Surface modification of silica nanoparticles: A new strategy for the realization of self-organized fluorescence chemosensorst

Article abstract of DOI:10.1039/b502052b

The self-organization of the proper subunits of a fluorescence chemosensor on the surface of silica nanoparticles allows the easy design and realization of new effective sensing systems. Commercially available silica particles (20 nm diameter) were functionalized with triethoxysilane derivatives of selective Cu(II) ligands and fluorescent dyes. Grafting of the sensor components to the particle surface ensures the spatial proximity between the sensor components and, as a consequence, binding of Cu(II) ions by the ligand subunits leads to quenching of the fluorescent units emission. In 9:1 DMSO-water solution, the coated silica nanoparticles (CSNs) selectively detect copper ions down to nanomolar concentrations. The operative range of the sensors can be tuned either by switching the ligand units or by modification of the components ratio. Sensors with the desired photophysical properties can be easily prepared by using different fluorescent dyes. Moreover, the organization of the network of sensor components gives rise to cooperative and collective effects: on one hand, the ligand subunits bound to the particle surfaces cooperate to form multivalent binding sites with an increased affinity for the Cu(II) ions; on the other hand, binding of a single metal ion leads to the quenching of several fluorescent groups producing a remarkable signal amplification. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b502052b

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Surface modification of silica nanoparticles: a new strategy for the
realization of self-organized fluorescence chemosensors{
a
Enrico Rampazzo, Elena Brasola, Silvia Marcuz, Fabrizio Mancin, Paolo Tecilla* and
a
a
a
b
a
Umberto Tonellato*
Received 8th February 2005, Accepted 18th April 2005
First published as an Advance Article on the web 5th May 2005
DOI: 10.1039/b502052b
The self-organization of the proper subunits of a fluorescence chemosensor on the surface of silica
nanoparticles allows the easy design and realization of new effective sensing systems.
Commercially available silica particles (20 nm diameter) were functionalized with triethoxysilane
derivatives of selective Cu(II) ligands and fluorescent dyes. Grafting of the sensor components to
the particle surface ensures the spatial proximity between the sensor components and, as a
consequence, binding of Cu(II) ions by the ligand subunits leads to quenching of the fluorescent
units emission. In 9 : 1 DMSO–water solution, the coated silica nanoparticles (CSNs) selectively
detect copper ions down to nanomolar concentrations. The operative range of the sensors can be
tuned either by switching the ligand units or by modification of the components ratio. Sensors
with the desired photophysical properties can be easily prepared by using different fluorescent
dyes. Moreover, the organization of the network of sensor components gives rise to cooperative
and collective effects: on one hand, the ligand subunits bound to the particle surfaces cooperate to
form multivalent binding sites with an increased affinity for the Cu(II) ions; on the other hand,
binding of a single metal ion leads to the quenching of several fluorescent groups producing a
remarkable signal amplification.
Introduction
strategy simply requires the design and synthesis of a limited
number of relatively simple building blocks which are then
2
allowed to self-organize spontaneously. As a result of the
The interest in fluorescence chemosensors, i.e. molecular
systems that recognize and signal the presence of a selected
substrate by a variation of their fluorescence emission, is
molecular organization into a supramolecular assembly, novel
properties and functions may result and lead to possible
important applications.
1
continuously growing. In fact, such systems provide a
sensitive and selective method to evaluate the presence and
the concentration of different substrates. In addition, their
molecular dimensions allow high response rate and spatial
resolution in the analyte detection, also making possible
intracellular monitoring of selected species in medical and
biochemical studies.
In view of that, exploiting the self-organization of receptors
and fluorescent dyes may, at least partially, allow the synthetic
problems connected to classical systems to be overcome and
provide an efficient strategy for the easy realization and
optimization of fluorescence chemosensors.
A few years ago, we reported a novel methodology to self-
assemble a fluorescent chemosensor for Cu(II) ions within
3
surfactant aggregates. The system (Scheme 1) was based on
In their typical design, fluorescent chemosensors are mole-
cules composed by one or more substrate binding units and
photoactive components, which generate the fluorescence
1
signal. The chemical complexity of such systems varies greatly
from case to case, as the sensor components can be integrated,
complementing each other, or simply connected by spacers.
However, the synthetic efforts involved in the realization and
in the subsequent optimization of chemosensors may be
demanding.
Self-assembling and self-organizing methodologies have
attracted increasing attention during recent years in the
chemistry of complex systems with functional properties.
Indeed, they are at the basis of the so-called ‘‘bottom-up’’
approach, as the building of complex structures with this
{ Electronic supplementary information (ESI) available: synthesis of
compounds 1–8; potentiometric titrations; quenching amplification.
See http://www.rsc.org/suppdata/jm/b5/b502052b/
*tecilla@dsch.univ.trieste.it (Paolo Tecilla)
umberto.tonellato@unipd.it (Umberto Tonellato)
Scheme 1 Surfactant aggregate based self-organized fluorescence
chemosensors.
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 2687–2696 | 2687

View Article Online
co-micelles made by a lipophilic dipeptide ligand (N-decylgly-
cylglycine), a fluorescent dye (8-anilinenaphthalensulfonic
acid, ANS) and a cationic surfactant (hexadecyltrimethyl-
ammonium bromide, CTABr). The concentration of the
species within the micellar aggregate ensured the spatial
proximity between the ligand and the dye so that the
complexation of the Cu(II) ions by the glycylglycine moiety
resulted in a signal generated by the quenching of the dye
fluorescence emission. This self-assembled chemosensor
allowed Cu(II) concentrations to be determined down to the
micromolar range in water at pH 7. The main advantages of
such a system are: a) selectivity, mainly due to the ligand
choice; b) simplicity: the sole mixing of the components (two of
them, CTABr and ANS, are commercially available) in water
was required to realize the sensor; c) modularity, which
allowed the modification or the optimization of the system by
simply substituting one of the components; d) the possibility to
tune the detection range just by the modification of the
3
,4
components ratio.
Then, analogous strategies, based on
5
Langmuir–Blodgett films and trialkoxysilane self-assembled
6
monolayers on quartz surfaces, were successfully explored.
7
However, the actual applicability of such systems is limited by
several factors. In particular, surfactant aggregates are delicate
objects due to their dynamic nature: they form only above the
critical micellar concentration, the fraction of non-micellized
components may be not negligible, and they are very sensitive
to environmental conditions, such as temperature and ionic
strength, which can affect the reproducibility of the sensor
response.
3
Chart 1 Ligands and fluorescent dyes (a: X 5 Si(OEt) ; b: X 5 H).
8
modification of the components ratio. Moreover, cooperation
of the ligand subunits bound to the particles surfaces to form
binding sites with an increased affinity for the substrate
To overcome such limitations, in a recent communication
we described a strategy for the realization of self-organized
fluorescence chemosensors for Cu(II) ions obtained by surface
modification of silica nanoparticles. Commercially available
particles (20 nm diameter) were functionalized (Scheme 2) with
the triethoxysilane derivatives of the selective Cu(II) ligand
picolinamide (1a, Chart 1) and the fluorophore dansylamide
(
Scheme 2) was demonstrated.
