- Email: [email protected]
Encapsulation of emitting CdTe QDs within silica beads to retain initial photoluminescence efficiency
Journal of Colloid and Interface Science 316 (2007) 420–427 www.elsevier.com/locate/jcis
Encapsulation of emitting CdTe QDs within silica beads to retain initial photoluminescence efficiency Ping Yang, Masanori Ando, Norio Murase ∗ Photonics Research Institute, National Institute of Advanced Industrial Science and Technology, Midorigaoka, Ikeda-city, Osaka 563-8577, Japan Received 23 March 2007; accepted 23 August 2007 Available online 31 August 2007
Abstract Highly luminescent silica beads (30 nm–2 µm ∅) incorporating CdTe quantum dots (QDs) were prepared via a two-step preparation procedure, namely a modified Stöber synthesis and a subsequent reverse micelle route. In the modified Stöber synthesis, the silica molecules are deposited on the surface of the QDs. After this first step, these coated QDs were incorporated into silica beads via a reverse micelle route. Inductively coupled plasma analysis revealed a red-emitting silica bead of 30 nm in diameter thus prepared encapsulated roughly 14 CdTe QDs. These glass beads (30–40 nm ∅) retained the initial photoluminescence (PL) efficiencies of the colloidal QDs (27 and 65% for the green- and red-emitting beads, respectively). The protection of QDs by a silica layer at the first step, together with the short total reaction time, is the main reason for the retention of the PL efficiency. The size of the glass beads can be easily controlled over the wide range by adjusting the injection speed and the ratio of chemicals used for the reverse micelle preparation. Since the original efficiency was maintained in the beads and is the highest ever reported for QD-containing silica beads, the method presented here is of significant importance for applications of silica beads to biological probes. © 2007 Elsevier Inc. All rights reserved. Keywords: CdTe; Nanocrystals; Silica bead; Luminescence
1. Introduction Highly luminescent semiconductor quantum dots (QDs) have been attracting much interest because they have sizedependent emission wavelength reflecting three-dimensional quantum confinement effect, narrow width of emission spectra, wide excitation wavelength range and fast emission decay that prevents the emission intensity from saturation. These features of the luminescent QDs would lead to various applications including fluorescent biomarkers and informational electronics such as displays and lighting. The development of biocompatible highly luminescent materials for cell staining and visualization has been an interesting objective in cellular biology and ultra-sensitive immunoassay, because luminescent probes for biomolecular labeling and recognition are of great importance in the fields of chemistry, biology, and medical sciences as well as in biotechnology. Semiconductor QDs covalently coupled to * Corresponding author. Fax: +81 72 751 9637.
E-mail address: [email protected] (N. Murase). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.08.058
biomolecules are expected to be used in ultra-sensitive biological detection because of their intense luminescence, narrow emission spectral width, and higher stability against photobleaching compared with organic dyes [1]. Recently, increasing attention has been directed to the coating of emitting QDs with a layer of transparent silica. Such coating is expected to bring many advantages because the thin silica layer as a protective capping material on the QDs increases the mechanical stability, enables a transfer into various organic and aqueous solvents, and protects QDs against oxidation and agglomeration. Furthermore, the QDs would benefit from having a silica shell to impart biocompatibility, because surface of silica can be easily modified to link bioconjugators [2,3]. There have been several reports to deposit a silica layer on semiconductor QDs [4–9]. To achieve noticeable signal increment in ultra-sensitive bioanalysis, silica spheres encapsulating highly luminescent CdTe QDs at high concentrations are a good candidate. It has recently been reported that thioglycolic acid (TGA)-capped CdTe QDs prepared by an aqueous solution method show very high photoluminescence (PL) efficiency
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
