Behavior of organic compounds confined in monoliths of sol-gel silica glass. Effects of guest-host hydrogen bonding on uptake, release, and isomerization of the guest compounds

Article abstract of DOI:10.1039/b005823h

Various proteins, catalysts, and other compounds can be encapsulated or diffused into porous sol-gel glasses, but little is known about their interactions with the glass matrix. We report unexpectedly large effects that hydrogen bonding between organic co

Full text of DOI:10.1039/b005823h

View Article Online / Journal Homepage / Table of Contents for this issue
Behavior of organic compounds con®ned in monoliths of sol±gel
silica glass. Effects of guest±host hydrogen bonding on uptake,
release, and isomerization of the guest compounds{
Jovica D. Badji c and Nenad M. Kosti cÂ
Department of Chemistry, Iowa State University, Ames, IA, 50011, USA
Received 19th July 2000, Accepted 10th October 2000
First published as an Advance Article on the web 3rd January 2001
Various proteins, catalysts, and other compounds can be encapsulated or diffused into porous sol±gel glasses,
but little is known about their interactions with the glass matrix. We report unexpectedly large effects that
hydrogen bonding between organic compounds and sol±gel silica has on equilibria and reactions involving
these guest compounds. Silica monolith immersed in a solution takes up the organic solute. Styrene, which is
incapable of hydrogen bonding, becomes evenly distributed between external solution and the glass. Aniline
and N,N-diethyl-p-methoxybenzamide, which are capable of hydrogen bonding, become extracted into the glass
4
when the solvent (neat CCl ) does not interfere with their hydrogen bonding with silica. They become evenly
distributed between the solution and the glass when a component of the solvent (DMF added to CCl ) or
4
chemical modi®cation (trimethylsilylation) of the silica surface suppresses hydrogen bonding of the guests with
the surface. Ultraviolet spectra show that silica±guest interactions are present when the guest uptake is
excessive and absent when this uptake is balanced. Ultraviolet spectra of aniline show that the hydrogen atoms
are donated by silica to the guest. Not only the extent, but also the rate, of uptake is enhanced when the guest
makes hydrogen bonds to the silica matrix; suppression of these bonds lowers the uptake rate. Silica monolith
4
extracts trans-3,3'-diacetylazobenzene from a CCl solution 1250-fold. Upon addition of DMF, hydrogen bonds
are broken, and the monolith completely releases the solute into the external solution. Five derivatives of
azobenzene (3,3'-dimethyl-, 3-acetyl-, 3,3'-diacetyl-, 3,5-diacetyl-, and 3,3',5,5'-tetraacetylazobenzene), which
differ in the propensity for accepting hydrogen atoms, served as photochromic probes and showed the effect of
hydrogen bonding on reactivity. Both the photoinduced (trans-to-cis) and the subsequent thermal (cis-to-trans)
isomerizations of the ®ve derivatives obey the ®rst-order law in glass as well as in free solution. When the
4
solvent (neat CCl ) allows hydrogen bonding, the proportion of the isomers in the photostationary state differs
between the glass and solution, and the rate constant for the thermal reaction is two to four times (in different
derivatives) smaller in the glass than in solution. Evidently, hydrogen bonding retards the rearrangement of the
probe molecules in the silica matrix. When hydrogen bonding is abolished (by addition of DMF to CCl
compositions of the photostationary state in the glass and in solution become equal, and so do the rate
4
), the
constants. Effects of hydrogen bonding on enzymes encapsulated in sol±gel glass and on the distribution of
analytes between the glass monolith and the sample solution should be taken into consideration when designing
accurate biosensors.
and guest molecules, which govern mobility and chemical
Introduction
reactivity of those guests, are poorly understood. To examine
these interactions, we have determined the af®nity of sol±gel
silica glass toward various organic compounds as guests, and
we have studied dynamical properties of some of these guests
inside the matrix. Other research of this kind has recently been
Controlled hydrolysis of alkoxides and polymerization of the
resulting oxyacids, the sol±gel method, is useful in several
branches of chemistry and engineering. Moderate processing
conditions enable encapsulation of various organic, inorganic,
and biological molecules in glasses having adjustable proper-
2
0±24
undertaken.
Work in our laboratory has shown that small molecules may
1
±8
ties.
catalyst supports, coatings, special polymers, and other new
The method is used for the fabrication of sensors,
behave quite differently in free solution and in the same solvent
25±27
9
±19
materials.
but con®ned in the pores of a hydrogel.
Soaking of silica
The guest compounds can be added at the onset of the
polymerization and encapsulated in the glass, or they can be
diffused into the porous glass after it has solidi®ed. In either
case interactions at the molecular level between glass matrix
glass in electrolyte solutions, no matter how prolonged, may
not ensure equal partitioning of certain ions between the glass
and the solution. At a pH value at which the pore walls are
3
2
negatively charged, anions, such as [Fe(CN)6] , are only
z
3
partially taken up, whereas cations, such as [Ru(NH
3
)
6
]
, are
excessively taken up by the glass from the surrounding
solution. In either case, internal and external concentrations
of the ion remain unequal after the equilibrium is reached. If
the electrostatic interactions are left unchecked, the unequal
distribution of an ionic analyte would invalidate a biosensor.
Hydrogen bonding is subtler and more speci®c than
{
One table (absorptivities of azobenzene derivatives) and nine ®gures
(
UV absorption spectra of styrene and N,N-diethyl-p-methoxybenza-
mide in solution and in silica glass; UV absorption spectra of aniline
and N,N-diethyl-p-methoxybenzamide in solution and in trimethylsi-
lylated silica glass; uptake of all three guests in silica glass pretreated
with HCl or NaOH or not pretreated; and H NMR, C NMR, and
mass spectra of 3,5-diacetylazobenzene and 3,3',5,5'-tetraacetylazo-
benzene) are available as supplementary data. For direct electronic
access see http://www.rsc.org/suppdata/jm/b0/b005823h/
1
13
2
8,29
electrostatic interactions.
this bonding in sol±gel silica glass has been reported.
Evidence for the existence of
2
1,30±33
408
J. Mater. Chem., 2001, 11, 408±418
This journal is # The Royal Society of Chemistry 2001
DOI: 10.1039/b005823h

View Article Online
However, to our knowledge, there is no published evidence for
the effects of hydrogen bonding on equilibria and reactions
involving the guest compounds within sol±gel silica.
Dehydration of silica monoliths
Two silica monoliths sized 868627 mm were soaked in
3
¯
00 mL of stirred acetonitrile in a capped 500 mL Erlenmeyer
ask. The solvent was twice replaced with fresh 300 mL
Because hydrogen bonds are relatively strong and direc-
tional, they may greatly affect physical and chemical processes
in doped glasses. Not much is known about these important
interactions, which are likely to affect properties and
performance of biosensors, immobilized catalysts, and various
composite materials prepared by the sol±gel method. These
phenomena deserve thorough investigation, and this study is a
piece of it. We detected hydrogen bonding, found examples of
its enormous effect on uptake of guests by silica, learned how to
enhance and abolish this bonding, and showed that it can affect
mobility of guest molecules and their rate of reaction. By
studying hydrogen bonding and its in¯uence on thermody-
namic and kinetic properties of absorbed and encapsulated
molecules in sol±gel silica monoliths, one can learn to control
chemical processes inside the matrix and to prepare silica
glasses with speci®c and desired properties.
portions, after 2 h and again after 21 h. This last soaking lasted
for an additional 24 h. The monoliths were then soaked in
1
50 mL of stirred CCl
The solvent was replaced with a fresh 150 mL portion after
4 h, and the soaking continued for an additional 24 h. In some
experiments, the second solvent was a 9 : 1 v/v mixture of CCl
4
in a capped 250 mL Erlenmeyer ¯ask.
