Inhibition of acid-catalyzed depolymerization of cellulose with boric acid in non-aqueous acidic media

Article abstract of DOI:10.1016/j.carres.2007.11.004

Boric acid inhibited the acid-catalyzed depolymerization of cellulose in sulfolane, a non-aqueous medium, at high temperature. Formation of the dehydration products such as levoglucosenone, furfural, and 5-hydroxymethyl furfural were also effectively inhibited. Similar inhibition was observed for cellooligosaccharides and starch, although the glucosidic bonds in methyl glucopyranosides and methyl cellobioside were cleaved to form α-d-glucofuranose cyclic 1,2:3,5-bisborate.

Full text of DOI:10.1016/j.carres.2007.11.004

Available online at www.sciencedirect.com
Carbohydrate Research 343 (2008) 249–255
Inhibition of acid-catalyzed depolymerization of cellulose
with boric acid in non-aqueous acidic media
Haruo Kawamoto,* Shinya Saito and Shiro Saka
Graduate School of Energy Science, Kyoto University, Yoshida-honmachi, Sakyo-ku, Kyoto 606-8501, Japan
Received 6 October 2007; accepted 1 November 2007
Available online 7 November 2007
Abstract—Boric acid inhibited the acid-catalyzed depolymerization of cellulose in sulfolane, a non-aqueous medium, at high tem-
perature. Formation of the dehydration products such as levoglucosenone, furfural, and 5-hydroxymethyl furfural were also effec-
tively inhibited. Similar inhibition was observed for cellooligosaccharides and starch, although the glucosidic bonds in methyl
glucopyranosides and methyl cellobioside were cleaved to form a-D-glucofuranose cyclic 1,2:3,5-bisborate.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Keywords: Cellulose; Starch; Acid-catalyzed depolymerization; Boric acid; Sugar borate; Cellooligosaccharide
1. Introduction
is produced.6 Such ester formation is also observed in
aqueous alkaline media.7–10 Although the pyranose,
furanose and open-chain structures are possible for the
esters of hexose, the furanose-type ester is known to
form preferentially.10–14
Cellulose is depolymerized in non-aqueous acidic media
through an acid-catalyzed transglycosylation reaction.1,2
Reduction in the degree of polymerization (DP) of
cotton cellulose at 150–200 ꢁC has been reported with
AlCl3 in CCl4.1 Addition of sulfuric acid has been
reported to accelerate the cellulose depolymerization in
sulfolane (tetramethylene sulfone) at 200 ꢁC, and the
depolymerized products were further converted to the
dehydration products including levoglucosenone, furfu-
ral, and 5-hydroxymethyl furfural (5-HMF), along with
colored substances.2 Similar acid catalysis was also
reported under pyrolysis conditions without using
solvent.3–5
Cyclic borate or boronate ester formation is a well-
known reaction of boric acid or boronic acid with the
hydroxyl groups of a sugar, especially 1,2- and 1,3-cis
diols.6 The ester formation is usually conducted by heat-
ing the carbohydrate in an organic solvent such as
benzene, 1,4-dioxane, pyridine, or N,N-dimethylform-
amide (DMF) with azeotropic removal of the water that
As for glucose, a-D-glucofuranose cyclic 1,2:3,5-
bis(phenylboronate) has been reported in heating with
1
phenylboronic acid in organic solvent.13–15 The H and
13C NMR spectra13–15 and the X-ray crystal structure15
have been reported for this boronate.
Kawamoto et al.16 have reported that boric acid
inhibited the formation of levoglucosenone, furfural,
and 5-HMF in the heat treatment of levoglucosan
(1,6-anhydro-b-D-glucopyranose) in 0.1% sulfuric acid–
sulfolane at 200 ꢁC through the formation of a-D-gluco-
furanose cyclic 1,2:3,5-bisborate (1). Phenylboronic acid
also had a similar effect. Thus, sugar–borate (or boro-
nate) ester was suggested to be fairly stable against
acid-catalyzed reactions at high temperature. It is also
interesting to know how such borate formation affects
the polysaccharide reaction in nonaqueous acidic media
at high temperature.
In this article, the influence of boric acid on the reac-
tions of polysaccharides (cellulose and starch) and
mono- and oligo-saccharides in acidic sulfolane is
investigated.
