Synthesis of butyl-α and β-D-glucopyranosides in the presence of dealuminated H-Y faujasites: Kinetic study, mechanism, stereoelectronic effects, and microreversibility principle

Article abstract of DOI:10.1006/jcat.1999.2477

The glycosylation reaction between D-glucose and n-butanol has been investigated over various dealuminated HY zeolites with Si/Al ratios from 2.5 to 100. In this way, butyl-α- and β-D-glucofuranosides and -pyranosides are readily synthesized at temperatures from 363 to 383 K and with a butanol/glucose ratio from 5 to 40. From the systematic study of the glycosylation reaction, a kinetic scheme involving both consecutive and competitive steps has been proposed. Butyl-D-glucofuranosides and butyl-D-glucopyranosides are primary products, butyl-D-glucofuranosides being then quantitatively converted into their pyranoside form. The use of microporous catalysts, associated with their shape selectivity properties, reduces the amount of oligomers and increases the amount of β-D-glucopyranosides in a significant manner. The β/α anomeric ratios observed are then discussed in terms of stereoelectronic effects which are, according to the principle of microreversibility of the same nature as those reported for the reverse reaction, i.e., hydrolysis of alkyl-D-glucopyranosides.

Full text of DOI:10.1006/jcat.1999.2477

Journal of Catalysis 185, 445–453 (1999)
Article ID jcat.1999.2477, available online at http://www.idealibrary.com on
Synthesis of Butyl- and -D-Glucopyranosides in the Presence
of Dealuminated H–Y Faujasites: Kinetic Study, Mechanism,
Stereoelectronic Effects, and Microreversibility Principle
Jean-Franc¸ois Chapat, Annie Finiels, Jacques Joffre, and Claude Moreau1
Laboratoire de Mate´riaux Catalytiques et Catalyse en Chimie Organique, UMR CNRS-ENSCM 5618, Ecole Nationale Supe´rieure de Chimie,
8, Rue de l’Ecole Normale, 34296 Montpellier Cedex 5, France
Received December 21, 1998; revised February 23, 1999; accepted February 23, 1999
have already shown that alkylglycosides could be synthe-
sized directly using heterogeneous catalysts as, for example,
macroporous sulfonated resins (9, 10) and acid clays (11).
Recently, it was shown that the use of acid zeolites (12)
could avoid the principal drawback of the direct Fischer
synthesis, i.e., the formation of oligomer species.
We report in the present paper the synthesis of butyl-D-
glucosides over dealuminated HY faujasites by acetaliza-
tion of D-glucose with n-butanol. A kinetic study of this
reaction has been carried out in order to clarify the mech-
anism. On the other hand, it was interesting to control the
stereoselectivity of the reaction since the tensioactive prop-
erties of and anomers are quite different (13, 14). The
influence of textural properties of various zeolites such as
The glycosylation reaction between D-glucose and n-butanol
has been investigated over various dealuminated HY zeolites with
Si/Al ratios from 2.5 to 100. In this way, butyl- - and -D-gluco-
furanosides and -pyranosides are readily synthesized at tempera-
tures from 363 to 383 K and with a butanol/glucose ratio from 5 to
40. From the systematic study of the glycosylation reaction, a kinetic
scheme involving both consecutive and competitive steps has been
proposed. Butyl-D-glucofuranosides and butyl-D-glucopyranosides
are primary products, butyl-D-glucofuranosides being then quanti-
tatively converted into their pyranoside form. The use of microp-
orous catalysts, associated with their shape selectivity properties,
reduces the amount of oligomers and increases the amount of
-
D-glucopyranosides in a significant manner. The anomeric ra-
/
tios observed are then discussed in terms of stereoelectronic effects
which are, according to the principle of microreversibility, of the
same nature as those reported for the reverse reaction, i.e., hydrol-
c
HY or HBEA on the anomeric
/ ratio of furanosides
and pyranosides is then discussed in terms of shape selec-
tive properties of zeolites and in terms of stereoelectronic
effects.
ysis of alkyl-D-glucopyranosides.
1999 Academic Press
Key Words: glycosylation; alkylglucosides; surfactants; zeolites;
stereoelectronic effects.
EXPERIMENTAL
INTRODUCTION
Materials
Alkylglycosides are a new class of nonionic surfactants
which find applications in cosmetics, food emulsifient, and
detergency (1, 2) because of their nontoxic and biodegrad-
able properties (3). As due to their composition, i.e., a glu-
cose hydrophilic moiety associated with a long hydropho-
bic alkyl chain of a fatty alcohol, alkylglycosides present
lyotropic and thermotropic properties as liquid crystals
(4). Although the first synthesis of alkylglycosides was de-
scribed a hundred years ago by Fischer (5), this reaction is
always under investigation (6–8). Generally, protection and
deprotection steps are required in order to obtain alkyl-
glycosides with a high selectivity. However, some authors
Anhydrous -D-glucose, n-butanol, and p-toluene sul-
fonic acid (PTSA) were supplied by Aldrich and used with-
out further purification. Butyl-D-glycosides were purified by
preparative chromatography on Chromagel 60A with ethyl
acetate/methanol 98/2 (v/v) as the eluent. They were iden-
tified by 13C NMR spectroscopy and by comparison with
the data from the literature (9).