2
4
and DMF. The Erlenmeyer ¯asks were always capped, to
protect the dry solvents from the moisture in the air.
Dehydrated silica monoliths were stored for further experi-
ments in a fresh portion of the same solvent in which they had
been soaked last, pure CCl or a 9 : 1 v/v mixture of CCl and
4
4
DMF.
Trimethylsilylation of silica monoliths
A silica monolith sized 868627 mm was exhaustively soaked,
with stirring, in a succession of relatively large volumes of
many solvents and solutions. These liquids, the volume of their
portions, the time of the soaking in each fresh portion, the
number of fresh portions, and the function of each liquid were
as follows: Acetonitrile, 350 mL, 24 h, one, for rinsing; a
mixture of 90 mL acetonitrile and 10 mL acetyl chloride, 2 h,
one, for removal of water; acetonitrile, 400 mL, 12 h, six, for
removal of acetyl chloride and reaction products; a mixture of
Experimental procedures
Instrumentation
1
3
Proton and C NMR spectra were recorded with a Varian
00 MHz instrument; infrared spectra, with a Bio Rad FTS-7
4
instrument; and UV absorption spectra, with a Perkin-Elmer
Lambda 18 spectrophotometer whose cuvette holder was
connected to the thermostated circulator bath Model 9500
from Fisher Scienti®c Co. Photochemical reactions were done
with a Rayonet 100 reactor, which has six ¯uorescent tubes
9
(
0 mL acetonitrile and 10 mL 1,1,1,3,3,3-hexamethyldisilazane
HMDSZ) under a nitrogen atmosphere, 24 h, ®ve, for
35
trimethylsilylation; acetonitrile, 400 mL, four, for removal
of HMDSZ and reaction products; and ®nally CCl , 150 mL,
4 h, two, for removal of previous solvents. Upon ®rst contact
4
Ê
designated 3500 A.
2
with HMDSZ the transparent monolith turned cloudy white,
but it became transparent again later.
This procedure was repeated with several monoliths. The
Chemicals
Distilled water was further demineralized to a resistivity greater
than 17 MV cm. Tetramethyl orthosilicate and acetonitrile
were purchased from Fisher Chemical Co; aniline, styrene,
thionyl chloride, p-anisic acid, nitrosobenzene, 3-aminoaceto-
phenone, 1,1,1,3,3,3-hexamethyldisilazane (HMDSZ), carbon
tetrachloride, benzene, and anhydrous N,N-dimethylforma-
mide (DMF), from Aldrich Chemical Co; and 3,3'-dimethyl-
azobenzene, from Trade TCI Mark.
trimethylsilylated monoliths were stored in CCl
experiments.
4
for further
Infrared spectra of dehydrated and trimethylsilylated silica
glasses
Silica monoliths of both kinds were ground in a mortar, the
powder was dried in a vacuum desiccator for 1 h, and the
spectroscopic sample was placed between NaCl plates.
Pretreatment of silica monoliths with acid and base
Hazards
Two silica monoliths sized 868627 mm were soaked in
4
Tetramethyl orthosilicate, benzene and CCl are highly toxic;
2
00 mL of a 0.10 M HCl solution, and the other two were
the last two chemicals may cause cancer. Protective gloves and
effective fume hood must be used in experiments with these
compounds.
2
4
soaked in 200 mL of a 3.2610 M NaOH solution, all in
capped 500 mL Erlenmeyer ¯asks, with stirring overnight. All
four monoliths were then dehydrated by the procedure
described above.
Preparation of silica monoliths
3
4
General procedure for uptake experiments
The silica sol was prepared by a standard procedure.
A
mixture of 15.76 g of tetramethyl orthosilicate, 3.38 g of water,
and 0.30 g of 0.040 M HCl was kept in an ice-cooled ultrasonic
bath for 30 min. Upon addition of 36.26 mL of a 10 mM
sodium phosphate buffer at pH 7.00, gelation began. The sol
was quickly poured into polystyrene cuvettes sized
4
Stock solutions of guest compounds in CCl were prepared as
2.05 mM in styrene, 0.48 mM in aniline, and 0.43 mM in N,N-
diethyl-p-methoxybenzamide, so that their UV absorbances
were not too high. Dehydrated silica monoliths sized
868627 mm were the hosts. Each monolith was soaked in
50.0 mL of a stock solution in a capped vial. The monolith
stood upright, with one of its 868 mm sides resting on the ¯at
bottom of the vial and the other ®ve sides exposed to the
solution. From time to time the monolith without the
surrounding solution was transferred to the 10 mm quartz
cuvette kept in the sample holder of the spectrophotometer. An
identical but undoped silica monolith made in the same batch
as the doped monolith was put in another 10 mm quartz
cuvette, and kept in the reference holder. Both monoliths were
1
0610640 mm, and left at 4 ³C. During the aging for 14
days, the samples were kept soaked and were washed with
water twice a day. The aged samples were exposed to air, still at
4
³C, for another 14 days of partial drying and shrinking. The
resulting transparent monoliths, prisms sized 868627 mm,
were soaked in acetonitrile and stored at 4 ³C for further use.
All monoliths used in this study had this size. Defective or
nonstandard monoliths were rejected, for the sake of
reproducibility and accuracy of the experiments.
J. Mater. Chem., 2001, 11, 408±418
409

View Article Online
held in the same place and the same position in all experiments,
so that their absorbances could be accurately compared. The
photostationary state (designated pss) for the isomerization
was reached. The relative concentrations of the trans and cis-
2
1
1
isomers were determined from the intensities of the H NMR
UV spectra were scanned at the rate of 120 nm min , and the
doped monolith was returned to the solution of the guest
compound (the dopant) and soaked further. The UV
measurements were made more frequently early in the
experiment, while the uptake was relatively fast, and less
frequently later, when the uptake slowed down. At the end of
the uptake experiment, the UV spectrum of the soaking
solution partially depleted of the guest compound was
compared with the UV spectrum of the stock solution of this
guest. The width in mm (w) of the soaked monolith was
measured with the caliper, and the absorbance of this monolith
multiplied by 10/w, so that it could be compared with the
absorbance of the stock solution.
resonances for the methyl groups in these isomers. Although
thermal isomerization is slow and the composition of the pss is
practically constant over a relatively short time (see below), for
the sake of accuracy the NMR spectrum was taken promptly.