*
Corresponding author. Tel./fax: +81 75 753 4737; e-mail: kawamoto@
energy.kyoto-u.ac.jp
0008-6215/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2007.11.004

250
H. Kawamoto et al. / Carbohydrate Research 343 (2008) 249–255
and a colorless sulfolane-soluble portion. The sulfolane-
R
O
soluble portion was neutralized with solid NaHCO3
O
B
1
O
(50 mg) and analyzed by HPLC, GPC, and H NMR
spectroscopy as described in the following. The residue
was sequentially washed with a satd aq NaHCO3 solu-
tion and water, then dried at 105 ꢁC for 24 h and
weighed. In the repeated treatment experiment, the res-
idue was treated again in a similar manner.
HO
O
O
B
OH
The sulfolane-soluble portion was mixed with a solu-
tion of p-dibromobenzene (an internal standard, 4.0 mg)
in tetrahydrofuran (THF, 1.0 mL), and the resulting
solution was analyzed by HPLC using a Shimadzu
LC-10A under the following conditions: column, STR
α-
D
-Glucofuranose cyclic 1,2:3,5-bisborate (R = -CH2OH) (1)
α-
D
-Xylofuranose cyclic 1,2:3,5-bisborate (R = -H) (2)
2. Experimental
ODS-II; flow rate, 1.0 mL minꢀ1
MeOH–H2O (5 min),
;
eluent, 20:80
(5 min),
2.1. Materials
20:80!30:70
30:70!100:0 (8 min); detector, UV220nm; column tem-
perature, 40 ꢁC; retention times, 5-HMF, 5.6 min; levo-
glucosenone, 6.2 min; furfural, 6.7 min.
Cellulose powder (cotton, 200–300 mesh, Toyo Roshi
Kaisha Ltd) and starch (corn, Nacalai Tesque Inc.) were
used as polysaccharide samples. D-Glucose, D-xylose, D-
cellobiose, methyl a-D-glucopyranoside, methyl b-D-
glucopyranoside were purchased from Nacalai Tesque
Inc. Cellopentaose and cellohexaose were obtained from
Sigma–Aldrich Japan K.K. Methyl b-D-cellobioside was
prepared from D-cellobiose by successive acetylation
(Ac2O/NaOAc/reflux), bromination (33% HBr in
AcOH/AcOH/rt), methyl glycosylation (MeOH/
Ag2CO3/molecular sieve/CHCl3/rt) and deacetylation
(28% NaOMe/MeOH/rt). Methyl 4-O-methyl b-D-
glucopyranoside was prepared from methyl b-D-gluco-
pyranoside via 4,6-benzylidene acetal formation (benzal-
dehyde dimethyl acetal/p-toluenesulfonic acid/DMF/
50 ꢁC/30 torr), benzylation (benzyl bromide/NaH/
DMF/rt), reductive cleavage of the 4,6-benzylidene to
the 6-O-benzyl derivative (Et3SiH/BF3ÆEt2O/CHCl3/
0 ꢁC), methylation (MeI/NaH/DMF/rt) and debenzyl-
ation (10% Pd–C/H2/MeOH/rt). Methyl 4-O-methyl b-
D-cellobioside was also prepared from methyl b-D-cello-
bioside by a similar procedure. Silica gel plates (Kiesel-
GPC analysis was conducted for the sulfolane-soluble
portion after dilution with a 4-fold volume of THF with
a Shimadzu LC-10A. The chromatographic conditions
were as follows: column, Shodex KF801 + KF802; flow
rate, 1.0 mL minꢀ1; eluent, THF; detector, RID; column
temperature, 40 ꢁC.
Since the borate ester was unstable during the work-
ing up process, the reaction mixture was directly ana-
1
lyzed by H NMR spectroscopy. The sulfolane-soluble
portion (0.15 mL) was mixed with CDCl3 or dimethyl
sulfoxide (DMSO-d6, 0.45 mL) with p-dibromobenzene
as an internal standard, and the resulting solution was
analyzed by 1H NMR spectroscopy. The 1H NMR spec-
tra were recorded with a Varian AC-400 (400 MHz)
spectrometer using tetramethylsilane (TMS) as an inter-
nal standard. Chemical shifts (d) and coupling constants
(J) are given in ppm and Hz, respectively.
