Dependence of amine-accelerated silicate condensation on amine structure

Article abstract of DOI:10.1039/b700514h

Diatoms are known to grow elaborate nano- and microstructured silica shells by depositing material from precursor-containing vesicles at mild temperature and pH. Oligo(1-methylazetane) and related moieties, in some cases attached to proteins, are believed

Full text of DOI:10.1039/b700514h

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Dependence of amine-accelerated silicate condensation on amine structure{
David B. Robinson, Judith L. Rognlien, Christina A. Bauer and Blake A. Simmons*
Received 12th January 2007, Accepted 27th February 2007
First published as an Advance Article on the web 16th March 2007
DOI: 10.1039/b700514h
Diatoms are known to grow elaborate nano- and microstructured silica shells by depositing
material from precursor-containing vesicles at mild temperature and pH. Oligo(1-methylazetane)
and related moieties, in some cases attached to proteins, are believed to facilitate this process. To
complement prior studies of more complex amines, we aim to understand why such a unique
structure has evolved through a systematic study of a set of simple model compounds. The degree
to which a series of diamines in solution enhances condensation of phosphate-buffered silicic acid
a
at neutral pH increases with increasing alkylation, a factor more important than amine pK . This
suggests why diatoms often use methylated oligomers. Bis(quaternary ammonium) salts result in
even greater reactivity enhancement, constituting a new class of compounds that promote
condensation under mild conditions. Methods are presented for incorporation of these new
moieties into artificial peptides or other template-forming molecules, which should allow for more
effective production of tailored silica nanostructures.
1
complexes; or partially hydrolyzed alkoxysilane.
6
17–21
The
Introduction
amines previously studied include large molecules such as
7,15,22
Inspired by the ability of diatoms to grow beautiful and com-
plicated skeletons of silica under conditions near room tem-
modified peptides extracted from diatoms,
polyamines
27
dendrimers,
12,23–26,
such as poly(allylamine) and polylysine,
16
1
,2
perature and neutral or mildly acidic pH, much recent work
has been reported on the chemical approach these organisms
use to achieve this, and on artificial schemes to produce similar
results. While other methods for growth of templated silica
oligo(N-methyl azetane)s,
and oligo(ethylene diamine)s
7,28
1
and/or spermine and spermidine.
Studies of smaller
2
0,29
molecules include long-chain diamines,
14,30
functional mono-
18
cyclic diamines, and quaternary
1
amines, amino acids,
9
3
2
1,31,32
33,34
exist, no technology exists for making controlled, complex
ammonium-containing surfactants
or polymers
have
three-dimensional structures on the nanometre and micro-
metre scales under mild synthetic conditions. This has been the
motivation of the previous work: such a technology would
open new paths toward sophisticated micro- and nanoscale
mechanical, optical, or electronic devices, as well as media for
also been undertaken. Most of these studies have focused on
producing silica with mesoscale order, and only a few have
attempted systematic studies relating structure to reactivity.
In an attempt to clarify the complex patterns observed, we
seek to identify a minimal amine moiety necessary to accelerate
silica condensation, and make small structural variations to
understand why this may be so. To accomplish this, we
have screened a few monoamines and a range of short
diamines, varying chain length, degree of alkylation, and
presence of adjacent functional groups. From this, we
have identified structural features that result in significant
changes to reactivity. This may help clarify the mechanism by
which amines interact with silicate to promote condensation
in diatoms, and facilitate creation of structured silica in
artificial settings.
4
chemical reactions and separations.
In diatoms, silica grows at the interfaces of precursor-
containing vesicles with the help of proteins. In the diatom
Cylindrotheca fusiformis these proteins are known to contain
lysines modified with oligo(1-methylazetane), as shown in
Fig. 1, terminated by a primary amine, as well as methylated
5
–8
lysines.
Other structures with intriguing similarities and
9
–11
differences have been observed in other species.
This has led
others to study the ability of various amines to react with silica
precursors such as silicic acid buffered by borate, tris(hydroxy-
1
2–15
methylaminomethane),
or
phosphate;
catecholate
Experimental
All materials were obtained from Aldrich and used as received,
unless otherwise specified. Solutions of the amines under study
were titrated to pH 7.5 with hydrochloric acid, giving a final
concentration of 62.5 mM (no titration was necessary for
quaternary ammonium compounds). Buffered silicic acid
solutions were prepared by neutralizing 0.5 M aqueous sodium
metasilicate (MP Biomedicals, nonahydrate) by addition to
Fig. 1 Monomer of oligoamine found in diatoms.
