Reduction of Self-Quenching in Fluorescent Silica-Coated Silver Nanoparticles
- First Online:
- Cite this article as:
- Viger, M.L., Live, L.S., Therrien, O.D. et al. Plasmonics (2008) 3: 33. doi:10.1007/s11468-007-9051-x
- 642 Views
This paper reports the development of spherical Ag@SiO2 nanocomposites in which fluorescein isothiocyanate molecules have been incorporated using a silane coupling agent and a straightforward microemulsion-based synthesis procedure. The photophysical characteristics of core-shell and coreless nanostructures with similar silica shell thickness and fluorophore densities are measured and compared, and show unequivocally that the presence of the silver core decreases the fluorophore lifetime by a factor as high as 4 and that the steady-state fluorescence intensity is increased by a factor as high as 3. The relationship between the enhancement in fluorescence yield and the influence of the silver core on resonance energy transfer processes was examined by fluorescence lifetime and anisotropy measurements. These Ag@SiO2 core-shell nanoparticles provide higher detectability and lower self-quenching, whereas the faster recycling time offers more robustness toward photobleaching.
Fluorescence techniques have been extensively used as sensitive tools in many areas of science and technology [1–3]. Fluorescence probes find applications as sensors for biomolecules [4–8], in medical diagnostic kits , as well as analytical tools in combating biological warfare [10–12]. Although fluorescence detection offers high sensitivity, most of the organic dyes used as luminescent probes suffer from drawbacks such as hydrophobicity, collisional quenching in aqueous media, and irreversible photodegradation under intense excitation light [13, 14].
The use of nanoparticle technology to overcome the limitations of classical fluorophore probes has increased in recent years, in particular, by the incorporation of organic dyes in nanoparticles (NP), including silica nanoparticles [15, 16]. These dye-doped NPs offer distinct advantages over commonly used dye molecules [15, 17], including high optical detection sensitivity, increased chemical and photostability, lower toxicity, and easier conjugation to target biomolecules.
However, it is well known that the quantum efficiency of most organic dyes is strongly impacted by self-quenching because of resonance energy transfer (RET) occurring between fluorophores when their relative concentration exceeds ~10−3 M [18–20], and one should expect this phenomenon to also affect the maximum fluorophore concentration usable in dye-doped silica NP technology. Lakowicz et al.  reported that the severity of self-quenching among overlabeled fluorescein-human serum albumin molecules decreased when the fluorophores were located in the proximity of subwavelength-size metallic silver islands. This phenomenon, termed metal-enhanced fluorescence (MEF), is based on interactions of the fluorophore with the free electrons in the metal responsible for the so-called plasmon absorption, which in turn can increase the local field and modify the excitation rate, enhance the emissive rates, and decrease the lifetimes of excited states [22–24].
Materials and methods
Fluorescein isothiocyanate (FiTC, isomer I), 3-(aminopropyl)triethoxysilane (APS), Igepal CO-520, cyclohexane, silver nitrate (AgNO3), hydrazine hydrate (N2H4.xH2O), ammonia (NH3, 30%), tetraethoxysilane (TEOS), potassium cyanide (KCN), Tris–HCL, and 40% Ludox solution were all purchased from Sigma Aldrich and were used as received.
Synthesis of Ag@SiO2 + FiTC
Synthesis of the colloidal system is a two-step process (Fig. 1a): FiTC is chemically bound to a silane-coupling agent, in this case APS, to form the reaction product APS–FiTC, and the silver core/silica-shell nanoparticles are then prepared in reverse micelles where the dye-labeled silane copolymerizes with TEOS to form the dye-doped silica shells. The core-shell NP micellar synthesis strategy was adapted from Adair et al. , whereas the inclusion of dye molecules in the silica shell follows a strategy proposed by Van Blaaderen and Vrij .
In a clean and dry test tube, 10.0 mg (27 µmol) of FiTC and 180.0 mg (814 µmol) of APS were dissolved in 2.6 ml of anhydrous ethanol, and the solution was stirred in darkness for a minimum of 12 h to allow the coupling reaction between the amine and isothiocyanate groups to proceed.