Starch and other model compounds were also treated
in a similar manner.
gel 60 F254
,
Merck) were used for thin-layer
chromatography (TLC). Other solvents and reagents
were purchased from Nacalai Tesque Inc.
3. Results and discussion
3.1. Cellulose
2.2. Heat treatment and characterization of the products
Figure 1 shows pictures of the solutions and residue
obtained from the 6-min heating of cellulose in 0.1% sul-
furic acid–sulfolane at 200 ꢁC in the presence and the
absence of boric acid (3 mol equiv against the glucose-
repeating unit in cellulose). The residue recoveries are
also included. In the absence of boric acid, cellulose
was completely converted to the sulfolane-soluble prod-
ucts, along with a change in the color of the solution to
dark reddish-brown. As listed in Table 1, levoglucose-
none (24.8 mol %), furfural (8.0 mol %), and 5-HMF
(2.7 mol %) (based on the glucose unit) were obtained
as the identified products. These results are consistent
with the results of the previous work.2 The boric acid
A sulfolane solution (1.0 mL) including 0.1% sulfuric
acid and boric acid (3 mol equiv for the glucose unit)
was added to cellulose (10 mg) in a Pyrex round-bottom
flask (volume: 20 mL). The flask was attached with a
condenser (a glass tube, 120 mm long and 14 mm in
diameter) and a nitrogen bag through a three-way tap.
After replacing the air in the reactor with nitrogen using
an aspirator, the flask was heated in a silicon oil bath at
200 ꢁC for 2 or 6 min. After the reaction, the reaction
mixture was immediately cooled with flowing air (15 s)
and subsequently by immersion of the flask in cold
water (3 min). The mixture was filtered to give a residue

H. Kawamoto et al. / Carbohydrate Research 343 (2008) 249–255
251
With boric acid
Without boric acid
*
Sulfolane
Solution
Boric acid
Residue
(%)
No
residue
a
92.0
0.0
Figure 1. Influence of the boric acid addition on the decomposition of
cellulose in 0.1% sulfuric acid/sulfolane (N2/200 ꢁC/6 min). Boric acid/
glucose unit in cellulose = 3:1 (mol/mol).
Table 1. Influence of the boric acid addition on the yields of
levoglucosenone, furfural, and 5-HMF in heat treatment of cellulose
in 0.1% sulfuric acid/sulfolane (N2/200 ꢁC/6 min)
b
Yield (mol %)
With boric acida
Without boric acid
10
15
20
Levoglucosenone
Furfural
5-HMF
0.2
0.2
0.1
24.8
8.0
2.7
Retention time (min)
Figure 2. GPC chromatograms of the sulfolane-soluble portions after
heat treatment of cellulose in 0.1% sulfuric acid/sulfolane in the
presence (a) and the absence (b) of boric acid (N2/200 ꢁC/6 min). Boric
Total
0.5
35.5
a Boric acid/glucose unit in cellulose = 3:1 (mol/mol).
acid/glucose unit in cellulose = 3:1 (mol/mol), retention time of a-D-
*
glucofuranose cyclic 1,2:3,5-bisborate.
addition substantially changed such reaction behavior of
cellulose. Solubilization of cellulose and colored sub-
stance formation were almost completely inhibited in
the presence of boric acid. Filtration gave a cream-
colored residue (92.0%) and a colorless sulfolane-soluble
portion. Furthermore, the formation of levoglucose-
none, furfural, and 5-HMF were also substantially
inhibited (total yield: 0.5%). Thus, the boric acid addi-
tion inhibited the acid-catalyzed depolymerization of
cellulose and further decomposition into dehydration
products.
Figure 2 shows the GPC chromatograms of these sul-
folane-soluble portions. In the absence of boric
acid (chromatogram b), several peaks, including a broad
signal in the higher molecular weight (MW) region, are
observed. Contrary to this, only the boric acid peak is
observed in the presence of boric acid (chromatogram
a), and the signal corresponding to a-D-glucofuranose
cyclic 1,2:3,5-bisborate (1, retention time: 15.5 min) is
not observed at all. From the 1H NMR signal
(6.5 ppm) assigned to the B-OH [sulfolane-soluble por-
tion/dimethylsulfoxide (DMSO)-d6 = 1/3, v/v], more
than 80% of the boric acid used was found to be recov-
ered. Since the formation of glucose borate 1 requires
2 mol equiv of boric acid, a large part of the boric acid
remains unreacted with the cellulose.