Sandia National Laboratories, 7011 East Avenue, Livermore,
California 94550 E-mail: basimmo@sandia.gov
Electronic supplementary information (ESI) available: Partial
titration curves of diamines; reactivity of phosphate-buffered silicic
acid induced by diamines at pH 6; SEM micrograph of dried gel using
DMAP Bz. See DOI: 10.1039/b700514h
{
1
.33 equivalents of phosphoric acid in a volume of water
necessary to give a final concentration of 90 mM silicate (i.e.,
after the amine is added). Both reagents were cooled in an ice
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 2113–2119 | 2113

View Article Online
bath before mixing, and then warmed in a room temperature
water bath for 5 min, resulting in a solution of pH 7. These
solutions were prepared immediately before use. An initial
absorbance spectrum was taken from 400 to 500 nm on a
Shimadzu UV-2501 spectrometer. Enough amine was added to
achieve a 5 mM concentration (10 mM for monoamines) and
the absorbance was measured as a function of time thereafter.
The poly(methyl methacrylate) cuvettes were shaken to achieve
initial mixing. Settling was sometimes observed when long time
intervals elapsed between measurements, in which case the
cuvette was inverted once before measurement. Absorbance
spectra increased gradually with decreasing wavelength. The
absorbance at 480 nm was taken as a measure of turbidity.
Scanning electron microscopy specimens were prepared by
washing 0.1 mL gel with 0.4 mL deionized water three times by
spinning for 15 min at 5000 rpm in a Millipore Microcon
filtration tube with a 30 kDa molecular weight cutoff, followed
by slow drying under reduced pressure, and sputtering with
p-Phenylenedimethylenebis(triethylammonium bromide) (X33)
.7 mmol a,a9-dibromo-p-xylene was dissolved in 18 mL
1
acetonitrile with stirring in a 40 uC water bath. 3.4 mmol
triethylamine was added dropwise. A precipitate formed
within 5 min, and the reaction was left to stir for an additional
hour. The product was cooled to room temperature and
isolated as above. A pyridinium version XPy was prepared
H 2
similarly. X33 d (500 MHz, D O, HDO = 4.66): 1.3 (t, 9H,
ethyl); 3.15 (q, 6H, ethyl); 4.39 (s, 2H, benzyl); 7.55 (s, 2H,
aryl). XPy: 5.76 (2H, benzyl); 7.45 (2H, aryl); 7.99 (2H, m-py);
8
.47 (1H, p-py); 8.8 (2H, o-py).
p-Phenylenedimethylenebis(2-acetamidoethyl
dimethylammonium bromide) (XE20 Ac)
2.0 mmol N,N-dimethyl ethanediamine was added dropwise to
2.0 mmol acetic anhydride and 2.0 mmol diisopropylethyl-
amine in 8 mL acetonitrile while stirring in a room temperature
water bath. This was left to stir overnight. This was added
without purification to 1.0 mmol a,a9-dibromo-p-xylene in
2
5
nm gold–palladium alloy. Specimens were examined at 2 and
kV on a Hitachi S-4500 microscope.
1
2 mL acetonitrile at 40 C, following the procedure for X33
thereafter. The N-methyl 2-acetamidoethyl version XE21 Ac
Preparation of quaternary ammonium dimers
and the 3-acetamidopropyl version XP20 Ac were prepared
N,N9-Dibenzyl N,N,N9,N9-tetramethyl propanediammonium
bromide (P22 Bz)
similarly. XE20 Ac dH (500 MHz, D O, HDO = 4.66): 1.9 (s,
2
+
3H, acetyl); 1.96 (s, 2H, acetate); 3.0 (s, 6H, N –CH ); 3.4 (t,
3
+
H, NH–CH ); 3.65 (t, 2H, N –CH ); 4.52 (s, 2H, benzyl);
2
7
2
2
2
.0 mmol N,N,N9,N9-tetramethyl diaminopropane and
.0 mmol benzyl bromide were added dropwise to 18 mL
.61 (s, 2H, aryl). XE21 Ac: 2.03 (s, 3H, acetyl); 3.03 (s, 9H,
4
+
N–CH
benzyl); 7.64 (2H, aryl). XP20 Ac: 1.88 (3H, acetyl); 2.01 (2H,
central propylene CH ); 2.98 (6H, N–CH ); 3.19 (2H, NH–
); 3.23 (2H, N –CH ); 4.48 (2H, benzyl); 7.58 (2H, aryl).