of 65% without post-preparative treatment when reaction conditions are optimized [10]. The Stöber synthesis of colloidal silica, by which silica particles with a diameter less than µm range are obtained, was first described in 1968 [11]. QD–SiO2 nanoparticles having so called raisin-bun-type composite structure were formed by addition of sodium silicate, with either CdTe, CdSe or CdSe/CdS QDs [12]. Bawendi’s group reported on the incorporation of luminescent CdSe/ZnS QDs into a SiO2 shell around SiO2 spheres of several µm in diameter, with PL efficiency of up to 13% [13]. Later, Nann and co-workers indicated that the silicacoating on organic solvent-dispersible QDs results in ordered structures under optimal conditions. They successfully obtained 30–120 nm ∅ silica spheres incorporating single QD in each sphere [7]. However, the formation of a silica shell on the QDs leads to a drastic reduction of PL efficiency compared with their initial value [13]. In contrast, Ow et al. prepared silica– organic dye composite spheres with 20–30 nm in diameter by a Stöber synthesis, and reported that the composite spheres exhibited 20 times brighter PL than that of the organic dye before coating silica layer [14]. These observations indicate that the PL efficiency of such composite spheres is extremely sensitive to incorporation conditions of fluorescent materials in silica spheres. Therefore, it is necessary to develop a novel preparation procedure for retaining the high PL efficiency of QDs in SiO2 beads. A reverse micelle approach, which was proposed 6 years ago by us [15], is a convenient and advantageous technique to coat a silica layer on QDs. Compared with the Stöber method, a reverse micelle route can easily yield more uniform sized silica spheres in the range of 30–150 nm in diameter [3,16] because the size of a water pool, which is a reaction field for the formation of silica spheres, is rather uniform. Furthermore, it is reported that the reverse micelle route normally does not depend very much on reaction conditions, leading to the formation of small silica particles with relatively low density of defects, and tunable size [8]. In this regard, several groups have developed different reverse micelle approaches to form QD–SiO2 composite beads. Recently, single QD encapsulated within a silica bead was prepared by Nann and co-workers [8]. Gao’s group prepared emitting silica beads incorporating CdTe QDs by a reverse micelle route with a PL efficiency of 7% [16]. Ying’s group reported on robust, non-cytotoxic, silica-coated CdSe QDs with 20% PL efficiency [4]. Bakalova’s group prepared single CdSe/ZnS QD with a SiO2 shell by dot-micelles, in which hydrophobic QDs were dispersed in detergent aqueous solution to form micelles. Their silica-coated QDs revealed about 64% PL efficiency [3]. To the best our knowledge, this value (64%) is the highest reported in previous literature. However, their PL efficiency of the silica-coated single QD micelles in aqueous solution is 30–50% lower than that of the initial QDs in chloroform. Therefore, when using a conventional Stöber synthesis or a conventional reverse micelle route, it would be rather difficult to obtain QD-incorporated silica spheres with retention of the initial PL efficiency, because a long reaction period would possibly cause surface deterioration, such as increase of surface defect density that slowly quenches the PL of
421
the QDs. In literature, strong PL was only observed in luminescent SiO2 spheres of several tens of nm in diameter doped with organic dye molecules [17]. They indicated dyes encapsulated within SiO2 spheres exhibited excellent photostability, facile surface modification for bioconjugation, and size uniformity and tunability. Since semiconductor QDs have advantages over organic dyes due to better long-term stability and wavelength tunability, development of highly luminescent QDs–SiO2 beads would provide wide applicability. According to our previous work [5], CdTe–SiO2 composite particles of a few tens of nm in diameter with 5–10% PL efficiency were obtained by a conventional reverse micelle route after reaction for 48 h. Therefore, it is a challenge to develop a novel preparation method with shorter reaction period to retain the initial PL efficiency of the QDs in SiO2 beads. In this paper, we report on successful preparation of highly luminescent, size-controlled (30 nm–2 µm ∅) SiO2 beads incorporating multiple CdTe QDs via a two-step preparation procedure. At the first step, SiO2 -coated CdTe QDs were prepared by a modified Stöber method. Subsequently, the SiO2 -coated CdTe QDs solution was injected into an organic solution to get sizecontrolled CdTe–SiO2 beads by a reverse micelle route. The emitting beads retained the initial PL efficiency of the QDs in colloidal solution. 2. Experimental 2.1. Starting materials All chemicals used were of analytical grade or of the highest purity available. Pure water was of a Milli-Q synthesis grade. TGA-capped CdTe (red- and green-emitting) QDs in aqueous solutions were prepared by a procedure using cadmium perchlorate and hydrogen telluride as described in a previous paper [10]. Green- and red-emitting TGA-capped CdTe QDs with the mean size 2.6 and 3.9 nm in diameter, respectively, were used for further preparation. The PL efficiencies of the green- and red-emitting QDs in aqueous solutions were 27 and 65%, respectively. 