To a clean 50-ml Erlenmeyer, 4 ml of Igepal CO-520, 10 ml of cyclohexane, and 325 µl of a 10−2 M AgNO3 solution were mixed and stirred at room temperature to form a homogeneous microemulsion. After 5 min of equilibration, 50 µl of 9 M hydrazine hydrate solution was added to reduce Ag+, and silver nanoparticles were formed within seconds. After completion of the silver cores, 16 µl of NH3 was introduced in the microemulsion as the catalyst in the synthesis of the silica shells. After another 10 min of equilibration, 35 µl of TEOS stock solution (consisting of 50% by weight of TEOS in cyclohexane) was added, immediately followed by the addition of APS–FiTC (0, 40, 80, 160, 320, 640 µl). After a 24-h aging period, the Ag@SiO2 nanocomposites were centrifuged and washed several times to remove unreacted APS–FiTC and resuspended in 15 ml of ethanol/water/tris–HCL solution (7:3:0.1). All measurements were performed on these stock suspensions. Three NP replicates were synthesized for each APS–FiTC concentration, making possible the average of the data.
Synthesis of nanoshell + FiTC
The nanoshells were obtained by mixing 40 µl of 0.25 M KCN solution and 5 ml of Ag@SiO2 nanocomposite suspension with agitation overnight to ensure the complete dissolution of silver cores (Fig. 1b). Each suspension was then washed and centrifuged to remove unreacted ions, followed by resuspension in 5 ml of ethanol/water/tris–HCL solution (7:3:0.1). This follows the procedure given in . All measurements were performed on these stock suspensions.
Particle size, morphology, structure, and polydispersity were determined by transmission electron microscopy (TEM, JEM-1230, JEOL). For TEM measurements, samples were placed on Formvar films (Fluka) supported by nickel hexagonal 200-mesh grids (SPI supplies).
The amount of FiTC incorporated into the silica matrix was established by dissolving the silica from both core-shell NPs and hollow nanoshells with sodium hydroxide. Links between the silica networks are easily destroyed in strongly alkaline solution, thus, liberating the FiTC molecules. Typically, 15 ml of 0.2 M NaOH solution was added to each NP sample. The mixtures became clear after a few hours under moderate stirring. Absorbance was measured at 495 nm (Hewlett-Packard spectrophotometer, 8452A) and compared to a calibration curve prepared by dissolving pure FiTC in a 50:50 v/v mixture of ethanol/water/tris–HCL solution (7:3:0.1) and 0.2 M NaOH solution.
Fluorescence measurements (steady-state and anisotropy) were performed on a Varian Cary Eclipse spectrofluorometer. In all cases, excitation was made at 495 nm. Fluorescence values used in Figs. 4 and 5 were taken as the integration of the fluorescence emission between 505 and 650 nm. Three-milliliter quartz cells with an optical length of 1.0 cm were used for all measurements.
Fluorescence lifetime data was obtained using a custom-made time-correlated single-photon counting (TCSPC) system comprising a violet (405 nm) picosecond pulsed diode laser source (Becker & Hickl, BDL-405), an actively quenched single-photon avalanche detector (id Quantique, id100-50) and a TCSPC acquisition module (Becker & Hickl, SPC-630). The source repetition rate was 20 MHz, and the instrumental response function (IRF) of the system was measured at 92 ps using a 5% Ludox solution. The fluorescence was detected at 530 nm using an interference filter (central wavelength, 528 nm; bandwidth, 38 nm; Semrock), and fluorescence lifetimes were extracted from decay curves using commercially available fluorescence lifetime analysis software (FluoFit, PicoQuant) and a double-exponential fitting model. The fluorescence lifetime values shown in Fig. 6 represent the average lifetime from the two values obtained from the double-exponential fit.
Results and discussion
Physical parameters of prepared nanoparticles
Nanoparticle diameter (nm)
FiTC concentration (mol kg−1)
Labeling efficiencies (%)
Average FiTC–FiTC distance (nm)
0.0E + 00
To evaluate the effects of the silver core on the self-quenching experienced by the fluorescein molecules within the silica matrix, we prepared control samples (nanoshells) from the core-shell NPs by dissolving the silver cores with cyanide (Fig. 2b), taking advantage of the porous nature of the silica shell matrix. It should be noted that the fluorophores incorporated into the silica shell by covalent bonding are not displaced by this etching process. Comparison of core-shell NPs with these nanoshells was chosen rather than with solid silica nanoparticles because silica nanoparticles and silica shells are grown by different synthetic routes. Whereas silica nanoparticles are formed by homogenous nucleation and growth, the NPs’ silica shells are grown after heterogeneous nucleation, and the labeling efficiencies are likely to be different for the two processes, making a comparison using these nanoshells in principle more accurate.
