Mixed cellulose-derived benzoates bonded on allylsilica gel as HPLC chiral stationary phases: Influence of the introduction of an aromatic moiety in the fixation substituent

Article abstract of DOI:10.1016/S0957-4166(03)00176-9

Several mixed alkenoxybenzoyl/benzoates and 10-undecenoyl/benzoates of cellulose were fixed onto allylsilica gel by the radical coupling of double bonds. The introduction of an aromatic group in the fixation substituent modifies the chiral recognition abi

Full text of DOI:10.1016/S0957-4166(03)00176-9

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 14 (2003) 1179–1185
Mixed cellulose-derived benzoates bonded on allylsilica gel as
HPLC chiral stationary phases: influence of the introduction of
an aromatic moiety in the fixation substituent
Jordi Garce´s,a,† Pilar Franco,a,‡ Laureano Oliverosb and Cristina Minguillo´na,*
aLaboratori de Qu´ımica Farmace`utica, Facultat de Farma`cia, Universitat de Barcelona, Avda. Diagonal s/n,
E-08028 Barcelona, Spain
bLaboratoire de Chimie Ge´ne´rale, Conservatoire National des Arts et Me´tiers, 292, rue Saint-Martin,
F-75141 Paris Cedex 03, France
Received 8 January 2003; accepted 14 February 2003
Abstract—Several mixed alkenoxybenzoyl/benzoates and 10-undecenoyl/benzoates of cellulose were fixed onto allylsilica gel by the
radical coupling of double bonds. The introduction of an aromatic group in the fixation substituent modifies the chiral recognition
ability of the resulting chiral stationary phases (CSPs) in comparison with the 10-undecenoate/benzoate cellulose derivatives.
Better enantioselectivity values are achieved when the electronic and geometric characteristics of both substituents, fixating and
derivatizing, are similar. © 2003 Elsevier Science Ltd. All rights reserved.
1. Introduction
resolving behaviour could be considered as a conse-
quence of two main factors: the lack of homogeneity in
the substitution of the cellulose derivative, containing
an aliphatic moiety on some of the hydroxyl groups,
and the influence that the process of fixation can cause
on the discrimination ability. It has been shown that
fixation of 10-undecenoyl/phenylcarbamates or ben-
zoates of polysaccharides to allylsilica gel takes place
either by heterogeneous coupling of the double bonds,
on the chromatographic matrix and on the polysaccha-
ride derivative, or by reticulation of the 10-undecenoyl
groups themselves.8 Both processes occur simulta-
neously and may affect the chiral discrimination ability
of the CSPs by deforming the secondary structure of
the polysaccharide. It is generally accepted that this
conformation plays an important role in the enantiose-
lectivity of the resulting CSP.1 Moreover, the presence
of a certain number of alkenoyl chains on the cellulose
derivative implies a decrease in the total number of aryl
residues, usually related with the chiral recognition
phenomenon.9
Chiral stationary phases (CSPs) based on polysaccha-
ride derivatives have been successfully used in liquid
chromatography for the separation of enantiomers.1
However, the solubility of these derivatives in certain
solvents limits their applicability in HPLC. The differ-
ent approaches for the fixation of the polysaccharide
derivative to the chromatographic matrix2 have allowed
the use of a broader variety of solvents as mobile phase
modifiers, in contrast to the classic hydrocarbon/alco-
hol mixtures which are compatible with the coated
phases.3–5
The 10-undecenoate/phenylcarbamates6 or benzoates7
of cellulose bonded on chromatographic matrices devel-
oped by our research group show certain differences in
their chiral discrimination ability from that of the
homosubstituted coated supports. These variations in
* Corresponding author. Present address: IRBB-Parc Cient´ıfic de
Barcelona, Josep Samitier, 1-5, E-08028 Barcelona, Spain. Tel.:
+34-93-403-7107; fax: +34-93-403-7114; e-mail: cminguillon@
pcb.ub.es
In the present study, in order to retain the structural
regularity in the polymer and, hence, the secondary
helical structure of the polysaccharide, while maintain-
ing the fixing capability, long chain alkenoxybenzoyl
groups have been introduced on cellulose instead of the
† Present address: AlconCus´ı R&D, Analytical Chemistry Unit, Camil
Fabra, 58, E-08320 El Masnou (Barcelona), Spain.
‡ Present address: Chiral Technologies Europe, Parc d’Innovation,
Bd. Gonthier d’Andernach, F-67400 Illkirch, France.
0957-4166/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0957-4166(03)00176-9

1180
J. Garce´s et al. / Tetrahedron: Asymmetry 14 (2003) 1179–1185
alkenoyl groups formerly used (Fig. 1). The synthesis
and the chromatographic results obtained with 4-
methylbenzoyl and 4-methoxybenzoyl derivatives of
cellulose are here compared to the analogous deriva-
tives fixed through 10-undecenoyl groups.
pared to that shown by 10-undecenoyl chloride under
the same conditions. Therefore, longer reaction times
were needed to obtain similar substitution degrees.
Different ratios of substituents were introduced on
derivatives A and A% to assess the effect of an increased
reticulation of the derivative on the chromatographic
behaviour of the resulting CSP. Derivatives containing
10-undecenoyl groups (C and D) were prepared for
comparative purposes (Scheme 2). All derivatives were
2. Results and discussion
1
2.1. Preparation of the chromatographic supports
characterised by their H NMR spectra and elemental
analyses (Table 1).13
4-(10-Undecenyloxy)benzoic acid was obtained as
described elsewhere.10 Thus, the 10-undecenyl p-toluen-
sulfonate, prepared from 10-undecen-1-ol, was used in
the alkylation of ethyl 4-hydroxybenzoate. The
obtained ester was later hydrolysed and the resulting
acid was transformed into the corresponding acyl chlo-
ride 1 by the treatment with thionyl chloride (Scheme
1).
The obtained cellulose derivatives were fixed on allylsil-
ica gel, previously prepared from spherical silica gel
(Nucleosil® 100-5) and allyltriethoxysilane. An addi-
tional treatment of ‘end-capping’ with hexamethyldisi-
lazane was performed on the chromatographic matrix
before fixation.7 The resulting CSPs were characterised
by their elemental analyses (Table 1). The new aromatic
alkenyl group were less efficient than 10-undecenoyl in
the fixation of the cellulose derivative. Only A%, with a
higher content of fixing groups, was immobilised in
comparable ratio to C. All CSPs obtained were slurry
packed into stainless-steel tubes in order to be chro-
matographically tested.
Cellulose derivatives were prepared following the
method described for the 10-undecenoate/3,5-
dimethylphenylcarbamate of cellulose.11 Thus, cellulose
was allowed to react successively with the prepared
alkenoxybenzoyl chloride 1 and the appropriate ben-
zoyl chloride (Scheme 2). Reaction time for the intro-
duction of the first reagent was set to yield substitution
degrees of fixing groups in the order of 0.2–0.3 groups
per glucose unit. This amount of fixing groups was
2.2. Chromatographic results
determined to be the optimal amount to fix 10-unde-
cenoyl derivatives onto 100 A allylsilica gel. The
aromatic reagent 1 showed a reduced reactivity com-
The resulting HPLC columns were tested in normal
phase conditions using heptane/2-propanol and hep-
tane/chloroform mixtures as the mobile phase. A selec-
12
,
Figure 1. Type of substitution of the synthesised cellulose benzoates.
Scheme 1. (a) K2CO3, butanone, reflux; (b) KOH, EtOH/H2O, 75°C; (c) Cl2SO.

J. Garce´s et al. / Tetrahedron: Asymmetry 14 (2003) 1179–1185
1181
tion of the chromatographic results for a series of
racemic compounds (Fig. 2) is presented in Table 2.
of more fixing groups on cellulose derivative A% allowed
the obtention of a CSP with a polymeric content com-
parable to that of CSPC. The higher number of
alkenoxybenzoyl groups on the cellulose derivative
(CSPA% versus CSPA) did not seem to have a negative
influence on the chiral resolution properties, as it has
been observed in previous studies with 10-undecenoyl
derivatives. In these latter the increase in the number of
fixing moieties on the polysaccharide resulted in an
important decrease in enantioselectivity and resolu-
tion,8,11 mainly attributable to the reticulation process
undergone by these groups. However, the effect of this
enhanced reticulation degree was not noticeable on
CSPA%. Selectivity values are only slightly reduced on
CSPA% regarding those on CSPA and CSPC (Fig. 3).
Only compound 6 is affected by a considerable reduc-
tion of selectivity in this CSP (Fig. 4). This fact suggests
a stabilisation of the secondary structure of the polysac-
charide derivative in spite of the reticulation process.
As a first approach, CSPA was prepared in order to be
compared with the already described CSPC as cellulose
derivative C was among the best chiral selectors previ-
ously studied.7 Due to the known strong effect of the
substituent position on enantioselectivity,14 the fact of
having a 4-substituted aromatic ring, as in the fixing
substituent, was considered an important factor to keep
a geometric regularity. In spite of the chemical differ-
ences between A and C, when the chromatographic
behaviour of CSPA and CSPC was compared, no
significant changes in enantioselectivity were observed.
Although cellulose derivative
A had substitution
degrees comparable to those of derivative C for both
types of substituents, the former yielded a support with
less cellulose content than the latter. The introduction
Scheme 2. (a) and (b) pyridine, reflux; (c) 2% w/w AIBN, 2 h, 100°C, without solvent. A different amount of reagent 1 was used
to obtain cellulose derivatives A and A%.
Table 1. Characterisation of cellulose derivatives and chiral stationary phases
Cellulose
derivative
Analyses of
cellulose derivatives
Substitution of glucose unitsa Chiral supports
Analyses of the final g cellulose derivative/
chiral supports
100 g phaseb
% C
% H
Unsaturat.
Ar
% C
% H
A
A%
B
C
D
69.26
71.45
64.41
69.11
64.87
5.91
6.80
5.40
5.99
5.70
0.2390.05
0.8890.12
0.2090.05
0.2990.06
0.2290.05
2.3790.16
1.9890.25
2.5190.25
2.4390.14
2.5590.21
CSPA
CSPA%
CSPB
CSPC
CSPD
11.32
14.08
9.59
14.19
9.51
1.65
2.00
1.52
1.65
1.39
12.5
15.9
10.4
16.2
10.5
a The maximum substitution degree of a glucose unit being 3 (number of OH). Based on 1H NMR and elemental analyses (calculations as reported
in Ref. 13).
b Based on the elemental analyses of the chiral supports.

1182
J. Garce´s et al. / Tetrahedron: Asymmetry 14 (2003) 1179–1185
Figure 2. Chemical structures of the racemic test compounds.
Table 2. Chromatographic results
Racemic
compd
CSPA%
CSPA
CSPC
CSPB
h
CSPD
Mobile
phase
k1
h
Rs
k1
h
Rs
k1
h
Rs
k1
Rs
k1
h
Rs
1
2
3
4
5
6
7
8
8.05 1.15
0.79
1.22
–
6.59 1.13
0.19
0.33
–
4.98
4.64
3.06
9.08
3.48
3.55
2.17
4.71
0.67
0.74
3.09
7.70
1.18
1.24
1.10
1.00
1.23
2.21
1.29
1.21
1.35
1.67
1.06
1.08
0.82
1.31
–
12.3
7.81 1.19
6.65 1.00
1.00
–
–
–
8.49 1.00
7.04 1.00
5.18 1.00
–
–
–
0.55
0.64
–
80:20a
80:20a
80:20a
98:2a
7.17 1.20
4.77 1.09
8.09 1.06
3.92 1.19
3.94 1.46
2.79 1.38
5.22 1.14
0.71 1.31
0.89 1.54
2.06 1.05
7.00 1.24
3.91 1.00
3.57 1.00
3.53 1.21
2.97 1.89
2.29 1.31
4.64 1.20
0.69 1.38
0.78 1.66
2.84 1.00
–
–
–
13.0
1.24
4.56 1.18
3.70 1.34
1.67d 1.00
5.85 1.12
0.71 1.45
0.79 1.64
2.61 1.12
1.74
1.27
0.97
–
0.96
1.77
2.33
–
10.6
1.15
3.73 1.19
2.36 1.00
1.59 1.00
4.79 1.10
0.63 1.43
0.63 1.66
2.15 1.00
2.05
2.45
1.69
1.72
1.85
3.50
–
1.54
2.41
–
1.65
1.73
2.95
–
1.90
3.98
0.85
2.21
1.83
2.94
–
98:2a
98:2a
–
98:2a
0.31
0.97
1.19
–
98:2a
9
10
11
98:2a
98:2a
98:2a
Naproxen 10.0
1.11
1.31
10.4
1.10
–
0.87
12.2
1.12
–
11.2
1.13
–
98:2:0.5b
1
2
5
7
9
10
15.3
1.07
0.86
1.72
–
0.53
1.23
1.81
11.2
1.00
–
1.33
–
–
–
9.09
4.82
6.71
2.31
0.52
0.53
1.05
1.23
1.04
1.12
1.19
1.34
–
2.03
–
0.94
0.77
1.40
15.2
1.13
0.98
1.35
–
–
1.50
2.17
6.39 1.12
3.35 1.33
7.06 1.00
1.50 1.00
0.47 1.00
0.40 1.46
0.71
1.20
–
–
–
75:25c
90:10c
95:5c
95:5c
95:5c
95:5c
5.46 1.16
5.08e 1.04
3.91 1.17
0.78 1.20
0.84 1.31
4.24 1.23
6.40 1.04
3.23 1.09
1.95 1.13
0.65 1.31
7.83 1.31
8.82 1.09
3.89 1.20
0.87 1.38
0.92 1.61
0.65
–
Mobile phase: a Heptane/2-propanol, b Heptane/2-propanol/TFA, c Heptane/chloroform, d Heptane/2-propanol (90:10), e Heptane/chloroform
(90:10).
The presence of the aromatic ring, together with a long
enough aliphatic chain, contributes to the maintainance
of the structural regularity of the chiral selector. Fur-
thermore, the presence of a higher number of fixing
groups led to an increase of cellulose derivative content
in the resulting CSP, which might be responsible of the
slight increase in retention for CSPA% compared to
CSPA.

J. Garce´s et al. / Tetrahedron: Asymmetry 14 (2003) 1179–1185
1183
Figure 3. Resolution of (I) naproxen (mobile phase: heptane/2-propanol/TFA (98:2:0.5); flow: 0.5 mL/min; u, 230 nm) and (II) 9
(mobile phase heptane/chloroform (95:5); Flow: 1 mL/min; u, 254 nm).
similarities with the coated material described by
Okamoto et al.15 and prepared with homogeneously
substituted 4-methoxybenzoate of cellulose.
In order to obtain polysaccharide derivatives with a
higher uniformity in terms of structural features, but
also of electronic properties of substituents, the prepa-
ration of two new cellulose derivatives B and D with a
4-alkoxyderivatization was undertaken. Both deriva-
tives were prepared and bonded on allylsilica gel,
although the percentage of cellulose on the CSPs was
not higher than 10.5 (see Table 1). This lower cellulose
content does not seem to be related to the bulkiness of
the new fixing agent, as this phenomenon was observed
for both derivatives. A reduced fixation degree was
already observed for some 10-undecenoate/benzoates of
cellulose7 (i.e. 10-undecenoate/3,5-methoxybenzoate),
which were fixed in a lower extent to the matrix than
their corresponding methyl-substituted analogues.
In general terms, the improvements in selectivity
achieved with CSPA/A% and CSPB compared to CSPC
and CSPD, respectively, allowed us to make an evalua-
tion of the effect of the use of 10-undecenoyl groups as
fixing substituents on selectivity. Although the existence
of a negative effect is shown by the obtained data, its
relevance depends on the derivative considered. Never-
theless, the ready availability and ease of use of 10-
undecenoyl chloride make it still a suitable choice for
the fixation of polysaccharide derivatives on chromato-
graphic matrices.
When the chromatographic behaviours of CSPB and
CSPD are compared, an improvement in the chiral
recognition ability can be observed as a result of the
introduction of an aromatic ring in the fixing sub-
stituent (Fig. 4). That is, when the electronic and
geometric characteristics of the main derivatizing and
the fixing agents are the same. This improvement is
even more noticeable due to the low recognition ability
shown by CSPD for some of the test compounds.
When comparison was possible, CSPD showed notable
3. Experimental
1H NMR spectra were recorded on a Varian (Palo
Alto, CA) GEMINI-300 Spectrometer with pyridine-d5
at 70°C. Elemental analyses were performed in a CE
Instruments apparatus (model EA 1108, Carlo Erba,
Milan, Italy) using standard conditions by the Serveis
Cient´ıfico-Te`cnics de la Universitat de Barcelona
(Spain). The CSPs were packed into stainless-steel tubes

1184
J. Garce´s et al. / Tetrahedron: Asymmetry 14 (2003) 1179–1185
Figure 4. Resolution of 6 using heptane/2-propanol (90:10) as mobile phase (Flow: 1 mL/min; u, 230 nm).
(150×4.6 mm I.D.) by the slurry method. The chro-
matographic experiments were performed on an HPLC
system consisting of a Waters 600E pump, a Waters 717
autosampler (Millipore, Milford, MA, USA) and
equipped with a Waters 996 photodiode array detector
reagents are indicated in Table 3. The appropriate
benzoyl chloride was added dropwise (4-methylbenzoyl
chloride for A and A%, or 4-methoxybenzoyl chloride
for B) and the mixture was refluxed (Table 3). The
resulting products were isolated as the insoluble frac-
tion in methanol. They were redissolved in chloroform,
reprecipitated in methanol and thoroughly washed in
this solvent. All derivatives were characterised by their
1H NMR spectra and elemental analyses (Table 1).
and
a Perkin–Elmer 241LC polarimetric detector
(Perkin–Elmer, Uberlingen, Germany). The volume of
sample injected was 3 mL. The void volume was deter-
mined using tri-tert-butylbenzene.
Avicel® and 4-methoxybenzoic acid were purchased
from Merck (Darmstadt, Germany). Tosyl chloride
(TsCl), 10-undecen-1-ol, 10-undecenoyl chloride, ethyl
4-hydroxybenzoate, thionyl chloride, a,a%-azoisobuty-
ronitrile (AIBN) and 4-methylbenzoic acid were sup-
plied by Fluka (Buchs, Switzerland). Nucleosil 100-5
Table 3. Molar ratio of reagents and reaction times for
the preparation of the cellulose benzoates
Cellulose derivative
Fixing reagenta
td (h)
Benzoate reagentb
Mc
Mc
td (h)
was
Germany).
a
product from Macherey-Nagel (Du¨ren,
A
A%
B
C
D
0.4
1.1
0.4
0.4
0.4
24
4
24
2
5.0
3.0
4.0
4.0
4.0
48
24
48
16
16
3.1. Preparation of the cellulose derivatives A, A% and
B
2
a Compound 1 for A, A% and B; 10-undecenoyl chloride for C and D.
b 4-Methylbenzoyl chloride for A, A% and C; 4-methoxybenzoyl chlo-
ride for B and D.
Cellulose (Avicel®, 1 g, 6.2 mmol of glucose units)
suspended in dry pyridine (50 mL) and 4-(10-undecenyl-
oxy)benzoyl chloride 1 in different proportions were
stirred under reflux. Reaction times and molar ratios of
c M=mol of reagent/mol glucose units.
d Time of reaction in pyridine at reflux temperature.

J. Garce´s et al. / Tetrahedron: Asymmetry 14 (2003) 1179–1185
1185
4-(10-Undecenyloxy)benzoate/4-methylbenzoate of cel-
lulose A 1H NMR (300 MHz, pyridine-d5, 70°C): l
1.2–1.9 (m, C3%H2-C8%H2); 2.03, 2.08 and 2.31 (3s+m,
2-Ar-CH3, 3-Ar-CH3, 6-Ar-CH3 and C9%H2); 3.72 (m,
C2%H2 and C5H); 4.12 (m, C4H); 4.45 (m, C6H2); 4.9–5.2
(m, C1H and C11%H2); 5.62 (m, C2H); 5.90 (m, C3H and
C10%H); 6.80, 7.01 and 7.19 (3d, 2-C3%%,5%%H, 3-C3%%,5%%H and
chloroform and heated to the refluxing temperature for
2 h. The resulting suspension was filtered off and the
solid washed with chloroform, tetrahydrofuran and
acetone. The materials thus obtained were characterised
by their elemental analyses (Table 1).
6-C3%%,5%%H); 7.84, 7.94 and 7.99 (3d, 2-C2%%,6%%H, 3-C2%%,6%%
and 6-C2%%,6%%H).
H
Acknowledgements
4-(10-Undecenyloxy)benzoate/4-methoxybenzoate
of
Financial support from the Direccio´n General de Ense-
n˜anza Superior, Ministerio de Educacio´n y Cultura of
Spain (Project No. PB96-0382) is acknowledged.
1
cellulose B H NMR (300 MHz, pyridine-d5, 70°C): l
1.2–1.9 (m, C3%H2-C8%H2); 2.12 (m, C9%H2); 3.56 (s, 2-
ArOCH3 and 3-ArOCH3); 3.70 (m, C2%H2 and C5H);
3.80 (s, 6-ArOCH3); 4.12 (m, C4H); 4.55 (m, C6H2);
4.9–5.2 (m, C1H and C11%H2); 5.65 (m, C2H); 5.95 (m,
C3H and C10%H); 6.61, 6.78 and 6.99 (3d, 2-C3%%,5%%H,
3-C3%%,5%%H and 6-C3%%,5%%H); 7.93, 8.06 and 8.08 (3d, 2-
C2%%,6%%H, 3-C2%%,6%%H and 6-C2%%,6%%H).
References
1. Yashima, E. J. Chromatogr. A 2001, 906, 105–125.
2. Franco, P.; Senso, A.; Oliveros, L.; Minguillo´n, C. J.
Chromatogr. A 2001, 906, 155–170.
3.2. Preparation of the cellulose derivatives C and D
3. Oliveros, L.; Lo´pez, P.; Minguillo´n, C.; Franco, P. J. Liq.
Chromatogr. 1995, 18, 1521–1532.
Derivatives C and D were synthesised analogously to A
and B by the reaction of cellulose (Avicel®, 1 g, 6.2
mmol of glucose units) with 10-undecenoyl chloride
(2.3 mmol) and 22.3 mmol of the appropriate benzoyl
chloride (4-methylbenzoyl chloride for C, or 4-
methoxybenzoyl chloride for D), as described
elsewhere.7
4. Oliveros, L.; Minguillo´n, C.; Serkiz, B.; Meunier, F.;
Volland, J. P.; Cordi, A. J. Chromatogr. A 1996, 729,
29–32.
5. Franco, P.; Minguillo´n, C.; Oliveros, L. J. Chromatogr. A
1998, 793, 239–247.
6. Oliveros, L.; Senso, A.; Franco, P.; Minguillo´n, C. Chi-
rality 1998, 10, 283–288.
10-Undecenoyl/4-methoxybenzoate of cellulose D 1H
NMR (300 MHz, pyridine-d5, 70°C): l 0.8–2.40 (m,
C2%H2-C9%H2); 3.56 and 3.57(s+s, 2-ArOCH3 and 3-
ArOCH3); 3.70 (m, C5H); 3.80 (s, 6-ArOCH3); 4.20 (m,
C4H); 4.51 (m, C6H2); 4.9–5.2 (m, C1H and C11%H2);
5.65 (m, C2H); 5.95 (m, C3H and C10%H); 6.61, 6.78 and
6.99 (3d, 2-C3%%,5%%H, 3-C3%%,5%%H and 6-C3%%,5%%H); 7.93, 8.06
and 8.08 (3d, 2-C2%%,6%%H, 3-C2%%,6%%H and 6-C2%%,6%%H).
7. Oliveros, L.; Senso, A.; Minguillo´n, C. Chirality 1997, 9,
145–149.
8. Franco, P.; Minguillo´n, C.; Oliveros, L. J. Chromatogr. A
1997, 791, 37–44.
9. Okamoto, Y.; Kaida, Y. J. Chromatogr. A 1994, 666,
403–419.
10. Kelly, S. M.; Buchecker, R. Helv. Chim. Acta 1988, 71,
461–466.
11. Minguillo´n, C.; Franco, P.; Oliveros, L. J. Chromatogr. A
1996, 728, 415–422.
3.3. Preparation of the chiral stationary phases (CSPs)
12. Minguillo´n, C.; Franco, P.; Oliveros, L.; Lo´pez, P. J.
Chromatogr. A 1996, 728, 407–414.
13. Senso, A.; Franco, P.; Oliveros, L.; Minguillo´n, C. Car-
bohydr. Res. 2000, 329, 367–376.
14. Francotte, E.; Wolf, R. M. J. Chromatogr. 1992, 595,
63–75.
A 2 g amount of allylsilica gel, prepared as previously
described,7 were added to a solution of 0.4 g of the
appropriate cellulose derivative in 20 mL of chloro-
form, containing a 2% w/w of AIBN (a,a%-azobisisobu-
tyronitrile). After evaporation of the solvent at room
temperature, the solid material was allowed to react for
2 h at 100°C. The CSP thus obtained was suspended in
15. Okamoto, Y.; Aburatani, R.; Hatada, K. J. Chromatogr.
1987, 389, 95–102.