From the known initial concentration and absorptivity of the
trans-isomer, the UV spectrum of the mixture of isomers, and
the composition of this mixture, the absorptivity of the cis-
isomer was determined. The results are given in Table S1 in the
supplementary information.
trans-to-cis photoisomerization of azobenzene derivatives
To a quartz cuvette sized 10610640 mm with all walls
transparent (so-called ¯uorescence cuvette) was fused a
ground-glass joint. The joint could be closed, to prevent
evaporation of the solvent held (together with the doped
monolith) in the cuvette. The following procedure was used for
all ®ve derivatives of azobenzene. A dehydrated silica monolith
sized 868627 mm was soaked in 50.0 mL of a solution of the
Uptake and release of trans-3,3'-diacetylazobenzene
This compound was chosen for the determination of the
maximum uptake of a guest compound by the sol±gel silica
glass. A dehydrated monolith sized 868627 mm was soaked
2
4
in 200 mL of a stirred 9.30610 M solution of the title
compound in CCl . The capped Erlenmeyer ¯ask was wrapped
4
trans-isomer in a 9 : 1 v/v mixture of CCl and DMF; DMF
ensured that the guest compound uniformly penetrated into
4
in aluminium foil, to prevent photoisomerization of the trans-
isomer. Decrease of its concentration in the external solution
was monitored occasionally by UV spectroscopy. Aliquots of
ca. 0.30 mL of the external solution were diluted by ca. 2.5 mL
silica (see the general description of uptake experiments). The
uptake was followed by UV spectrophotometry. When the
absorbance of the trans-isomer in silica reached that in the
external solution, the monolith and 1.50 mL of the soaking
solution were transferred to the special cuvette and always kept
in the same position inside it. The capped cuvette was
irradiated, always in the same position with respect to the
lamps. The walls transmitted UV light from all sides. The
ultraviolet spectrum of the monolith was taken occasionally.
When it ceased changing, the photostationary state (pss) was
reached. Absorbances were converted to concentrations
according to eqns (1) and (2), in which A is absorbance, e is
absorptivity, subscript o denotes initial concentration, and lack
of a subscript denotes the concentration at the time of
of CCl , to lower their absorbance below 1.0; both the aliquot
4
and the solvent were weighed with an analytical balance, so
that the accurate dilution factor be known. Then the UV
spectrum was recorded, and from it the concentration of the
title compound in the external solution was determined. Slight
decreases in the volume of the soaking solution after each
successive removal of aliquots were taken into account in the
successive determinations of its concentration. When this
concentration almost ceased decreasing with time, the uptake
experiment was ended.
The release of the title compound by the monolith was
triggered by adding 20.0 mL of DMF to the external solution.
Increase in the concentration of the title compound in the
external solution was followed by UV spectroscopy, as
described above. When this concentration nearly reached the
4
1
measurement. The experimental results were ®tted by the
linear least-squares method, eqn. (3), in which t is time and k is
1
the apparent ®rst-order rate constant for the intramolecular
4
photoisomerization.
1
2
4
initial value of 9.30610 M, the experiment was ended.
A
Ao
1
{
‰
cisŠ~
‰transŠo
(1)
(2)
(3)
ecis
Calculation of the volume required for trans-to-cis isomerization
1{
etrans
We accepted a published mechanism for isomerization of
6±39
3
azobenzene.
Conformation of the trans-isomer was
‰
transŠ~‰transŠ {‰cisŠ
optimized with the program Hyperchem 4.5. All ®ve stages
of the isomerization were examined separately for each
0
4
0
azobenzene derivative. The structures at each of the ®ve
stages were superimposed and treated as one structure, the van
der Waals volume of which was then calculated by the program
GRASP 1.2. This volume can accommodate the entire
isomerization process of the given derivative. This procedure
was repeated for each of the ®ve derivatives of azobenzene.
‰transŠ {‰transŠ
transŠ{‰transŠ
‰transŠ
0
pss
0
.
k1 t
ln
~
‰
‰transŠ {‰transŠ
pss
0
pss
cis-to-trans thermal isomerization of azobenzene derivatives
Determination of absorptivities of azobenzene derivatives
The following procedure was used for all ®ve derivatives of
azobenzene. A dehydrated silica monolith sized 868627 mm
was soaked in 50.0 mL of a 5610 M solution of the pure
trans-isomer. Two solvent systems were used. In one series of
The following procedure was used for the trans-isomers of all
2
5
®
0
¯
ve azobenzene derivatives. A sample of ca. 17 mg, weighed to
.01 mg, was dissolved in 500.0¡0.2 mL CCl , in a volumetric
ask. The UV spectrum of this solution allowed direct
4
experiments neat CCl
v/v mixture of CCl
4
was used and in the other series, a 9 : 1
4
and DMF was used. As before, when
calculation of absorptivities of the trans-isomer at various
wavelengths.
Absorptivities of the cis-isomer were determined indirectly.
A sample of ca. 1.0 mg of the trans-isomer was dissolved in
absorbance of the monolith reached that of the surrounding
solution, the monolith was transferred to the special cuvette for
UV irradiation. This time, the photoisomerization was just a
method for preparing the cis-isomer, whose thermal isomer-
ization was the reaction of interest. When the solvent was neat
1
0.0 mL of a 95 : 5 v/v mixture of CCl4 and CDCl , in a
3
volumetric ¯ask. The UV spectrum was recorded, and the
solution was irradiated (by the procedure to be described next)
until its UV spectrum ceased to change, i.e., until the
CCl , the monolith was transferred without the external
4
solution, to avoid continued uptake of azobenzene during
4
10
J. Mater. Chem., 2001, 11, 408±418

View Article Online
irradiation. When mixed solvent system was used, the monolith
was transferred with 1.50 mL of the soaking solution, because
in this case there was no danger of continued uptake.
The capped cuvette was irradiated in the photochemical
reactor (as described above) until the UV spectrum ceased
changing, a sign that the photostationary state was reached. To
the solvent-free monolith was then added 1.5 mL of neat CCl
and the monolith that had been immersed in the mixture of
CCl and DMF was left so. The liquid around the monolith is
4
,
4
needed to conduct the heat to the monolith. In both cases, the
sealed cuvette was immersed above the level of the monolith
and the surrounding liquid in the water bath at 50.0¡0.5 ³C;
thermal isomerization at a lower temperature would be
impractically slow. The cuvette would occasionally be trans-
ferred to the spectrophotometer, and thermal isomerization
would be followed by recording UV spectra. When the
isomerization was completed, in experiments employing neat
CCl its spectrum was also taken, to detect possible leakage of
4
the azobenzene from the monolith.
Ultraviolet bands of the ®ve azobenzene derivatives reach
their maxima at different wavelengths. The maximum absor-
bance was converted to the concentration of the trans-isomer
according to eqns. (1) and (2). The change of concentration in
time was ®tted by the linear least-squares method to eqn. (4), in
which k
tion. Other quantities were de®ned in connection with
2
is the ®rst-order rate constant for thermal isomeriza-
4
1
eqn. (3).
‰
transŠ {‰transŠ
0
pss
.
ln
~k2 t
(4)
‰
transŠ {‰transŠ
0
Fig. 1 Infrared spectra of pulverized sol±gel silica monoliths (a) before
and (b) after treatment with hexamethyldisilazane. The highlighted
bands in (b) are characteristic of the Si(CH
3 3
) groups.
Preparation of N,N-diethyl-p-methoxybenzamide
42 1
3
reported; H NMR in CDCl : d 1.17 (br, 6 H), 3.40 (br, 4 H),
To a solution of 17.4 g (0.11 mol) of p-anisic acid in 50.0 mL of
benzene were added 17.6 mL of thionyl chloride and several
drops of DMF. The mixture was re¯uxed in a strong fume
hood until gas evolution ceased, for ca. 2 hours. Benzene was
cautiously evaporated under reduced pressure, and p-anisoyl
chloride was obtained in almost theoretical yield. It was used
without further puri®cation.
3.82 (s, 3 H), 6.89 (m, 2H) and 7.34 (m, 2H).
