The silica beads were prepared using the Stöber method, and the sizes of these beads were controlled by the amount of ammonium in the preparation. The silica beads were identified from transmission electron micrographs (TEMs). It was shown that the beads became large with an increase in the quantity of ammonium (, all data were listed in ). On the basis of the emission spectral change from Ru(bpy)32+ in solution before and after the bead formation, we suggest that most of the Ru(bpy)32+ molecules in the initial solution were incorporated into the silica beads, and the concentration of Ru-(bpy)32+in the silica beads was estimated to be near 1 × 10−4 M (), almost independent of the size of the silica beads. According to the bead size, the concentration was stated as an average number of dye molecules per bead (). This concentration is not high, so the self-quenching between the dyes was negligible in the silica bead. We found that there was almost no Ru(bpy)32+ leaching to solution even when the dye-labeled beads were stored in water for a week as well as during the operations of shell growth and purification as discussed later, implying that the absorption of Ru(bpy)32+ in the silica matrix was strong.
TEM images for (a) silica bead made by 1 mL of ammonium, (b) silica bead made by 10 mL of ammonium, (c) silver particle seeded bead, and (d) silver nanoshells with silica cores made by 5 mL of ammonium and a metal shell thickness of 30 nm.
Correlation of Ammonium Amount Used in the Preparation of Dye-Labeled Silica Beads and Their Properties
The silica beads were aminated using 3-aminopropyltri-methoxy silane and then conjugated by citrate-protected silver colloids. Such conjugation was verified by the TEM image shown in . These metal colloids on the silica beads acted as seeds for a smooth growth of metal shells.9c
The metal shells on silica beads were grown layer by layer by chemical reduction of silver nitrate and verified by TEM images. The thickness of the metal shell was estimated by subtracting the silica bead diameter from the total diameter of the shell. Besides the first layer, which was about 5 nm, each repetitive silver accumulative operation could lead to about a 10 nm thick metal shell accumulation on the silica bead under the current conditions. The final thickness of the metal shell reached 55 nm after six layers of silver deposition. Compared with the consuming quantity of silver salt in the reduction, the growth of the silver shell was almost 2 times thicker, implying that the forming metal shells might be porous and continuous. This viewpoint was also verified by their TEM images (), on which the margin of metal shell was quite rough.
The dye-labeled silica beads displayed a strong scattering even though it was an orange color, so the absorption of the dye is barely present on the extinction profile. When the silver colloids were seeded onto the silica beads, a metal plasmon absorbance rose near 400 nm (), which was from the conjugated metal particles. The plasmon absorbance band became stronger and broadened with the silver shell growth, but the wavelength did not have significant shifting. This result was consistent with the observation from Liu et al.9c
The plasmon wavelength from the silver nanoshell was red-shifted from 400 to 450 nm when the size of the silica beads was increased from 40 to 600 nm. These shifting values are much shorter than those theoretically predicted,8
which is principally due to the porous property of silver shell as well as the dielectric property of silica bead.
Absorbance spectrum of silver shells with 200 nm diameter silica cores but growing metal shells.
The dye-labeled silica beads displayed an emission maximum at 605 nm upon excitation at 450 nm, 7 nm blue-shifted from free dye in water. It was interesting to notice that the emission band become narrower when the silver shell was deposited on the silica bead (taking an example for the diameter of silica bead =200 nm in ). This narrowness was not observed clearly when the beads were only seeded by silver colloids or coated by thin metal shells but became obvious with growth of the silver shell. The narrower emission spectra for the dye-labeled metal nanoshell have been predicted by theoretical calculation.16
We believe herein that the narrowed emission may be due to a coupling of the excitation energy of the dye with the surface plasmon in the metal shell. The observed emission is the result of plasmon radiating light into the far field. For a thin shell, the coupling is weak so the narrowness is not obvious. With the metal shell growth, the thick shell is more suitable for coupling between exited fluorophore and surface plasmon. This observation, we believe, is an important discovery in fluorophore-metal coupling, related to radiative decay engineering (RDE)1a,b
and surface plasmon-coupled emission (SPCE).18
When the silver shell was removed by dissolution with sodium cyanide in solution, the emission band reverted to that from the dye-labeled bead without a metal shell, implying that the narrowness of the emission spectrum can indeed be attributed to the presence of the metal shell.
Emission spectra of Ru(bpy)32+-labeled silica beads (diameter 200 nm) and their silver nanoshells with different metal thicknesses upon excitation at 450 nm. The emission band was more narrow when the metal shell became thick.
Using the same concentration of dye-labeled bead, we observed that the emission spectrum became stronger with shell growth (taking an example for the diameter of silica bead = 200 nm in ). The enhancement efficiency, estimated to be a ratio of the intensity from the silver shell over that of the silica bead at the same bead concentration, was plotted against (TEM-identified) shell thickness (). It was shown that, for all five different sized silica beads, the emission intensity could not be enhanced efficiently at a thin metal shell below 5 nm but then rose with an increase of metal shell thickness. The enhancement reached saturates at 40 nm. The emission spectrum began to decrease slightly when the shell thickness was over 40 nm. The maximum enhancement efficiency was 15.7 times for the smallest silica bead coated by a 40 nm thick metal shell. This value can be compared with theoretical prediction of 30-fold enhancement for rhodamine 6G in an optimal size and thickness of silver nanocavity.16b
Also, it was shown that the maximum enhancement efficiency was decreased with an increase of silica core size ().