Indeed, modification of nanoparticle surfaces offers an
attractive approach to the realization of non-dynamic orga-
nized assemblies of functional subunits which would not
9
suffer from the limitations of surfactant-based systems. Silica
(3a, Chart 1). As in the case of surfactant-based systems,
nanoparticles, in particular, are suitable for the realization of
fluorescence chemosensors: they are transparent to light,
photophysically inert and their surface can be easily modified
grafting of the sensor components to the particle surface
ensures the spatial proximity required to signal Cu(II) by
quenching of the fluorescence emission. In 9 : 1 DMSO–water
solution, the coated silica nanoparticles (CSNs) selectively
detect copper ions down to micromolar concentrations and the
operating range of the sensor can be tuned by the simple
1
0
by reaction with alkoxysilane derivatives.
The use of polymer nanoparticles as chemically inert
matrices to entrap fluorescent chemosensors for intracellular
applications has been proposed by the groups of Kopelman
1
1
and Rosenzweig. The nanoparticle matrix not only protects
the sensor from interference with the cellular content, but also
allows the realization of multicomponent systems. Montalti
and coworkers investigated fluorophore functionalized silica
nanoparticles and reported evidence that collective processes
can arise from the organization of the components in an
1
0
extended network. In a recent report, Larpent and coworkers
described a study on latex nanoparticles containing a Cyclam
ligand and a BODIPY fluorescent dye: they found significant
signal amplification in Cu(II) detection due to the quenching of
1
2
4
4 fluorophores by a single metal ion.
In this paper, we highlight the potentiality of this strategy
for the realization of self-organized fluorescence chemo-
sensors, showing that a library of sensor subunits can be
easily realized and used to prepare different sensing systems
Scheme 2 Coated silica nanoparticles based self-organized fluores-
cence chemosensors.
2
688 | J. Mater. Chem., 2005, 15, 2687–2696
This journal is ß The Royal Society of Chemistry 2005

View Article Online
with the most suitable features for the desired application.
New collective effects due to the spatial proximity of a great
number of subunits, which can give rise to an improvement
of the sensor performance, have been recognized and will be
1
3
described.
Results and discussion
Scheme 3
-aminopropyl-triethoxysilane (APTES) or propylamine.
Design, synthesis and characterization
3
The Cu(II) selective ligands and fluorescent dyes designed for
the present study are shown in Chart 1. Alkoxysilane
derivatives 1a–8a, needed for silica nanoparticle functionaliza-
tion, were prepared. The corresponding compounds not
bearing the triethoxysilane moiety (1b–8b) were also synthe-
sized for comparison purposes.
Fluorescent derivatives 3, 4 and 8 were directly prepared
from commercially available dansyl chloride, NBD chloride
(4-chloro-7-nitrobenzofurazan)
and
4-bromomethyl-7-
methoxycoumarin, while ligand 1 and fluorophores 5 and 7
were prepared respectively from picolinic acid, coumarin-3-
carboxylic acid and 6-methoxynaphthalen-2-carboxylic acid
via thionyl chloride activation.
The choice of receptors 1 and 2 was based on the scrutiny of
several Cu(II) selective ligands, usually based on aminoamides
A two-step synthesis (Scheme 3) was required for ligands 2:
picolinamine was reacted with 2-chloroacetyl chloride to yield
the corresponding amide 9 which was then condensed with
APTES or propylamine.
1
4
or dipeptides, reported in the literature. The key recognition
motif of such ligands is the presence of an amidic group in a
central position. As a matter of fact, the interaction between
the deprotonated amidic group and Cu(II) ions is particularly
strong and such a unique property ensures that, at neutral pH
conditions, Cu(II) is the only metal ion capable of strongly
binding to amide based ligands by promoting their deprotona-
tion. Other transition metals show little or no affinity at all for
the same ligands, unless much more basic conditions are used.
Based on these premises, we selected picolinamide (1) and
In the case of the fluorophore derivatives 6 an alternative
synthetic route not requiring APTES was followed (Scheme 4).
1,8-Naphthalic anhydride was condensed with 3-amino-
propene to yield the corresponding imide 10, which was then
reacted with triethoxysilane in the presence of hydrogen
hexachloroplatinate to obtain 6.
In any case, purification of the triethoxysilane derivatives
was accomplished using flash chromatography on silica gel.
Coated silica nanoparticles (CSNs) were prepared from the
triethoxy derivatives 1a–8a following the procedure recently
(
2-pyridinmethyl)-glycinamide (2) as Cu(II) selective recogni-
tion elements for our sensors. The presence, when possible, of
the pyridine group instead of more widely employed alkyl
amines implies several advantages: i) no protection of the
amino group is required during the synthesis of the ligand,
ii) the chromatographic purification of the alkoxysilane
derivatives is made easier due to the increased lipophilicity
and a lesser basicity, iii) the solubility in organic solvents is
improved.
9
reported by Montalti and coworkers. Commercially available
LUDOX silica nanoparticles were heated at 70 uC in a 1 : 1 : 1
mixture of water, ethanol and acetic acid in the presence of the
desired coating subunits. Subsequent purification involved
precipitation, extraction in organic solvent and size exclusion
chromatography.
1
5
Picolinamide is reported to be a selective ligand for Cu(II)
at neutral pH. Ligand 2 was newly designed, in order to further
increase the Cu(II) affinity of the sensor recognition units, by
adding a third chelating site. Binding constants were evaluated
by potentiometric titrations in water (see ESI{). In the case of
CNS characterized by different ligand to dye ratios, here-
after defined by the ligand molar fraction x, were prepared
using ligands 1a and 2a and fluorophore 3a simply by
modulation of the concentration of the alkoxysilane deriva-
tives in the reaction mixture. In the case of ligand 1a (and dye
2
ligand 1b, precipitation of Cu(OH) prevented the titration of
3
a) the composition of the coating was always found to be
solutions containing equimolar amounts of ligand and metal
close to that of the reaction mixture, while using ligand 2a, a
substantially more abundant dye fraction than that of the
reaction mixture was observed in the resulting CSNs. This is
likely to be ascribed to the fact that the secondary amine of
ligand 2a is protonated in the acidic reaction medium, at
variance with the weakly basic pyridine and aniline nitrogens
of ligand 1a and fluorophore 3a. Hence, electrostatic repulsion
between the positively charged ligand 2a subunits may disfavor
ion. Experiments performed with a 3-fold excess of 1b over
Cu(NO
3
)
2
revealed the formation of complexes with both 1 : 2
values
]?[H ] /[LH] ?[M], see ref. 16) were respectively
8.6 and 217.6. In the case of ligand 2b, titration of an
equimolar solution of ligand and metal yield a log K value of
and 1 : 3 metal to ligand stoichiometry, and the log K
n
+
n
n
n n
(K 5 [ML
2
1
3
.3 for the 1 : 1 complex. Apparent binding constants at a
n
fixed pH value (K (app) 5 [ML ]/[LH] ?[M]) allow an easier
n
n
comparison of the binding strength of the ligands: at pH 7,
log K
log K
2
3
(app) 5 5.4 and log K (app) 5 3.4 for 1b and
1
6
1
(app) 5 10.3 for 2b.