Catalysts
Dealuminated H-form zeolites were obtained from De-
gussa, PQ Zeolites, and Uetikon. The different data con-
cerning the catalysts used, origin, and Si/Al ratios as well
as their acidity measured by thermodesorption of ammonia
are reported in Table 1.
1 To whom correspondence should be addressed. Fax: +33-(0)467 144
349; E-mail: cmoreau@cit.enscm.fr.
445
0021-9517/99 $30.00
c
Copyright
1999 by Academic Press
All rights of reproduction in any form reserved.

446
CHAPAT ET AL.
TABLE 1
Zeolites Used for Glycosylation of D-Glucose with n-Butanol
1000 rpm agitation speed was therefore used in all further
experiments. The initial reaction rate increases proportion-
ally with catalyst weight up to a maximum for 3 g cata-
lyst weight. Further experiments were then achieved in the
presence of 3 g of zeolite (6% of the total weight).
1
Zeolite
Si/Al
Acidity (meq H+
g )
HY (UETIKON)
HY (PQ-CBV 712)
HY (PQ-CBV 720)
HY (PQ-CBV 740)
HY (PQ-CBV 760)
HY (PQ)
HY (PQ)
HY (DEGUSSA)
HBEA (PQ-CP811)
2.5
5.7
15
21
25.5
29
40
100
12.5
0.94
0.82
0.56
0.30
0.22
0.16
0.15
0.06
0.92
(b) Pore diffusion. The importance of internal diffu-
sion can be estimated from the criterion of Weisz (18) 8 =
dN/dt 1/C0 R2/Deff, in which dN/dt is the reaction
3
rate (mol cm
s
1), C0 is the initial reactant concentra-
tion (mol cm 3), R is the particle radius (cm), and Deff is
the effective diffusivity coefficient (cm2 s 1). The intracrys-
talline diffusion limitation can be ruled out when 8 < 0.1.
7
3
1
Under our conditions, dN/dt = 3.08 10 mol cm
s ,
C0 = 0.533 10 mol cm 3, and R = 0.25 10 cm. This
requires Deff > 10 12 cm2 s 1, a reasonable value for the dif-
fusivity coefficient of that type of molecule in microporous
systems.
3
4
Work-up Procedure
As described in a preliminary work, the procedure was
typically as follows: 50 ml of n-butanol (0.54 mol), 4.80 g of
-D-glucose (0.027 mol), and 3 g of freshly calcined zeolite
(773 K for 6 h) were poured into a magnetically stirred
(1000 rpm) 100 ml glass reactor operating in a batch mode.
Zero time was taken to be when the temperature reached
383 K. The reaction was thus performed with an optimal
butanol/glucose molar ratio of 20 in the presence of 6 wt%
of catalyst. D-glucose was completely dissolved in butanol
during the formation of butyl-D-glucosides so that there is
no kinetic limitation due to glucose solubilization.
(c) Energies of activation. The influence of the temper-
ature on the reaction rates was investigated in the range
363–383 K. The energies of activation calculated by the
Arrhenius law are reported in Table 2. These values are on
the same order of magnitude ( 30 kcal mol 1) and show
that the reaction is not governed by external diffusion limi-
tations (17). Moreover, it can be noticed that the energies of
activation are slightly lower for furanosides than for pyra-
nosides, thus confirming that, as in homogeneous catalysis,
furanosides are the kinetic products and pyranosides the
thermodynamic ones.
Analyses
Samples were withdrawn periodically from the reactor,
filtered, and analyzed by HPLC. Analyses were performed
using a Shimadzu LC-6A pump and a refractive index RID-
6A detector. The column used was a BIORAD HPX-87C
(300 7.8 mm) thermostated at 353 K. Deionized water
was the eluent (0.5 ml/min), and D-sorbitol was the external
standard.
(d) Reaction scheme. Under the experimental condi-
tions used (4.80 g of D-glucose, 50 ml of n-butanol, 3 g of
freshly calcined zeolite, 383 K, and 1000 rpm of agitation
speed), it has been shown that the reaction could be taken as
obeying chemical regime conditions. A typical concentra-
tion vs time profile for the glycosylation of butanol at 383 K
is presented in Fig. 1. This figure shows the disappearance
of D-glucose and appearance of the various identified prod-
Kinetic Modeling
ucts (butyl- and -D-glucofuranosides and butyl- and
-
The initial rate constants were deduced from the experi- D-glucopyranosides) as a function of time.
mental plots of concentrations in all species formed vs reac-
tion time by curve fitting and simulation using the AnaCin
software (16). This kinetic treatment implies all reaction
TABLE 2
Initial Reaction Rates ( 105 mol s 1) for the Disappearance of
D-Glucose and the Appearance of Butyl-D-Glucosides at Different
Temperatures and Corresponding Energies of Activation
steps to be first order in reactants.