2.2. Step 1: Synthesis of SiO2 -coated CdTe QDs A two-step preparation procedure was used to derive the CdTe–SiO2 beads (Scheme 1). In the first step, the SiO2 -coated CdTe QDs were prepared by a modified Stöber technique. Typically, CdTe colloidal solution (1 mL), water (1 mL), diluted ammonia (50 µL, the volume ratio of water to ammonia (25%) is 3:1), and tetraethyl orthosilicate (TEOS, 0.2 mL, [H2 O]/[TEOS] ∼ 120) were mixed in a beaker covered with aluminum foil to reduce the evaporation of ammonia during coating. In this stage, the pH of the precursor is about 10. After continuous stirring for 3 h, the solution of the SiO2 -coated CdTe QDs was obtained. 2.3. Step 2: Preparation of CdTe–SiO2 beads CdTe–SiO2 beads were prepared by a reverse micelle route using the above mentioned SiO2 -coated CdTe QDs as a starting
422
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
Scheme 1. Encapsulation of emitting CdTe QDs within a SiO2 bead via a modified Stöber synthesis (Step 1) and a reverse micelle route (Step 2).
material. A reverse micelle employing Igepal CO-520 as a surfactant, and cyclohexane as a hydrophobic organic solvent was prepared with different water/oil ratios (W/O). Typically, Igepal CO 520 (2.25 g) added in cyclohexane (12.5 g) was stirred until the solution remained clear. Then, the SiO2 -coated QDs solution (pH ∼ 9, 2 mL) obtained in Step 1 was injected into the stock solution dropwise to form the reverse micelles under vigorous stirring. Except for W/O, the size of a water pool in the reverse micelle depended strongly on the injection speed of water under an alkaline condition. To encapsulate the multiple SiO2 -coated CdTe QDs into a SiO2 bead, TEOS (0.3 mL) was injected into the microemulsion. After that, diluted ammonia was injected to adjust pH to ∼10. The microemulsion was then stirred for 4 h. 2.4. Characterization Absorption and PL spectra were taken using conventional spectrometers (Hitachi U-4000 and F-4500). The PL efficiencies of the emitting of CdTe–SiO2 beads and CdTe QDs in solutions were estimated by the method previously reported [18]. Briefly, the absorbance dependence of PL intensity was measured by using standard solution (quinine sulfate in 0.05 M H2 SO4 ). After that, the absorbance and PL of the sample solution in interest were measured. The efficiency was derived by comparing those with the previously measured ones for the standard solution. To measure precisely the absorption and PL spectra of the emitting beads of several tens nm in size, a blank microemulsion without emitting beads was prepared as a reference, and the absorption and PL spectra of the reference were subtracted from those of the sample microemulsion. Transmission electron microscopy (TEM) observation and energy dispersive X-ray (EDX) analysis were performed on a Hitachi H-9000NA (300 kV) microscope and a FEI Technai G2 F20 (200 kV) microscope, respectively. For preparing a TEM specimen, the beads were first precipitated by ace-
tonitrile (CH3 CN) from the microemulsion, and centrifuged at 4000 rpm for 15 min. After that, the powder sample was dried at 40 ◦ C for 1 h. The green-emitting CdTe–QDs contain sulfur ions solely on the surface, whereas the red-emitting ones impregnate them due to hydrolysis of surfactant (TGA) during lengthy reflux upon synthesis [19,20]. Therefore, we used green-emitting QDs to prepare samples for measuring the S/Cd ratio of the samples before and after Step 1. The molar ratio of Cd/Si of red-emitting CdTe–SiO2 beads was measured with a Horiba Jobin Yvon JY-238U Ultima inductively coupled plasma (ICP) spectrometer. The optical images and mean size of the CdTe–SiO2 beads of few micrometers in diameter were obtained with a Nikon Eclipse 80i fluorescent microscope. For the beads with sub-micrometer in diameter, size and its distribution in solution were obtained with a dynamic light scattering particles analyzer (Nanotrac, Nikkiso). Zeta-potential was measured by an electrophoretic light scattering spectrophotometer (ELS-8000, Otsuka Electronic) at pH 10. 