and glucose yields, which were determined by the GC
analysis of the alditol acetates prepared from the water
extracts of the residues are listed in Table 2. In this
repeated treatment, 1st and 2nd cycles showed very sim-
ilar results. About 10% (1st cycle: 8.0%, 2nd cycle:
10.6%) of the amount of the cellulose used were
removed after each treatment. These results and high
boric acid recovery indicate that only the cellulose mole-
cules near the surface area are reactive with boric acid
under these conditions. These molecules would depoly-
merize to the low-MW products, which are further
converted to the borate complexes. About 10% of the
cellulose would be converted to such borate com-
plexes and removed as soluble substances during wash-
ing with water. Low yields of glucose (1st cycle:
0.023%, 2nd cycle: 0.032%) indicate that these mole-
cules are not depolymerized up to the monomer
level. The cellulose molecules covered with the borate
Table 2. Cellulose recovery (%) and glucose yield (mol %) in the
repeated heat treatment of cellulose in 0.1% sulfuric acid/sulfolane in
the presence of boric acid (boric acid/glucose unit in cellulose = 3:1,
mol/mol) (N2/200 ꢁC/6 min)
Cellulose recovery (%)
Glucose yield (mol %)
First cycle
Second cycle
92.0
0.023
The residue obtained from the heat treatment of cellu-
lose in the presence of boric acid was further treated
under the similar conditions. The cellulose recoveries
89.4a
0.032a
a Based on the amount used in the second cycle.

252
H. Kawamoto et al. / Carbohydrate Research 343 (2008) 249–255
1
complex layer would be stabilized against the acid-cata-
lyzed transglycosylation.
Xylose also showed similar results. From the H NMR
spectrum (Fig. 3), a-D-xylofuranose cyclic 1,2:3,5-bisb-
orate (2) was suggested to form in 92.9 mol % yield.
The structure was confirmed by comparing the 1H
NMR spectrum with that of a-D-xylofuranose cyclic
1,2:3,5-bis(phenylboronate) from the literature.17 Glu-
cose borate 1 and xylose borate 2 were fairly stable dur-
ing heat treatment in acidic sulfolane. High yields of
these esters also indicate that the OH-6 in glucose borate
1 does not influence its stability.
These results are totally different from the previous
results with levoglucosan. Under similar reaction condi-
tions, the C-1–O bond of levoglucosan was completely
cleaved through acid-catalyzed transglycosylation and
gave a-D-glucofuranose cyclic 1,2:3,5-bisborate 1 in
71.5 mol % yield.
3.2. Model compounds
Contrary to these results, cellopentaose and cello-
hexaose showed the cellulose-type behavior. These com-
pounds were recovered as cream-colored substances
with colorless sulfolane solutions.
To investigate the origin of the different behaviors
between cellulose and levoglucosan in more detail, vari-
ous low-MW model compounds and starch (listed in
1
Table 3) were treated under similar conditions. The H
Methyl a- and b-D-glucosides, methyl b-D-cellobioside
exhibited the similar results (levoglucosan-type behav-
ior) to those of glucose and cellobiose. After acid-cata-
lyzed glycosidic bond cleavage, similar borate
formation would occur as observed for the reducing
sugars. Methyl 4-O-methyl-b-D-glucopyranoside and
methyl 40-O-methyl-b-D-cellobioside gave colored sulfo-
lane-soluble portions, which indicated the progress of
some degradation reactions. The 1H NMR signals
assigned to levoglucosenone (h), furfural (s), and 5-
HMF (n) are observed in these compounds, and their
total yields were 16.8 and 15.2 mol % (based on the glu-
cose unit), respectively. These two NMR spectra are
very similar except for the large signals (d) assigned to
glucose borate 1 in the spectrum from 40-O-methyl-b-
D-cellobioside. Glucose borate 1 is considered to arise
from the glucose unit without a 4-O-methyl group. This
NMR spectra (sulfolane-soluble portion/DMSO-d6 or
CDCl3 = 3/1, v/v) are shown in Figures 3 and 4. Table
3 summarizes the reaction behavior (solubilization,
color of solution) and the yields of borate, furfural, 5-
HMF, and levoglucosenone along with the solubility
of the model compounds in sulfolane. Starch, a polysac-
charide, exhibited the cellulose-type behavior. The influ-
ences of boric acid on the model compounds varied
depending on their structures.