3 2 2
); 3.44 (2H, NMe–CH ); 3.83 (2H, N –CH ); 4.57 (2H,
acetonitrile in a room temperature water bath with stirring,
and left overnight. The precipitated product was filtered,
rinsed with ethyl acetate, in which the reactants and mono-
ammonium salts are soluble, and dried under vacuum. Ethane,
butane, DABCO, and DMAP versions (E22 Bz, B22 Bz,
DABCO Bz, DMAP Bz) were prepared similarly, as were
products of E22 with allyl bromide (allyl E22), and products
with ethyl bromoacetate (P22 AcOEt, E22 AcOEt, DABCO
AcOEt). To induce precipitation of P22 AcOEt, most of the
solvent was removed under vacuum and ethyl acetate was
added. DABCO is 1,4-diazabicyclo[2.2.2]octane. DABCO Bz
was prepared in N,N-dimethylformamide, and the product was
treated with hot isopropanol. DMAP is 4-(dimethylamino)-
pyridine; its reaction product is a monosubstituted pyridinium
2
3
+
CH
2
2
Acetamides
P22 AcNHPr, E22 AcNHPr, DABCO AcNHPr, and P22 AcPrd
were prepared by reacting P22 AcOEt, E22 AcOEt or DABCO
21
AcOEt with 10 mL mg
neat propylamine or pyrrolidine
overnight on a 37 uC shaker. P22 AcNHPr dissolved; excess
amine was decanted from the others. P22 AcPrd was sonicated
briefly in acetonitrile, which was decanted, removing impurities.
Products were dried under nitrogen and then under vacuum.
P22 AcNHPr dH (500 MHz, D O, HDO = 4.66): 0.8 (t, 3H,
2
salt. E22 Bz dH (500 MHz, D O, HDO = 4.66): 3.1 (s, 6H,
2
propyl CH ); 1.43 (sext, 2H, propyl CH –CH –CH ); 2.31
3
2
2
3
N–CH ); 3.9 (s, 2H, N–CH ); 4.6 (s, 2H, benzyl); 7.46 (d, 4H,
3
2
(pent, 1H, central CH ); 3.11 (t, 2H, propyl NH–CH ); 3.21 (s,
2
2
+
+
aryl); 7.53 (m, 1H, aryl). Allyl E22: 3.1 (s, 6H, N–CH
H, N–CH ); 4.0 (d, 2H, allyl N–CH ); 5.7 [t (asym.), 2H, allyl
CH ]; 6.0 (sept, 1H, allyl CH). P22 Bz: 2.36 (pent, 1H, central
CH –); 3.0 (s, 6H, N–CH ); 3.3 (t, 2H, N–CH ); 4.47 (s, 2H,
benzyl); 7.45 (d, 4H, aryl); 7.5 (m, 1H, aryl). B22 Bz: 1.86
3
); 3.8 (s,
6
0
3
2
H, N –CH
.8 (t, 3H, propyl CH
.11 (t, 2H, propyl NH–CH
3
); 3.56 (t, 2H, bridging N –CH
2
). E22 AcNHPr:
2
2
2
3
); 1.43 (sext, 2H, propyl CH
2
–CH
); 3.28 (s, 6H, N –CH ); 4.24 (s,
). DABCO AcNHPr: 0.8 (t, 3H, propyl
); 1.43 (sext, 2H, propyl CH –CH –CH ); 3.11 (t, 2H,
propyl NH–CH ); 4.21 (s, 6H, bridging N –CH ). P22 AcPrd
H NMR: 1.8 (pent, 2H, ring N–CH –CH ); 1.9 (pent, 2H, ring
2 3
–CH );
+
2
2
3
+
–
2
3
2
H, bridging N –CH
2
CH
3
2
2
3
+
(
(
(
2 3 2
2H, central CH ); 2.96 (6H, N–CH ); 3.3 (2H, N–CH ); 4.4
2
2
1
2H, benzyl); 7.47 (4H, aryl); 7.5 (1H, aryl). DABCO Bz: 3.88
s, 6H, bridging); 7.47 (m, 4H, aryl); 7.53 (m, 1H, aryl); the
2
2
N–CH –CH ); 2.25 (pent, 1H, central CH ); 3.25 (s, 6H,
2
2
2
+
N –CH ); 3.35 (d, 2H, N –CH ); 3.4 (t, 2H, ring N–CH ); 3.66
+
benzyl H probably exchange with the solvent. P22 AcOEt: 1.22
3H, ethyl); 2.3 (1H, central CH ); 3.25 (6H, N–CH ); 3.6 (2H,
3
2
2
(
2
3
(t, 2H, ring N–CH ). Acyl protons do not appear in these
2
N–CH ); 4.22 (2H, ethyl); 4.3 (2H, acyl). E22 AcOEt: 1.22 (t,
2
spectra, presumably due to exchange with the solvent.
3
H, ethyl); 4.25 (s, 6H, N–CH
bridging); 4.37 (2H, acyl). DABCO AcOEt: 1.22 (3H, ethyl);
.24 (2H, ethyl); 4.25 (6H, bridging); 4.54 (2H, acyl). DMAP
Bz: 3.06 (6H, N–CH ); 5.18 (2H, benzyl); 6.73 (2H, m-py); 7.24
2H, aryl); 7.34 (3H, aryl); 7.91 (2H, o-py).