RESULTS AND DISCUSSION
Kinetic Studies
D-
Butyl- -D- Butyl- -D- Butyl- -D- Butyl- -D-
T (K)
glucose furanoside furanoside pyranoside pyranoside
(a) Mass transfer. In order to establish the kinetic rate
law, the chemical regime conditions must be first deter-
mined. For the study of the external diffusion of the re-
actants to the catalyst surface (17), the agitation speed and
the weight of the catalyst were varied.
363
373
383
0.130
0.440
1.540
0.075
0.235
0.675
0.050
0.145
0.585
0.010
0.030
0.130
0.005
0.015
0.075
1
Ea (kcal mol
)
35
31
35
36
38
The initial reaction rate for the disappearance of glucose
Note. 4.8 g of D-glucose, 50 ml of n-butanol, 6 wt% of freshly calcined
is independent of the agitation speed above 400 rpm. A HY15, 1000 rpm.

SYNTHESIS OF BUTYL-D-GLUCOPYRANOSIDES OVER ZEOLITES
447
FIG. 2. AnaCin modeling plot for glycosylation of D-glucose (4.8 g)
with n-butanol (50 ml) in the presence of freshly calcined HY15 (6 wt% )
at 383 K and 1000 rpm of agitation speed.
FIG. 1. Run profile for glycosylation of D-glucose (4.8 g) with n-
butanol (50 ml) in the presence of freshly calcined HY15 (6 wt% ) at
383 K and 1000 rpm of agitation speed (᭡, -D-glucose; ᭹, butyl- -D-gluco-
pyranoside; ❍, butyl- -D-glucopyranoside; ᭿, butyl- -D-glucofuranoside;
, butyl- -D-glucofuranoside).
agreement is obtained between the two modes of calcula-
tion of the initial reaction rates (Table 4).
(f) Influence of the initial concentration of D-glucose.
We have measured initial rates at different initial concen-
trations in D-glucose while the concentration of butanol
is kept constant. Figure 3 shows the initial reaction rate
evolution with the initial concentration of D-glucose. As it
can be seen from Fig. 3, the initial rate increases, reaches a
maximum when the D-glucose concentration is about 0.3–
0.5 M, and then decreases. This maximum may indicate a
Langmuir–Hinshelwood type mechanism with both reac-
tants adsorbed (15).
From Fig. 1, it is shown that butyl-D-glucofuranosides are
the major primary products of the reaction (kinetic prod-
ucts). Their concentration passes through a maximum and
then decreases with time to give the thermodynamically
more stable butyl-D-glucopyranosides (19, 20). However, it
is worth noting that butyl-D-glucopyranosides are also pri-
mary products. According to these observations, and to pre-
viously published works (21, 22), a general reaction scheme
can be advanced for the glycosylation reaction, involving
both parallel and successive reactions (Scheme 1).
(g) Kinetic law. As already proposed in the case of hy-
droxymethylation of furfuryl alcohol in the presence of ze-
olites (23), adsorption of reactants could take place on two
types of catalytic sites. Indeed, in the hypothesis of adsorp-
tion on Brønsted sites for both alcohol and glucose, two
species positively charged could not react together in an
acetalization reaction. Consequently, we assume that the
molecules of glucose and butanol are adsorbed on adja-
cent but different places: glucose on the protonic sites and
butanol on the framework by hydrophobic interactions be-
tween hydrocarbon chain of alcohol and silicon atoms. An
enhancement of the nucleophilicity of the hydroxyl group
of the alcohol could not be ruled out by interaction between
the hydrogen of the hydroxyl group and an oxygen atom of
the framework. With such an assumption, the rate for the
(e) Kinetic results. Such a reaction scheme has been
confirmed by the use of the kinetic AnaCin modeling soft-
ware by providing the best fit between calculated curves and
experimental data (Fig. 2). The mass balance is assumed to
be oligoglucosides without further distinction. The pseudo
first-order rate constants resulting from AnaCin curves are
reported in Table 3. From the rate constant values, it is
then possible to calculate the initial disappearance rate of
D-glucose, r0 = kC0. On the other hand, initial reaction rates
can also be calculated as the tangents, at zero time, to the
experimental concentration vs time curves. A fairly good
TABLE 3
1
Calculated Pseudo First-Order Rate Constants (min
from AnaCin Kinetic Software
)
k1
k-1
k2
k-2
k3
k-3
k4
k-4
0.036
0.005
0.004
0.001
0.013
0
0.005
0.001
SCHEME 1. Simplified reaction scheme for glycosylation of D-glucose
with n-butanol.
Note. 4.8 g of D-glucose, 50 ml of n-butanol, 6 wt% of freshly calcined
HY15, 383 K, 1000 rpm.

448
CHAPAT ET AL.
