1. Introduction

Surface plasmon, excited by the interaction between light and electron plasma waves at the metal surface, has attracted considerable attention for its great potential applications in solar cells, nanophotonics, nanophosphors [1–3]. Furthermore, the enhanced local electromagnetic field resulted from the surface plasmon resonance (SPR) has been widely investigated and utilized in surface-enhanced Raman spectroscopy (SERS) [4] and the fluorescence enhancement of semiconductors located near the metal particles. For example, the photoluminescence (PL) of CdSe QDs located near Au nanoparticles (NPs) is highly sensitive to the spectral overlap between them, the QD/Au NPs ratio, and the mean interparticle separation [5]. Guo et al. [6] have reported a fivefold increase in the fluorescence of CdS-SiO₂ nanowires with Au nanoparticles attachment. According to the previous studies, the following two conditions are essential for fluorescence enhancement of semiconductor nanostructures via surface plasmon resonance. On the one hand, the surface-plasmon energy of metal should match the luminescent photon energy of semiconductor [7]. The energy match provides an effective energy transfer channel between the luminescent semiconductor and the metal surface or vice versa. On the other hand, the distance between semiconductor and metal surface should be large enough to effectively suppress the energy loss induced by the inevitable nonradiative energy transfer from semiconductor to metal [8].

Benefitted from the photo-stable, size-dependent optical properties and chemically functionalizable surfaces, fluorescent semiconductor quantum dots have been widely used for biological label [9]. Compared to CdS and CdSe QDs, CdTe QDs have shown several advantages in fluorescent biological label. For instance, the excitation photon energy of CdTe QDs used for biological label is fundamentally lower than that of CdS and CdSe QDs with the same size as a result of its much smaller band gap at room temperature, which is helpful for reducing damage of the living cells. Additionally, CdTe QDs exhibit better photoluminescent properties with weak defect luminescence and high luminescence efficiency than CdS and CdSe [10]. In view of these factors, enhancing the fluorescence efficiency of CdTe QDs is of great importance for biolabeling applications. In fact, lots of schemes have been proposed for realization the fluorescence enhancement of CdTe QDs. For example, several groups have reported the proper surface modification with various organic and inorganic functional molecules [11] and surface coating [12] to enhance the fluorescence efficiency of CdTe QDs. Komarala group have demonstrated that the plasmons could enhance the emission of CdTe QDs placed in close proximity to the layers of Au NPs on a quartz substrate [13]. In this work, we show that the fluorescence of CdTe QDs could be fine-tuned in Au-SiO₂-CdTe composite NPs synthesized in water solution. By controlling the plasmon coupling between the Au NPs and CdTe QDs via optimizing the size of the Au NPs and the thickness of SiO₂, the fluorescence could be enhanced up to 10 times compared to bare CdTe QDs. With the systematic study on this composite system by correlating fluorescence and lifetime measurement, we provide further physical insight on the interaction of metal and semiconductor nanostructures via plasmons.

2. Experimental section

Au NPs were prepared according to the standard sodium citrate (TSC) reduction method [14]. Namely, Au NPs (18 nm) were synthesized by reduction of the HAuCl₄ (1% in w/v, 500 µL) with the TSC (1% in w/v, 1850 µL). The average diameters of Au NPs increased from 18 nm to 51 nm as the amount of the TSC decreased from 1850 µL to 400 µL. To deposit a SiO₂ shell, we used modified Stöber method [15]. The tetraethoxysilane (TEOS, 40 µL) was added to the solution of Au NPs (10 mL) and stirred for 20 h at room temperature so as to obtain the Au-SiO₂ core-shell NPs (SiO₂ thickness: 1 nm). The thickness of the silica layer was changed from 1 nm to 11 nm with the concentration of the TEOS increased from 40 µL to 440 µL while keeping the same coating time. Then, the (3-aminopropyl)trimethoxysilane (APTMS, 30 µL) was added into the solution of Au-SiO₂ core-shell NPs (20 mL) and the mixture solution was stirred for 1 h to modify the silica surface. CdTe QDs prepared according to the hydrothermal method [16] were mixed with the modified Au-SiO₂ core-shell NPs and stirred for at least 3 h. The as-synthesized ASC composite NPs were collected by centrifugation and sequential washed with ethanol and water for several times.