The synthetic route for the alkoxysilane derivatives 1a–8a
and the model compounds 1b–8b was straightforward. With
the exception of ligands 2 and fluorophores 6, the synthesis of
the desired compounds just required a one-step condensation
of a properly activated derivative with commercially available
Scheme 4
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 2687–2696 | 2689

View Article Online
Fig. 1 TEM images of LUDOX silica nanoparticles (a), CSNs with
ligand 1a and dye 3a (b, x 5 0.5), CSNs with ligand 2a and dye 3a
(c, x 5 0.25). The bars correspond to 100 nm.
the attachment to the nanoparticle surface in favor of the dye
subunit.
Nanoparticles covered by the sole fluorescent subunits
4
a–8a turned out to be not dispersible in any solvent and
only mixed coated nanoparticles with an excess of ligand 1a
were prepared from these compounds.
2
1
Fig. 2 Spectrofluorimetric titration of CSNs (0.0015 mg mL ) with
The resulting particles were investigated by transmission
electron microscopy (TEM, Fig. 1). Uncoated LUDOX silica
particles show a mean diameter of 15 ¡ 6 nm, while CSNs
have diameters which range from about 16 ¡ 5 nm, when
functionalized with mixtures containing ligand 1a, to 18 ¡
Cu(NO in 10% water–DMSO, HEPES buffer 0.01 M pH 7, 25 uC,
3 2
)
l
exc 5 340 nm, lem 5 520 nm (m: CNSs coated by only 3a; &: CNSs
coated by 1a and 3a, x 5 0.5; $: CNSs coated by 2a and 3a, x 5 0.5).
and 3a (x 5 0.5), is reported in Fig. 2. In each case, on
increasing the metal ion concentration, the fluorescence
emission decreases. The effect is quite modest in the case of
nanoparticles coated by 3a alone, while almost complete
quenching (down to about 5% of the initial value) is reached
for the CSNs coated with ligands 1a and 2a. The systems are
stable and reproducible and different titration experiments
gave identical results within the experimental error (,2%).
The observed fluorescence quenching is the result of the
binding of the Cu(II) ions to the nanoparticles. This is
supported by the saturation behavior of the profiles reported
in Fig. 2, diagnostic of non-collisional quenching, and by the
fact that the emission recovers its initial value after the
addition of an excess of EDTA, a strongly competitive ligand.
As a consequence, the sensitivity of CSNs containing 2a is
much higher than that of the nanoparticles containing 1a,
which is a weaker ligand for Cu(II). Fig. 2 indicates also that
CSNs coated with 3a alone show a fluorescence decrease upon
Cu(II) addition (Fig. 2, m). A similar effect is obtained by
adding Cu(II) to a 3b solution indicating that, in this solvent
system, the dansylamide unit alone can provide weak Cu(II)
binding sites. However, the presence of the ligand subunits is
necessary to achieve a strong binding of the substrate and
authenticate the CSNs as sensors.
6
nm in the case of ligand 2a. These results indicate,
particularly in the case of CNS coated with ligand 2a, that
the growth of a multilayer thin film of the organic silane
precursors on the surface of the LUDOX silica nanoparticles
is more likely than a monolayer coating. The possibility of
self-condensation of the silane derivatives in the acidic reaction
conditions, leading to the formation of nanoparticles lacking
the LUDOX silica core, was investigated by carrying out the
coating reaction in the absence of the LUDOX nanoparticles.
TEM analysis of the reaction product revealed the absence of
any recognizable nanoparticles.
The coated silica nanoparticles can be stored as dry powders
and are dispersible in non-polar organic solvents (CHCl
3
,
CH Cl ) and in DMSO containing water up to 10%. UV-Vis
2
2
and fluorescence spectra of the fluorophores immobilized on
the silica nanoparticles are similar to those of the reference
1
dyes 3b–8b. H NMR analysis showed the signal broadening
typical of immobilized subunits on the nanoparticle surfaces
and, in most cases, allowed the determination of the ratio of
the components present in the coating.
The number of components present on a single nanoparticle
1
0,17
can be roughly estimated
the dye and the ligand concentration in the CSNs solutions
determined by absorbance and NMR measured component
on the basis of four parameters:
As highlighted in Fig. 2, the sensitivity gain obtained by
using ligand 2a is excellent: if the amount of Cu(II) necessary to
(
1
8
ratio), the particle radii and the density of LUDOX silica. In
our case such estimation gives an average of 7000 subunits per
particle, similar to that calculated by Montalti and coworkers
decrease the initial fluorescence emission by 10% ([Cu(II)]10%
)
is taken as the detection limit, such a limiting Cu(II) con-
2
6
28
1
0
centration decreases from 6.8 6 10 M to 3.0 6 10 M on
going from 1a to 2a containing CNSs.
for 1a coated CSNs. In the approximation of a single
monolayer of subunits on a smooth sphere, this value would
correspond to an average distance between two subunits of
The sensitivity of the CNS to Cu(II) is notably related to the
concentration of the nanoparticles. Fig. 3 shows the depen-
dence of the [Cu(II)]10% values on the amount of 2a/3a CNSs
present in the solution: the sensitivity improves by decreasing
the nanoparticle concentration to reach, between 0.0015 and
˚
about 4 A.
Cu(II) sensing
2
1
The effect of the addition of an aqueous Cu(NO ) solution on
3
0.0006 mg mL , a limiting value close to 30 nM. Further
dilution of the CNSs leads to a worsening of sensitivity. A
similar trend is observed also in the case of 1a/3a nano-
particles, but the benefits deriving from CSNs dilution are
2
the fluorescence emission of CSNs solutions (10% water–
2
DMSO, CSNs 0.0015 mg mL , HEPES buffer 0.01 M,
1
pH 7.0) coated respectively by 3a alone, 1a and 3a (x 5 0.5), 2a
2
690 | J. Mater. Chem., 2005, 15, 2687–2696
This journal is ß The Royal Society of Chemistry 2005

View Article Online
Fig. 3 [Cu(II)]10% values for CSNs coated by 2a and 3a (x 5 0.5)
as a function of the total particle concentration. Conditions: 10%
water–DMSO, HEPES buffer 0.01 M pH 7, 25 uC, lexc 5 340 nm,
lem 5 520 nm.
much less important. Such behavior can be explained as the
result of two opposite effects: on one hand, by decreasing the
nanoparticle concentration the amount of Cu(II) ions neces-
sary to quench the emission of all the fluorophores is reduced
and, hence, the sensitivity is improved; on the other hand, the
decrease of the ligand concentration leads to a decrease of the
overall Cu(II) binding ability of the nanoparticles and to a
lower sensitivity. With the strong ligand 2a the positive effect
of dilution prevails down to very low nanoparticle concentra-
tions while, in the case of the weaker ligand 1a, the two effects
almost cancel each other over a large concentration range.