We determined the FiTC concentration in the NP silica shells by dissolving the silica with sodium hydroxide and measuring the released FiTC by absorption spectrometry. These are indicated in Table 1 as molar concentrations, assuming a typical silica density of 2.0 g/cm3 and complete polymerization of TEOS. From these absorption measurements, the fraction of FiTC molecules added that were covalently entrapped inside the silica matrix (labeling efficiency) was calculated as ~20%. Effective dye concentrations inside the silica shells as high as 25 mM were calculated, which are considerably higher than those reported in previous work by other research groups, which are typically at the micromolar level [19, 30, 39]. Interestingly, the average fluorophore-to-fluorophore distance in the silica matrix, calculated from the concentration data, is on the order of the critical Förster distance for fluorescein, i.e., ~42 Å [13, 21]. As dye molecules are randomly distributed in the silica matrix, both the total self-quenching from close-range interactions between aggregated dye molecules and the longer-range intermolecular energy transfer processes (e.g., homotransfer between separated dye molecules and transfer between excited monomers and neighboring quenched aggregates) may be expected in these colloids [18, 40].
The lower fluorescence intensity, as well as the red shift in the fluorescein absorption band (Fig. 3) observed for the nanoshells, is a manifestation of self-quenching. A similar red shift of the emission of fluorescein was also observed by spectrofluorimetry (data not shown). As the dye concentration inside the silica matrix increases, proximity between the dye molecules increases, and the stronger interaction between them lowers the energy of the excited states. The higher fluorescence signal observed for the core-shell NPs can be explained by the interaction of excited fluorescein molecules with the metal core, which competes with RET between neighboring fluorescein molecules, resulting in the mitigation of self-quenching. For the highest FiTC loading used in this study, we observed a decrease in fluorescence emission for both core-shell and nanoshell NPs, probably as a consequence of the EtOH-induced aggregation of nanoparticles discussed above.
Resonance energy transfer between fluorescein molecules should also be observable by fluorescence lifetime measurements; i.e., a decrease in lifetimes of excited states should be observed at high labeling densities [18, 21, 43]. We measured and compared fluorescence lifetimes for core-shell NPs and nanoshells using TCSPC.
Fluorescence lifetime decay curves measured for the two most lightly labeled nanostructures (τ1 and τ2) were fitted using a double exponential, and an average lifetime value was calculated from the two components for each decay curve. Fluorescence lifetime decay curves for the remaining systems (τ3–τ5) contained a wide distribution of lifetimes (a characteristic behavior of systems where fluorophores are randomly distributed over space and orientation) and could not be fitted adequately with a double exponential. The lifetime value corresponding to each sample is an average of all decay times computed within the decay curves using the same fitting routine.
We have reported the observation of a reduction of self-quenching in spherical Ag@SiO2 nanocomposites in which FiTC molecules have been incorporated using a silane-coupling agent. The absence in these nanostructures of a spacer shell to locate the fluorophores at a precise distance from the core  and the distribution of the FiTC molecules throughout the silica shell prevents a quantitative interpretation of the lifetime measurement data; nevertheless, the comparison of photophysical characteristics of both core-shell and coreless nanostructures of similar silica shell thickness and fluorophore densities shows unequivocally that the presence of the silver core decreases the fluorophore lifetime by a factor as high as 4 and that the steady-state fluorescence intensity is increased by a factor as high as 3. The relationship between the enhancement in fluorescence yield and the influence of the silver core on resonance energy transfer processes was examined by fluorescence lifetime and anisotropy measurements. These Ag@SiO2 core-shell nanoparticles provide higher detectability and lower self-quenching, whereas the faster recycling time offers more robustness toward photobleaching. The improved photophysical properties and the small size of such tailored fluorescent nanocomposites could be useful for many sensing and imaging applications, including intracellular sensing of important chemical species.
We are grateful to the Natural Sciences and Engineering Research Council of Canada (NSERC) and the “Fonds Québécois de Recherche sur la Nature et les Technologies” (FQRNT) for the financial support of this research. The authors express their gratitude to the “Service de Microscopie de l’Université Laval” for their assistance with the TEM measurements used in this work. Furthermore, comments and suggestions by members of the Ritcey Group at Laval University are also acknowledged.