TABLE 4
Influence of the Si/Al Ratio on the Initial Rate
of Disappearance of D-Glucose
Zeolite
(Si/Al ratio)
r0 AnaCin
105 mol s
r0 Graphic
1
a
1 b
(
)
(
105 mol s
)
HY (2.5)
HY (5.7)
HY (15)
HY (21)
HY (25.5)
HY (29)
HY (40)
HY (100)
0.020
1.290
1.780
0.730
0.560
0.530
0.300
0.130
0.020
1.080
1.540
0.690
0.525
0.420
0.270
0.110
FIG. 4. Plot of f([G]) versus reaction time (min) for glycosylation of
D-glucose (4.8 g) with n-butanol (50 ml) in the presence of freshly calcined
HY15 (6 wt% ) at 383 K and 1000 rpm of agitation speed.
a Calculation from AnaCin software.
b Calculation from the slope of the tangent at the origin.
glucose concentrations between 0.3 and 0.5 M, with a reac-
1
tion relative order close to zero, r0 = k = 1.5 10 5 mol s
.
reaction can be written as
At low initial D-glucose concentrations, the rate equation
G[G]
A[A]
is reduced to k G[G]0. A
value can thus be calculated,
G
r = k
,
equal to 5, from known values of [G]0 (0.044 mol l 1), r0
(1 + G[G]) (1 + A[A])
(0.31 10 5 mol s 1), and k (1.5 10 5 mol s 1).
where [G] and [A] are the concentrations of glucose and
of butanol and and their respective adsorption con-
By plotting now f ([G]) = 1/ G ln([G]0/[G]) + ([G]0
[G]) as a function of reaction time (Fig. 4), a straight line
is obtained and the rate constant deduced from the slope
G
A
stants. Taking into account that butanol is in large excess
with respect to glucose, the following simplified equation is
obtained,
(1.7 10 mol s 1) is in good agreement with the value
5
obtained from experimental data (1.5 10 5 mol s 1).
d[G]
G[G]
r =
= k
,
(h) Influence of the Si/Al ratio. As previously men-
tioned, using zeolites in glycosylation decreases the forma-
tion of by-products such as oligoglucosides (12). Another
interesting feature of using zeolites is the possibility to con-
trol their acidity by dealumination. The acid strength of a
given site increases when the number of Al atoms in the
next nearest neighbor positions (NNN) decreases. A com-
pletely isolated aluminum tetrahedron has zero NNN and
is consequently the strongest type of framework Brønsted
site. This is the case for Faujasite-type zeolites with Si/Al
ratios higher than 6.8 (24). As shown in the kinetic part,
the glycosylation reaction requires Brønsted acid sites for
adsorption of glucose through protonation of the anomeric
oxygen. Acetalization then occurs by a nucleophilic attack
of butanol.
dt
(1 + G[G])
which finally leads, after integration, to the kinetic rate
equation below:
1
[G]0
[G]
ln
+ ([G]0 [G]) = kt.
G
As shown in Fig. 3, the initial reaction rate is maximum
and approximatively constant (1.5 10 5 mol s 1) for initial
Glycosylation of butanol was therefore studied over a se-
ries of dealuminated HY zeolites with different framework
Si/Al ratios (from 2.5 to 100). The initial reaction rates for
the disappearance of glucose are reported in Table 4. By
plotting those reaction rates as a function of the Si/Al ratio
(Fig. 5), it can be seen that the activity increases up to a
Si/Al ratio equal to 15 and then decreases significantly. As
already reported for other reactions, this Si/Al ratio of 15 is
representative of the balance between the number and the
strength of the acid sites for HY zeolites (25).
FIG. 3. Plot of initial rate of disappearance of D-glucose versus initial
concentration in D-glucose (10% error bars) for glycosylation of D-glucose
(4.8 g) with n-butanol (50 ml) in the presence of freshly calcined HY15
(6 wt% ) at 383 K and 1000 rpm of agitation speed.
In Fig. 6, the turn-over frequencies (TOF), calculated
by dividing the initial reaction rate by the number of acid

SYNTHESIS OF BUTYL-D-GLUCOPYRANOSIDES OVER ZEOLITES
TABLE 5
449
Comparison of the Different Types of Catalysts
Used for Glycosylation
1
Catalyst
HY 15
HBEA 12.5
PTSA
r0 glucose ( 105 mol s
)
Mass balance (% )
1.540
1.500
1.460
91a
75b
85
Note. Products adsorbed on the catalyst (a5 and b10% ) were recovered
by extraction of the solid with water.
be an efficient catalyst for glycosylation (12). Para-toluene
sulfonic acid (PTSA) was used as the reference to homo-
geneous catalysis. The experimental results obtained are
shown in Table 5. Except the amount of p-toluene sulfonic
acid (0.05 g), all other parameters are unchanged (4.80 g
of D-glucose, 50 ml of n-butanol, 6 wt% zeolite, 383 K,
1000 rpm). Figures 7 and 8 illustrate glucose conversion
vs time plots for HBEA and PTSA catalysts, respectively.