As for the reducing sugars (Fig. 3), the DP determined
whether the influence is of the cellulose- or levoglucosan
type. Glucose and cellobiose formed a-D-glucofura-
nose cyclic 1,2:3,5-bisborate (1) in 85.1 and 71.9 mol %
yields (based on the glucose unit), respectively, and the
formation of the dehydration products and colored sub-
stances were completely inhibited (levoglucosan-type).
Table 3. Reactivities of various sugars in acidic sulfolane with boric acid at 200 ꢁC [boric acid (3 mol equiv for the monosaccharide-unit)/0.1%
sulfuric acid/sulfolane/N2/2 min]
Solubility in
sulfolane (20 ꢁC)a
Reaction behavior
Yieldd (mol %)
Solubilizationb Color of Borate Furfural 5-HMF Levoglucosenone
solutionc
g
D-Glucose
D-Xylose
D-Cellobiose
Cellopentaose
n
n
n
·
s
s
s
·
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
+
+
85.1e
92.9f
71.9e
—
—
—
—
—
—
—
—
—
—
—
9.5
6.7
—
—
—
—
—
—
—
—
—
2.3
3.5
Trace
—
—
—
—
—
—
—
5.0
5.0
Cellohexaose
·
·
—
Methyl b-D-glucopyranoside
Methyl a-D-glucopyranoside
Methyl b-D-cellobioside
Levoglucosan
Methyl 4-O-methyl-b-D-glucopyranoside
Methyl 40-O-methyl-b-D-cellobioside
s
s
s
s
s
s
s
s
s
s
s
s
68.9e
79.8e
59.6e
71.5e
—
36.6e
Cellulose
Starch
·
·
·
·
ꢀ
ꢀ
—
—
—
—
—
—
—
—
a s: soluble, n: partially soluble, ·: insoluble.
b s: solubilized, ·: not solubilized.
c +: dark reddish brown, ꢀ: colorless.
d Based on the monosaccharide unit.
e a-D-Glucofuranose cyclic 1,2:3,5-bisborate.
f a-D-Xylofuranose cyclic 1,2:3,5-bisborate.
g Not detectable.

H. Kawamoto et al. / Carbohydrate Research 343 (2008) 249–255
253
Figure 3. 1H NMR spectra of the sulfolane-soluble portions obtained from some reducing sugars [boric acid (3 mol equiv for the monosaccharide-
unit)/0.1% sulfuric acid/sulfolane/N2/200 ꢁC/2 min, sulfolane-soluble portion /DMSO-d6 = 1/3, v/v]. *1 Impurity in sulfolane, *2 internal standard (p-
dibromobenzene).
is also supported from the yield (36.6 mol %), which is
almost half of that (71.9 mol %) from cellobiose. The
4-O-methyl glucose unit would further decompose to
dehydration products and colored substances.
depending on the sugar type. Such solubility differences
would be important in the reaction behavior in acidic
sulfolane with boric acid. Soluble sugar undergoes the
acid-catalyzed transglycosylation to form the monosac-
charide borate (pathway a). The following reaction
behavior depends on the stability of the borate complex.
The less-stable complex (e.g., 4-O-methyl derivatives)
further decomposes to the products including levo-
glucosenone, furfural, 5-HMF, and colored substances,
while the stable complex remains unreacted in the
solution.
As for the insoluble sugar, pathway a becomes com-
petitive with the reaction with boric acid at the surface
molecules (pathway b). When the reactivity is a > b,
sugar is converted to the soluble borate complex
(glucose, xylose, and cellobiose). In the case of the reac-
tivity a < b, the sugar particle is covered with the borate
complexes, and these serve as a protective layer for the
inner sugar molecules against the acid-catalyzed trans-
glycosyl- ation and other degradation reactions (cello-
pentaose, cellohexaose, cellulose, and starch), although
the protection mechanism of the coated layer is not
known at the moment.