3. Results and discussion 3.1. Formation of SiO2 -coated QDs and CdTe–SiO2 beads The formation of sol–gel silica is based on complicated hydrolysis–condensation reactions of silicon alkoxide (Si(OR)4 ) in alcoholic solutions. The polymerization of silica sol in an alkaline solution occurs by internal condensation and crosslinking to give dense silica particles. We prepared the SiO2 -coated CdTe QDs in aqueous solution with ammonia as a catalyst and without addition of ethanol via a modified Stöber method (Step 1). Since ethanol is not added, TEOS is not mixed homogeneously with the aqueous solution of CdTe QDs, and therefore the hydrolysis of TEOS takes place slowly at the interface of TEOS and water. In addition, it is known that ammonia catalyst accelerates the hydrolysis of TEOS [8]. In the case of the present experiment, the hydrolysis of TEOS was slow because of a relatively low concentration of ammonia (0.04–0.09 M) and without ethanol added. This slow hydrolysis speed resulted in low silica sol concentration. This low concentration prevents silica from their own nucleation. Consequently, silica was deposited onto CdTe QDs to form the SiO2 -coated CdTe QDs. The QDs seem to act as nuclei for deposition of a silica layer. Similar mechanism was reported on the formation of ZnS–SiO2 composite particles (∼50 nm ∅) [6]. The authors reported that at low ammonia concentration (0.1 M), any nucleation of pure SiO2 particles did not occur during silica coating. In addition, Gao’s group reported on SiO2 -coated QD using TGA and thioglycerol as capping agents via a reverse micelle method without any further surface modification on the QDs [16]. Previously, a mechanism with ligands exchange was proposed to explain the formation of a SiO2 layer on the surface of QDs. Nann and co-workers prepared QD–SiO2 beads [8], in which homogeneous silica layer was coated on the QDs by the exchange between TEOS and hydrophobic ligands. However, as describe below, we propose a mechanism without ligands exchange to explain the formation of SiO2 -coated CdTe QDs. For
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
423
Table 1 Effect of injection speed of H2 O phase, W/O and O/S ratios on mean size of glass beads with reaction time for 4 h and 0.06 M TEOS added at pH 9 during Step 2
Scheme 2. Formation mechanism of SiO2 -coated CdTe QDs during Step 1.
TGA-capped CdTe QDs, the sulfur in mercapto group linked to the Cd2+ on the surface of the QDs. The QDs act as nuclei for deposition of a silica layer. TGA was maintained on the surface of the QDs after Step 1. A proposed mechanism of a thin silica layer on the surface of the QDs was shown in Scheme 2. The ≡Si–O–from the hydrolysis of TEOS linked to the surface of the QDs because hydrolysis products are negatively charged [6, 16] and there are Cd2+ ions on the surface of the QDs. A similar mechanism was proposed by Bakalova and co-workers [3] to indicate the formation of single CdSe/ZnS QD with a SiO2 shell. The measurements of S/Cd ratio and zeta-potential at pH 10 supported our arguments. The EDX analysis revealed the molar ratio of S/Cd of the green-emitting QDs and the green-emitting SiO2 -coated CdTe QDs are 0.65± 0.27 and 0.70± 0.36, respectively. These values are very closed to the value (0.63 ± 0.01) derived from bulk powder analysis [20]. This means S/Cd ratio was maintained before and after Step 1. The zeta-potential of the SiO2 -coated CdTe QDs solution (−41 mV) is similar with that of the QDs colloidal solution (−40 mV). Furthermore, the zeta-potential of the SiO2 -coated CdTe QDs solution is changed into −53 mV after the TGA in the solution was removed, while the zeta-potential of a blank sample (SiO2 nanoparticles in an aqueous solution without the QDs prepared by a same preparation procedure) is −55 mV. This small increase of zeta-potential is ascribed to TGA molecules retained on the SiO2 shell. The SiO2 nanoparticles are much more negatively charged compared with TGA in alkaline solution because base-catalyzed hydrolysis and condensation of TEOS monomer provides the resulting particles with negative surface charge in water due to the ionization of the surface hydroxyl groups [21]. In addition, because of the slow hydrolysis and condensation speed of TEOS and the short reaction time during Step 1, there are many –OH and –OC2 H5 groups retained within the thin silica layer. Therefore, the glass network of thin silica layer is not developed very well after Step 1. Subsequently, the SiO2 -coated CdTe QDs were used to grow a SiO2 shell by a reverse micelle route (Step 2). In a typical preparation, cyclohexane served as a continuous oil phase where TEOS and Igepal CO-520 were dissolved. The SiO2 coated CdTe QDs solution and ammonia were then added. The hydrolysis of TEOS takes place at the interface of water and oil
Mean size
W/O (molar ratio)a
O/S (molar ratio)b
Injection speed (mL/min)
500 nm 1 µm 2 µm
0.51 0.51 0.51
37 37 37
0.09 0.30 1.40
30 nm 40 nm 1 µm 2 µm
0.50 0.40 0.56 0.75
30 35 30 30
0.05 0.10 0.40 0.50
a Molar ratio of water to oil in microemulsion. b Molar ratio of oil to surfactant in microemulsion.