3
); 4.24 (q, 2H, ethyl); 4.21 (2H,
Results and discussion
4
3
In an effort to make this work relevant to the chemistry of
diatoms, we use a reactant comparable to the material found in
(
2
114 | J. Mater. Chem., 2007, 17, 2113–2119
This journal is ß The Royal Society of Chemistry 2007

View Article Online
natural waters and accumulated by organisms, silicic acid (the
lowest pK of which is 9.8). Sodium metasilicate, Na SiO
dissolves in water to form Na SiO , which can be acidified.
a
2
3
,
2
H
2
4
This must be done with care, because silicic acid spontaneously
condenses to form a gel at near-neutral pH, especially at high
concentrations or if unbuffered. A phosphate buffer was used,
which establishes mild pH, and affords a solution stable on a
timescale of hours while allowing amine-accelerated reactions
to occur on a rapid but measurable timescale. Measurements
presented here are at pH 7 unless otherwise noted. The
enhanced stability of the silicic acid in these solutions is likely
3
5
to be due to both buffering and complexation, and this is an
1
important factor in the chemistry of condensation. This is
5
3
also suspected to be the case inside a diatom, where silicic
6
acid concentrations are known to be much higher than in
3
seawater. The actual pH in the regions of silica deposition
7
3
8
within a diatom are expected to be as low as 5, but for
reasons made apparent below, these would not be ideal condi-
tions for the full range of our experiments. A disadvantage of
the use of phosphate is that it interferes with the colorimetric
Fig. 3 Reactivity of phosphate-buffered silicic acid induced by
various diamines, as measured by the same method as in Fig. 2, but
with 5 mM of a specified diamine from a stock solution titrated to pH
7.5 by hydrochloric acid. The structures of the diamines are described
in Table 1. The three similar diamines are listed from most to least
turbid.
3
9
molybdate test. However, the turbidity of the solution
increases during gelation, and this property has previously
been shown to be useful for rapid screening of silicate
1
5,40
reactivity under various conditions.
We have found that
as the amine is more alkylated, although no major change is
seen as quaternary alkyl chains become longer than ethyl.
To study substances that bear closer structural resemblance
to the oligoamines found in diatoms, we studied a similar series
of diaminoethanes and diaminopropanes, varying the number
of N-methyl groups on each, as shown in Fig. 3. We abbreviate
the names of each amine by indicating ‘‘E’’ for diaminoethane,
9
0 mM is a silicic acid concentration that results in an optically
uniform gel over a wide range of the amines used, and reacts
on an appropriate timescale for rapid reactivity screening.
To identify the simplest effective analog of the oligo(N-
methylazetane) catalysts present in diatoms, we examined the
reactivity of silicic acid with a variety of mono- and diamines
and quaternary ammonium salts. The absorbance curves of
representative monoamines are illustrated in Fig. 2. All amines
produce more rapid turbidity increases and higher final
turbidities than a control sample using an equal volume of
deionized water in place of the amine. A clear increase is seen
‘‘P’’ for diaminopropane, or ‘‘B’’ for diaminobutane, followed
by a digit indicating the number of methyl groups on each
amine. The structures of those labelled in the figures appear in
Table 1. All combinations of 0–2 methyl groups were tested for
E and P, and one B substance was tested. Again, increased
alkylation results in significant increases in the rate of reaction
and final turbidity; there is a clear correlation between
reactivity and the number of methyl groups on a diamine. In
most cases, the reactivity is near that of similarly alkylated
monoamines, but diamines with more than three methyl
groups are at least as reactive as tetraethylammonium. P22 is
dramatically more reactive than E22, but no major change is
seen between P22 and B22. The oligoamine identified in
Table 1 Structures of linear diamines
n
R
1
R
2
R
3
R
4
E21
E22
P11
P20
P21
P22
IP11
B22
0
0
1
1
1
1
1
2
CH
CH
CH
CH
CH
CH
3
3
3
3
3
3
CH
CH
H
3
CH
CH
CH
H
3
3
3
H
CH
H
H
H
CH
H
CH
3
3
Fig. 2 Reactivity of phosphate-buffered silicic acid induced by
various ammonium salts, as measured by monitoring absorbance at
CH
CH
CH
H
3
3
3
480 nm as a function of time after preparation of a solution of 90 mM
CH
CH
3
3
3
sodium metasilicate, 120 mM phosphoric acid, and 10 mM of a
CH(CH
CH
3
)
2
CH(CH
3 2
)
specified alkylammonium chloride (bromide, in the case of Et
top three are listed in decreasing order of their intial slopes.