The morphologies and microstructural information of the samples were obtained using a field-emission scanning electron microscopy (SEM, Hitachi S-4800) and transmission electron microscopy (TEM, FEI Tecnai F20). The energy dispersive spectroscopy (EDS) attached to TEM was used to analyze the composition of the products. The UV-Vis absorption spectra were recorded on a double beam UV-VIS spectrophotometer (TU-1901). The photoluminescence (PL) measurements were performed using a spectrofluorimeter (QM4) with a Xe lamp under the same excitation wavelength (488 nm). The time-resolved PL decay curves were measured using a pico-second laser system (TR-PL, SP2500, ACTON).

3. Results and discussion

3.1 The synthesis mechanism of ASC composite NPs

The multistep synthesis mechanism of the ASC composite NPs were schematically shown in Fig. 1. Firstly, Au NPs with various diameters were synthesized by adjusting the amount of the sodium citrate. The APTMS was hydrolyzed into silanol and amino groups, which was bonded with gold due to the large stability constant of gold amines complex and the silanol groups pointed to solution. This procedure is a surface activation process of Au NPs. After the surface activation, the TEOS was used as a silica precursor. It hydrolyzed under the catalysis of ammonia and reacted with the resulting silanol groups to form a homogeneous silica layer on the Au NPs. The silica surface was then modified with the adding more APTMS, which reacted further with the surface silanol groups to produce silica surface functionalized with -NH₂ groups. In the final stage, CdTe QDs were attached onto the surface of the Au-SiO₂ core-shell NPs after the surface modification because of strong binding interactions between the -NH₂ groups and the CdTe QDs.

 

Fig. 1 Schematics of the synthesis process of Au-SiO₂-CdTe composite NPs.

Download Full Size | PDF

3.2 The morphology and microstructure

Figure 2(a) shows the typical SEM image of the as-prepared mono-disperse spherical Au NPs. The representative TEM image of Au NPs after shell coating [Fig. 2(b)] reveals that an amorphous shell was homogeneously coated across the whole surface of the Au NPs, and the diameter of Au NPs was 29 ± 5 nm. From the high resolution TEM (HRTEM) image [Fig. 2(c)] of the area marked in orange in Fig. 2(b), it can be clearly observed that the NPs were indeed a core-shell structure with a uniform amorphous shell of ~4 nm. The corresponding EDS pattern confirms that the sample was mainly composed of Au, Si, and O elements, except for the Cu element which came from the Cu micro grid used for TEM test. Therefore, the coated layer of the Au NPs was an amorphous SiO₂ shell. Figure 2(d) further shows the magnified image of the area marked in pink rectangle in Fig. 2(c), which indicates that the Au NPs were single crystalline and well-crystallized with clear lattice fringes. The lattice fringes are spaced 0.196 nm apart, which is consistent with the (200) lattice planes of cubic gold (0.203 nm). As confirmed by the HRTEM image in Fig. 2(e), the average diameter of QDs (sky blue circle) was about 12 nm and the QDs were directly connected with the surface of Au-SiO₂ core-shell. The corresponding EDS pattern demonstrates that the QDs are the CdTe QDs [inset in Fig. 2(f)]. Based on the above results, the ASC composite NPs were successfully obtained. The single crystallinity and well-crystallized CdTe QDs were demonstrated by the enlarged HRTEM image shown in Fig. 2(f). The lattice fringes with interplanar spacing of 0.220 nm is consistent with the (220) plane of the cubic phase of CdTe (0.229 nm).