phases. Hydrolyzed products are subsequently added to form a dense silica shell around the SiO2 -coated CdTe QDs, providing a shielding. Because a water pool in a reverse micelle leads to the formation of a bead, the number of the QDs in a SiO2 bead depends strongly on the size of a water pool and the concentration of the QDs. It is found that the size is governed by the ratio of oil to surfactant (O/S), the ratio of water to oil (W/O), the injection speed of water phase under alkaline condition, the amount of TEOS, and reaction time. Especially for the large beads, their sizes depend strongly on the injection speed. Our experiment demonstrated the mean sizes of emitting beads are 2 µm, 1 µm, and 500 nm when the injection speeds of a water phase are 1.40, 0.30, and 0.09 mL/min, in which the molar ratios of W/O and O/S were kept at 0.51 and 37 (pH ∼ 9). Table 1 summarized these situations together with other typical results. The size distributions of samples measured by a dynamic light scattering method are shown in Fig. 1. The size of the green-emitting QDs shown in Fig. 1a is 2.6 ± 0.4 nm. This result is almost same as that derived from the first absorption peak of a colloidal solution (2.6 nm). The size of the green-emitting SiO2 -coated CdTe QDs is 3.8 ± 0.4 nm (Fig. 1b) because a thin silica layer (∼0.6 nm) was grown on the surface of the QDs after Step 1. EDX analysis of these specimens showed the ratio of S/Cd remained un-change during Step 1 as explained before. The thickness of the SiO2 layer depended strongly on the amount of TEOS added and reaction time in Step 1. After Step 2, the size of green-emitting CdTe–SiO2 beads in microemulsion is 31 ± 7 nm (Fig. 1c). Similarly, the TEM photograph and size distribution of red-emitting beads are shown in Fig. 2. The results from TEM observation (31 ± 8 nm) and the measurement of size distribution (30 ± 7 nm) are almost same. Furthermore, the TEM pictures of red-emitting SiO2 beads with different size (100 nm–1 µm) were shown in Fig. 3. The QDs were observed in the bead in Fig. 3a. Fusion of sol–gel derived silica at edge (right) of the bead was generated because of the focused electron beam (300 keV, ca. 8 µA) during observation. We can identify at least 40 QDs having the mean size of 4 nm in the bead. The size is almost the same as that derived from their first absorption peak (3.9 nm).
424
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
Fig. 1. Size distributions of green-emitting CdTe QDs (a), SiO2 -coated CdTe QDs (b), and CdTe–SiO2 beads (c) measured by a dynamic light scattering method.
Fig. 2. TEM photograph (a) and size distribution (b) of red-emitting CdTe–SiO2 beads.