4
N). The
3
CH
3
CH
3
3
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 2113–2119 | 2115

View Article Online
C. fusiformis bears a strong resemblance to P22, which
could be considered a small model compound for this larger
structure.
The second pK values of some diamines approach that of
a
the second pK value of phosphate, so it could be suggested
a
that reactivity is related to acid–base chemistry of the amine.
a
Our titration data showed that the second pK of P00 is 9.0,
and that of E00 is 7.1, and that of the others could be found by
subtracting 0.5 for each tertiary amine (see ESI{). Increased
methylation results in lower pK
increased number of methylenes between the amines raises the
second amine pK , which opposes the reactivity trend.
Furthermore, the pK depends on which amines the methyl
a
, but more importantly, an
a
a
groups are on, while the reactivity does not; as Fig. 3 shows,
the reactivity induced by P20, with a pKa of 8.5, is nearly
identical to that of P11, with a pK of 9.0. It is apparent from
a
these results that the reactivity is not simply a buffering
effect by the amine. More alkylated amines are also more
nucleophilic, but if this were important, we would again expect
P20 to be more reactive than P11. Also, both are essentially
fully protonated under the conditions used, which would
inhibit this property. This and the observed effectiveness of
quaternary ammonium suggest that nucleophilicity, considered
Fig. 4 Comparison of the reactivity of phosphate-buffered silicic acid
induced by bis(quaternary ammonium) salts to previously discussed
amines, as measured by the same method as in Fig. 3 (quaternary salts
were not titrated). The structures of the diamines are described in
4
Tables 1 and 2. Et N is tetraethylammonium bromide.
1
7
important in the case of alkoxysilanes, is not important to
the mechanism of accelerated condensation of phosphate-
buffered silicic acid.
the P22 and P22 Bz cases is very fast compared to the timescale
of the experiment. As seen for the tertiary amines, the propyl
spacer is optimal among the three measured. A test using allyl
bromide instead of benzyl bromide shows that the results are
not unique to benzyl derivatives.
Instead, we expect the mechanism to be related to
electrostatic interactions. While silicic acid is essentially fully
protonated at neutral pH, its oligomers (which form sponta-
Conformational constraints in quaternary salts further
enhance reactivity. In Fig. 5, benzylated DABCO shows
significantly more reactivity than its less constrained analog,
E22 Bz. A diammonium salt with a xylyl spacer behaves
similarly to those with alkyl spacers, but when the salt is
composed of conformationally constrained pyridinium instead
4
1,42
neously) have lower pK values.
a
This creates the possibility
of ionic interactions with cationic amines. As observed
here, and is well known, the condensation of silicic acid is
4
1
very sensitive to concentration. To explain why polylysine
behaves differently from lysine with respect to silicic acid,
Coradin et al. have suggested that neighboring cationic amines
help to preconcentrate anionic oligomers in a configuration
Table 2 Structures of bis(quaternary ammonium) bromide salts with
aliphatic spacers
2
3
favorable for reaction. We believe that alkylation of an
amine enhances this effect by decreasing solvent interactions
4
3
with it, favoring association with silicate oligomers. A
version of P11 containing isopropyl side groups instead of
methyl groups (‘‘IP11’’) reacts like P21 instead of P11.
Dependence on such subtle structural effects suggests that
close-range association between the diamine and silicate
occurs. Our results warrant further study of diamines,
including study of the approach presented here, by methods
such as nuclear magnetic resonance, to learn about their
E22 (n = 0), P22 (n = 1), B22 (n = 2)
DABCO
a
R =
E22 Bz
P22 Bz
B22 Bz
DABCO Bz
1
7,37
configuration while interacting with silicate.
allyl E22
The above results lead to the question of the reactivity of
bis(quaternary ammonium) salts. Such structures would
depart from those observed in diatoms, but approach the field
E22 AcOEt
P22 AcOEt
DABCO AcOEt
4
4
of gemini surfactants, which have been applied to silica in
4
5
conventional sol–gel syntheses and surface studies. These are
easily synthesized by reaction of a tertiary amine with an alkyl
bromide. When a highly reactive substrate is used, such as
benzyl bromide, this proceeds quickly and cleanly enough to
facilitate rapid screening of a variety of products. We observe
from Fig. 4 that, whereas P22 is about twice as effective as
tetraethylammonium at inducing turbidity, its benzylated
version is twice as effective again, and the reaction in both
E22 AcNHPr
P22 AcNHPr
DABCO AcNHPr
P22 AcNHPrd
a
R groups attach through the leftmost bond
2
116 | J. Mater. Chem., 2007, 17, 2113–2119
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Fig. 6 Scanning electron micrographs of dried gels formed using
representatives of the several types of amines studied. Top row, from
left to right: control (H
2
O), triethylammonium chloride (Et
3
N),
Fig. 5 Reactivity of phosphate-buffered silicic acid induced by
conformationally constrained bis(quaternary ammonium) salts, as
measured by the same method as Fig. 4. Chemical structures are shown
in Tables 2 and 3.