Decomposition of the 4-O-methylated glucose-moiety
is explained with difficulty in the stable borate ester
formation. Foster18 compared the borate complexation
of various methylated glucoses in aqueous alkaline med-
ia and reported that methylation of the OH-1, OH-2, or
OH-4 substantially reduced the complexation ability,
while the influence of the methylation of OH-3 or OH-
6 was small. The 4-O-methylation inhibits the formation
of the glucofuranose-type structure. Preferential forma-
tion of the furanose-type esters has been reported in the
literature.10–14
3.3. Mechanism
Based on the present results, the origin of the different
actions of boric acid on sugar is discussed with a pro-
posed mechanism shown as in Figure 5. As shown in
Table 3, the solubility of the sugar in sulfolane varied

254
H. Kawamoto et al. / Carbohydrate Research 343 (2008) 249–255
1
Figure 4. H NMR spectra of the sulfolane-soluble portions obtained from methyl 4-O-methyl-b-D-glucopyranoside and methyl 40-O-methyl-b-D-
cellobioside [boric acid (3 mol equiv for the monosaccharide-unit)/sulfolane/N2/200 ꢁC/2 min, sulfolane-soluble portion/CDCl3 = 1/3, v/v].
Impurity in sulfolane, s: furfural, n: 5-HMF, h: levoglucosenone, d: a-D-glucofuranose cyclic 1,2:3,5-bisborate.
*
Further decomposition products
Levoglucosenone
c
d
Furfural-type products
Colored substances
Depolymerization
(Transglycosylation)
Mono-
saccharide
borate
(Unstable borate)
(soluble)
a
Stabilized against the
further decomposition
(Stable borate)
Sugar
b
Borate complex
Protection layer
against acid-catalyzed
reaction
Reaction
with boric acid
(at the surface)
Figure 5. A proposed mechanism for the different actions of boric acid on sugar under heat treatment in non-aqueous acidic media at high
temperature.
Acknowledgments
21st COE program ‘Establishment of COE on Sustain-
able-Energy System’ (2002.4-2007.3) supported by The
Ministry of Education, Culture, Sports, Science and
Technology, Japan.
This work was supported by a Grant-in-Aid for Scien-
tific Research (c)(2) (No. 16580132, 2004.4-2006.3) and

H. Kawamoto et al. / Carbohydrate Research 343 (2008) 249–255
255
10. van den Berg, R.; Peters, J. A.; van Bekkum, H.
Carbohydr. Res. 1994, 253, 1–12.
References
11. Foster, A. B.; Stacey, M. J. Chem. Soc. 1955, 1778–
1781.
1. Ivanov, V. I.; Afanasev, V. A.; Sarybaeva, R. I. Dokl.
Akad. Nauk SSSR 1970, 192, 1043–1045.
2. Kawamoto, H.; Saito, S.; Hatanak, W.; Saka, S. J. Wood
Sci. 2006, 53, 127–133.
3. Halpern, Y.; Riffer, R.; Broido, A. J. Org. Chem. 1973, 38,
204–209.
12. Benner, K.; Klufers, P. Carbohydr. Res. 2000, 327, 287–
¨
292.
13. Norrild, J. C.; Eggert, H. J. Am. Chem. Soc. 1995, 117,
1479–1484.
14. Draffin, S. P.; Duggan, P. J.; Fallon, G. D.; Tyndall, E. M.
Acta Crystallogr., Sect. E 2005, E61, o1733–o1735.
15. Wood, P. J.; Siddiqui, I. R. Carbohydr. Res. 1974, 36, 247–
256.
16. Kawamoto, H.; Saito, S.; Saka, S. J. Anal. Appl. Pyrolysis,
submitted for publication.
17. Wood, P. J.; Siddiqui, I. R. Carbohydr. Res. 1974, 32, 97–
104.
18. Foster, A. B. J. Chem. Soc. 1953, 982–986.
4. Fung, D. P. C. Wood Sci. 1976, 9, 55–57.
5. Dobele, G.; Rossinskaja, G.; Telysheva, G.; Meier, D.;
Faix, O. J. Anal. Appl. Pyrolysis 1999, 49, 307–317.
6. Ferrier, R. J. Adv. Carbohydr. Chem. 1978, 35, 31–80.
7. van Duin, M.; Peters, J. A.; Kieboon, A. P. C.; van
Bekkum, H. Tetrahedron 1984, 40, 2901–2911.
8. Chapelle, S.; Verchere, J.-F. Tetrahedron 1988, 44, 4469–
4482.
9. Chapelle, S.; Verchere, J.-F. Carbohydr. Res. 1989, 191,
63–70.