The most striking feature between those two catalysts is the
FIG. 5. Plot of initial rate of disappearance of D-glucose versus Si/Al
ratio (10% error bars) for glycosylation of D-glucose (4.8 g) with n-butanol
(50 ml) in the presence of freshly calcined HY (6 wt% ) at 383 K and
1000 rpm of agitation speed.
sites as determined by TPD of ammonia (Table 1), are now
expressed as a function of the Si/Al ratio. It can be seen
that the activity per site is maximum and nearly constant
from Si/Al = 15 to 29. For higher Si/Al ratios, the catalyst is
too hydrophobic for a good adsorption of D-glucose and the
TOF slightly decreases. On the contrary, for HY 2.5, glucose
is preferentially adsorbed and the reaction may then be
limited by both diffusion and adsorption of n-butanol. As
recently reported by Corma et al. (26), “a highly hydrophilic
reactant (glucose) and a much more hydrophobic one (n-
butanol) have to diffuse through the pores and adsorb in
the zeolite.” Consequently, zeolites having too low or too
high Si/Al ratios are not efficient catalysts for the reaction
considered.
important difference observed for the
/ ratio of pyra-
nosides in the first two hours of reaction. Concerning the
formation of butyl - and -D-glucofuranosides, the stere-
oselectivity of the reaction is not controlled because the
,
anomerization is too fast and the thermodynamic ra-
tio is rapidly obtained (27, 28), so that the most stable
-
furanoside is always preferentially formed with a / ratio
around 1.2. Concerning the formation of butyl - and -D-
glucopyranosides, plots of
/ ratio of pyranosides vs glu-
cose conversion for HY15, HBEA12.5, and PTSA catalysts
are reported in Fig. 9. The / ratio ofpyranosidesisaround
1.5 for HY15 and around 0.7 for HBEA12.5, whereas PTSA
shows an intermediate / ratio around 1. At nearly a com-
Selectivity of Zeolites
(a) Comparison of different catalysts. The behavior of
the HY zeolite with a Si/Al ratio of 15 was compared with
that of a HBEA zeolite with a Si/Al ratio of 12.5, known to
plete conversion of glucose, the thermodynamic ratio
of 0.5 is obtained.
/
FIG. 7. Run profile for glycosylation of D-glucose (4.8 g) with n-
butanol (50 ml) in the presence of freshly calcined HBEA12.5 (6 wt% )
at 383 K and 1000 rpm of agitation speed (᭡, -D-glucose; ᭹, butyl- -
D-glucopyranoside; ❍, butyl- -D-glucopyranoside; ᭿, butyl- -D-gluco-
furanoside; , butyl- -D-glucofuranoside).
FIG. 6. Turn-over frequency versus framework Si/Al ratio (10% error
bars) for glycosylation of D-glucose (4.8 g) with n-butanol (50 ml) in the
presence offreshly calcined HY (6 wt% ) at 383K and 1000 rpm of agitation
speed.

450
CHAPAT ET AL.
TABLE 6
Energies and Critical Diameters of Reactants and Products
Energy
(kcal mol
Critical diameter
1
a
b
˚
Reactants and products
)
(A)
-D-glucofuranose
-D-glucofuranose
-D-glucopyranose
-D-glucopyranose
11.9
11.5
5.0
5.4
5.9
6.4
6.4
5.3
-butyl-D-glucofuranoside
-butyl-D-glucofuranoside
-butyl-D-glucopyranoside
-butyl-D-glucopyranoside
13.5
13.2
6.3
5.1
5.7
6.3
6.3
6.6
FIG. 8. Run profile for glycosylation of D-glucose (4.8 g) with n-
a
Hyperchem calculated energies must be considered as relative ener-
butanol (50 ml) in the presence of PTSA (0.05 g) at 383 K and 1000 rpm
of agitation speed (᭡, -D-glucose; ᭹, butyl- -D-glucopyranoside; ❍,
butyl- -D-glucopyranoside; ᭿, butyl- -D-glucofuranoside; , butyl- -D-
glucofuranoside).
gies.
b Estimated critical diameter after projection of molecules in different
˚
plans ( 0.3 A).
In a previous paper (15), we explained that the increase
in the selectivity to butyl -D-glucopyranosides over HY15
zeolite could result from the shape selective properties of
that kind of catalyst, thus favoring the formation of the
-anomer with an equatorial OR group over that of the
expected more sterically hindered -anomer with an axial
OR group. However, the results obtained with the HBEA
zeolite do not agree with that simple explanation since the
selectivity should have increased as due to the more con-
fined crystalline network in HBEA catalyst (tridirectional
ics force field AmberS. This force field was designed for
modeling oligosaccharides (30) presenting glycosidic link-
ages as alkylglucosides.
We have first determined the most stable conforma-
tions of the different glucosides by conformational anal-
ysis. The following stability order, from the more to the
less stable, is obtained: -pyranose> -pyranose> -fura-
nose> -furanose (Table 6), six-membered rings (pyrano-
sides) being more stable than five-membered rings (furano-
sides) (31). Although the molecules are considered under
vacuum at 0 K for conformational analysis calculations, the
computed energy orders for free sugars are in good agree-
ment with the experimental results. The -pyranose form is
more stable because of the endoanomeric effect (32) and,
on the contrary, the -furanose form is less stable because
of cis interactions between the alkoxy group on C1 and the
hydroxyl group on C2 (20). The satisfactory results obtained
for molecular modeling of free sugars prompted us to in-
vestigate further the butyl-D-glucoside forms. Once again,
calculations agree quite well with the literature data. The
same stability order is obtained (Table 6), and the differ-
ences in calculated energies for the different anomers are
conventional ones (33).