The ratio between the ligand and the dye in the CNSs
coating, i.e. the x value, is another relevant factor that allows
the tuning of the sensor sensitivity. Fig. 4 shows the variation
of the fluorescence emission upon addition of Cu(II) to
different solutions containing nanoparticles coated by 2a/3a
2
1
Fig. 4 Spectrofluorimetric titrationof: a) 2a/3a CSNs(0.003 mg mL
with Cu(NO
(&: x 5 0; %: x 5 0.02; $: x 5 0.04; #: x 5 0.07; m:
x 5 0.26; n: x 5 0.50); b) 1a/3a CSNs (0.003 mg mL ) with Cu(NO
&: x 5 0; %: x 5 0.15; $: x 5 0.4; #: x 5 0.55; m: x 5 0.81).
Conditions: 10% water–DMSO, HEPES buffer 0.01 M pH 7, 25 uC,
exc 5 340 nm, lem 5 520 nm. The lines represent the fit of the data
see text).
)
3 2
)
2
1
3 2
)
(
(
Fig. 4a) and by 1a/3a (Fig. 4b) in different ligand to
l
fluorescent ratios. In both cases the increase of the x value,
i.e. the increase of the ligand content of the CSNs coating,
leads to an improvement of the sensitivity. In the case of 2a/3a
CSNs, the quenching profiles show a clear biphasic behavior:
first, the strong quenching of part of the initial fluorescence
emission is observed at low Cu(II) concentrations; subse-
quently, further addition of Cu(II) produces a minor emission
decrease which still reaches a plateau value at about 5% of
the initial emission for very high Cu(II) concentrations.
Interestingly, 2a/3a CNSs with different x values display a
similar Cu(II) affinity at low metal ion concentrations,
differing only in the amount of fluorescence quenching
produced. The extent of the initial emission decrease is roughly
proportional to the amount of ligand subunit present in the
nanoparticles and reaches complete quenching for x 5 0.5.
Further increments of the amount of ligand in the CSNs
coating do not produce any sensitivity improvement.
(
reaches about 70%. With these CSNs, however, the biphasic
behavior of the quenching profiles is less evident and different
Cu(II) affinity is shown by CSNs with different x values.
The different behavior of the two CSNs series is the result of
the different Cu(II) binding ability of the ligands subunits 1a
and 2a. In the case of the stronger Cu(II) chelator 2a,
complexation of the metal ions by the ligand subunits occurs
at much lower concentration than the binding to the dansyl
subunits. The two processes are hence well distinguishable in
the titration profile. In the first part, the Cu(II) ions are
presumably complexed by the ligand subunits on the
nanoparticle surface and quench the fluorescence emission of
the surrounding dansyl groups. Once all the ligand subunits
have been saturated by Cu(II) ions, the CSNs show a residual
emission that is probably due to those dye subunits which are
not close enough to a binding site to interact with the metal
ion. Further addition of Cu(II) causes the quenching of this
residual emission due to the low affinity interaction of the
dansyl units themselves with the metal ions, as observed in the
case of the CSNs coated by 3a only (Fig. 2). In the case of
2a/3a CSNs, containing only 2% of ligand subunits, the extent
As already observed in the previous experiments, the 1a/3a
CSNs are generally less sensitive to Cu(II) than 2a/3a CSNs,
probably due to the minor affinity of these particles for the
metal ion. In addition, in the case of the 1a/3a CSNs the
sensitivity variation spans over a larger x interval, and levels
off only when the amount of ligand subunits in the coating
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 2687–2696 | 2691

View Article Online
of the initial quenching is about 20% (Fig. 4a). This value
suggests that a single Cu(II) ion bound to a 2a subunit can
quench the emission of about 10 surrounding dyes.
The easy interchange of the sensor component is one of
the major advantages of the self-assembling strategies. The
results so far reported show how it is possible to switch the
ligand subunit of the CSNs in order to modulate the sensor
operative range. By the same means, the signaling unit, i.e. the
fluorescent dye, can also be easily changed to set up a series of
nanoparticle-based sensors with different photophysical fea-
tures. CSNs coated by 1a and all the trialkoxysilane dye
derivatives 3a–8a have been prepared. As highlighted by
Fig. 5b, the emission spectra of these CSNs span over a large
wavelength interval from 300 to 600 nm (correspondingly,
excitation wavelengths range from 285 to 466 nm), allowing
the choice of the more suitable sensor for the desired
application. Titration (Fig. 5a) of the different nanoparticles
with Cu(II) shows almost identical behaviour for all the CSNs,
confirming that the recognition of the metal ion by the sensor
is due essentially to its interaction with the binding subunit.
Then, we turned our attention to selectivity, which is the
main feature to validate the CSNs as sensors. The effect of the
addition of other divalent metal ions, such as Ca(II), Mg(II),
Zn(II), Cd(II), Co(II), Ni(II), Fe(II) and Pb(II), was evaluated in
This hypothesis is supported by interpolation of the titration
profiles. A simplified model involving the presence of two
independent and different binding sites for metal ions with 1 : 1
stoichiometry was used. The interpolation of each profile
reported in Fig. 4a gave a good fit and average log Kapp values
of 6.1 and 3.7 were determined respectively for the 2a and 3a
subunits. A log K value of 3.5 was obtained by the fit of the
titration profile obtained for CSNs coated with 3a alone. Thus,
using these apparent binding constants, it is possible to
calculate the metal ion content of the nanoparticles. In the
case of 2a/3a CSNs containing a 2% of ligand subunits
2
1
28
(
0.003 mg mL , [2a]tot 5 8.9 6 10
M, [3a]tot 5 4.4 6
2
6
1
0
2
M), the presence of a Cu(II) concentration of 1.0 6
M causes a 10.3% quenching of the initial fluorescence
6
1
0
emission. The amount of metal ions bound to the particles in
2
8
these conditions is 4.73 6 10 M: hence, each Cu(II) ion
causes the quenching of about 9.5 dansyl subunits.
As 20% of the dansyl subunits are quenched in the 2% ligand
containing CSNs, a quenching radius of about 0.7 nm can be
1
9
estimated for a single Cu(II) ion. With the CSNs containing
%, 7% and 26% of ligand subunit the initial quenching
increases respectively to 35%, 55% and 70%, and reaches
00% for a ligand content of 50%. Of course, as the ligand
4
1
content of the coating increases, the number of dye units
surrounding each ligand decreases and the ‘‘spheres of influ-
ence’’ of the different binding sites overlap reducing the
number of dyes effectively quenched by a single Cu(II) ion.
When ligand content reaches x 5 0.5 all the fluorescent groups
are close enough to a binding site to interact with it and by
consequence no benefit is obtained by further increments of
the x value.
In the case of CSNs containing the weaker ligand 1a, the
biphasic shape of the titration profiles is much less evident and,
most noticeably, the Cu(II) affinity of the nanoparticles
apparently increases with increasing x values. The ligand
strength is quite modest and binding of Cu(II) ions occurs
almost simultaneously both at 1a and 3a subunits. In fact,
fitting of the titration profiles for 1a/3a with a x value of 0.15
gave log K values respectively of 3.86 and 3.1. At a constant
nanoparticle concentration, the increase of the x value implies
an increase of the total ligand concentration in the solution
and, as a consequence, a larger extent of binding of the metal
ions. However, also in this case, cooperative effects have been
shown to play an important role. This was highlighted by the
results of experiments carried out on nanoparticle solutions
with different x values but with a constant overall concentra-
8
tion of ligand 1a. These results indicated that CSNs with large
x have an intrinsic greater affinity for the metal ion and this
can reasonably be ascribed to the surface-organization of the
ligand subunits. At low x values, the picolinamide groups are
far from each other on the particle surface and operate as
individual ligands. As the x value increases, also the relative
density of binding units on the CSNs surface increases and this
may lead to the formation of multivalent binding sites (e.g.
with 2 : 1 or 3 : 1 ligand to metal stoichiometries) with a greater
Cu(II) affinity, as shown in Scheme 2.
Fig. 5 Spectrofluorimetric titration with Cu(NO
3 2
) (a) and normalized
emission spectra (b) of CSNs coated by 1a and dyes 3a–8a (x 5 0.90,
21
0
4
l
.003 mg mL , &: 3a, lexc 5 340 nm, lem 5 520 nm; %: 4a, lexc
5
66 nm, lem 5 538 nm; $: 5a, lexc 5 295 nm, lem 5 345 nm; #: 6a,
exc 5 333 nm, lem 5 470 nm; m: 7a, lexc 5 285 nm, lem 5 357 nm; n:
8a, lexc 5 322 nm, lem 5 394 nm). Conditions: 10% water–DMSO,
HEPES buffer 0.01 M pH 7, 25 uC.
2
692 | J. Mater. Chem., 2005, 15, 2687–2696
This journal is ß The Royal Society of Chemistry 2005

View Article Online
Cu(II) with a very large affinity and Ni(II) ions with a much
1
4
lower one. In fact, after the addition of Cu(II) (50 mM) to the
solutions containing each other metal ion, the emission is
reduced to the same value (20%) observed with Cu(II) alone,
thus showing that there is no interference due to these ions in
the Cu(II) determination. Moreover, titration with Cu(II) of a
solution containing the CSNs and all the above metal ions,
each at a 50 mM concentration, results in a quenching profile
very similar to that obtained in the absence of any interferent
(
Fig. 7, squares).
CSNs coated with 2a and 3a (x 5 0.5), at lower metal ion
concentrations, behave in a very similar way, showing an even
better selectivity toward Ni(II) (Fig. 6b). In fact, while the
emission of the CSNs is almost completely quenched (88%
decrease) in the presence of 2.5 mM of Cu(II), the same amount
of Ni(II) ions produced no (or moderate, up to 10%) emission
decrease. Again, competitive titration with Cu(II) of a solution
containing the CSNs and all the above metal ions (each 25 mM)
results in a quenching profile identical to that obtained in the
absence of any interferent (Fig. 7, circles). Such results
underline the fact that the CSNs respond selectively to Cu(II)
ions even in the presence of an excess of several other divalent
metal ions. Using these CSNs, however, remarkable fluores-
cence quenching is observed at higher concentrations of metal
ion: in the presence of a 50 mM concentration of Ni(II) the
emission is reduced to 50%, and the same amount of Co(II),
Ca(II) and Zn(II) induces a 30% emission decrease. Further
investigation is needed to explain such effects, although it is
likely that the presence of several neighboring secondary
amino groups bound to the 2a ligand subunits may lead to the
formation of non-specific metal ion binding sites on the
nanoparticle surfaces.
Fig. 6 Relative fluorescence intensity of: a) 1a/3a (x
5
0.86,
.003 mg mL ) CSNs in the presence of the indicated metal ions
alone (50 mM, back row) and with Cu(II) (50 mM, front row); b) 2a/3a
21
0
2
1
(
x 5 0.50, 0.0015 mg mL ) CSNs in the presence of the indicated
metal ions alone (2.5 mM, back row) and with Cu(II) (2.5 mM, front
row). Conditions: 10% water–DMSO, HEPES buffer 0.01 M pH 7,
25 uC, lexc 5 340 nm, lem 5 520 nm.
the competition experiments reported in Figs. 6 and 7. In the
case of 1a/3a CSNs, with x 5 0.86 (Fig. 6a), the presence of
5
0 mM of Cu(II) quenches the emission of CSNs to 20% of the
Conclusions
initial value, while the presence of the same concentration
of the other cations does not produce any effect on the
fluorescence of the CSNs, with the exception of Ni(II) which
brings about a 45% decrease of the emission. This result is in
line with the reported ability of picolinamide ligand to bind
Although Cu(II) sensing does not attract great interest by itself,
as numerous analytic methods are already available for this
metal ion, the results presented in this work highlight the
potential of the use of coated silica nanoparticles in the
realization of self-assembled fluorescence chemosensors for
transition metal ions and, in future developments, for other
substrates. When compared to the sensors in the surfactant
aggregates, the CSNs based systems retain most of their
peculiar advantages. High sensitivity, down to the nanomolar
range in the case of the 2a/3a nanoparticles, and excellent
selectivity are obtained. Realization of the sensor just requires
the mixing of the selected components in the reaction vessel
and, unlike the micellar systems, a short work-up. Tuning of
the operating range can be obtained by the simple modifica-
tion of the components ratio. Optimization or modification of
the sensor features can be obtained by simply changing one of
the coating components for a different one.
Fig. 7 Spectrofluorimetric titration of 1a/3a (squares, x 5 0.86,
In addition, CSNs based sensors present several new
advantages. Particularly noteworthy is the possibility to store
the sensor as a dry powder, to simply prepare and calibrate
batch solutions and, as a consequence, achieve a much better
reproducibility than the ‘‘on demand’’ prepared micellar
systems. Furthermore, due to their non-dynamic nature,
CSNs are not sensitive to environmental conditions, such as
2
1
21
0
.003 mg mL ) and 2a/3a (circles, x 5 0.50, 0.0015 mg mL
CSNs with Cu(NO in the absence (filled symbols) and in the
presence (empty symbols) of CaCl , MgCl , Zn(NO , Cd(NO
CoCl , Ni(NO , Fe(NO and Pb(NO each 50 mM (in the case of
a/3a CSNs) or 25 mM (in the case of 2a/3a CSNs). Conditions: 10%
water–DMSO, HEPES buffer 0.01 M pH 7, 25 uC, lexc 5 340 nm,
em 5 520 nm.
)
3 2
)
2
2
3
)
2
3 2
) ,
2
3
)
2
3
)
2
3 2
)
1
l
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 2687–2696 | 2693

View Article Online
the composition of the medium in which the analysis is carried
out, in contrast with the case of surfactant aggregates
performed using Merck 60 F254 (0.25 mm) precoated silica
gel glass plates and Machery-Nagel Poligram SIL G/UV254
precoated plastic sheets (0.25 mm). Preparative column
chromatography was carried out on glass columns packed
with Macherey-Nagel 60 (70–230 mesh) and on Kieselgel 60,
0.063 mm Merck silica gel.
Besides all these important advantages, the possibility of
cooperative effects due to the grafting of the sensor com-
ponents to the nanoparticle surface is to be duly emphasized.
In the case of 1a/3a nanoparticles, cooperation between the
ligand subunits was previously shown to lead to the formation
of Cu(II) binding sites with a greater metal ion affinity.
Moreover, in the case of 2a/3a CSNs here reported, the
binding of a single Cu(II) ion apparently quenches the emission
of about ten surrounding dansyl dyes. This gives rise to a
signal amplification which contributes, along with the high
Cu(II) affinity of the 2a ligand, to the achievement of the
remarkable sensitivity of these CSNs. Similar collective
phenomena have been observed in dansyl functionalized
NMR spectra were recorded with a Bruker AC 250F
spectrometer. Chemical shifts are reported relative to tetra-
methylsilane as internal standard. Signals in NMR spectra are
reported as follows: s 5 singlet; d 5 doublet, t 5 triplet, q 5
quartet, m 5 multiplet, b 5 broadened. ESI-MS mass spectra
were obtained with a Navigator ThermoQuest-Finningan
mass spectrometer. GC-MS analyses were performed on an
Agilent 6850 Series GC System, equipped with a quadrupolar
mass analyzer and with a fused silica capillary column EC-1
Alltech (stationary phase polymethylsiloxane, length 30 m,
internal diameter 0.25 mm and film thickness 0.25 mm).
Generally, unless otherwise specified a 250 uC injector tem-
perature was used.
2
0
10
dendrimers, in dansyl coated silica nanoparticles and in
1
2
fluorescent dyes impregnated polymer nanoparticles. These
observations indicate that the role of the nanoparticle scaffold
is more subtle than merely keeping close the sensor compo-
nents: the spatial proximity and organization of the coating
components can eventually lead to the tuning or modification
of the intrinsic properties of the insulated subunits.
Elemental analyses were performed by the Laboratorio di
Microanalisi of the Department of Chemical Sciences of the
University of Padova. Transmission Electron Microscopy
Of course, there are still some limitations of this approach to
be addressed. The first one is related to the synthesis of the
derivatives required for the nanoparticle surface functionaliza-
tion. Handling the trialkoxysilane derivatives is more ticklish
than that of the lipophilic derivatives required for micellar
sensors and not all the sensor components, which can be
designed, are easily synthesized. However, due also to the
growing interest in materials science and nanotechnology,
several trialkoxysilane derivatives, such as 3a, are already
commercially available and many others, predictably soon,
will be put on the market. Once the library of the available and
feasible components is defined, the possible combination
and opportunities to realize diverse systems will become easier
and stimulating.
(
TEM) experiments were performed at the CSPA of the
University of Trieste. TEM images of the particles were
obtained with a Philips EM 208 transmission electron
microscope operating at 100 KeV. Samples for TEM were
prepared by spreading a drop of nanoparticles solution in
21
CHCl (y5 mg mL ) onto standard carbon-coated copper
grids (200 mesh). Dimensional analysis of nanoparticles
from TEM images was made with the Image J software,
developed by the Research Services Branch (RSB) of National
Institute of Mental Health (NIMH), USA (http://rsb.info.nih.
gov/nih-image/about.html).
3
Potentiometric titrations were performed using a Metrohom
7
16 DMS Titrino dynamic titrator. UV-VIS absorption
measurements were performed at 25 uC by means of Perkin-
Elmer Lambda 16 45 spectrophotometers equipped
A more delicate problem involves the solubility of the CSNs.
The solvent system employed in this study, 10% water–DMSO,
is suitable for chemical analysis, as Cu(II) water solutions can
be added directly to the CSNs solution, but it is, of course, not
suitable for more desirable applications of such nanosensors,
as in the case of intracellular analysis. Moreover, water
solubility would allow more practical nanoparticle purification
methods such as dialysis or ultrafiltration. A possible solution
e
with thermostated cell holders. Quartz cells with optical
pathlength of 1 cm were used. Fluorescence spectra were
recorded at 25 uC with a LS-50B spectrofluorimeter equipped
with a Hamamatsu R928 photomultiplier and thermostated
cell holder, quartz cells with optical pathlength of 1 cm
were used.
2
1
The synthetic procedures for compounds 1–8 are reported in
the ESI.{
to this limitation could be, as reported for Au nanoparticles,
the addition of a third component to the coating, such as
poly(ethylene glycol) derivatives, capable of improving the
water solubility of the CSNs. Finally, extension of this strategy
to substrates that do not intrinsically quench fluorescence
emission as Cu(II) or Ni(II) is to be addressed.
Potentiometric titrations
Protonation constants and Cu(II) complex formation con-
stants for ligands 1b and 2b were determined by pH
Work is in progress in our laboratories to better define the
scope, the application and, possibly, overcome the limitations
of this strategy.
potentiometric titrations (25 uC, 0.10 M NaCl). Solutions
23
approximately 1 6 10
M of the hydrochloride salts of
3 2
the ligands and, when necessary, Cu(NO )
were titrated
using a 0.1 M sodium hydroxide solution. The electrode
system was calibrated by titrating a 0.01 M solution of HCl
so that the pKw value was 13.78. The pH and the volume
of added NaOH data were fitted with the computer program
Experimental
General
2
2
All commercially available solvents and reagents were used as
received without further purification. TLC analyses were
BEST to obtain the desired protonation and complex
formation constants.
2
694 | J. Mater. Chem., 2005, 15, 2687–2696
This journal is ß The Royal Society of Chemistry 2005

View Article Online
Spectrophotometric titrations
Cu(NO , Zn(NO , Ni(NO
Pb(NO CaCl and MgCl were analytical grade products.
obtained after the evaporation of the solvent was finally
purified by gel permeation chromatography (Sephadex LH20,
3
)
2
3
)
2
3 2 2 3 2 2
) , CoCl , Fe(NO ) , CdCl ,
CH
1
H-NMR analysis.
2 2
Cl ). For this batch a x value of 0.50 was determined by
3
)
2
2
2
Metal ion stock solutions were titrated against EDTA
following standard procedures. Solution used during the
spectrophotometric measurements and titrations were pre-
pared using deionized water (R . 18 MV), obtained with a
Milli-Q (Millipore) purification system, or HPLC grade
dimethylsulfoxide. The buffer 4-(2-hydroxyethyl)-1-piperazi-
neethanesulfonic acid (HEPES, Alrich) was used as received
and stock solutions (0.2 M) were prepared using Milli-Q water.
Stock solutions of CSNs were prepared in HPLC grade
1
a/3a CSNs (precursor ratio 1 : 1)
1
H-NMR (250 MHz, CDCl
,
3
25 uC) d: 0.67 (4H,
NCH CH CH -Si, 1a and 3a); 1.70 (4H, NCH CH CH -Si,
2
2
2
2
2
2
1
a and 3a); 2.86 (8H, NCH CH CH -Si and N(CH ) , 3a);
2 2 2 3 2
3
.42 (2H, NCH
CH CH -Si, 3a); 7.12, 7.38 (3H, CH(DNS) and
CH(py)); 7.79 (1H, CH(Py)); 8.22, 8.39, 8.48 (7H,
NHCH CH CH -Si and Harom, 1a and 3a). FT-IR (KBr):
370 (OH); 2928, 2869 (CH); 1666 (CLO); 1523 (HNCO); 1304
2 2 2
CH CH -Si, 1a); 6.19 (1H, very broad,
NHCH
2
2
2
CHCl
determined from the absorbance at 340 nm using the e value
3
. The total molar concentration of 3a subunits was
2
2
2
3
2
1
.
2
1
21
4650 ¡ 30 M cm ) of N-propyl-dansylamide. The total
(SLO); 1112 (Si–O) cm
(
concentration of 1a and 2a subunits was calculated using the
1
coating components ratio determined by H-NMR spectro-
2a/3a CSNs (precursor ratio 1 : 1)
1
scopy. The desired amount of CHCl3 solution was then
transferred into a volumetric flask, the solvent was evaporated
H-NMR (250 MHz, CDCl3, 25 uC) d: 0.59 (4H,
NCH CH CH -Si, 2a and 3a); 1.60 (4H, NCH CH CH -Si,
2
2
2
2
2
2
under a gentle N stream and the CSNs were redissolved in
2
2a and 3a); 2.64 (2H, NCH CH CH -Si, 2a; this signal is
2 2 2
DMSO (90%) and water (10%) buffered at pH 7.0 (HEPES,
strongly overlapped with the next); 2.85 (8H, NCH CH CH -
2
2
2
0
.01 M). 2 mL aliquots of these solutions were transferred
Si and N(CH ) , 3a); 3.36 (2H, (CO)CH NH, 2a),
3 2 2
into fluorescence quartz cells, small volumes (up to 50 mL) of
concentrated metal ions solutions were added and the
fluorescence or absorption spectra were recorded.
4.54 ((CO)NHCH -(pyr), 2a), 6.2 (1H, very broad,
2
NHCH CH CH -Si, 3a); 7.10, 7.44, 8.18, 8.34, 8.46 (11 H
2
2
2
Harom and (CO)NHCH -(pyr) 2a and 3a).
2
Fitting of the titration curves was performed with the
2
3
software package Scientist 2.01. A model involving the
formation of 1 : 1 complexes of Cu(II) with two different
ligands was used. Total concentrations of ligand and dye
subunits, as determined by UV-Vis absorbance and NMR
ratios, were set as invariable parameters. Errors on the fits
were always less than 10%.
1a/4a CSNs (precursor ratio 10 : 1)
1
H-NMR (250 MHz, CDCl
a and 4a); 1.72 (NCH
NCH CH CH -Si, 1a and 4a); 6.16, 7.37, 7.80, 8.19, 8.48
7H, Harom and NHCH CH CH -Si, 1a and 4a).
3
, 25 uC) d: 0.70(NCH
2 2 2
CH CH -Si,
1
2
CH CH -Si, 1a and 4a); 3.42
2
2
(
(
2
2
2
2
2
2
1
a/5a CSNs (precursor ratio 10 : 1)
Preparation of coated silica nanoparticles
1
H-NMR (250 MHz, CDCl , 25 uC) d: 0.72 (NCH CH CH -
3
2
2
2
Preparation of CSNs 1a/3a with x 5 0.50 is reported as an
example of a general procedure. All the other CSNs were
prepared following the same procedure with different amounts
of the precursors 1a–8a (the total amount of the two
alkoxysilane derivatives was fixed at 0.25 mmol). The effective
relative amount of the precursors on the surface of the CSNs
Si, 1a and 5a); 1.75 (NCH CH CH -Si, 1a and 5a); 3.44
2
2
2
(
NCH CH CH -Si, 1a and 5a); 7.36, 7.61, 7.82, 8.21. 8.48, 8.82
2 2 2
(H
arom and NHCH
2 2 2
CH CH -Si, 1a and 5a).
1
a/6a CSNs (precursor ratio 10 : 1)
1
1
H-NMR (250 MHz, CDCl , 25 uC) d: 0.72(NCH CH CH -Si,
was determined by H-NMR spectroscopy.
3
2
2
2
1
a and 6a); 1.74 (NCH
2
CH
2
CH
2
-Si, 1a and 6a); 3.42
CH CH -Si, 6a); 7.35,
CH CH -Si, 1a and 6a).
(
NCH CH CH -Si, 1a); 4.10 (NCH
2
2
2
2
2
2
Preparation of 1a/3a CSNs (c 5 0.5)
7
.78, 8.14, 8.47 (Harom and NHCH
2
2
2
1
a (42 mg, 0.13 mmol) and 54 mg (0.13 mmol) of 3a were
placed in a round bottom flask and then dissolved with
.25 mL of ethanol, 1.25 mL of water and 1.25 mL of acetic
acid. A water suspension (125 ml) of silica nanoparticles
LUDOX AS-30 Colloidal Silica, Aldrich) was added. The
reaction mixture was warmed at 80 uC under stirring for
8 hours. After this time ethanol was evaporated under
1
a/7a CSNs (precursor ratio 10 : 1)
1
1
H-NMR (250 MHz, CDCl
a and 7a); 1.74 (NCH
NCH CH CH -Si, 1a and 7a); 3.90 (OCH , 7a); 7.07, 7.36,
3
, 25 uC) d: 0.72(NCH
2 2 2
CH CH -Si,
1
2
CH CH -Si, 1a and 7a); 3.44
2
2
(
(
2
2
2
3
7
.78, 8.20, 8.47 (Harom and NHCH CH CH -Si, 1a and 7a).
2 2 2
4
reduced pressure and solid NaHCO was added to the suspen-
3
1
a/8a CSNs (precursor ratio 10 : 1)
sion to reach a pH value between 7 and 8. The precipitate was
filtered and washed with a borate buffer solution (pH 9.5,
1
H-NMR (250 MHz, CDCl , 25 uC) d: 0.71(NCH CH CH -Si,
3
2
2
2
5 6 2 mL) and with water (5 6 2 mL). The solid was dried
under vacuum and redissolved with 100 mL of dichloro-
2 2 2
1a and 8a); 1.72 (NCH CH CH -Si, 1a and 8a); 3.41
2 2 2 3 2
(NCH CH CH -Si, 1a); 4.25 (OCH and NHCH -arom, 8a;
methane. The organic solution was washed with water
7.33, 7.75, 8.16, 8.46 (Harom and NHCH CH CH -Si, 1a
2
2
2
(
100 mL) and dried over Na
2
SO
4
. The crude solid material
and 8a).
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 2687–2696 | 2695

View Article Online
den Berg, D. N. Reinhoudt and M. Crego-Calama, J. Am. Chem.
Soc., 2004, 126, 7293–7299.