tetraethylammonium bromide (Et
4
N); bottom row, from left to right:
P22, P22 Bz, and benzyl viologen (BzV). The scale bar shown in the
upper center of the figure applies to all the SEM images.
addressed by spectroscopic methods, as noted above, and by
molecular dynamics simulations. The pH inside the silica
deposition vesicle of a diatom is reportedly lower than those
of alkylammonium, the resulting turbidity is substantially
higher. In fact, very high turbidities are rapidly attained with
the use of alkylated 4,49-bipyridinium salts (viologens) or an
ethylene-bridged 2,29-bipyridinium salt (diquat). Benzyl violo-
gen becomes twice as turbid as P22 Bz.
3
8
used here; this probably helps prevent spontaneous condensa-
tion, or helps produce material with a smoother texture. The
diatom’s use of longer oligoamines may compensate for the
diminished reactivity of tertiary oligoamines as pH is reduced.
Diatoms make limited use of quaternary salts, which have
not been observed in nearby pairs in the polyamines found in
For the amines in Fig. 2 and 3, a reaction resulting in higher
turbidity also approaches that value more quickly, so both
features of the data can be taken as qualitative measures of
reactivity. Consistent correlation between the speed of the
reaction and the final turbidity was observed. In the cases of
P22 and most quaternary salts, the reaction occurs on the
timescale of mixing, and we have assumed that this correlation
still holds. However, DMAP Bz, which contains both a
pyridinium and a tertiary amine, shows quite high turbidity
but slow kinetics compared to others in Fig. 5, and casts
some doubt on the generality of the correlation. To gain an
understanding of the correlation between reactivity and
turbidity, electron micrographs were obtained on representa-
tive dried gels. These appear in Fig. 6. The gels are composed
of fused particles, the size of which increases for more rapidly
reacting amines. The roughness and porosity of the resulting
materials also increases. Particles can become larger if their
growth is rapid enough to outpace nucleation of nearby
particles, and larger particles scatter more light, so the
connection between reactivity and turbidity is plausible.
To gain further evidence for this correlation, several
reactions were run at pH 6 (by increasing the phosphoric acid
ratio from 1.33 to 1.78) and the rates were seen to slow down
to measurable values (see ESI{). Under these conditions,
methyl viologen reached 60 percent of its final value in 3 min,
P22 Bz in 10 min, and P22 in 65 min, suggesting that the
correlation still holds in most fast reactions. The decrease
in reaction rate with decreasing pH helps corroborate the
involvement of anionic silicate species in the reaction.
Constrained quaternary ammonium salts may be more reactive
by providing fixed targets to which silicate can associate more
closely and for longer periods of time. This hypothesis could be
1
0
diatoms. It may be that the more reactive moieties studied
here are avoided by diatoms due to their inhibitory effect on
metabolic functions. In particular, viologens are widely used as
herbicides, so it seems unlikely that diatoms would make use of
them from a metabolic standpoint. However, this does not
preclude their use for artificial silica growth in vitro. In fact,
viologen polymers have been studied for this purpose using a
3
4
conventional sol–gel route.
To learn how bis(quaternary ammonium) groups may
behave in more complex chemical environments, we have
studied the effect of nearby functional groups, particularly
those that may enable attachment of these moieties to larger
molecules. As shown in Fig. 7, the reaction product of P22
with ethyl bromoacetate reacts rapidly with silicate and
contains active ester groups that, while stable in pH 7 aqueous
4
6
solutions, can be readily reacted with neat amines to form
amides. This could provide a simple way to incorporate
bis(quaternary ammonium) into more complex molecules.
Amides in this position have a significant effect on reactivity.
Propyl amides develop turbidity more slowly than their parent
esters, but the final values may be higher or lower. DABCO
derivatives were omitted from Fig. 7 for clarity; the ester
rapidly reaches its final absorbance of 1.0, and the propyl
amide trace is close to that of E22 AcOEt. The tertiary amide
prepared with P22 EtOAc and pyrrolidine rapidly becomes
more than twice as turbid as the ester. This approach is well
suited for incorporation into solid-phase combinatorial
approaches, which would allow for more thorough screening
of the effect of closely neighboring groups on reactivity.
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 2113–2119 | 2117

View Article Online
absorbance curve of XP20 Ac in Fig. 7 is comparable to the
amide-free X33. XP20 Ac has three methylene groups
separating the amide from the diammonium; the XE20 Ac
product, with two methylenes, gives similar results. The
tertiary amide XE21 Ac shows slightly diminished reactivity
compared to the secondary amide. These approaches should
allow easy incorporation of silicate-reactive groups into more
complex molecules such as surfactants or peptides.