˚
framework with straight 7.6 6.4 A channels connected by
˚
5.5 A channels) than in HY15 (tridirectional framework
˚
˚
with 7.4 A windows connecting 11.8 A spherical cavities)
(29). In order to account for this unexpected stereoselectiv-
ity order, a molecular modeling simulation was performed
to appreciate the size of the molecules involved.
(b) Molecular modeling. Molecular modeling simula-
tions were performed with Hyperchem 4.5, a well-known
molecular modeling software, using the molecular mechan-
Second, from the most stable conformations determined
for glycosides, it was then possible to estimate the size of
the different conformers from projections in different plans.
The corresponding ring diameter in which the conformers
can enter was assessed as the critical diameter (Table 6).
As expected for five-membered rings compared to six-
membered rings, the critical diameters of furanosides are
less important than those of pyranosides. It clearly appears
from these calculations that butyl-D-glucopyranosides have
a critical diameter close to that of D-glucose and that there
is no significant difference between - and -pyranosides.
FIG. 9. Plot of the
/
ratio of butyl-D-glucopyranosides versus D-
glucose conversion in the presence of different catalysts (᭹, HY15; ❍,
PTSA; ᭡, HBEA12.5) for glycosylation of glucose (4.8 g) with n-butanol
(50 ml) at 383 K and 1000 rpm of agitation speed.
˚
With a mean value of 6.3 A for the critical diameter of
both conformers, the stereoselectivity of the glycosylation

SYNTHESIS OF BUTYL-D-GLUCOPYRANOSIDES OVER ZEOLITES
451
FIG. 10. Illustration of endo- and exoanomeric effects (R1, hydroxyl
groups).
reaction over H–Y zeolites cannot result from products
shape selective properties only. Indeed, transition states se-
lective properties cannot be ruled out if we consider the
influence of stereoelectronic effects.
Let us now focus our attention on the different anome-
ric effects which take place in alkyl-D-glucopyranosides
(Fig. 10). An endoanomeric effect takes place if an elec-
tron lone pair on the endocyclic oxygen is antiperiplanar to
the exocyclic C–O bond. In this manner, a stabilizing inter-
action exists between this electron lone pair and the anti-
bonding
orbital of the C–O anomeric bond (34). Such
FIG. 11. Reaction pathwayfor hydrolysisofalkyl-D-glucopyranosides
taking into account the relative stabilities from the more stable (1) to the
less stable (7) intermediate.
a stabilization, only occurring with the -pyranoside form,
leads to a decrease in free energy of about 0.5 kcal mol 1 as
already calculated with AmberS by Still et al. (33). It is also
worth mentioning the existence of an exoanomeric effect
based on the same delocalization between the electron lone
of the -anomer
to give the intermediate
the reverse
3
2
pair of the exocyclic oxygen and the antibonding
of the endocyclic C–O bond (35).
orbital
anomeric effect occurs when the leaving group is gauche to
the two electron lone pairs of the endocyclic oxygen atom.
Then it would have to undergo a conformational change to
(c) Stereoelectronic effects. Stereoelectronic effects are
a key element in order to understand the chemical reactivity
of organic reactions (36). It is generally accepted now that
stereoelectronic effects play an important role in glycoside
hydrolytic processes (37).
the -twist-boat form
for stereoelectronic assistance to
6
the cleavage of the leaving group. For the -anomer , the
1
stereoelectronic assistance can directly occur from the in-
termediate
but it will undergo a conformational change
5
to the more stable -twist-boat form
because of the re-
4
For homogeneous acid-catalyzed hydrolysis of alkyl-D-
glucopyranosides, it has already been shown that the
verse anomeric effect. Both anomers must therefore form
-
the same half-chair intermediate to hydrolyze. The result
7
anomer hydrolyzes 2 or 3 times faster than the -anomer
(38). In order to understand why the -anomer, which
does not present the required electron lone pair orienta-
tion to hydrolyze with stereoelectronic assistance due to
endo-anomeric effect, hydrolyzes faster than the -anomer;
the formation of a twist-boat transition state must be con-
sidered and then the occurence of reverse anomeric effect
(Fig. 11).
The reverse anomeric effect is a stabilizing electronic ef-
fect taking place in charged species and leading to preferred
conformations which have the positively charged anomeric
substituent in the equatorial position as it occured with the
pyridinium group (39). When the positively charged atom
is gauche to the two electron lone pairs, a maximum sta-
bilizing effect, like in the -anomer, is obtained. As it was
rationalized by Deslongchamps (37) in the case of proto-
nated alkoxy groups in hydrolysis of alkylglucopyranosides
of the competitive formation of the two possible intermedi-
ates and to hydrolyze with stereoelectronic assistance
4
7
in the -anomer then leads to a hydrolysis rate lower for
5
the -anomer than for the -anomer.