Preparation of azobenzenes
4
3,44
See Scheme 1.
synthesized by oxidation of the corresponding derivatives of
All derivatives of nitrosobenzene were
4
5
aniline with Caro's acid. A solution containing 3.0 mmol of
nitrosobenzene or its derivative and 3.0 mmol of a derivative of
aniline in 15 mL of glacial acetic acid was heated at 60 ³C for
A solution of 21.8 g of p-anisoyl chloride (0.13 mol) in
0.0 mL of diethyl ether was added slowly in a dropping funnel
5
1
2 h, poured into an ice-cold, saturated aqueous solution of
to a solution of 40 mL of diethylamine (0.39 mol) in 150 mL of
diethyl ether, kept at 0 ³C. Upon complete addition of p-anisoyl
chloride, the reaction mixture was stirred for 2 h at room
temperature and extracted with 200 mL of diethyl ether. The
organic layer was washed with three 150 mL portions of water,
NaHCO , and extracted with benzene. Evaporation of benzene
3
gave a crude solid. Its puri®cation on a silica-gel column
yielded a pure azobenzene derivative. Chromatographic
conditions are given below.
0
.10 M HCl, 0.10 M NaOH and one portion of brine and dried
with magnesium sulfate. The organic solvent was evaporated,
and the crude title compound was recrystallized from
methanol. The yield was 18.9 g or 70%; mp 42±43 ³C, as
3-Acetylazobenzene. Eluent was a 9 : 1 v/v mixture of
benzene and ethyl acetate. Orange needles, mp 88±89 ³C (cf.
4
3 1
88±90 ³C); H NMR in CDCl : d 2.68 (s, 3 H), 7.51 (m, 3 H),
3
1
.61 (t, 1 H), 7.93 (m, 2 H), 8.08 (m, 2 H) and 8.46 (s, 1 H); C
3
7
NMR in CDCl : d 26.9, 122.9, 123.1, 127.0, 129.2, 129.5, 130.3,
3
1
31.5, 138.1, 152.5, 152.7 and 197.6.
3
,5-Diacetylazobenzene. Eluent was a 6 : 4 v/v mixture of
1
benzene and ethyl acetate. Orange needles, mp 87±89 ³C. H
NMR in CDCl : d 2.72 (s, 6 H), 7.54 (m, 3 H), 7.96 (m, 2 H),
3
1
.62 (t, 1 H) and 8.64 (d, 2 H); C NMR in CDCl
3
8
3
: d 27.0,
1
LRMS-CI: calcd for M of C16
23.3, 126.5, 129.3, 129.5, 132.0, 138.5, 152.3, 153.0 and 196.9;
z
14 2 2
H N O 266, found 266.
3
,3'-Diacetylazobenzene. Eluent was a 8 : 2 v/v mixture of
4
6
benzene and ethyl acetate; mp 133±134 ³C (cf. 131±134 ³C);
1
3
H NMR in CDCl : d 2.67 (s, 6 H), 7.61 (m, 2 H), 8.07 (m, 4 H)
1
and 8.46 (s, 2 H). C NMR in CDCl : d 26.9, 123.1, 127.2,
3
3
Scheme 1
129.6, 130.8, 138.2, 152.5, and 197.5.
J. Mater. Chem., 2001, 11, 408±418
411

View Article Online
1
1
3
mixture of benzene, ethyl acetate, and chloroform; mp 221±
,3',5,5'-Tetraacetylazobenzene. Eluent was a 5 : 2 : 3 v/v/v
dilution factor of ca. 10 . The trimethylsilylation procedure
was applied to the dehydrated glass and included chemical as
well as extractive dehydration. The additional dilution factor in
1
2
8
24 ³C; H NMR in CDCl : d 2.75 (s, 12 H), 8.67 (t, 2 H) and
3
13
.70 (d, 4 H); C NMR in CDCl
43
3
: d 27.0 , 126.7 , 130.3 , 138.6,
52.5 and 196.7; LRMS-CI: calcd for M of C H N O 350,
the trimethylsilylation procedure reached ca. 10 . Much
z
1
found 350.
smaller factors probably would have been suf®cient, but we
strove for very accurate experiments.
Trimethylsilylated silica does not show any bands in the
2
0
18
2
4
-
1
1
650±1790 cm region in the infrared spectrum (Fig. 1). The
Results and discussion
absence of carbonyl groups proves that acetyl chloride does
not react with silanol hydroxy groups during the dehydra-
tion preceding treatment with hexamethyldisilazane.
Hydrogen bonding between silica glass and guest compounds
Hydrogen bonds between sol±gel silica and guest compounds
embedded in it (during formation of the glass) or absorbed in it
(
exist, but were directly detected only recently.
by diffusion into the preformed glass) are widely assumed to
2
Balanced and excessive uptake of the guest compounds by silica
glass
1,30±33
Not much
is known about these important interactions, which are likely
to affect properties and performance of biosensors, immobi-
lized catalysts, and various composite materials prepared by
the sol±gel method. Beyond detecting hydrogen bonding, we
wish to understand their effects on equilibria and reactions of
guest compounds within a sol±gel silica monolith.
Because sol±gel silica is porous, it is widely assumed that
molecules smaller than the pores can diffuse in and out of the
matrix and that their concentrations in the matrix and in the
surrounding solution will become equal, i.e., that the uptake
will be balanced. Research in this laboratory has already
refuted this assumption in cases when both the silica matrix and
2
5±27
the guest are charged.
If these charges are alike, the host
Treatment of the silica monolith
matrix partially rejects the guest; if the charges are opposite, the
matrix takes up the guest excessively. We term this phenom-
enon imbalanced uptake.
After aging and partial drying, the pores in the silica monoliths
contain water that must be removed lest it interfere with the
hydrogen bonds under study. We extracted water with a
succession of solvents chosen so that each dilutes and removes
the previous one. These exhaustive procedures took consider-
able time. The dehydration procedure achieved the overall
2
5±27
Electrostatic interactions, which we studied before,
are
strong and isotropic. In this study we examine the con-
sequences of subtler interactions. Hydrogen bonding is
relatively strong but also anisotropic and selective. These
characteristics make hydrogen bonding extremely important
2
8,47±49
for equilibria and chemical reactivity.
The three compounds serving as guests are shown below.
Styrene is incapable of hydrogen bonding; aniline can both
accept and donate hydrogen atoms; and N,N-diethyl-p-
methoxybenzamide can only accept them. These three
compounds as probes allowed us to examine the capacity of
sol±gel silica monoliths as both acceptor and donors of
hydrogen. We also managed to enhance and abolish hydrogen
Fig. 3 Uptake of (a) aniline, and (b) N,N-diethyl-p-methoxybenzamide
by sol±gel silica glass when it is incapable of hydrogen bonding to the
guest compounds. Glass monolith sized 868627 mm is soaked in
Fig. 2 Uptake of (a) styrene, (b) aniline, and (c) N,N-diethyl-p-
methoxybenzamide by sol±gel silica glass when it is capable of
hydrogen bonding to the guest compounds. Glass monolith sized
50.0 mL of a 9 : 1 v/v mixture of CCl and DMF; the latter component
4
868627 mm is soaked in a 50.0 mL solution of each solute in CCl
4
. As
in the mixture prevents hydrogen bonding between silica and the guest.
the guest enters, its absorbance in the glass increases. The dashed line
marks the absorbance of the external solution. Uptake is balanced in
As the guest enters, its absorbance in the glass increases. The dashed
line marks the absorbance of the external solution. Uptake is balanced
in both cases.
(
a) and excessive in (b) and (c).
412 J. Mater. Chem., 2001, 11, 408±418

View Article Online
bonding simply by adjusting the solvent properties, as will be
explained below.
that the spectrum of N,N-diethyl-p-methoxybenzamide does
change when its uptake is excessive (from neat CCl ), but does
4
not change when its uptake is balanced (in the presence of
DMF). Perturbation of the chromophore is diagnostic of its
interaction with silica. That perturbation and excessive uptake
occur together probably is more than a coincidence; the two
phenomena likely exist in a causal relationship.
Because the UV spectrum of aniline is well understood, this
compound is used as a reliable probe for hydrogen bonding.
Blue shift signi®es hydrogen acceptance; red shift, hydrogen
5
2±55
donation.
taken up from the solution in neat CCl
Fig. 4 shows a large blue shift when aniline is
and no spectral change
Fig. 2 shows that when the solvent was neat CCl
styrene was balanced, whereas the uptake of the other two
guests was excessive. Fig. 3 shows that when CCl contained
0% by volume of DMF, the uptake of these two guests became
balanced as well.
4
, uptake of
4
when aniline is taken up from the solution containing also
DMF. Evidently, aniline accepts hydrogen from the silica
matrix when the silanol groups on its surfaces are available.
When, however, silanol groups are saturated with DMF, a
good acceptor, hydrogen bonding between silica and aniline is
abolished.
4
1
Excessive uptake of relatively nonpolar organic molecules
into the highly polar matrix such as silica is unexpected on the
basis of the well-known principles of extraction and partition
chromatography. When excessive uptake occurs, there must
exist special attractive interactions between the host and the
Large uptake and complete release of trans-3,3'-diacetylazo-
benzene by silica glass
5
0
guest. Interaction of CCl
These interactions are to be expected because chlorine atoms in
4
with silica has just been reported.
5
1
As Fig. 2 shows, the concentration of hydrogen-bonding guests
becomes approximately three times greater in the glass than in
the external solution after approximately three hours of
soaking, when these experiments were ended. To ®nd out the
maximum capacity of silica for a guest, we chose the title
compound, which contains two hydrogen-accepting acetyl
groups, because we had a good supply of it. The experiment
lasted for 11 days, until the uptake practically ceased. The
results are shown in Fig. 5.
organic molecules are known to form weak hydrogen bonds.
Styrene, which lacks heteroatoms and is incapable of hydrogen
bonding, lacks special attraction to the silica surface and
cannot displace CCl from the silica surface. Consequently, its
4
uptake by the silica monolith is balanced. The other two guests
have nitrogen and oxygen atoms, which can form relatively
strong hydrogen bonds to silica, and therefore displace CCl
4
from the silica surface. Consequently, their uptake by silica is
excessive. When the hydrogen-bonding solvent DMF is present
in large excess over the guests, it abolishes excessive uptake of
these two guests.
We de®ne the uptake coef®cient as the ratio of the guest
concentrations in the glass and in the solution surrounding the
glass. Concentration of the guest in 200 mL of solution
2
4
24
decreased from 9.3610 M to 1.2610 M in a sealed
container. The ``departed'' quantity of it must have accumu-
lated in the silica monolith, whose overall volume is
Ultraviolet-spectroscopic evidence for hydrogen bonding
Fig. S1 in the supplementary information shows that the
absorption spectrum of styrene does not change when this
compound enters the silica matrix. The absence of interaction
with the matrix is consistent with the balanced uptake observed
for styrene. Fig. S2 in the supplementary information shows
25
.7 mL. Because only ca. 65% of this volume is pores and
1
therefore accessible to the guest, the aforementioned ratio of
21
concentration ``inside'' and ``outside'' is 1.5610 M to
24
1
.2610 M, or 1250. We did not thoroughly study the
other derivatives of azobenzene, but they too are likely to show
greatly excessive uptake.
This surprisingly large number shows how important
hydrogen bonding between silica matrix and guests compounds
can be. Clearly, this bonding must be taken into consideration
in studies of diffusion and partition of solutes between sol±gel
materials and the surrounding media and in careful studies of
equilibria and rates of reactions inside sol±gel materials.
Upon addition of DMF to the depleted external solution, the
absorbed guest was fully released by the silica monolith in ca.
Fig. 5 Uptake (the downward plot) and release (the upward plot) of
trans-3,3'-diacetylazobenzene (designated C) by sol±gel silica glass,
followed by the UV spectrophotometry of C in the external solution.
The glass monolith sized 868627 mm was soaked in 200.0 mL of a
stirred 9.30610 M solution of C in CCl . For release, 20 mL of
4
DMF was added to the external solution.
Fig. 4 Absorption spectra of aniline in solution (solid line) and inside
sol±gel silica glass (dashed line). The solvent is (a) CCl
v/v mixture of CCl and DMF. Note the large blue shift in (a) as aniline
4
and (b) a 9 : 1
4
2
4
enters the glass from the solution. Hydrogen bonding between aniline
and silica is present in (a) and absent in (b).
J. Mater. Chem., 2001, 11, 408±418
413

View Article Online
Fig. 6 Uptake of (a) aniline and (b) N,N-diethyl-p-methoxybenzamide
by sol±gel silica glass chemically modi®ed with Si(CH groups, which
3 3
)
preclude hydrogen bonding between the guest compounds and silica.
For conditions, see caption to Fig. 2.
75 h, much faster than it had been taken up. Clearly, hydrogen
bonds between silica and the guest can be broken by a
competing hydrogen-acceptor.
Our ®nding holds promise for practical applications of sol±
gel silica. This material can ef®ciently extract from solution a
solute with complementary hydrogen-bonding ability and then
release this solute upon addition of a small amount of a suitable
solvent. This cycle is reminiscent of molecular sieves, analytical
and preparative separations, and controlled delivery of drugs.
Nearly-balanced uptake of the guest compounds by chemically-
modi®ed silica glass
Fig. 7 Kinetics of photoinduced (trans-to-cis) isomerization of (a) 3,3'-
dimethylazobenzene, designated A, (b) 3-acetylazobenzene, designated
B, (c) 3,3'-diacetylazobenzene, designated C, (d) 3,5-diacetylazoben-
zene, designated D, and (e) 3,3',5,5'-tetraacetylazobenzene, designated
E, in solution (&) and inside sol±gel silica glass ($) when hydrogen
bonding between the azobenzene derivatives and silica is suppressed.
The solvent in all experiments was a 9 : 1 v/v mixture of CCl and DMF;
4
the latter component in the mixture makes hydrogen bonds to silica.
Solid lines are ®ttings to ®rst-order rate law.
Upon exhaustive treatment of silica monolith with hexa-
methyldisilazane, hydroxy groups on the surface were
``capped'' with the Si(CH
organic groups are evident by the infrared bands at ca. 3000
3 3
) groups. As Fig. 1 shows, these
2
1
and 1260 cm , corresponding to C±H and Si±C vibrations,
respectively.
Both compounds that show excessive uptake by pristine
silica (Figs. 2b and 2c) show nearly-balanced uptake by the
modi®ed silica (Figs. 6a and 6b) even though the solvent, CCl
4
,
surfaces of porous silica and prevents the guest molecules from
forming hydrogen bonds with these surfaces.
allows hydrogen bonding. We attribute the absence of this
bonding to trimethylsilylation. Hydroxylic hydrogen atoms to
be donated to the guests are absent, and the oxygen atoms that
might accept hydrogen atoms from aniline are protected by the
Rate of uptake of the guest compounds by silica monoliths
bulky Si(CH
3
)
3
groups. Fig. 6 shows that concentration of each
The time at which the concentration of the guest compound
inside the silica monolith becomes equal to the concentration in
the external solution may be taken as an approximate measure
of the rate of uptake if, as in our study, the external
concentrations are similar in all experiments. In the case of
balanced uptake, internal concentration levels off with time. In
the case of excessive uptake, internal concentration continues
to increase.
As Fig. 3 shows, the equalization time was ca. 400 min for
both guests that are capable of hydrogen bonding. When this
bonding was suppressed, they were taken up at the same rate.
As Fig. 2b and 2c show, the two rates became greater and quite
different when hydrogen bonding was allowed; then the
equalization times were 75 and 35 min.
guest in the silica host levels off slightly below the concentra-
tion in the external solution. If real, this small ``under-
achievement'' may be the result of the steric bulk of the capping
groups.
The chromophores that were markedly perturbed by the
pristine silica (Fig. 4) are only slightly affected by the modi®ed
silica (Fig. S3 in the supplementary information). Because
hydrogen bonding is suppressed, the slight change presumably
is a result of a nonspeci®c effect of the altered medium. Again,
excessive uptake coincides with marked spectral change,
whereas nearly or completely balanced uptake coincides with
little or no spectral change.
Evidently, excessive uptake of hydrogen-bonding guests can
be abolished not only by adding a competing solvent but also
by modifying the silica surface. This ®nding corroborates our
view, stated above, that the competing solvent DMF lines the
In attempts to alter the uptake rates, we treated silica
monolith with acid or with base before dehydration. Fig. S4 in
the supplementary information shows, however, that the
414
J. Mater. Chem., 2001, 11, 408±418

View Article Online
Fig. 9 Thermal (cis-to-trans) isomerization of (a) 3,3'-dimethylazo-
benzene, designated A, (b) 3-acetylazobenzene, designated B, (c)
Fig. 8 Thermal (cis-to-trans) isomerization of (a) 3,3'-dimethylazoben-
zene, designated A, (b) 3-acetylazobenzene, designated B, (c) 3,3'-
diacetylazobenzene, designated C, (d) 3,5-diacetylazobenzene, desig-
nated D, and (e) 3,3',5,5'-tetraacetylazobenzene, designated E, in
solution (&) and inside sol±gel silica glass ($) when hydrogen bonding
between the azobenzene derivatives and silica is suppressed. The
solvent in all experiments was a 9 : 1 v/v mixture of CCl and DMF; the
4
latter component in the mixture makes hydrogen bonds to silica. Solid
lines are ®ttings to ®rst-order rate law.
3
,3'-diacetylazobenzene, designated C, (d) 3,5-diacetylazobenzene,
designated D, and (e) 3,3',5,5'-tetraacetylazobenzene, designated E,
in solution (&) and inside sol±gel silica glass ($) when hydrogen
bonding between the azobenzene derivatives and silica is present.
The solvent in all experiments was CCl
rst-order rate law.
4
. Solid lines are ®ttings to
®
When the silica surface was modi®ed with the bulky
Si(CH ) groups, the uptake slowed down; the equalization
time for both guest compounds exceeded 1000 min. Again,
suppression of hydrogen bonding seems to lower the rate of
uptake.
differently-treated monoliths resembled each other and
untreated monoliths in the rates of uptake of the same
3
3
hydrogen-bonding guest in CCl
bonding.
4
, the solvent that allows this
Clearly, rate of uptake varies greatly when hydrogen
bonding is allowed or abolished and when the silica surface
is chemically modi®ed. Even before quantitative studies of the
rate are done, our semiquantitative ®nding should be taken
into account in studies of equilibria and chemical reactions
involving sol±gel glasses doped with enzymes and other
chemicals.
Table 1 Estimated volume needed for thermal (cis-to-trans) isomeriza-
tion of ®ve azobenzene derivatives and ®rst-order rate constants k in
the absence of hydrogen bonding in solution and inside sol±gel silica
glass. The solvent in both cases is a 9 : 1 v/v mixture of CCl and DMF
2
4
3
0 k
21
1
2
/min
3
Symbol Volume/A In solution In silica
Ê
Compound
3
,3'-Dimethyl-
azobenzene
A
334
3.02¡0.02 2 .73¡0.02
Kinetics of azobenzene isomerization
Isomerization of azobenzene has been used to explore the
5
microstructure and rigidity of solid polymers.
3
3
-Acetylazobenzene
,3'-Diacetylazo-
benzene
B
C
392
425
1.72¡0.02 1.56¡0.01
1.21¡0.01 1.29¡0.01
6±58
In principle,
this process can be used in optical storage systems, nanode-
vices, and molecular switches. Other researchers have applied
this elegant method to assess the steric constraint of sol±gel
3
3
,5-Diacetylazo-
benzene
,3',5,5'-Tetraacetyla-
zobenzene
D
E
462
505
2.10¡0.08 1.80¡0.03
0.78¡0.02 0.78¡0.02
41,59,60
glass on molecules contained in it. The function of this
glass in various applications depends on the mobility and
J. Mater. Chem., 2001, 11, 408±418 415

View Article Online
Table 2 Percentage of the cis isomer in the photostationary state
reached in trans-to-cis isomerization of ®ve azobenzene derivatives in
solution and inside sol±gel silica glass when hydrogen bonding between
these derivatives and silica is absent and present
variability of the free volume felt by the photochromic
molecules.
The azobenzene derivatives
In silica
We purchased the compound designated A and synthesized
and puri®ed those designated B to E. Their propensity for
accepting hydrogen atoms increases with the number of acetyl
substituents. We studied the effect of these substituents, and
therefore of hydrogen bonds, on the thermal isomerization
shown in eqn. (5).
Without With
Symbol In solution H-bonding H-bonding
a
b
a
Compound
3
3
3
3
3
a
,3'-Dimethyl-
azobenzene
-Acetylazo-
benzene
,3'-Diacetylazo-
benzene
,5-Diacetylazo-
benzene
,3',5,5'-Tetraacetyl- E
azobenzene
A
B
92¡4
89¡4
78¡4
70¡4
49¡3
92¡4
89¡4
82¡4
70¡4
48¡3
80¡4
82¡4
75¡4
68¡4
19¡2
C
D
b
. Solvent is a 9 : 1 v/v mixture of CCl
Solvent is CCl
4
4
and DMF.
2
Table 3 Fist-order rate constants k and their ratio for thermal (cis-to-
ꢀ
0†
trans) isomerization of four azobenzene derivatives in solution and
inside sol±gel silica glass when hydrogen bonding between azobenzene
and silica is present. The solvent in both cases is CCl
4
3
21
1
2
0 k /min
solution
2
silica
k
Compound
Symbol
In solution
1.63¡0.02
2.09¡0.08
0.83¡0.02
In silica
k
2
3
3
3
,3'-Diacetyl-
azobenzene
,5-Diacetyl-
azobenzene
,3',5,5'-Tetraacetyl-
azobenzene
C
D
E
0.46¡0.02
0.60¡0.02
0.41¡0.02
3.5
3.5
2.0
Uptake and isomerization of azobenzene derivatives in silica
glass when hydrogen bonds are absent
In these experiments the solvent was a 9 : 1 v/v mixture of CCl4
and DMF. The latter component, a hydrogen acceptor, binds
to silica and forestalls its hydrogen bonding to azobenzene
derivatives. All ®ve guests showed balanced uptake into the
silica monoliths. Both photoinduced and thermal reactions in
eqn. (5) for all ®ve compounds taken up in silica were
monophasic, as linear plots in Fig. 7 and 8 show. Table 1
shows that the rate constants k for thermal isomerization (the
2
second reaction in eqn. (5)) are virtually the same in solution
and in the glass. Table 2 shows that the photoisomerization
2
3
chemical reactivity of the dopant. Because hydrogen bonding
greatly affects the extent and the rate of uptake and release of
guest compounds by the host glass, we study quantitatively the
in¯uence of hydrogen bonding on a reaction of a series of guest
molecules absorbed in the host matrix. Interconversion of
geometric (cis and trans) isomers of azobenzene is shown in
eqn. (5). It is a simple reaction, amounting to motion of
molecular parts. We chose it because its kinetics are well
6
1±63
known.
(
the ®rst reaction in eqn (5)) occurs to the same extent in
hn,k1
D,k2
solution and in the glass. Evidently, in the absence of hydrogen
bonding these two media are equal as environments for the
isomerization. When azobenzene cannot make hydrogen bonds
to the matrix, this photoprobe does not differentiate between
bulk solvent and solvent con®ned in the glass pores. A similar
conclusion was reached in a study of isomerization in which
hydrogen bonding was not possible because the protic solvent
trans� azobenzene ?/ cis� azobenzene A
trans� azobenzene
(5)
The so-called forward reaction is effected by UV light, and it
is relatively fast. The so-called back reaction is spontaneous at
room temperature but slow. We studied it at 50.0 ³C, in the
dark.
Study of the forward reaction presents pitfalls. The ®rst-
order rate constant for photoisomerization changes when the
amount of light absorbed by the irradiated sample changes.
The absorbed amount, in turn, depends on the concentration of
the sample and on the irradiation wavelength. To avoid errors
in kinetic analysis, we determined the order of the reactions of
the various azobenzene derivatives, but did not calculate or
4
1,59,60
was present. Clearly, the glass matrix itself does not
restrict the motion required for the isomerization.
The largest photoprobe, the one designated E, requires the
3
volume of ca. 505 A for isomerization. A sphere of this volume
Ê
Ê
has a diameter of ca. 10 A. Evidently, the pores in sol±gel silica
are larger than this size.
Uptake and isomerization of azobenzene derivatives in silica
glass when hydrogen bonds are present
compare the rate constants k . Fortunately, study of the back
1
reaction is straightforward. This was shown with various
photochromic molecules, which proved useful for exploration
4
In these experiments the solvent was neat CCl . All ®ve
azobenzene derivatives showed excessive uptake by silica
monoliths because of hydrogen bonding to silica. In the case
of the compound C, which we studied in detail (see above), this
excess is 1250-fold.
The thermal back reaction (the second step in eqn. (5)) of all
®ve photoprobes was studied separately in solution and inside
4
0,41,56,59,60,64±69
of polymers.
Above the glass-transition tem-
perature the back reaction inside the polymer is monophasic, as
in solution. Below this temperature, however, the reaction
inside the polymer is biphasic or multiphasic. This change in
kinetics has been attributed to heterogeneity of the matrix and
416
J. Mater. Chem., 2001, 11, 408±418

View Article Online
the silica. Because these probes are different compounds, which
differ in the substituents and in electron delocalization,
7
8
M. Reetz, A. Zonta and J. Simpelkamp, Angew. Chem., Int. Ed.
Engl., 1995, 34, 301.
D. Avnir, S. Braun, O. Lev and M. Ottolenghi, Chem. Mater.,
6
2,70±73
``vertical'' comparisons in Table 3 are unwarranted.
1994, 6, 1605.
We make only ``horizontal'' comparisons, those of the same
compound in two environments. As Fig. 9 shows, isomeriza-
tion of all ®ve photoprobes in solution, and of all except B in
silica, were monophasic. The reaction of the compound B in
silica appeared biphasic.
At the end of the long kinetic experiments, the neat CCl4
surrounding the monoliths was checked by UV spectro-
photometry. The compounds A and B leaked from the silica,
and the compounds C, D, and E did not. The biphasic reaction
of B probably is caused by the simultaneous isomerization, at
different rates, of this compound in silica and in the
surrounding solution. The reaction of A may appear mono-
phasic if the rates in silica and in solution are comparable. For
the sake of accuracy, we omitted compounds A and B from
Table 3.
Kinetic results for the compounds C, D, and E are not
complicated by leakage. They are given in Table 3. Thermal
isomerization is slower in silica than in solution. Mobility of
azobenzene within silica seems to be hindered by hydrogen
bonding to the pore surfaces. Both photoprobes having two
acetyl groups (C and D) show similar rate of isomerization and
the same degree of hindrance, 3.5-fold. The photoprobe having
four acetyl groups (E) shows a lesser hindrance, but this fact
may be an indirect, not a direct consequence of hydrogen
bonding. The more extensive bonding of E with silica hindered
already the photoinduced reaction, the trans-to-cis isomeriza-
tion, so that only ca. one-half of the trans isomer initially
present in silica was converted to the cis isomer. Only that half
of the total compound E within the glass underwent the
thermal conversion back to the trans con®guration.
9
0
C. Sanchez, F. Ribot and B. Lebeau, J. Mater. Chem., 1999, 9, 35.
E. H. Lan, B. C. Dave, J. M. Fukuto, B. Dunn, J. I. Zink and
J. S. Valentine, J. Mater. Chem., 1999, 9, 45.
1
11 F. Gelman, D. Avnir, H. Schumann and J. Blum, J. Mol. Catal. A:
Chem., 1999, 146, 123.
12
J. Blum, A. Rosenfeld, F. Gelman, H. Schumann and D. Avnir,
J. Mol. Catal. A: Chem., 1999, 146, 117.
R. Obert and B. C. Dave, J. Am. Chem. Soc., 1999, 121, 12192.
N. Husing and U. Schubert, Angew. Chem., Int. Ed., 1998, 37, 22.
1
1
3
4
15 D. Levy, F. Del Monte, J. M. Oton, G. Fiksman, I. Matias,
P. Datta and M. Lopez-Amo, J. Sol±gel Sci. Technol., 1997, 8, 931.
16
17
18
D. Levy, Chem. Mater., 1997, 9, 2666.
D. A. Loy and K. J. Shea, Chem. Rev., 1995, 95, 1431.
R. Reisfeld and C. K. Joergensen, Struct. Bonding (Berlin), 1992,
77, 207.
L. L. Hench and J. K. West, Chem. Rev., 1990, 90, 33.
19
20 C. Rottman, G. Grader, Y. D. Hazan, S. Melchior and D. Avnir,
J. Am. Chem. Soc., 1999, 121, 8533.
21 S. Spange, Y. Zimmermann and A. Graeser, Chem. Mater., 1999,
11, 3245.
22
23
24
M. S. Rao and B. C. Dave, J. Am. Chem. Soc., 1998, 120, 13270.
B. Dunn and J. I. Zink, Chem. Mater., 1997, 9, 2280.
M. Ueda, H.-B. Kim, T. Ikeda and K. Ichimura, J. Mater. Chem.,
1995, 5, 889.
25 C. Shen and N. M. Kosti c , J. Electroanal. Chem., 1997, 438, 61.
26 C. Shen and N. M. Kosti c , J. Am. Chem. Soc., 1997, 119, 1304.
2
2
7
8
J. D. Badji c and N. M. Kosti c , Chem. Mater., 1999, 11, 3671.
M. D. Joesten and L. J. Schaad, Hydrogen Bonding, New York,
M. Dekker, 1974.
M. C. Etter, Acc. Chem. Res., 1990, 23, 120.
K. Matsui, K. Nozawa and T. Yoshida, Bull. Chem. Soc. Jpn.,
1999, 72, 591.
2
3
9
0
31 L. Sieminska and T. W. Zerda, J. Phys. Chem., 1996, 100, 4591.
3
3
2
3
L. Nikiel and T. W. Zerda, J. Phys. Chem., 1991, 95, 4063.
K. Matsui, T. Matsuzuka and H. Fujita, J. Phys. Chem., 1989, 93,
4991.
3
4
S. A. Yamanaka, F. Nishida, L. M. Ellerby, C. R. Nishida,
B. Dunn, J. S. Valentine and J. I. Zink, Chem. Mater., 1992, 4, 495.
J. Cossy and P. Pale, Tetrahedron Lett., 1987, 28, 6039.
Conclusion
3
3
5
6
Speci®c interaction between the host matrix and the guest
molecules may restrict or broaden the applicability of doped
sol±gel glasses in chemistry, biochemistry, and materials
science. Hydrogen bonding is an important interaction,
whose implications have yet to be explored.
R. Cimiraglia, T. Asano and H.-J. Hofmann, Gazz. Chim. Ital.,
1
996, 126, 679.
R. Cimiraglia and H.-J. Hofmann, Chem. Phys. Lett., 1994, 217,
30.
3
7
4
38 S. Kobayashi, H. Yokoyama and H. Kamei, Chem. Phys. Lett.,
1987, 138, 333.
3
This study shows that sol±gel silica can donate hydrogen
atoms to acceptor molecules occluded in the silica monolith.
Hydrogen bonding can cause enormous excess in uptake of
hydrogen-bonding solutes from solution and can modulate
reactivity of the compounds trapped in glass monoliths. When
hydrogen bonding is suppressed, the uptake becomes balanced,
and the reactivity becomes normal. These results should
caution those who design biosensors and other devices that
depend on uptake and other equilibria involving sol±gel
glasses. The same results, however, encourage those who
seek new applications of sol±gel glasses as molecular sieves and
means for gradual or controlled delivery of substances such as
drugs.
9
N. Nishimura, T. Tanaka and Y. Sueishi, J. Chem. Soc., Chem.
Commun., 1985, 903.
J. G. Victor and J. M. Torkelson, Macromolecules, 1987, 20, 2241.
M. Ueda, H.-B. Kim and K. Ichimura, Chem. Mater., 1994, 6,
4
4
0
1
1771.
42 E. T. McCabe, W. T. Barthel, S. I. Gertler and S. I. Hall, J. Org.
Chem., 1954, 19, 493.
4
3
G. M. Badger, C. P. Joshua and G. E. Lewis, Aust. J. Chem., 1965,
8, 1639.
1
4
4
4
5
P. Ulrich and A. Cerami, J. Med. Chem., 1984, 27, 35.
B. G. Gowenlock, J. Pfab and V. M. Young, J. Chem. Soc., Perkin
Trans. 2, 1997, 1793.
J. Rosengaus and I. Willner, J. Chem. Soc., Chem. Commun., 1993,
1044.
J. W. Wijnen and J. B. F. N. Engberts, J. Org. Chem., 1997, 62,
4
4
6
7
2039.
Acknowledgements
48 P. J. Kropp, G. W. Breton, S. L. Craig, S. D. Crawford,
W. F. Durland Jr., J. E. Jones III and J. S. Raleigh, J. Org. Chem.,
1995, 60, 4146.
This work was supported by the U.S. National Science
Foundation.
49 F. D. Lewis, C. L. Stern and B. A. Yoon, J. Am. Chem. Soc., 1992,
114, 3131.
5
5
0
1
S. A. Ruetten, J. Phys. Chem. B, 1999, 103, 9285.
S. Patai, ed., in The Chemistry of the Carbon-Halogen Bond, parts 1
and 2, 1973.
References
1
2
3
4
J. Livage, Bull. Mater. Sci., 1999, 22, 201.
52 R. Nakagaki, I. Aoyama, K. Shimizu and M. Akagi, J. Phys. Org.
Chem., 1993, 6, 261.
L. Schmid, M. Rohr and A. Baiker, Chem. Commun., 1999, 2303.
I. Gill and A. Ballesteros, J. Am. Chem. Soc., 1998, 120, 8587.
A. B. Wojcik and L. C. Klein, Appl. Organomet. Chem., 1997, 11,
53 T. K. Pal, G. K. Mallik, S. Laha, K. Chatterjee, T. Ganguly and
S. B. Banerjee, Spectrochim. Acta, Part A, 1987, 43A, 853.
54 P. Suppan, J. Photochem. Photobiol., A, 1990, 50, 293.
55 C. W. N. Cumper and A. Singleton, J. Chem. Soc., B, 1968, 649.
56 S. Y. Grebenkin and B. V. Bol'shakov, J. Photochem. Photobiol.,
A, 1999, 122, 205.
1
29.
J. Livage, C. R. Acad. Sci., Ser. IIb: Mec., Phys., Chim., Astron.,
996, 322, 417.
D. Avnir, Acc. Chem. Res., 1995, 28, 328.
5
6
1
J. Mater. Chem., 2001, 11, 408±418
417

View Article Online
5
5
5
7
8
9
I. Mita, K. Horie and K. Hirao, Macromolecules, 1989, 22, 558.
C. S. Paik and H. Morawetz, Macromolecules, 1972, 5, 171.
M. Ueda, H. B. Kim, T. Ikeda and K. Ichimura, Chem. Mater.,
66 C. S. P. Sung, L. Lamarre and K. H. Chung, Macromolecules,
1981, 14, 1839.
67 L. Lamarre and C. S. P. Sung, Macromolecules, 1983, 16, 1729.
68 M. Kryszewski, B. Nadolski, A. M. North and R. A. Pethrick,
J. Chem. Soc., Faraday Trans. 2, 1980, 76, 351.
69 W. J. Priest and M. M. Sifain, J. Polym. Sci., Polym. Chem Ed.,
1971, 9, 3161.
70 N. R. King, E. A. Whale, F. J. Davis, A. Gilbert and
G. R. Mitchell, J. Mater. Chem., 1997, 7, 625.
71 N. Siampiringue, G. Guyot, S. Monti and P. Bortolus,
J. Photochem., 1987, 37, 185.
72 N. Nishimura, T. Sueyoshi, H. Yamanaka, E. Imai, S. Yamamoto
and S. Hasegawa, Bull. Chem. Soc. Jpn., 1976, 49, 1381.
73 D. G. Whitten, P. D. Wildes, J. G. Paci®ci and G. Irick Jr., J. Am.
Chem. Soc., 1971, 93, 2004.
1
992, 4, 1229.
6
0
1
M. Ueda, H.-B. Kim, T. Ikeda and K. Ichimura, J. Non-Cryst.
Solids, 1993, 163, 125.
N. J. Bunce, G. Ferguson, C. L. Forber and G. J. Stachnyk, J. Org.
Chem., 1987, 52, 394.
T. Asano and T. Okada, J. Org. Chem., 1984, 49, 4387.
N. Nishimura, S. Kosako and Y. Sueishi, Bull. Chem. Soc. Jpn.,
6
6
6
2
3
1
984, 57, 1617.
S. Y. Grebenkin and B. V. Bol'shakov, Chem. Phys., 1998, 234,
39.
Z. G. Gardlund, J. Polym. Sci., Polym. Lett. Ed., 1968, 6, 57.
6
6
4
5
2
418
J. Mater. Chem., 2001, 11, 408±418