Figure 5 Dependence of emission spectrum of dye-labeled silver shells with silica beads (diameter 200 nm) on metal thickness upon excitation at 450 nm. The Ru(bpy)32+ concentrations were normalized to 4.4 × 10−7 M in solution. The control was done (more ...)
Dependences of the emission enhancement efficiency of silver nanoshells with different silica core beads on the silver shell thickness.
Dependences of maximal emission enhancement efficiency for the silver shells on the silica bead sizes.
Emission enhancement is regarded to come from the EM enhancement mechanism, in which plasmon excitation from the metal particle creates an enhanced EM field nearby.1,4
For the small metal shells (10 nm size), the EM field stays uniform inside the cavity.8
However, for the large metal shells in this study, the finite wavelength effect leads to depolarization, and thus, the field inside the shell is heterogeneous. According to a discrete dipole approximation (DDA) model suggested by Schatz and Hupp, the maximal EM enhancement field occurs close to the shell surface along the x
We expect that the volume of efficient EM enhancement field becomes relatively smaller compared with the whole interior volume of the nanocavity with increasing size of the silica core. It means that the number of dye molecules localized in the efficient EM enhancement field was decreased with increasing size of the silica core, so that the maximal enhancement efficiency was shown to reduce with an increase of silica bead size in this study. Forthermore, it is also expected that the efficient EM enhancement field becomes larger and stronger with an increase of metal shell thickness so that the enhancement efficiency increases with the metal shell thickness.8
When the metal shell creates a saturated efficient EM enhancement field in the cavity, a thicker metal shell is only able to block the emission of dye due to strong absorption at excitation and emission wavelengths so that the enhancement efficiency decreases after a maximal value.
Different from some polymer matrixes that are composed of hydrophilic/hydrophobic regions such as Nafion,18
the silica beads are quite hydrophilic gels. When the Ru(bpy)32+
are physically adsorbed into the silica beads, they are expected to disperse homogeneously into the whole beads. Because the silica beads were relatively large, it was suggested that only a portion of dyes localized in the efficient EM field were enhanced but the others were not. The shell in this study exhibited strong emission enhancement, implying that the enhancement of emission from the dyes in the “hot” region should be much more efficient than the apparent value. Compared with the 1 order of magnitude enhancement for the dye localized on the outside of the metal particle, the interior enhancement is stronger due to a more intensive interior electromagnetic field.
Emission enhancement on metal surfaces is believed to occur through an increase of the intrinsic decay rate for the fluorophore near the metal surface. As a result, lifetime is an important parameter through which to evaluate this mechanism.3
To our knowledge, there are no publications to describe the different hydrophilic/hydrophobic regions in the silica bead. Intensity decays were fit with a double-exponential function model for the Ru(bpy)32+
in the silica beads in the absence or presence of silver shells in this study (), and some data are listed in . It was shown that all chi
values ranged between 1.1 and 1.5, indicating that the double-exponential function fit well. The average
lifetime of Ru(bpy)32+
in beads was 497 ns, longer than that in water (382 ns), which was ascribed to the immobilization of Ru(bpy)32+
in the solid matrix. After coating the silver shell on the silica bead, the lifetime decreased progressively with shell growth, which could be shown in . The shortest lifetime obtained was 131 ns for a 40 nm thick shell on the 200 nm silica bead, which was only about ¼ of that from bare beads. The observed shortest lifetime also correlates to the strongest intensity enhancement for the silver shell sample, implying that the increase of the intrinsic decay rate of the fluorophore near the metal surface was an important factor in controlling emission enhancement. The silver shell on the larger silica shell resulted in a longer lifetime, consistent with the decreased emission intensity. It is known that the decrease in lifetime can lead to the increased photostability as measured in terms of excitation-radiation cycles.3
This suggests that the number of detectable dyes may increase on the order of 50-fold (4 × 13.5) for the particle as compared to a bare bead.
Emission decay curves fits for the Ru(bpy)32+ in bare silica bead, 10 nm thick silver shell, and 40 nm thick silver shells on the silica beads with an average diameter of 200 nm.
Lifetime Data Obtained Using the Double-Exponential Model for the Ru(bpy)32+ Incorporated into Silica Beads and Then Coating with Silver Shells
Today, many assays are based on the use of nanoparticles to obtain brighter emission than a single fluorophore.19
Hence, it is of interest to compare the brightness of these particles to a single Ru(bpy)32+
molecule. This comparison can be made by dividing the intensity of dye-labeled bare bead by the interior number of Ru(bpy)32+
per particle, and we estimate that the particles are at least 1000-fold brighter than a single Ru(bpy)32+
molecule for the small size dye-labeled silver shell. We are currently working on metal shells with an optimal sized silica bead, in which the dyes can be induced more completely. We also expect that the total enhancement effect will depend on the fluorophore used in the core.