Such an assumption also took into account the reactivity
of conformationally rigid bicyclic acetals (Fig. 12) for which
the -anomer hydrolyzes1.5timesfaster than the -anomer
(40). The formation ofa twist-boat form isno longer favored
because of the rigidity of the bicyclic system. Consequent-
ly, the -anomer, which has the required orientation for
FIG. 12. Bicyclic acetals used as hydrolysis models in homogeneous
(Fig. 11), it results that, after protonation of the OR group catalysis.

452
CHAPAT ET AL.
twist-boat formation is no longer favored because of steric
hindrance within the intracrystalline network.
By virtue of the microreversibility principle, the same
transition states should be invoked for the formation
of alkyl-D-glucopyranosides and hydrolysis thereof, with
the formation of a cyclic carboxonium ion intermediate
(Scheme 2). In this manner, it is possible to explain why
the HY zeolite favors the selectivity, whereas the HBEA
zeolite will favor the one. Within the large pores of HY
˚
zeolite (cavities with a diameter of 11.8 A), the formation of
twist-boat forms can occur and the -anomer will be prefer-
entially formed via the
In the more confined HBEA zeolite with only 7.6 6.4 A
channels, the formation of twist-forms will be no longer fa-
→
→
→
route (Fig. 11).
2
7
6
3
˚
vored, and the
→
→
route (Fig. 11) leading to the
1
7
5
more stable -anomer will be preferred. It should be noted
that the anomeric / ratio will drop to its thermodynam-
ical value at a nearly complete conversion of glucose.
CONCLUSION
Several conclusions can be drawn from this work: (i) the
direct synthesis of butyl-D-glucosides starting from glucose
and n-butanol is easily achieved in the presence of acid
dealuminated Y-faujasites as catalysts, (ii) the amount of
oligomers is significantly reduced compared to the ho-
mogeneous catalyzed reaction, (iii) the selectivity to the
-anomer is higher for H–Y compared to H–BEA zeo-
lite, (iv) the kinetic reaction scheme proposed, involving
both consecutive and competitive steps, accounts for the
SCHEME 2. Glycosylation pathway involving microreversibility
principle.
stereoelectronic assistance after protonation, hydrolyzes
faster than the -anomer which cannot adopt a twist-boat
conformation favoring stereoelectronic assistance.
higher
/ observed with Y-faujasites as due to the result
of stereoelectronic effects associated with the shape selec-
tive propertiesofthe catalystsin the transition states, (v) the
factors of reactivity invoked in the direct glycosylation reac-
tion are of the same nature as those involved in the reverse
reaction, i.e., hydrolysis of alkyl-D-glucopyranosides, thus
allowing for the first time the consideration of the principle
of microreversibility for reactions catalyzed by solids, and,
finally, (vi) such a process could be easily transposed to the
synthesis of long chain alkyl-D-glucosides.
For hydrolysis of methyl - and -D-glucopyranosides
in the presence of protonic zeolites, a similar behavior is
found (41). In the absence of steric constraints, stereo-
electronic effects will apply. Moreover, an additional effect
resulting from the presence of an electron deficient surface
is observed. The favorable orientation of the electron lone
pair(s) leads to both easier protonation and adsorption of
oxygen atoms. In that way, methyl -D-glucopyranoside was
found to hydrolyze faster than methyl -D-glucopyranoside
by a factor of around 6 in the presence of a HY zeolite with
a Si/Al ratio of 15 (41). However, if the internal pore diam-
eter is reduced as in H-form mordenites, the formation of
twist-boat forms is more difficult and the / ratio is equal
to 1 (42).
When the methyl substituent is replaced by a bulkier one
such as a glucose moiety as in disaccharides, maltose ( -
1-4 linkage), and cellobiose ( -1-4 linkage), it was shown
that maltose ( -anomer) hydrolyzed 3 times faster than
cellobiose ( -anomer) in the presence of the HY15 zeo-
lite (43). The rigidity of the disaccharide system is then re-
inforced because of the presence of microporous cavities,
and then the -linked anomer will hydrolyze faster since
REFERENCES
1. Hensen, H., Busch, P., Kra¨chter, H. U., and Tesmann, H., Tensid. Surf.
Det. 30, 116 (1993).
2. Mentech, J., Beck, R., and Burzio, F., in “Carbohydrates as Raw Or-
ganic Materials” (G. Descotes, Ed.), p. 185. VCH, Weinheim, 1993.
3. Hughes, F. A., and Lew, B. W., J. Am. Oil Chem. Soc. 47, 162
(1970).
4. Jeffrey, G. A., Acc. Chem. Res. 19, 168 (1986).
5. Fischer, E., Ber. Deutsch. Chem. Ges. 26, 2400 (1893).
6. Groen, J., WO 95/07915 (1995).
7. Van Bekkum, H., De Goede, A. T. J. W., Van Der Leij, I. G.,
Van Der Heijden, A. M., and Van Rantwijk, F., WO 96/36640 (1996).