Peptide-bound oligoamines in diatoms are tethered by a
lysine residue. The butylene group in lysine, and perhaps the
first few amine monomers, may serve as a spacer to isolate
amines from interference by groups such as primary amides
along the peptide backbone. The strategies noted above should
allow incorporation of bis(quaternary ammonium) moieties
into artificial peptides that mimic those of diatoms while
retaining their exceptional reactivity, perhaps creating a path
toward new artificial silica morphologies.
Fig. 7 Reactivity of phosphate-buffered silicic acid induced by
functionalized bis(quaternary ammonium) salts, as measured by the
same method as Fig. 4. Chemical structures are described in Table 2
and 3. P22 derivatives are shown by circles, E22 derivatives by squares,
and xylyl dibromide derivatives by triangles.
Conclusion
Short, highly alkylated diamines constitute
a class of
minimally complex moieties that strongly accelerate condensa-
tion of silicic acid to silica under mild conditions. However,
nearby functional groups can modify their effectiveness, so one
could expect a larger structure, such as an oligomer, to be
more effective in complex environments. Our observations
support a previously proposed mechanism in which neighbor-
ing cationic amines bring silicate oligomers into a configura-
tion favoring condensation. To this is added a preference for
amines with a high degree of alkylation, which we expect
would facilitate close association with silicate reactants. The
More consistent properties can be obtained when functional
groups are spaced farther from the bis(quaternary ammonium)
moiety. By acylating asymmetric diamines E20 and P20, and
forming a diammonium salt with a xylyl group between them,
we can see the effect of isolating an amide by additional
methylene groups (see Table 3, XE20 Ac and XP20 Ac). The
Table 3 Structures of pyridinium bromide salts and bis(quaternary
ammonium) bromide salts with aromatic spacers. R groups attach
through the leftmost bond
4
2
oligoamines in diatoms exhibit these important features.
However, they are not known to use bis(quaternary ammo-
nium) salts, which are even more reactive than tertiary
diamines and for which we have devised strategies for
incorporation into larger molecules. The application of
these materials in structured environments, such as surfactant
or oligopeptide assemblies, may allow for low-cost and
large-scale synthesis of silica structures with specific and
programmable nano- and microscale geometries, to a degree of
sophistication rivalling that of diatoms.
Viologen:
R = methyl or benzyl
Diquat
DMAP Bz
Acknowledgements
a
This work was performed by Sandia National Laboratories, a
multiprogram laboratory operated by Sandia Corporation, a
Lockheed Martin Company, for the United States Department
of Energy’s National Nuclear Security Administration under
contract DE-AC04-94AL85000.
R =
XPy
X33
References
1
2
A. Falciatore and C. Bowler, Annu. Rev. Plant Biol., 2002, 53, 109.
M. Hildebrand, E. York, J. I. Kelz, A. K. Davis, L. G. Frigeri,
D. P. Allison and M. J. Doktycz, J. Mater. Res., 2006, 21, 2689.
I. Park, Z. Wang and T. J. Pinnavaia, Chem. Mater., 2005, 17, 383.
F. Schuth, Annu. Rev. Mater. Res., 2005, 35, 209.
XE20Ac (n = 0, R
XE21Ac (n = 0, R
1
= H)
= CH
= H)
1
3
)
3
4
5
XP20Ac (n = 1, R
1
N. Kroger, R. Deutzmann and M. Sumper, Science, 1999, 286,
1
129.
6 N. Kroger, R. Deutzmann, C. Bergsdorf and M. Sumper, Proc.
Natl. Acad. Sci. U. S. A., 2000, 97, 14133.
a
R groups attach through the leftmost bond.
2
118 | J. Mater. Chem., 2007, 17, 2113–2119
This journal is ß The Royal Society of Chemistry 2007

View Article Online
7
N. Kroger, R. Deutzmann and M. Sumper, J. Biol. Chem., 2001,
76, 26066.
30 D. Belton, G. Paine, S. V. Patwardhan and C. C. Perry, J. Mater.
Chem., 2004, 14, 2231.
2
8
9
M. Sumper and N. Kroger, J. Mater. Chem., 2004, 14, 2059.
M. Sumper and E. Brunner, Adv. Funct. Mater., 2006, 16, 17.
31 D. H. W. Hubert, M. Jung, P. M. Frederik, P. H. H. Bomans,
J. Meuldijk and A. L. German, Adv. Mater., 2000, 12, 1286.
32 H.-P. Hentze, S. R. Raghavan, C. A. McKelvey and E. W. Kaler,
Langmuir, 2003, 19, 1069.
33 Y. Jia, G. M. Gray, J. N. Hay, Y. Li, G.-F. Unali, F. L. Baines and
S. P. Armes, J. Mater. Chem., 2004, 15, 2202.
1
1
1
0 M. Sumper and G. Lehmann, ChemBioChem, 2006, 7, 1419.
1 N. Poulsen and N. Kroger, J. Biol. Chem., 2004, 279, 42993.
2 T. Mizutani, H. Nagase, N. Fujiwara and H. Ogoshi, Bull. Chem.
Soc. Jpn., 1998, 71, 2017.
13 T. Mizutani, H. Nagase and H. Ogoshi, Chem. Lett., 1998, 133.
14 T. Coradin and J. Livage, Colloids Surf., B, 2001, 21, 329.
15 M. Sumper, S. Lorenz and E. Brunner, Angew. Chem., Int. Ed.,
34 M. J. Adeogun and N. Hay John, Polym. Int., 1996, 41, 123.
35 S. D. Kinrade, R. J. Hamilton, A. S. Schach and C. T. G. Knight,
J. Chem. Soc., Dalton Trans., 2001, 961.
2
003, 42, 5192.
36 S. D. Kinrade, A. M. E. Gillson and C. T. G. Knight, J. Chem.
Soc., Dalton Trans., 2002, 307.
37 K. Thamatrakoln and M. Hildebrand, J. Nanosci. Nanotechnol.,
2005, 5, 158.
38 E. G. Vrieling, W. W. C. Gieskes and T. P. M. Beelen, J. Phycol.,
1999, 35, 548.
39 R. K. Iler, The Chemistry of Silica, Wiley-Interscience, New York,
1979.
1
6 V. V. Annenkov, S. V. Patwardhan, D. Belton, E. N. Danilovtseva
and C. C. Perry, Chem. Commun., 2006, 1521.
7 K. M. Delak and N. Sahai, Chem. Mater., 2005, 17, 3221.
8 S. M. Jones, J. Non-Cryst. Solids, 2001, 291, 206.
9 K. M. Roth, Y. Zhou, W. Yang and D. E. Morse, J. Am. Chem.
Soc., 2005, 127, 325.
1
1
1
2
0 P. T. Tanev, Y. Liang and T. J. Pinnavaia, J. Am. Chem. Soc.,
1
997, 119, 8616.
40 J. L. Trompette and M. Meireles, J. Colloid Interface Sci., 2003,
263, 522.
41 P. Schindler and H. R. Kamber, Helv. Chim. Acta, 1968, 51, 1781.
42 C. C. Harrison and N. Loton, J. Chem. Soc., Faraday Trans., 1995,
91, 4287.
2
2
1 P. T. Tanev and T. J. Pinnavaia, Science, 1996, 271, 1267.
2 R. R. Naik, P. W. Whitlock, F. Rodriguez, L. L. Brott,
D. D. Glawe, S. J. Clarson and M. O. Stone, Chem. Commun.,
2
003, 238.
2
2
3 T. Coradin, O. Durupthy and J. Livage, Langmuir, 2002, 18, 2331.
4 K. Lutz, C. Groger, M. Sumper and E. Brunner, Phys. Chem.
Chem. Phys., 2005, 7, 2812.
43 D. H. Aue, H. M. Webb and M. T. Bowers, J. Am. Chem. Soc.,
1976, 98, 318–329.
44 Gemini Surfactants: Synthesis, Interfacial and Solution-Phase
Behavior, and Applications, ed. R. Zana and J. Xia, Marcel
Dekker, New York, 2003.
2
2
5 S. V. Patwardhan and S. J. Clarson, Silicon Chem., 2002, 1, 207.
6 C. A. Bauer, D. B. Robinson and B. A. Simmons, Small, 2007, 3,
5
8.
45 R. Ryoo, I.-S. Park, S. Jun, C. W. Lee, M. Kruk and M. Jaroniec,
J. Am. Chem. Soc., 2001, 123, 1650; R. Atkin, V. S. J. Craig,
E. J. Wanless and S. Biggs, J. Phys. Chem. B, 2003, 107, 2978;
M. Benjelloun, P. Van Der Voort, P. Cool, O. Collart and
E. F. Vansant, Phys. Chem. Chem. Phys., 2001, 3, 127.
46 P. Haberfield and D. Fortier, J. Org. Chem., 1983, 48, 4554.
2
2
7 M. R. Knecht and D. W. Wright, Langmuir, 2004, 20, 4728.
8 D. J. Belton, S. V. Patwardhan and C. C. Perry, J. Mater. Chem.,
2
005, 15, 4629.
9 D. Belton, S. V. Patwardhan and C. C. Perry, Chem. Commun.,
005, 3475.
2
2
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 2113–2119 | 2119