 

Fig. 2 (a) SEM image of the as-prepared Au NPs. (b) TEM image and (c) high-resolution TEM image of the as-prepared Au-SiO₂ particles. (d) The magnified image of the area marked with pink rectangle in (c). (e) The high-resolution TEM image of the as-prepared ASC composite NPs (CdTe QDs: sky blue circles, SiO₂: the red lines or the black arrows). (f) The enlargement HRTEM image of the CdTe QDs. The inset in (c) and (f) (bottom left) are the EDS patterns of the Au-SiO₂ particles and ASC composite NPs, respectively. The inset in the top right of (d) and (f) are the FFT images of Au-SiO₂ particles and CdTe, respectively.

Download Full Size | PDF

3.3 Optical properties

The Au NPs, Au-SiO₂ core-shell NPs, CdTe QDs and ASC composite NPs were dispersed in water for UV-vis absorption, PL and lifetime measurements. During the PL and lifetime measurements, the concentration of ASC composite NPs was essentially equal to that of the bare CdTe QDs (4 × 10⁻⁵ mol/L), and the unlinked QDs could not affect the enhancement or quenching in fluorescence intensity [17]. The solid lines in Fig. 3(a) show the UV-vis absorption spectra of the as-prepared Au NPs with different size distribution. The red shifts of these spectra were observed from 522.5 nm to 541.5 nm when the diameters of Au NPs increased from 18 nm to 51 nm. An emission peak at 531 nm (dash line) was observed in a typical PL spectrum of bare CdTe QDs. This emission peak (2.335 eV) is much higher than the band gap of CdTe crystal at room temperature (1.5 eV), which is due to a typical quantum confinement effect [18]. It can be seen that the absorption spectrum of the Au NPs with diameter of 29 nm (peak wavelength at 528 nm) has the largest overlap with the emission spectrum of the CdTe QDs. Therefore, the 29 nm-Au NPs are fundamentally expected to have the most significant enhancement effect for CdTe QDs [7]. In this regard, the diameter of the Au NPs used in this work is fixed to ~29 nm.

 

Fig. 3 (a) The UV-vis absorption spectra of the as-prepared Au NPs with different size distributions (solid lines) and the PL spectrum of bare CdTe QDs (dash line). (b) The UV-vis absorption spectra of Au NPs (29 ± 5 nm, black), CdTe QDs (green), Au-SiO₂ particles (red) and ASC composite NPs (SiO₂ thickness: 4 nm, blue).

Download Full Size | PDF

Figure 3(b) shows the typical UV-vis absorption spectra of the Au NPs (~29 nm), CdTe QDs, Au-SiO₂ particles and ASC composite NPs (SiO₂ thickness: 4 nm), respectively. As can be seen, Au NPs have a characteristic surface plasmon peak at 528 nm, while CdTe QDs exhibit an absorption peak at around 434 nm. The absorption spectrum position of ASC composite NPs is consistent with the characteristic peak of CdTe QDs and the broad surface plasmon band of the Au NPs. When the Au NPs are coated with SiO₂ shell, the intensity of the plasmon absorption increases and the red-shift in the absorption spectrum is observed, which is due to the increase of local refractive index around the particles after the formation of SiO₂ shells [19].