8. Biermann, M., Schmid, K., and Schulz, P., Starch/Sta¨rke 45, 281
(1993).

SYNTHESIS OF BUTYL-D-GLUCOPYRANOSIDES OVER ZEOLITES
453
9. Straathof, A. J. J., Romein, J., Van Rantwijk, F., Kieboom, A. P. G.,
and Van Bekkum, H., Starch/Sta¨rke 39, 362 (1987).
25. Richard, F., Carreyre, H., and Pe´rot, F., J. Mol. Catal. A Chemical 103,
51 (1995).
10. Straathof, A. J. J., Van Bekkum, H., and Kieboom, A. P. G.,
Starch/Sta¨rke 40, 229 (1988).
26. Camblor, M. A., Corma, A., Iborra, S., Miquel, S., Primo, J., and
Valencia, S., J. Catal. 172, 76 (1997).
11. Brochette, S., Descotes, G., Bouchu, A., Queneau, Y., Monnier, N.,
and Petrier, C., J. Mol. Catal. A Chemical 123, 123 (1997).
27. Smirnyagin, V., and Bishop, C. T., Can. J. Chem. 46, 3085 (1968).
28. Mowery, D. F., J. Am. Chem. Soc. 77, 1667 (1955).
12. Corma, A., Iborra, S., Miquel, S., and Primo, J., J. Catal. 161, 713 29. Meier, W. M., Olson, D. H., and Baerlocher, Ch., “Atlas of Zeo-
(1996).
lite Structure Types,” 4th revised version. Butterworth-Heinemann,
London,1996.
30. Homans, S. W., Biochem. 29, 9110 (1990).
13. Matsumura, S., Imai, K., Yoshikawa, S., Kawada, K., and Uchibori, T.,
J. Am. Oil Chem. Soc. 67, 996 (1990).
14. Goodby, J. W., Haley, J. A., Mackenzie, G., Watson, J., Plusquellec, D., 31. Angyal, S. J., Adv. Carbohydr. Chem. Biochem. 42, 15 (1984).
and Ferrieres, V., J. Mater. Chem. 5, 2209 (1995).
32. Senderowitz, H., Parish, C., and Clark Still, W., J. Am. Chem. Soc. 118,
15. Chapat, J.-F., and Moreau, C., Carbohydr. Lett. 3, 25 (1998).
2078 (1996).
16. (a) Joffre, J., Geneste, P., Guida, A., Szabo, G., and Moreau, C., Stud. 33. Senderowitz, H., and Clark Still, W., J. Org. Chem. 62, 1427 (1997).
Phys. Theor. Chem. 71, 409 (1989); (b) Joffre, J., and Joffre F., AnaCin, 34. Lemieux, R. U., and Chu¨, N. J., Abstr. Pap. Am. Chem. Soc. 133, 31N
Langage et informatique, Toulouse (1995).
(1958).
17. Bond, G. C., “Heterogeneous Catalysis: Principle and Applications,”
Oxford Chemistry Series No. 34 (P. W. Atkins, J. S. E. Holker, and
A. K. Holliday, Eds.), 2nd Ed., Clarendon Press, Oxford, 1987.
18. Weisz, P., Chem. Eng. Prog. 55, 29 (1957).
35. Lemieux, R. U., Pavia, A. A., Martin, J. C., and Watanabe, K. A., Can.
J. Chem. 47, 4427 (1969).
36. Deslongchamps, P., “Stereoelectronic Effects in Organic Chemistry”
(J. E. Baldwin, Ed.), Organic Chemistry Series, Vol. I. Pergamon Press,
Oxford, 1983.
19. Capon, B., Chem. Rev. 69, 440 (1969).
20. Collins, P., and Ferrier, R., “Monosaccharides.” Wiley, Chichester,
1995.
37. Deslongchamps, P., Pure App. Chem. 65, 1161 (1993).
38. Be Miller, J. N., and Doyle, E. R., Carbohydr. Res. 20, 23 (1971).
21. Levene, P. A., Raymond, A. L., and Dillon, R. T., J. Biol. Chem. 95, 39. Lemieux, R. U., and Morgan, A. R., Can. J. Chem. 43, 2205 (1965).
699 (1932).
40. Van Eickeren, P., J. Org. Chem. 45, 4641 (1980).
41. Le Strat, V., and Moreau, C., Catal. Lett. 51, 219 (1998).
42. Le Strat, V., and Moreau, C., unpublished results.
43. Moreau, C., Durand, R., Duhamet, J., and Rivalier, P., J. Carbohydr.
Chem. 16, 709 (1997).
22. Shafizadeh, F., Adv. Carbohydr. Chem. 13, 9 (1958).
23. Lecomte, J., Finiels, A., Geneste, P., and Moreau, C., J. Mol. Catal. A
Chemical 133, 283 (1998).
24. Barthomeuf, D., Mater. Chem. Phys. 17, 49 (1987).