Acknowledgements
7
An alternative approach to the realization of self-assembled
chemosensors, called ‘‘chemosensing ensemble’’, has been pro-
posed and involves the displacement of a non-covalent dye–
receptor complex by the substrate. See for example: (a)
S. L. Wiskur, H. A ¨ı t-Haddou, J. J. Lavigne and E. V. Anslyn,
Acc. Chem. Res., 2001, 34, 963–972; (b) M. A. Hortal a` ,
L. Fabbrizzi, N. Marcotte, F. Stomeo and A. Taglietti, J. Am.
Chem. Soc., 2003, 125, 20–21.
E. Brasola, F. Mancin, E. Rampazzo, P. Tecilla and U. Tonellato,
Chem. Commun., 2003, 3026–3027.
Selected reviews: (a) R. Shenhar and V. M. Rotello, Acc. Chem.
Res., 2003, 36, 549–561; (b) C. M. Niemeyer, Angew. Chem., Int.
Ed., 2001, 40, 4128–4158.
The authors thank Mr. Claudio Gamboz and Prof. Maria
Rosa Soranzo (CSPA, University of Trieste), for kind help
with TEM analysis. Financial support for this research has
been partly provided by the Ministry of Instruction, University
and Research (MIUR contracts 2003030309 and 2003037580)
and by University of Padova (University Reasearch Project
CPDA034893).
8
9
a
a
a
Enrico Rampazzo, Elena Brasola, Silvia Marcuz, Fabrizio Mancin,
a
b
Paolo Tecilla* and Umberto Tonellato*
Dipartimento di Chimica Organica and Istituto CNR di Tecnologia
delle Membrane - Sezione di Padova, Universit a` di Padova, via Marzolo
, I -35131, Padova, Italy. E-mail: umberto.tonellato@unipd.it;
a
a
1
1
0 M. Montalti, L. Prodi, N. Zaccheroni and G. Falini, J. Am. Chem.
Soc., 2002, 124, 13540–13546.
1 (a) S. M. Buck, Y. E. L. Koo, E. Park, H. Xu, M. Brasuel,
M. A. Philbert and R. Kopelman, Curr. Opin. Chem. Biol., 2004, 8,
1
Fax: +39 0498275239; Tel: +39 0498275269
Dipartimento di Scienze Chimiche, Universit a` di Trieste, via Giorgieri
b
5
40–546; (b) S. M. Buck, H. Xu, M. Brasuel, M. A. Philbert and
1, I-34127, Italy. E-mail: tecilla@dsch.univ.trieste.it;
R. Kopelman, Talanta, 2004, 63, 41–59; (c) J. Lu and
Z. Rosenzweig, Fresenius J. Anal. Chem., 2000, 366, 569–575.
2 R. M e´ allet-Renault, R. Pansu, S. Amigoni-Gerbierb and
C. Larpent, Chem. Commun., 2004, 2344–2345.
3 Other examples of nanoparticles based self-organized fluorescence
sensors have been recently reported: (a) K. M. G a´ ttas-Asfura and
R. M. Leblanc, Chem. Commun., 2003, 2684–2685; (b) Y. Chen and
Z. Rosenzweig, Anal. Chem., 2002, 74, 5132–5138.
Fax: +39 0405583903; Tel: +39 0405583925
1
1
References
1
(a) Special Issue on ‘‘Luminescent Sensors’’, ed. L. Fabbrizzi,
Coord. Chem. Rev., 2000, 205; (b) A. P. de Silva,
H. Q. N. Gunaratne, T. Gunnlaugsson, A. J. M. Huxley,
C. P. McCoy, J. T. Rademacher and T. E. Rice, Chem. Rev.,
1
1
4 H. Siegel and R. B. Martin, Chem. Rev., 1982, 82, 385–426.
5 H. L. Conley and R. B. Martin, J. Phys. Chem., 1965, 69,
1
997, 97, 1515–1566; (c) Fluorescent Chemosensors for Ion and
Molecule Recognition, ed. A. W. Czarnik, ACS Symposium Series
38, American Chemical Society, Washington DC, 1993.
2
6 Complex formation involves deprotonation of amide nitrogen: the
914–2923.
1
5
+ n
n
n n
overall binding constant is hence: K 5 [ML ]?[H ] /[LH] ?[M],
2
3
(a) S. Mann, Chem. Commun., 2004, 1–4; (b) S. Hecht, Angew.
Chem., Int. Ed., 2003, 42, 24–26.
(a) P. Grandini, F. Mancin, P. Tecilla, P. Scrimin and U. Tonellato,
Angew. Chem., Int. Ed., 1999, 38, 3061–3064; (b) M. Berton,
F. Mancin, G. Stocchero, P. Tecilla and U. Tonellato, Langmuir,
where M is the metal ion and LH the neutral ligand.
7 N. M. Okun, T. M. Anderson and C. L. Hill, J. Am. Chem. Soc.,
003, 125, 3194–3195.
8 A. van Blaaderen and A. P. M. Kentgens, J. Non-Cryst. Solids,
992, 149, 161–178.
9 In the hypothesis of a homogeneous distribution of the subunits,
the quenching of 20% of the dyes implies that 20% of the particle
surface area is the quenching ‘‘dominion’’ of the Cu(II) ions. A
single particle (radius 9 nm) contains about 7000 subunits and
hence 140 ligand units.
1
1
1
2
1
2
001, 17, 7521–7528.
Other examples of the same approach have been recently reported:
a) Y. Diaz-Fernandez, A. Perez-Gramatges, V. Amendola, F. Foti,
C. Mangano, P. Pallavicini and S. Patroni, Chem. Commun., 2004,
650–1651; (b) Y. Diaz-Fernandez, A. Perez-Gramatges,
4
(
1
S. Rodriguez-Calvo, C. Mangano and P. Pallavicini, Chem. Phys.
Lett., 2004, 398, 245–249.
Y. Zheng, J. Orbulescu, X. Ji, F. M. Andreopulos, S. M. Pham and
R. M. Leblanc, J. Am. Chem. Soc., 2003, 125, 2680–2686.
(a) M. Crego-Calama and D. N. Reinhoudt, Adv. Mater., 2001, 13,
20 F. Vogtle, S. Gestermann, C. Kauffmann, P. Ceroni, V. Vicinelli
and V. Balzani, J. Am. Chem. Soc., 2000, 122, 10398–10404.
21 P. Pengo, S. Polizzi, M. Battagliarin, L. Pasquato and P. Scrimin,
J. Mater. Chem., 2003, 13, 2471–2478.
22 A. E. Martell and R. J. Motekaitis, Determination and Use of
Stability Constants, 2nd edn., VCH, New York, 1992.
23 Scientist 2.01, Micromath Scientific Software, Salt Lake City, 1995.
5
6
1171–1174; (b) L. Basabe-Desmonts, J. Beld, R. S. Zimmerman,
J. Hernando, P. Mela, M. F. Garc ´ı a Paraj o´ , N. F. van Hulst, A. van
2
696 | J. Mater. Chem., 2005, 15, 2687–2696
This journal is ß The Royal Society of Chemistry 2005