical procedure (Figure 3d), SnCl2·2H2O (0.171 mmol) was
combined with HAP (0.171 mmol) in CD3CN (1.0 mL) in a flask at
room temperature. The flask was then immersed in an oil bath.
The mixture was stirred with a magnetic stir bar at 408C for 0.5 h
under ambient air and then cooled to room temperature. Then,
the mixture was transferred directly to an NMR tube.
[3] a) T. Sakai, M. Ishizaki, K. Kobayashi, M. Goto, Bull. Chem. Soc. Jpn. 1981,
54, 2083–2088; b) T. Sakai, M. Ishizaki, M. Goto, Bull. Chem. Soc. Jpn.
1982, 55, 2409–2414.
[4] For previous reports on the synthesis of HBL, see: a) T. Kagayama, S. Sa-
kaguchi, Y. Ishii, Tetrahedron Lett. 2005, 46, 3687–3689; b) S. Yamaguchi,
K. Motokura, Y. Sakamoto, A. Miyaji, T. Baba, Chem. Commun. 2014, 50,
4600–4602; c) S. Yamaguchi, T. Matsuo, K. Motokura, Y. Sakamoto, A.
Miyaji, T. Baba, ChemSusChem 2015, 8, 853–860; d) S. Van de Vyver, C.
Odermatt, K. Romero, T. Prasomsri, Y. Romµn-Leshkov, ACS Catal. 2015,
5, 972–977.
[5] For examples of the utilization of HBL as a starting material in the field
of pharmaceutical syntheses, see: a) S. Shiury, J. Partridge, M. Uskokovic,
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. Öhler,
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.
Analysis
1H NMR spectra were recorded in CDCl3 with an Avance 400 spec-
trometer (Bruker) operated at 400 MHz. 119Sn NMR spectra were re-
corded by using an Avance 400 spectrometer operated at
186.46 MHz. The chemical shift values were determined relative to
an external reference, SnMe4 (d=0 ppm). A Shimadzu QP2010 Plus
system equipped with a DB-1 column was used for GC–MS analy-
sis. The temperature program was as follows: 323 K for 2 min, fol-
lowed by a linear ramp of 20 KminÀ1 to 553 K, and then 553 K for
20 min. Materials were purchased from Wako Pure Chemicals,
Tokyo Kasei Co., Kanto Kagaku Co., and Aldrich Inc. and were used
without further purification. The detailed conditions for the XAFS
measurements are described in the Supporting Information.
[6] S. Yamaguchi, T. Matsuo, K. Motokura, Y. Sakamoto, A. Miyaji, T. Baba,
ChemSusChem 2015, 8, 3661–3667.
Acknowledgements
[7] Y. Hayashi, Y. Sasaki, Chem. Commun. 2005, 2716–2718.
[8] a) M. S. Holm, Y. J. Pagµn-Torres, S. Saravanamurugan, A. Riisager, J. A.
Dumesic, E. Taarning, Green Chem. 2012, 14, 702–706; b) M. Dusselier,
P. Van Wouwe, F. de Clippel, J. Dijkmans, D. W. Gammon, B. F. Sels,
ChemCatChem 2013, 5, 569–575; c) 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; d) R. De Clercq, M. Dusselier, C.
Christiaens, J. Dijkmans, R. I. Iacobescu, Y. Pontikes, B. F. Sels, ACS Catal.
2015, 5, 5803–5811.
[9] Saccharides are converted into 1,3-dihydroxyacetone via a retro aldol
reaction followed by an intermolecular aldol reaction with formalde-
hyde to provide HBL, whereas HBL is not generated from HAP.
[10] a) J. C. Barnes, H. A. Sampson, T. J. R. Weakley, J. Chem. Soc. Dalton
Trans. 1980, 949–953; b) X. A. Mao, X. Z. You, A. B. Dai, Inorg. Chim.
Acta 1989, 156, 177–178.
This study was carried out with the approval of the Photon Fac-
tory Advisory Committee (PAC No. 2014G046). We thank Dr.
Yohei Uemura from the Institute for Molecular Science for valua-
ble assistance with XAFS. We also thank Prof. Dr. Wang-Jae Chun
from the International Christian University for his help in analyz-
ing the XAFS results.
Keywords: domino reactions
reaction · homogeneous catalysis · lactones
· formaldehyde · formose
[1] a) A. M. Butlerov, C. R. Acad. Sci. 1861, 53, 145–147; b) O. Loew, J. Prakt.
Chem. 1886, 34, 51–55; c) R. Breslow, Tetrahedron Lett. 1959, 1, 22–26;
d) R. F. Socha, A. H. Weiss, M. M. Sakharov, J. Catal. 1981, 67, 207–217;
e) J. Deng, T. Pan, Q. Xu, M. Y. Chen, Y. Zhang, Q. X. Guo, Y. Fu, Sci. Rep.
2013, 3, 1244; f) I. V. Delidovich, A. N. Simonov, O. P. Taran, V. N. Parmon,
ChemSusChem 2014, 7, 1833–1846.
[2] Conventional methods for the synthesis of unprotected sugars under
basic conditions: a) R. Breslow, J. Am. Chem. Soc. 1958, 80, 3719–3726;
b) T. Matsumoto, S. Inoue, J. Chem. Soc. Perkin Trans. 1 1982, 1975–
1979; under mineral conditions: c) A. G. Cairns-Smith, P. Ingram, G. L.
Walker, J. Theor. Biol. 1972, 35, 601–604; d) Y. Shigemasa, S.-i. Akagi, E.
Waki, R. Nakashima, J. Catal. 1981, 69, 58–68; e) Y. Shigemasa, A. Matsu-
da, N. Ueda, R. Nakashima, K.-i. Harada, N. Takeda, M. Suzuki, S. Saito,
Carbohydr. Res. 1984, 134, C4–C6; f) A. W. Schwartz, R. M. de Graaf, J.
Mol. Evol. 1993, 36, 101–106; g) R. Krishnamurthy, S. Pitsch, G. Arrhe-
nius, Origins Life Evol. Biospheres 1999, 29, 139–152; h) S. Pitsch, R.
[11] See Table S1 and Figure S3 for more information.
[12] T. Hara, M. Hatakeyama, A. Kim, N. Ichikuni, S. Shimazu, Green Chem.
2012, 14, 771–777.
[13] M. T. Reetz, K. Harms, W. Reif, Tetrahedron Lett. 1988, 29, 5881–5884.
[14] A detailed route to GA and GLA from formaldehydes with a complex of
SnCl2·2H2O and an accelerator, including 1,2-hydride and alkyl shifts, is
shown in Scheme S2.
[15] A Sn-catalyzed aldol reaction followed by a hydride shift to the other
Carbon atom, for example, see Ref. [8a–d].
[16] T. Mizuno, A. H. Weiss, Adv. Carbohydr. Chem. Biochem. 1974, 29, 173–
227.
Received: December 17, 2015
Published online on February 25, 2016
ChemCatChem 2016, 8, 1386 – 1391
www.chemcatchem.org
1391
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim