Reduced shrinkage of sol-gel derived silicas using sugar-based silsesquioxane precursors

Article abstract of DOI:10.1039/b502959g

Monolithic siliceous materials were prepared, using sol-gel based methods, from mixtures of trifunctional silanes based on sugar lactones, including silyl-modified gluconamide GLS and maltonamide MLS, and a tetrafunctional silane derived from glycerol. The tri- and tetrafunctional compounds cured at different rates, which led to an enhanced presence of sugar moieties at the external surface of the pores in the monoliths. The resulting silicas exhibited dramatically reduced degrees of shrinkage (<10%) when compared to silica monoliths prepared in the absence of trifunctional silanes (up to 85%). The sugars also alter the morphology of the material, with significant reductions in both micropore volume and surface area for materials containing GLS. The reduced shrinkage, presence of sugars on the silica surface, and altered morphology are likely to be important factors in providing such materials with the ability to stabilize entrained proteins. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b502959g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Reduced shrinkage of sol–gel derived silicas using sugar-based
silsesquioxane precursors{
Yang Chen, Zheng Zhang, Xihua Sui, John D. Brennan and Michael A. Brook*
Received 28th February 2005, Accepted 25th May 2005
First published as an Advance Article on the web 22nd June 2005
DOI: 10.1039/b502959g
Monolithic siliceous materials were prepared, using sol–gel based methods, from mixtures of
trifunctional silanes based on sugar lactones, including silyl-modified gluconamide GLS and
maltonamide MLS, and a tetrafunctional silane derived from glycerol. The tri- and
tetrafunctional compounds cured at different rates, which led to an enhanced presence of sugar
moieties at the external surface of the pores in the monoliths. The resulting silicas exhibited
dramatically reduced degrees of shrinkage (,10%) when compared to silica monoliths prepared in
the absence of trifunctional silanes (up to 85%). The sugars also alter the morphology of the
material, with significant reductions in both micropore volume and surface area for materials
containing GLS. The reduced shrinkage, presence of sugars on the silica surface, and altered
morphology are likely to be important factors in providing such materials with the ability to
stabilize entrained proteins.
shrinkage levels during drying of TEOS-derived silica (up to
Introduction
85% shrinkage) to as low as 50% over 2 months under ideal
conditions. While this represents a significant improvement in
the sol–gel synthesis, it is insufficient when the protein-doped
gels are to be used in applications such as chromatography, for
which void spaces resulting from shrinkage or cracking are
significantly detrimental to performance. In addition, while
DGS based materials provided improved stability for many
proteins relative to TEOS based materials, there were still
significant losses in activity for many proteins with aging, and
a complete lack of activity for other proteins.11
The use of sol–gel processing methods provides an exceptional
degree of control over the composition and morphology of
silicate materials. Thus, total porosity, pore size and shape,
regularity of pore distribution, etc., can be manipulated using a
variety of starting materials, reaction conditions and dopants.1
Many of these conditions, however, are incompatible with
the incorporation of fragile compounds such as biomolecules,
proteins in particular. Either the synthetic conditions are
damaging to protein structure (e.g., pH conditions, the
presence of denaturants such as ethanol) or the final curing
conditions require elevated temperatures. Furthermore, many
materials undergo significant changes as a result of aging (i.e.,
shrinkage, pore collapse, alterations in internal polarity and
water content, etc.) that lead to poor long-term stability of
entrapped proteins. Thus, it is of interest to prepare improved
materials that can incorporate active biomolecules in silica
matrices to create materials that can serve as biosensors,
immobilized enzyme reactors or affinity chromatography
supports.2–7
One method that was recently reported for creating highly
biocompatible sol–gel derived materials was to include trifunc-
tional silanes bearing covalently bound sugars into DGS
derived silica to produce sugar-modified silica. Sugars are
known to be stabilizing toward proteins,9,12 and indeed sugar-
modified materials were shown to be amenable to entrapment
of fragile enzymes including urease, Factor Xa, luciferase
and Src kinase.13 The contributing factors for protein
stabilization included the presence of the glycerol, and absence
of denaturants, lower monolith shrinkage and higher water
retention.13 More recent studies of entrapped human serum
albumin (HSA) demonstrated that initial conformation,
accessibility to external analytes, thermal stability, long-term
stability and degree of ligand binding to HSA were best in
DGS-derived materials that contained covalently tethered
gluconamidylsilane (GLS) moieties relative to unmodified
DGS-derived materials, TEOS or TEOS–GLS-derived
materials. Measurement of protein rotational dynamics
showed that entrapment led to an immediate loss of global
motion in all materials. However, the restriction of motion
was most dramatic in GLS doped materials, suggesting
preferential interactions of the protein with the sugar coated
surfaces.
Several strategies have been developed to mitigate the
effects of aging on sol–gel derived silica. Gill and Ballesteros
described poly(glycerylsilane)s (PGS) as precursors for silica
that exhibited reduced shrinkage and increased pore sizes.8
Brook et al. have reported a series of sugar, sugar alcohol,
oligo- and polysaccharide-derived silanes as precursors to
protein compatible silicas.9,10 The use of these latter com-
pounds as starting materials, particularly diglycerylsilane
(DGS) and monosorbitylsilane, reduced the normally high
Department of Chemistry, McMaster University, 1280 Main St. W.,
Hamilton, ON, Canada L8S 4M1. E-mail: mabrook@mcmaster.ca;
Fax: +1 905 522 2509; Tel: +1 905 525 9140 ext. 23483
{ Electronic supplementary information (ESI) available: Figures of
ATR FT-IR, FT-IR, 13C NMR and 29Si NMR data for samples 3–5
and 8–10. See http://dx.doi/10.1039/b502959g/
Trifunctional silanes form silsesquioxanes upon hydro-
lysis,14,15 which generally show a lower degree of crosslinking
3132 | J. Mater. Chem., 2005, 15, 3132–3141
This journal is ß The Royal Society of Chemistry 2005

View Article Online
than materials derived from tetrafunctional silanes. In addi-
tion, trifunctional silanes are frequently used as surface
modifying materials, as coupling agents or adhesion pro-
moters.16 In cases where trifunctional and tetrafunctional
silanes are co-hydrolyzed and condensed, the trifunctional
materials will generally cure more slowly than the tetraalk-
oxysilane counterparts, depending on the specific kinetics at
the pH used for the hydrolysis/condensation sequences.17,18 As
a result, the non-hydrolyzed group on the trifunctional
silane is ultimately expected to be preferentially located on
the exposed surfaces of pores and microchannels in the
resulting siliceous material, which would change the nature
of the interactions between these surfaces and the entrapped
biomolecules.
Results
Preparation of coupling agents
Two classes of coupling agents based on saccharides were
prepared, those derived from a mono-GLS (from glucon-
amide), and a disaccharide MLS (from maltose), respectively
(Table 1). Although a variety of approaches exist to graft
sugars to silanes, we chose to modify the anomeric hemiacetal
centre at the terminus of the saccharidic chains. Oxidation of
either of the sugars,23 normally performed with iodine,24
converts the anomeric hemiacetal into the lactone that can
then be opened by an amino-modified alkoxysilane to produce
a sugar-modified coupling agent.21 These materials were
characterized by standard spectroscopic techniques including
IR, 13C NMR and MS as summarized in Scheme 1. It should
be noted that while both sugars studied are reducing sugars,
the general protocol described herein should be amenable to
both reducing and non-reducing sugars.
In order to test this hypothesis, and with the desire to
develop matrices suitable for protein entrapment, we have
prepared materials containing mixtures of trifunctional
silanes based on sugars [GLS and the analogous malton-
amidylsilane (MLS)],19–21 and tetrafunctional silanes based
on DGS, and have examined the morpohology, shrinkage,
surface composition, shrinkage and retention of water by such
materials. The material properties are related to the protein
stabilizing properties of the materials, providing new insights
into why such materials are capable of maintaining protein
activity.22
Silica–silsesquioxane composites
Unlike many trifunctional silanes based on ethoxy and
methoxy groups, the compounds GLS and MLS are soluble
in water. Also unlike normal organically modified trialkoxy-
silanes,16 the sols of neither of these compounds alone gelled,
even after 6 months.
Table 1 13C NMR, FT-IR and ESI spectral data of gluconamide silane and maltonamide silane
Compound
13C
Glulactone
e 172.6
GLS
Maltlactone
MLS
Carbonyl
Anomeric
Methyne
e 172.9
e 172.9
e 172.6
k 101.4
f, g, i, l, m, p
72.5–73.8a
h 80.6
n 69.8
j1 61.3
j2 63.4
a 41.3
k 100.5
f, g, i, l, m, p
70.5–73.5a
h 80.4
n 69.8
j1 61.0
j2 61.3
f 81.9
g 72.0
h 68.4
i 74.3
j 60.8
f 74.2
g 72.1
h 70.7
i 73.0
j 64.0
a 41.5
b 23.3
c 7.8
Methylene
b 23.4
c 7.8
OCH2CH3
18.7 y 18.9
58.3 (overlapped)
1646
422.2 (100)(M + Na)+
400.2 (15)(M + 1)+
18.4 y 19.1
58.3a
1646
IR (neat, KBr) n(CLO) cm21
MS (ESI) m/z
1728
1747 (d-lactone)
584.3 (30)(M + Na)+
562.4 (20)(M + 1)+
a
A series of overlapped peaks.
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 3132–3141 | 3133

View Article Online
Scheme 1 Preparation of sugar coupling agents, shown for MLS.
Fig. 1 Shrinkage data of samples 1–18 over 45 d. DGS (1), DGS +
GLS (3–6), DGS + MLS, TEOS (11), TEOS + GLS (12–15), TEOS +
MLS (16–18). Recipes for these monoliths are provided in Tables 2–4.
A series of gels derived from the co-hydrolysis of DGS with
compounds GLS or MLS, respectively, led to organically
modified silicas possessing very different properties than silica
derived from DGS alone. Curing kinetics for the composites
depended upon the ratio of starting materials. There was a
trend to slower curing with an increase in the proportion of
the trifunctional component (Table 2). For practical reasons,
the amount of compounds GLS had to be limited to about
75 mol% (75 mol% GLS # DGS : GLS 1 : 6 w/w), and MLS
to about 40 mol%, above which curing did not occur
within about 1 week. In all cases, optically clear monolithic
silsesquioxane-modified silicas resulted after aging for one
week post gelation.
environment (i.e., not under water), occurs to a level of up to
approximately 65% (see sample 1, Fig. 1), much less than
TEOS-derived gels, which shrink approximately 85%.9 By
contrast, the incorporation of the sugar-modified trifunctional
groups GLS or MLS into a DGS gel dramatically reduces
shrinkage of the modified silicas, in some cases by an order of
magnitude to less than 10%, after drying in air over the same
lengthy (45 d) time period (samples 3–6, 8–10, Fig. 1).
For comparison, the effects of GLS or MLS on shrinkage
in TEOS gels were examined (Fig. 1, 11–18). After washing
and drying for 57 d, there was a marked decrease in shrinkage
(about 26) over TEOS alone (80%), 11, but the degree of
shrinkage in the TEOS–GLS material (42%) was still
significantly greater than was observed with the combination
of DGS and the sugar-based coupling agents (e.g., compare
TEOS–GLS sample 16 42% shrinkage vs. DGS–GLS sample 6
6% shrinkage, Fig. 1). This demonstrates that glycerol (in
DGS) and the saccharidic amides (from GLS or MLS) act
cooperatively to reduce shrinkage.
The physical behavior of the resulting silica–silsesquioxane
composites was studied. The most significant change between
the monolithic silicas derived from DGS alone, and those that
also included sugar-based coupling agent GLS and MLS, was
the degree of shrinkage of the siliceous products. Shrinkage is
directly linked to the degree of hydration of the monolith and
the presence of sugars in the gel. Unwashed samples 1, 3–6,
and 8–10, which still contained entrained glycerol, showed
no shrinkage over 45 d in air.25 Washed samples (see the
Experimental section), which contained the sugar (derived
from GLS and MLS) but not glycerol (see TGA results below),
similarly did not shrink if stored under water in closed
containers over 6 months. Normally, shrinkage of washed
samples of DGS-derived gels, when allowed to rest in an open
The reduced shrinkage of these monoliths could be a
consequence of the entrained sugars, and/or the accompanying
water. Thermogravimetric analysis (TGA) was utilized to
better understand the degree to which GLS and MLS and the
glycerol product of hydrolysis were physically and chemically
constrained within the monolith. A comparison was initially
made between unwashed gels and those that had been first
washed with water to remove ungrafted material.
All the samples contained significant amounts of water, as is
to be expected from the presence of hygroscopic sugar-
containing moieties. For example, the TGA of an unwashed
sample 1 showed loss of about 15% weight of water and a
further 12% weight change associated with loss of glycerol
above 150 uC (Fig. 2). After washing with water, the same
sample 1 had little residual weight loss (2.7%) after the
temperature was raised to 150 uC (data not shown). These
observations confirm that most of the glycerol formed by
hydrolysis of DGS is initially available in accessible micro- or
mesopores, or channels connecting the two, but can then be
easily removed by washing.
Table 2 Preparation of TQ resins
Sample Ratio of DGS : GLS (w/w) Gelation time/min Aged/d
1
2a
3
4
5
1 : 0
DGS : sorbitol, 4 : 1
4 : 1
3 : 1
2 : 1
1 : 1
10
60
65
70
90
90
7
7
7
7
20
20
6
Sample Ratio of DGS : MLS (w/w) Gelation time/min Aged/d
7b
8
DGS : maltose, 16 : 1
55
60
70
70
7
7
7
7
16 : 1
8 : 1
4 : 1
9
The kinetics of water loss were investigated from samples 1
and 4 (samples were washed, but not freeze dried). TGA, at
fixed temperature of 40 uC in air over an experimental period
of 75 min (Fig. 3), showed surprisingly linear drying kinetics
for physically absorbed water.26 Sample 4, containing grafted
10
a
This model system contained only DGS and sorbital, a sugar
b
analogue of GLS that cannot covalently bind into the matrix. This
model system contained only DGS and maltose, a sugar analogue of
MLS that cannot covalently bind into the matrix.
3134 | J. Mater. Chem., 2005, 15, 3132–3141
This journal is ß The Royal Society of Chemistry 2005

View Article Online
Table 3 TGA data on TEOS and sugar–TEOS gels
Sample
Sugar–TEOS (mol %)
TGA data (weight loss, %)
11
12
13
14
15
16
17
18
TEOS
GLS–TEOS
MLS–TEOS
2.15
10.3
16.9
25.7
31.5
11.2
27.9
37.7
5
10
15
20
5
10
15
carbohydrates in proporation to their concentration in the sol
was clearly demonstrated by TGA (Table 3). Significantly
more organic material could be pyrolyzed from samples 3–5
and 8–10 (10–38% wt, Fig. 2, Table 4) than from the silica
derived solely from DGS. Colour (yellow A dark brown)
evolved in samples 3–5 and 8–10 starting about 200 uC due to
a caramelization of the grafted sugar. In the case of TEOS
sols 11, to which the sugar-based coupling agents were added
(12–18, Fig. 2), residual sugars were also present in the mono-
lith in proportion to their composition in the sol (Table 3).
The difference between the theoretical and measured weight
loss provides guidance as to the efficiency of incorporation of
GLS and MLS during condensation (samples 3–5 and 8–10),
(Table 4). These values were confirmed by combustion
elemental analysis data of the gels (Table 5). The quantity of
nitrogen, in particular, is diagnostic of the amount of
saccharide present. The efficiency of sugar incorporation,
from about 30–70%, was related to a degree both to the type of
Fig. 2 TGA of unwashed samples 1 (after washing, the weight loss
was 2.7%), 3–5, and 8–10.
Table 4 Thermogravimetric analyses of samples 1, 3–5, 8–10
Ratio of
DGS : Sugar
Theoretical Weight Efficiency
coupling agent weight loss loss by of T
Sample (w : w)
(%)a
TGA
incorporationb
1
3
4
5
8
9
10
a
100 : 0
0.0
2.17
29.9
38.2
31.5
10.3
19.7
33.4
—
60
69
51
31
41
53
80 : 20 (GLS)
75 : 25 (GLS)
66 : 33 (GLS)
94 : 06 (MLS) 33.4
89 : 11 (MLS) 48.5
80 : 20 (MLS) 62.7
50.0
55.4
62.1
Fig. 3 TGA comparison of the drying kinetics of wet samples 1 and 4
by measured weight loss at 40 uC over 75 min.
saccharide that holds tenaciously onto water, showed slower
release kinetics.
Assuming all added GLS or MLS is incorporated the monolith
b
and C,N,H are lost as gaseous byproducts. Assuming alkoxysilane
hydrolysis is complete, and the only residual source of carbon arises
from the saccharide residues in GLS or MLS.
To normalize the degree to which all samples were hydrated,
and also to remove ungrafted sugars, the remaining samples
were allowed to age for 7 d, were ground up, washed with
water (see experimental section), freeze dried for 2 d and
characterized by TGA.
Table 5 Elemental analyses of sample 3–5, 8–10
Sugar based
on Ca (wt%)
Sugar based
on Na (wt%)
Silica gels containing grafted GLS and MLS, both of which
are hygroscopic, showed loss of significant amounts of water
on heating. Surprisingly, however, there was no correlation
between the sugar content and the amount of entrained water
in the monoliths: after washing, all monoliths exhibited weight
losses due to water of 5–10%.
Sample
C (%)
N (%)
3
4
5
8
9
10
a
14.87
14.97
16.76
6.13
9.65
13.18
1.70
1.73
1.98
0.54
0.51
0.94
39.7
39.9
44.7
15.3
24.0
32.8
35.0
35.6
40.7
17.3
16.3
30.1
Once bonded into the matrix, sugar moieties derived from
GLS or MLS could not be cleaved by normal hydrolysis and/
or removed by washing. After washing the gels to remove
residual, ungrafted material including glycerol and ungrafted
GLS and MLS, respectively, the presence of the residual
Using the molecular formula for sugar as sugar-CLO–
NH(CH2)3SiO3/2. The amount of residual unhydrolyzed alkoxysilane
groups is reflected in the difference between the N and C values and,
as can be seen, is rather limited.
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 3132–3141 | 3135

View Article Online
Table 6 Solid-state 13C and 29Si CPMAS NMR spectral data of
sample 3 and 10
Sample
13C, d/ppm)
29Si, d/ppm
3
9.8, 22.8, 41.9, 63.4,
72.7, 174.5
257.4, 266.1, 292.9,
2101.2, 2110.5
10
9.3, 22.2, 41.9, 63.4,
72.9, 102.6, 174.8
257.9, 267.2, 292.6,
2101.4, 2109.8
Table 7 Surface characteristics of monoliths prepared with GLS, and
MLS, respectively (see also the supplementary information for pore
size distributions)
Surface Total pore
volume/
Ratio of DGS : sugar area/
Sample coupling agent w : w m2 g21 cm3 g21
Average pore
diameter/nm
1
3
4
8
9
10
100 : 0
635.0
43.4
3.3
362.7
332.5
18.9
0.467
0.035
0.006
0.661
0.528
0.023
2.9
3.2
7.2
7.3
6.4
5.0
80 : 20 (GLS)
75 : 25 (GLS)
66 : 33 (GLS)
94 : 06 (MLS)
89 : 11 (MLS)
found in the solution (DMSO) 13C NMR spectra of the
starting materials (Table 1), while the resonances around
18.8 and 58.3 ppm assigned to ethoxy groups disappeared
following hydrolysis and condensation.
In the 29Si CP-MAS NMR spectra of samples 3 and 10,
respectively (see the supplementary information), the well-
known resonances of Q2, Q3, and Q4 from the monolithic
silica gel are accompanied by T2 and T3 groups in the 257 to
266 ppm range for sample 3, and 258 to 267 ppm for sample
10, respectively. These data are consistent with the formation
of Si–O–Si linkages between sugar-based trifunctional silane
and the underlying silica. That is, the monoliths are comprised
of QT resins that have a higher abundance of T groups at the
surface than within the structure.
Fig. 4 XPS and elemental analysis results for powder samples 3–5,
and 8–10.
T-functional silane, GLS vs. MLS, and their respective
concentration in the sol (Table 4).
It was initially unclear whether the T units derived from
compounds GLS and MLS were homogenously dispersed
throughout the modified silica, located in domains within the
silica, or were located primarily on the internal pore and
microchannel surfaces. Dried and ground samples of the
monoliths were examined by XPS to assess the nature of the
surface at boundaries. Fig. 4 shows the XPS results, presented
as the C/Si atom% ratio, for powdered samples 3–5 (prepared
from DGS with increasing proportions of GLS) and 8–10
(prepared from DGS with increasing proportions of MLS)
at takeoff angles of 90u. These data represent the elemental
The monoliths were further characterized by nitrogen
absorption measurements. Monoliths prepared from DGS
alone had much higher surface areas and total pore volumes
than any of those prepared from recipes that included GLS
(Table 7). MLS affected the system quite differently. The
average pore diameter was changed in an unanticipated
manner: at low concentrations, the pores and the total
pore volumes are larger than monoliths prepared from DGS
alone. However, with increasing concentrations, the average
pore size dropped.
˚
composition of the uppermost 50–100 A of the exposed
surface, and can be compared with elemental analysis results
by combustion (Table 5), which represents the constitution of
the entire sample. The higher carbon : silicon atomic ratios in
the XPS than in the combustion analysis data are consistent
with exposed siliceous surfaces that are enriched with the
sugar-based coupling agents.
Discussion
The presence of the bound sugar coupling agents into the
monolith per se was demonstrated by both 29Si and 13C CP-
MAS NMR spectra, which reports on the bulk sample. The
increased concentration of the saccharides at the air interface
was demonstrated by ATR FT-IR. The latter, in particular, is
diagnostic because of the amide linkages that appear in the
region between 1650 and 1700 cm21 (refer to the electronic
supplementary information). The resonances from sugar
fragments of the trifunctional component in the 13C CP-
MAS NMR spectra of 3 and 10, respectively (Table 6, also see
the supplementary information), are comparable to those
Under carefully controlled conditions, trifunctional silanes
can be hydrolyzed to give a series of well-defined regular
prisms with Si–O–Si edges. The best known of these are the
polyhedral silsesquioxanes (POSS),27 as exemplified by T8
Cy
derivatives (e.g., T8 Scheme 2), a variety of which is now
commercially available. More generally, however, trifunctional
silanes are used to generate amorphous organofunctional
layers on surfaces, where they function as ‘‘coupling agents’’ or
‘‘adhesion promoters’’.16 With some notable exceptions,28 it is
very difficult to prepare monolayers on surfaces. Commercial
applications also exist for trifunctional resins.29
3136 | J. Mater. Chem., 2005, 15, 3132–3141
This journal is ß The Royal Society of Chemistry 2005

View Article Online
ultimately lead to silica (Scheme 3b–c). Second, and perhaps
more important, any polyol-containing silane moiety will be
much more soluble in water than silicon species derived from
monofunctional alcohols and siloxane linkages. As a result, the
essentially irreversible precipitation reaction of silica is at least
retarded and, in the limit, completely suppressed (Scheme 3a–b
is favored over 3c). The presence of bound and unbound
sugars may control the reaction in additional ways, including
increasing the viscosity (retarding the rate of substitution), and
sequestering water (which favors silanols over disiloxanes).
Nature uses this tactic of polyol concentration to control silica
formation in diatoms, plants and other organisms: sugar or
catechol-rich domains sequester silicon as polyol complexes to
prevent silica formation until desired, at which point, water is
added to reestablish the normal equilibrium.30,31
The hydrolysis and condensation reactions leading to
monolith formation are solution equilibria upon which is
superimposed a heterogeneous equilibrium between soluble
species and the insoluble silica ultimately formed. Unlike
TEOS, which undergoes primary particle formation, and then
particle aggregation, the precipitation of DGS can be con-
trolled, and limited to the formation of nanoparticles because
of the enhanced solubility of the DGS-derived materials.32 The
glycerol in DGS thus changes not only the equilibrium
cure chemistry, but also the subsequent surface chemistry of
colloidal silica materials.
Scheme 2 Examples of polyhedral silsesquioxanes.
The relationship between the chemistry of tetrafunctional
and trifunctional silanes is key in determining the resulting
monolith properties. The sugar-derived trifunctional silanes
GLS and MLS did not hydrolyze and cure effectively on their
own: too many polyol groups are present in the sol. In the case
of DGS : GLS or DGS : MLS mixtures, sols formed from
mole ratios of DGS : GLS of less than 1 : 3 similarly did not
lead to curing within one week.
The addition to a DGS sol of trifunctional compounds
based on the sugar lactones GLS or MLS similarly and further
significantly changes the behavior of the resulting cure
processes and the characteristics of the final material. Their
presence increases the load of saccharidic materials that, for
the reasons enunciated above, suppress gelation. First, the
presence of free or complexed (as alkoxysilanes) bidentate or
multidentate alcohols slows the reaction progress. This can be
seen in the efficiency of gelation; the curing reaction is retarded
with increased amounts of the sugar amide in solution (Table 2,
Scheme 3a). While free sugars can retard curing (Table 2,
entries 2 and 6), the sugar-based alkoxysilanes are more
effective at reducing the rate of gelation at a given concentra-
tion, presumably because they are directly and, in the case of
GLS or MLS, irreversibly linked into the evolving matrix.
A comparison of the two alkoxysilanes GLS or MLS shows
that differences exist between the two that extend beyond the
simple presence of the extra sugar moieties in the latter case.
Even at comparable loadings, based on saccharide units,
the efficiency of incorporation was lower in the latter case
(Table 4). To understand the origin of this, one must consider
both the local and global concentration of saccharides. Adding
free saccharides to any of the sols reduces the rate of gelation,
by pushing the equilibrium reaction towards starting
materials (Scheme 3). Different silicon species will be affected
differently by the proximity of sugars. Tetrafunctional silicon
atoms, such as those derived from DGS, exhibit reduced
reactivity when the sugar is chemically bound to the silicon,
rather than when simply in solution nearby (e.g., TEOS +
glycerol). In the case of GLS or MLS, the silicon can never
escape the sugar. Thus, these materials will always be less
reactive than tetrafunctional silanes. The effect of the local
sugar concentrations is manifested in the much lower reactivity
‘‘Normally’’, tetraalkoxysilanes undergo an efficient series
of pH sensitive hydrolysis and condensation reactions leading
to silica.1 The term ‘‘normally’’, however, almost always refers
to monofunctional alkoxysilanes such as Si(OEt)4, TEOS and
Si(OMe)4, TMOS. The presence of polyols changes the cure
behavior of silica sols. The hydrolysis and gelation of TEOS in
the presence of glycerol greatly retards the rate of gelation.10
DGS, constituted directly from the key elements—a tetra-
alkoxysilane and glycerol—undergoes curing even more
slowly. This demonstrates that the presence of a polyol is not
the only feature affecting curing: the proximity of the polyol to
the silicon center, as in DGS, retards the gelation process.
It has been previously proposed that the presence of
chemically bound or free sugars distorts the normal hydro-
lysis/condensation equilibria of alkoxysilanes.10 Two factors
affect this change in equilibrium. First, many more non-
productive nucleophiles are present in the sugar sol. A complex
cascade of intramolecular and intermolecular nucleophilic
substitutions can occur (Scheme 3a), in a sugar syrup, to
compete with the hydrolysis and condensation reactions that
Scheme 3 Model hydrolysis/condensation reaction scheme for DGS.
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 3132–3141 | 3137

View Article Online
of MLS than GLS, which results in a much lower capacity to
be chemically bound into the monolithic TQ silica gel.
The addition of the trifunctional coupling agent to the
tetrafunctional silane starting material also impacts on the
surface chemistry of the resulting gel: even at much lower
loadings in the sol, much more MLS was found at the silica
surfaces than GLS (Fig. 1). The pH profiles of hydrolysis/
condensation for tri- and tetrafunctional silanes follow similar,
but not identical trends. It had been anticipated that the
trifunctional silanes would undergo hydrolysis/condensation
more slowly, based on the work of Osterholz and Pohl.33 That
is, sequential curing would take place. Initially, DGS would
undergo hydrolysis and condensation reactions. MLS and
GLS would later on begin to participate in the process and,
towards the end, much higher levels of the sugar coupling
agents would reside at the surface (Fig. 4). A corollary of this
proposal is that the concentration of MLS and GLS would
gradually increase in the sol and on the evolving gel surface
over time such that, at higher concentrations further reaction
will be suppressed. This view is supported by the inability to
entrain all the sugarsilane coupling agents in the gel and also
explains the observation that the situation in this respect is
worse with MLS than with GLS: the former intrinsically cures
less effectively into the silica structure and much of it will
eventually be washed, unreacted or only partly reacted, from
the monolithic gel.
ethoxy to glyceroxy: TEOS gels shrink significantly more than
DGS-based gels, even in the presence of additional entrained
sugars derived from GLS or MLS. This likely arises from
morphological differences between monoliths derived from the
two tetraalkoxysilanes, which arise from different rates of
curing in DGS/GLS (or MLS) sols vs. TEOS–GLS (or MLS)
sols. Such differences arise from the change in equilibrium
when sugars are present and, in unwashed samples, are
exacerbated in the base of TEOS-derived monoliths by the
rapid loss of ethanol through evaporation. This can lead to
significant pore collapse, which will not occur when the non-
volatile glycerol leaving group is present.
TEOS hydrolyses, essentially irreversibly, to give silanols
and ethanol, which can then diffuse away from the reactive
centre. Based on local viscosity, local water concentration, or
adventitious non-productive nucleophiles, there should be no
retardation of the rate. By contrast, as shown in model form in
Scheme 3, DGS is fundamentally less efficient at condensation.
By virtue of the covalent linkage to the saccharide groups,
curing occurs even more slowly in the presence of GLS or
MLS: as noted above, neither will cure independently, they
must be co-constituents with tetrafunctional silanes in order
to gel.
The differences in shrinkage between gels containing GLS or
MLS derived from TEOS or DGS, respectively, are thus
related to the differences in curing of the tetrafunctional and
trifunctional silanes. TEOS will cure much more rapidly than
either GLS or MLS leading to gel structures comprised of local
domains of high silica content, coated with sugar rich layers
derived from GLS or MLS (core shell structures). While the
sugars are hygroscopic, the morphology is such that they
reduce the shrinking on only a small part of the monolith.
By contrast, the rates of cure of DGS and GLS or MLS,
respectively, are much closer. As a consequence, there is
preferential concentration of the slower curing coupling agents
at the surface, but they also exist throughout the matrix where
they can act as a barrier to micropores (or an agent that
suppresses their formation) and as hydrated plasticization
agents (ripple structure).
The more significant difference observed upon adding
modified trifunctional silanes GLS or MLS to the siliceous
precursor, DGS,9 is the dramatic reduction of shrinkage in the
gel (Fig. 1). However, the ultimate degree of shrinkage is also
strongly dependent on the nature of the tetrafunctional
precursor. For comparable degrees of GLS or MLS incorpora-
tion, DGS-based gels shrank much less than TEOS derived
gels: DGS 68%; TEOS 80%; DGS–GLS 10%; TEOS–GLS
ca. 40% (Fig. 1). Two issues thus need to be addressed: the
differences in shrinkage engendered by the use of DGS vs.
TEOS, and the origins in the decrease in shrinkage when GLS
or MLS are present in the sol.
Unlike the case with TEOS or DGS alone, washing a formed
monolith with water cannot remove the sugars associated
with GLS or MLS; they are covalently linked into the matrix,
preferentially at the water-exposed interfaces. The water
associated with these sugars can assist in plasticizing the
monolith and, by holding onto the water, resist the evolution
of capillary forces that normally occur on drying. Moreover,
the addition of GLS or MLS to DGS sols leads to significant
decreases in total available surface area (in non-thermally
treated silicas). One explanation for this change is a loss of
micropores which, if correct, would mean a dramatic reduction
in the evolving capillary forces. This proposition is strongly
supported by the significant loss of total surface area when
GLS is present in the sol: it does not address the characteristics
of the MLS-containing materials, which are currently under
further investigation. Thus, the decrease in shrinkage could
thus be associated both with better plasticization and an
overall reduction of the magnitude of the forces responsible for
shrinkage in the first place.
In addition to the consequences of shrinkage, the presence
of the sugarsilane coupling agents completely changes the
nature of the siliceous surface. Dramatic losses in available
surface area (by nitrogen adsorption, Table 7) accompany the
presence of sugar at the interface. Much more important,
however, for the ultimate purpose of these monoliths, is the
presentation of sugar moieties at the water interface to aid in
stabilizing entrapped proteins. When proteins are present,
there is a direct correlation between the lifetime of entrapped
enzymes and the degree of shrinkage of the gel. Sugars are
known to be stabilizing environments for proteins.34,35
These monoliths take advantage of sugar chemistry in a
variety of ways: (i) the sugars lead to competitive curing rates
between tetra- and trifunctional silanes, (ii) the resulting
monoliths have sugars throughout the matrix, but preferen-
tially at interfaces, (iii) this morphology leads to dramatically
reduced shrinking, and, (iv) the surface is enriched in sugars.
The latter two factors strongly contribute to enhanced stability
of proteins when trapped in a silica material derived from DGS
and GLS or MLS.13,36 In particular, the presence of sugars at
This explanation does not address the changes engendered
by changing the alkoxygroup of the tetrafunctional silane from
3138 | J. Mater. Chem., 2005, 15, 3132–3141
This journal is ß The Royal Society of Chemistry 2005

View Article Online
the surface alters the propensity for electrostatic adsorption of
proteins to the silica,22 while at the same time presenting a
biocompatible environment that may play a role in altering
the hydration state,34 and hence the dynamics,22 of the protein.
Ultimately, this leads to better long-term stability of the
entrapped proteins.
Preparation of silsesquioxane precursors
GLS: To a solution of D-gluconolactone (0.91 g, 5.2 mmol) in
DMSO (10 mL) and EtOH (5 mL) was added 3-aminopropyl-
triethoxylsilane (1.11 g, 5.0 mmol). The mixture was stirred at
60 uC for 20 h. The solvents were evaporated under vacuum
and the oil residue was dissolved in dichloromethane.
Unreacted D-gluconolactone was filtered off, the filtrate was
concentrated and added to a large amount of pentane. The
white precipitate was collected and dried in vacuo to give
GLS as a pale yellow solid, 1.83 g (92% yield). 1H NMR
(200.2 MHz, d6-DMSO): d 0.50 (m, 2H, SiCH2), 1.12 (t, 9H,
J = 6.98 Hz, SiOCH2CH3), 1.45 (m, br, 2H, SiCH2CH2), 3.04
(m, 2H, CH2NHCO), 3.74 (q, J = 6.98 Hz, 6H, SiOCH2CH3),
3.40–5.32 (m, 11 H, glucose CH and CH2, and OH), 7.61 (s,
br, 1H, NHCO). 13C NMR (50.3 MHz, d6-DMSO): d 7.8
(SiCH2), 18.7–18.9 (SiOCH2CH3), 23.3 (SiCH2CH2), 41.5
(CH2NHCO), 58.3 (SiOCH2CH3), 64.0, 70.7, 72.1, 74.2, 73.0
(glucose CH and CH2), 172.9 (NHCO). FT-IR (KBr):
1646 cm21 (n(CLO)). MS-ESI (ES+): 422.2 (M + Na, 100)+,
400.2 (M + 1, 15)+, 354 (5), 236 (18).
Conclusion
Siliceous materials (TQ resins) were prepared using sol–gel
techniques from the tetrafunctional silane DGS and trifunc-
tional silanes based on sugar lactones. Comparison of carbon :
silicon atomic ratios in the XPS and elemental analyses
demonstrated that, under the conditions used, trifunctional
sugar-based silane coupling agents are primarily covalently
bound to the solvent exposed surface of the siliceous
material. The additional use of the sugar-based, trifunctional
silane coupling agents, in addition to tetrafunctional silanes,
led to significantly less shrinkage of the resulting crosslinked
siliceous gels and the controlled modification of the internal
surfaces.
MLS: D-Maltonolactone was prepared using literature
procedures in which maltose was oxidized with iodine.24 To
a solution of D-maltonolactone (0.75 g, 2.2 mmol) in DMSO
(10 mL) and EtOH (5 mL) was added 3-aminopropyltriethoxy-
silane (0.44 g, 2.0 mmol). The mixture was stirred at 60 uC
for 20 h. The solvents were evaporated under vacuum and
oil residue was dissolved in dichloromethane. Unreacted
D-maltonolactone was filtered off, the filtrate was concen-
trated and added to a large amount of pentane. The white
precipitate was collected and dried in vacuo to give MLS
as pale yellow solid, 0.98 g (87% yield). 1H NMR (200.2 MHz,
d6-DMSO): d 0.49 (m, br, 2H, SiCH2), 1.08 (t, J = 6.96 Hz, 9H,
SiOCH2CH3), 1.43 (m, br, 2H, SiCH2CH2), 3.70 (q, J =
6.96 Hz, 6H, SiOCH2CH3), 3.05–5.47 (m, 22 H, CH2NHCO
and maltose CH and CH2, and OH), 7.60 (s, br, 1H, NHCO)
ppm. 13C NMR (50.3 MHz, d6-DMSO): d 7.8 (SiCH2), 18.4–
19.1 (SiOCH2CH3), 23.4 (SiCH2CH2), 41.3 (CH2NHCO), 56.6
(SiOCH2CH3), 61.3, 63.4, 69.8, 72.5–73.8 (overlapped), 80.6,
101.4 (maltose CH and CH2), 172.9 (NHCO) ppm. FT-IR
(KBr): 1643 cm21 (n(CLO)). MS-ESI (ES+): 584.3 (M + Na,
30)+, 562.4 (M + 1, 20)+.
Experimental
Materials and methods
D-Gluconolactone (glulactone), D-maltose monohydrate,
iodine, silver carbonate, 3-aminopropyltriethoxylsilane and
anhydrous methyl sulfoxide (Aldrich Chemical Co.) were
used as received. Diglycerylsilane DGS was prepared as
previously reported.10 The strong cationic exchange resin
Amberlite IR-120 (Aldrich Chemical Co.) was rinsed with
distilled water before use. D-Maltonolactone (maltlactone),
was prepared according to literature procedures (iodine
oxidation of the polysaccharide).21,24 Dichloromethane and
pentane were distilled from CaH, EtOH was distilled from
Mg before use.
1H and 13C NMR were recorded at room temperature on a
Bruker AC-200 spectrometer; solid state 13C and 29Si CPMAS
NMR spectra were recorded on a Bruker AC-300 at 75.47 and
59.62 MHz, respectively. FTIR-ATR spectra were recorded
on Bruker, Tensor 27. Electrospray mass spectra were
recorded on a Micromass Quattro LC, triple quadruple
MS. Thermogravimetric analyses were obtained using a
Thermowaage Sta STA 409 at a heating rate of 3 uC min21
from room temperature to 600 uC. X-ray photoelectron
spectroscopy (XPS) was performed at Surface Interface
Ontario (Toronto ON). The surfaces of the samples were
analyzed using a Leybold Max 200 X-ray photoelectron
spectrometer with a Mg Ka non-monochromatic X-ray source.
The spot size used in all cases was 2 6 4 mm. Survey scans
were performed from 0 to 1000 eV. Both low resolution and
C(1s) high resolution analyses were performed. The raw data
were analyzed and quantified using the software Specslab
(Specs Gmbh, Berlin). Data were obtained at takeoff angles of
90u. Binding energies were referenced to the C(1s) signal for
the carbon-silicon bond that was assigned a binding energy
of 284.4 eV. The raw atomic% for Si(2p) and C(1s) signals
were converted into ratios (Fig. 4). Elemental analyses were
performed by Guelph Chemical Labs.
Preparation of samples 1–10 and shrinkage measurements
All of the samples below were treated in the following way after
gelation: fresh sol–gels were aged in a closed container at 5 uC
for 20 h, then further aged at room temperature for 7 or 20 d.
Aged hydrogels were washed with water 5 6 5 mL. This was
done by soaking the whole aged gel (1 mL initial volume) in
5 mL water at room temperature for 4 h. The water was replaced
4 times, the last time the gel was kept over 8 h, for a total of 24 h.
The gels were then allowed to dry at room temperature in an
open container for 45 d. Shrinkage was recorded against the
initial volumes of the sample sols on a v/v% basis (Figure 1).
Sample 1 (DGS gel). To a solution of DGS (240 mg,
1.1 mmol) in H2O (0.50 mL) was added Tris buffer (0.50 mL,
50 mM, pH = 8.4). The mixture was left at room temperature
to gel (Tables 2–4). After washing (see above) and drying for
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 3132–3141 | 3139

View Article Online
Anal. Chem., 2000, 366, 635–645; (d) V. I. Muronetz, M. Sholukh
and T. Korpela, J. Biochem. Biophys. Methods, 2001, 49, 29–47; (e)
T. Baczek and R. Kaliszan, J. Biochem. Biophys. Methods, 2001,
49, 83–98; (f) R. R. Burgess and N. E. Thompson, Curr. Opin.
Biotechnol., 2002, 13, 304–308.
3 (a) D. S. Hage, T. A. G. Noctor and I. W. Wainer, J. Chromatogr.,
A, 1995, 693, 23–32; (b) H. Hofstetter, O. Hofstetter and
V. Schurig, J. Microcolumn Sep., 1998, 10, 287–291; (c)
O. Hofstetter, H. Lindstrom and H. Hofstetter, Anal. Chem.,
2002, 74, 2119–2125; (d) L. Fitos, J. Visy and M. Simonyi,
J. Biochem. Biophys. Methods, 2002, 54, 71–84.
4 (a) Y. L. Hsieh, H. Wang, C. Elicone, J. Mark, S. A. Martin and
F. Regnier, Anal. Chem., 1996, 68, 455–462; (b) C. Wang,
R. Oleschuk, F. Ouchen, J. Li, P. Thibault and D. J. Harrison,
Rapid. Commun. Mass Spectrom., 2000, 14, 1377–1383; (c) S. Wang
and F. E. Regnier, J. Chromatogr., A, 2001, 913, 429–436; (d)
D. S. Peterson, T. Rohr, F. Svec and J. M. J. Frechet, J. Proteome
Res., 2002, 1, 563–568; (e) D. S. Peterson, T. Rohr, F. Svec and
J. M. J. Frechet, Anal. Chem., 2002, 74, 4081–4088; (f) G. W. Slysz
and D. C. Schriemer, Rapid Commun. Mass Spectrom., 2003, 17,
1044–1050.
45 d, the magnitude of shrinkage was recorded. Freeze drying
gave a clear, colorless solid.
Sample 2 (DGS + sorbitol). To a solution of DGS (240 mg,
1.1 mmol) in H2O (0.50 mL) was added sorbitol (60 mg,
0.33 mmol in 0.50 mL (50 mM, pH = 8.4) Tris buffer). The
mixture was left at room temperature to gel (Tables 2–4). The
hydrogel was aged at 4 uC for 20 h in a closed container, then
further aged and dried in air at room temperature for 6 d.
After washing (see above) and drying for 45 d, the magnitude
of shrinkage was recorded. Shrinkage was then recorded.
Freezing dried gave white powder.
Samples 3–10 (DGS + GLS (3–6) or + MLS (7–10).
Prepared in a similar manner. The reaction conditions are
listed in Table 2 and Table 4.
5 D. Prazeres, F. Miguel and J. M. S. Cabral, in Multiphase
Bioreactor Design, ed. J. M. S. Cabral, M. Mota and J. Tramper,
Taylor & Francis Ltd., London, UK, 2001, pp. 135–180.
6 (a) D. C. Schriemer, D. R. Bundle, L. Li and O. Hindsgaul, Angew.
Chem., Int. Ed., 1998, 37, 3383; (b) B. Zhang, M. M. Palcic,
D. C. Schriemer, G. Alvarez-Manilla, M. Pierce and O. Hindsgaul,
Anal. Biochem., 2001, 299, 173–182.
7 (a) M. T. Baynham, S. Patel, R. Moaddel and I. W. Wainer,
J. Chromatogr., B, 2002, 772, 155–161; (b) R. Moaddel, L. Lu,
M. Baynham and I. W. Wainer, J. Chromatogr., B, 2002, 768,
41–53; (c) R. Moaddel, J.-F. Cloix, G. Ertem and I. W. Wainer,
Pharm. Res., 2002, 19, 104–107; (d) R. Moaddel and I. W. Wainer,
J. Pharm. Biomed. Anal., 2003, 30, 1715–1724.
TGA analysis
Thermogravimetric analysis (TGA) was performed under air,
with a flow rate of 50 cm2 min21. The heating rate was
5 uC min21 from room temperature to 750 uC. The TGA of
unwashed samples is shown in Fig. 2. Washed samples were
obtained by crushing the monolith; washing with deionized
water for about 2 h with stirring using a magnetic stirring bar,
at which point the water was removed by filtration. The
washing and filtering was repeated 3 times, and in total,
approximately 200 mL H2O was used. A comparison of the
drying of washed samples 1 and 4 is shown in Fig. 3. For the
TGA of washed and freeze dried samples, shown in Table 4,
the washed samples were further freeze dried at 0 uC for 20 h at
0.5–1 torr, after which the TGA was performed.
8 I. Gill and A. Ballesteros, J. Am. Chem. Soc., 1998, 120,
8587.
9 (a) M. A. Brook, J. D. Brennan and Y. Chen, Polyol-Modified
Silanes as Precursors for Silica, US Patent Application, 60/384084,
McMaster University; (b) Y. Chen, J. D. Brennan and M. A. Brook,
Controlled Shrinkage Sol–Gel Silsesquioxane Precursors, US Patent
Application 60/405308, McMaster University.
10 (a) M. A. Brook, Y. Chen, K. Guo, Z. Zhang and J. D. Brennan,
J. Sol-Gel Sci. Technol., 2004, 31, 343–348; (b) M. A. Brook,
Y. Chen, Z. Zhang and J. D. Brennan, J. Mater. Chem., 2004, 14,
1469–1479.
11 T. R. Besanger, Y. Chen, A. K. Deisingh, R. Hodgson, W. Jin,
S. Mayer, M. A. Brook and J. D. Brennan, Anal. Chem., 2003, 75,
2382–2391.
12 T. Arakawa and S. N. Timasheff, Biochemistry, 1982, 21, 6536.
13 (a) J. A. Cruz-Aguado, Y. Chen, Z. Zhang, M. A. Brook and
J. D. Brennan, Anal. Chem., 2004, 76, 4182–4188; (b) J. A. Cruz-
Aguado, Y. Chen, Z. Zhang, N. H. Elowe, M. A. Brook and
J. D. Brennan, J. Am. Chem. Soc., 2004, 126, 6878–6879.
14 R. H. Baney, M. Itoh, A. Sakakibara and T. Suzuki, Chem. Rev.,
1995, 95, 1409.
15 (a) For other reviews, see: M. G. Voronkov and V. I. Lavrent’yev,
Top. Curr. Chem., 1982, 102, 199; (b) J. D. Lichtenhan, in
Polymeric Materials Encyclopedia: Synthesis, Properties and
Applications – Silsesquioxane-Based Polymers, ed. J. C. Salsmone,
CRC Press, Boca Raton, Florida, 1996, p. 7768.
Monolith porosity
The nitrogen sorption isotherm, surface area and pore
radius were measured with NOVA 2000 from Quantachrome
Instruments. The samples were first prepared as were the TGA
samples. The samples were then degassed with a final vacuum
in the order of 10 millitorr (or less) for 5 h at 100 uC. BET
surface area was calculated by the BET (Brunauer, Emmett
and Teller) equation; the pore diameter from nitrogen
adsorption–desorption isotherms was calculated by the
Wheeler formula. All the data were calculated using the
software provided with the instruments.
Acknowledgements
We acknowledge with gratitude the financial support of this
research from MDS Sciex and the Natural Sciences and
Engineering Research Council of Canada (NSERC). JDB
holds the Canada Research Chair in Bioanalytical Chemistry.
MAB thanks the Canada Council for the Arts for a Killam
Fellowship (2003–2004).
16 E. P. Plueddemann, Silane Coupling Agents, 2nd ed., Plenum,
New York, 1991.
17 (a) D. A. Loy, A. Straumanis, D. A. Schneider, B. Mather,
A. Sanchez, C. R. Baugher and K. J. Shea, Polym. Prepr., 2004, 45,
591; (b) E. R. Pohl and F. O. Osterholz, in Silanes, Surfaces and
Interfaces, Vol. 1, ed. D. E. Leyden, Gordon and Breach,
Amsterdam, 1986, p. 481.
18 (a) Y. Chujo and T. Saegusa, Adv. Polym. Sci., 1992, 100, 11; (b)
Y. Schubert, in Organically Modified Silica Aerogels from
RSi(OR9)3/Si(OR9)4 Mixtures in Tailor-Made Silicon-Oxygen
Compounds, ed. R. Corriu and P. Jutzi, Vieweg, Wiesbaden,
Germany, 1996, p. 263 (from a workshop at the University of
Bielefeld, 1995).
References
1 C. J. Brinker and G. W. Scherer, Sol–gel Science, Academic Press,
San Diego, 1990.
2 (a) D. S. Hage, J. Chromatogr., B, 1998, 715, 3–38; (b) D. S. Hage,
Clin. Chem., 1999, 45, 593–615; (c) M. G. Weller, Fresenius’ J.
19 It should be noted that previous work has produced sugar modified
silica by first modifying a silica surface with a coupling agent, and
3140 | J. Mater. Chem., 2005, 15, 3132–3141
This journal is ß The Royal Society of Chemistry 2005

View Article Online
then adding a saccharide moiety, or directly using saccharidic
coupling agents to modify preexisting silica.21 This is distinct from
the ‘‘one-shot’’ use of saccharidic trifunctional silanes as siliceous
monolith precursors. These silanes can be both internal and
external constituents of the produced gel, which modify both bulk
and surface properties.
27 (a) S. H. Phillips, T. S. Haddad and S. J. Tomczak, Curr. Opin.
Solid State Mater. Sci., 2004, 8, 21–29; (b) F. J. Feher, R. Terroba,
R.-Z. Jin, S. Lucker, F. Nguyen, R. Brutchey and K. D. Wyndham,
Mater. Res. Soc. Symp. Proc., 2001, 628 (Organic/Inorganic
Hybrid Materials), CC2.1.1–CC2.1.6; (c) J. J. Schwab and
J. D. Lichtenhan, Appl. Organomet. Chem., 1998, 12, 707–713.
28 (a) Y. Koide, Q. Wang, J. Cui, D. D. Benson and T. J. Marks,
J. Am. Chem. Soc., 2000, 122, 11266–11267; (b) T. J. Marks, Ji. Cui,
J. E. Malinsky, Q. Wang, G. E. Jabbour, S. E. Shaheen,
J. D. Anderson, P. A. Lee, A. G. Richter, B. Kippelen, P. Dutta,
N. Peyghambarian and N. R. Armstrong, Polym. Prepr., 1999, 40,
2, 1196–1197.
20 (a) M. Kanno and S. Kusakaishi, Japanese Patent 3046964,
Tohoku Munekata K. K., Japan, 1999; (b) Japanese Patent
Application 79-169422, Showa Denko K. K., Japan, 1979; (c)
T. Uehara and Y. Tajima, Japanese Patent Application 92-264622,
Sekisui Chemical Co Ltd, Japan, 1992.
21 Gluconamides: (a) R. J. DePasquale and M. E. Wilson, US Patent
4591652, SCM Corp., USA, 1986; (b) M. Haupt, S. Knaus,
T. Rohr and H. Gruber, J. Macromol. Sci., Pure Appl. Chem.,
2000, A37, 323. A related compound, the maltopentaose-modified
silane: N. Enomoto, S. Furukawa, Y. Ogasawara, H. Akano,
Y. Kawamura, E. Yashima and Y. Okamoto, Anal. Chem., 1996,
68, 2798.
22 X. Sui, J. A. Cruz-Aguado, Y. Chen, Z. Zhang, M. A. Brook and
J. D. Brennan, Chem. Mater., 2005, 17, 1174–1182.
23 We note that the same methodology has been applied to the
preparation of dextran-modified silanes. However, the curing
chemistry of these materials is more complex, much slower, and
not discussed further in this paper.
29 Many such patents exist. See, for example: R. P. Boisvert,
G. T. Burns, T. C.-S. Chao, D. E. Katsoulis and S. K. Sarmah,
US Patent 5972512, Dow Corning, 1999.
30 (a) D. Belton, G. Paine, S. V. Patwardhan and C. C. Perry,
J. Mater. Chem., 2004, 14, 2231–2241; (b) C. C. Perry, Rev.
Mineral. Geochem., 2003, 54, 291–327; (c) C. C. Perry and
T. Keeling-Tucker, J. Biol. Inorg. Chem., 2000, 5, 537–550.
31 (a) S. D. Kinrade, R. J. Hamilton, A. S. Schach and C. T. G. Knight,
J. Chem. Soc., Dalton Trans., 2001, 961; (b) S. D. Kinrade,
K. J. Maa, A. S. Schach, T. A. Sloan and C. T. G. Knight, J. Chem.
Soc., Dalton Trans., 1999, 3149.
32 (a) D. Tleugabulova, A. M. Duft, Z. Zhang, Y. Chen, M. A. Brook
and J. D. Brennan, Langmuir, 2004, 20, 5924–5932; (b)
D. Tleugabulova, Z. Zhang, Y. Chen, M. A. Brook and
J. D. Brennan, Langmuir, 2004, 20, 848–854.
24 (a) K. Hamshimoto, S. Imanishi, M. Okada and H. Sumitomo,
J. Polym. Sci., Polym. Chem. Ed., 1991, 29, 1271; (b) T. Zhang and
R. E. Marchant, Macromolecules, 1994, 27, 7302.
25 We have compared the shrinkage of all samples at least up to 57 d.
Little shrinkage occurs after this time, as long as the conditions are
held the same (e.g., under water, in air, etc.). As an example,
unwashed DGS gels underwent no further shrinkage from 57 d to
two years when aged in air in an open container.
33 (a) E. R. Pohl and F. D. Osterholtz, Polym. Sci. Technol., 1985, 27,
157; (b) Y. Sugahara, S. Okada, S. Sato, K. Kuroda and C. Kato,
J. Non-Cryst. Solids, 1994, 167, 21; (c) K. A. Smith, Macro-
molecules, 1987, 20, 2514; (d) F. D. Osterholtz and E. R. Pohl,
J. Adhes. Sci. Technol., 1992, 6, 127.
26 (a) J. A. Hockey, Chem. Ind., 1965, 57; (b) R. K. Iler, The
Chemistry of Silica, John Wiley & Sons, New York, 1979; (c)
E. F. Vansant, P. Van Der Voort and K. C. Vrancken,
Characterization and chemical modification of the silica surface,
Elsevier, Amsterdam, 1995.
34 D. W. Bolen, Methods, 2004, 34, 312–322.
35 D. K. Eggers and J. S. Valentine, J. Mol. Biol., 2001, 314, 911–922.
36 R. J. Hodgson, Y. Chen, Z. Zhang, D. Tleugabulova, H. Long,
X. Zhao, M. Organ, M. A. Brook and J. D. Brennan, Anal. Chem.,
2004, 76, 2780–2790.
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 3132–3141 | 3141