SiO₂ was introduced to adjust the distance between Au NPs and CdTe QDs due to its optically transparent and dielectric properties, and it has no effect on the optical properties of the adjacent fluorescent molecules [20]. Figure 4(a) shows the PL spectra of ASC composite NPs with various thickness of SiO₂ layer. Apparently, all the spectra show the same characteristic CdTe emission band (531 nm), and the emission intensity of these ASC composite NPs increase remarkably except for the ASC-1 nm sample compared with the bare CdTe QDs. The inset in Fig. 4(a) shows the real-color photographs of the corresponding mono-disperse ASC water solutions in cuvette, which were excitated by the same laser (405 nm, 5 mW). It is quite clear that all ASC solutions exhibited the same green color but with different brightness, which was further confirmed by their corresponding PL spectra [Fig. 4(a)]. In order to quantify the enhancement effect, we define an enhancement factor as: F = I_ASC/I_CdTe, where I_ASC and I_CdTe are the intensity of PL spectra of ASC composite NPs and bare CdTe QDs, respectively. Figure 4(b) plots the F value as a function of SiO₂ layer thickness, which implies that the optimal SiO₂ thickness is ~4 nm and the enhancement factor decreases when the SiO₂ thickness increased thicker than 4 nm. The enhancement of fluorescence is induced by the enhanced electromagnetic field of localized surface plasmon resonance (LSPR). And LSPR has multiple enhancement ways between CdTe QDs and Au NPs. One way is called “excitation enhancement [21],” which refer to the increase in the numbers of photoexcitation and photoexcited fluorophores per unit time induced by the enhanced electromagnetic field from LSPR [17]. Another way is an increase in the radiative decay rate during the process of energy transferring from CdTe QDs to Au NPs caused by the spectral overlap between the UV-vis absorption spectra of Au NPs and the emission spectra of the CdTe QDs, which is called “emission enhancement [21].” Moreover, the nonradiative decay rate also exists in the process of energy transferring from the CdTe QDs to Au NPs. The excitation wavelength 488 nm is away from the LSPR, so the excitation enhancement of LSPR might only have a small contribution in the observed PL enhancement. Moreover, the absorption spectrum of the Au NPs has a large overlap with the emission spectrum of the CdTe QDs, which would lead to the increase of the radiative decay rate. Therefore, the observed enhancement of fluorescence maybe primarily occurred through the emission enhancement of the LSPR. When the SiO₂ layer reaches an appropriate thickness, the interaction between the “excitation enhancement” and “emission enhancement” is the strongest, leading to the largest enhancement factor. The radiative decay rate is decreased while the nonradiative decay rate is increased with decreasing of the thickness of SiO₂. So, when the CdTe QDs are too close to the Au NPs (1 nm), nonradiative energy transfer from the CdTe QDs to the Au NPs would increase considerably, which finally result in the quenching of the emission [8].

 

Fig. 4 (a) The PL spectra of bare CdTe QDs and ASC composite NPs with various thickness of SiO₂ layer. The inset is the real-color photographs of the monodisperse water solution of ASC composite NPs with various thickness of SiO₂ layer in cuvette. (b) SiO₂ layer thickness dependent F value, indicated by the black squares. The red line is used to guide eyes to show the data trend. (c) Representative room temperature TR-PL decay curves of bare CdTe QDs (black rectangles) and ASC composite NPs (29 nm-Au NPs, 4 nm-SiO₂ layer, red circles), along with the exponentially fitted decay curves.

Download Full Size | PDF

In order to further confirm the above discussed enhancement mechanism, the time-resolved PL (TR-PL) measurements were performed by time-correlated photon counting (TCSPC) using a supercontinuum light source with a repetition rate of 10 MHz. Figure 4(c) shows the lifetime decay profiles of the bare CdTe QDs and the ASC composite NPs (SiO₂ thickness: 4 nm) under the excitation of the same picosecond laser (488 nm, 1 mW). All of these decay profiles can be well fitted with a single exponential model. The exponentially fitted life times are 13.0 ns and 2.8 ns for the bare CdTe QDs and the ASC composite NPs, respectively. The shortened lifetime of the ASC composite NPs indicates that the recombination rate of photogenerated electrons and hole is increased, which results in the increase of the radiative decay rate from photoexcited fluorophores. The increase of the radiative decay rate leads to the higher emission quantum efficiency and enhances PL signal [22], which well verifies the above discussed enhancement mechanism.

4. Conclusions

In summary, Au-SiO₂-CdTe composite NPs were successfully synthesized by a multistep approach in water solution. Based on the large overlap between the emission spectrum of CdTe QDs and the absorption spectrum of the Au NPs, the fluorescence effect can be efficiently tuned by adjusting the thickness of SiO₂. A maximum enhancement over 10 folds was achieved in the presence of 4 nm SiO₂ interlayer. These kinds of metal-dielectric-QD nanostructures would be useful for understanding the exciton dynamics of QDs interacting with localized surface plasmon resonance and for enhancing the performance of CdTe QDs as fluorescence probe in applications of biolabeling.