3.2. Intense PL of CdTe–SiO2 beads Fig. 4 shows PL and absorption spectra of the red- and greenemitting CdTe–SiO2 beads with about 30 nm in diameter. The first absorption and PL peaks of the red-emitting beads did not change comparing with those of the as-prepared QDs in aqueous solution. However, a little blue shift was observed for the green-emitting beads. This phenomenon is associated with the different surface state of the red- and green-emitting QDs [20,23]. Because the mean size of the emitting beads is about 30 nm, a scattering effect was observed in their absorption spectra. The PL efficiencies of the prepared green- and red-emitting beads in microemulsion were estimated to be as high as 27 and 65%, respectively, when the mean size of these beads is 30– 40 nm. These values are same with those of initial colloidal so-
lutions. To the best of our knowledge, this PL efficiency (65%) is the highest compared with those reported in previous literature. Gao’s group indicated the PL efficiency of SiO2 -coated CdTe QDs redispersed in water is 7%. The PL efficiency of the CdTe QDs as-prepared in aqueous solution is 20% [16]. Bakalova and co-workers reported SiO2 -coated single QD micelles in water reveal 64% PL efficiency, even though this value is 30–50% lower than the initial QDs in chloroform [3]. An ICP analysis demonstrated the molar ratio of Cd/Si of red-emitting beads is 2.4 ± 0.1%. This corresponds to the volume ratio of CdTe QDs in the silica beads of approximately 3.5%. Therefore, a red-emitting bead with 30 nm ∅ encapsulates roughly 14 QDs. The present preparation method provides an effective way to enhance PL intensity from a SiO2 bead by encapsulating multiple QDs within it.
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
425
Fig. 3. TEM photographs of red-emitting SiO2 beads with different size. In case of (a), originally spherical shape of the beads was changed during the observation of QDs inside of the beads due to the irradiation of focused electron beam.
Fig. 5 shows the color images of the green- and red-emitting CdTe–SiO2 beads in micrometer range. Their emission spectra are shown in Fig. 6 together with the excitation spectrum of the red-emitting beads. The broad excitation band and intense PL were observed. 3.3. Reason for retaining high PL efficiency In general, the decreased PL efficiency of QDs is attributed to the surface deterioration that slowly quenches the PL during incorporation. Solutions of hydrophobic or hydrophilic QDs have been reported to be used for the preparation of QD–SiO2 beads [8,16]. In any case, exchange of ligands and surface deterioration occurred during the incorporation. Because the cap-exchange usually leads to diminish PL efficiency, low PL efficiency was observed in the previous works. For example, Bawendi and co-workers indicated that the original surface ligands were changed into 5-amino-1-pentanol and 3-aminopropyltrimethoxysilane. Their demonstrated PL efficiency of QDs within SiO2 beads was 13% whereas initial core-shell CdS/ZnS QDs before the incorporation had a PL efficiency of 38% [13]. In contrast, we observed PL efficiency as high as 65% by retaining the initial PL efficiency even in SiO2 beads. The possible reasons for the efficiency retention are as follows.
One reason is the proper pH control during incorporation of the QDs. To obtain high PL efficiency of QDs in SiO2 beads, it is important to retain the stability of the QDs during incorporation. We have indicated that TGA-capped CdTe QDs in aqueous solution exhibits high PL efficiency and good stability in the pH range of 6–10 because carboxyl group in TGA molecule has a dissociated form in such alkaline to weak acidic conditions reflecting the dissociation constant (pKa = 3.67) of TGA [22]. When the pH is less than 5, the carboxyl groups in TGA are changed from the dissociation form with a negative charge to a non-dissociation form without a charge, which makes the QDs unstable and promotes the agglomeration. In the preparation of CdTe–SiO2 beads, we used alkaline solution with pH 8–10 during the Stöber synthesis (Step 1) and the reverse micelle route (Step 2). The second reason is that a thin silica layer was coated on the surface of the QD in Step 1 without removing ligands (TGA). This thin layer plays an important role to retain high PL efficiency and increase the concentration of the QDs in the SiO2 bead. Surface deterioration, such as exchanging or removing of capping ligands, quenched slowly the PL. Such surface reaction was almost negligible in the subsequent preparation procedure because a thin SiO2 shell was coated on the surface of the QDs during Step 1. In addition, no addition of ethanol during incor-
426
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
Fig. 4. Absorption (a) and PL (b) spectra of CdTe–SiO2 beads in microemulsions and CdTe QDs in aqueous solutions. Black, red, green, and blue curves are from green-emitting beads, green-emitting QDs, red-emitting beads, and red-emitting QDs, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
poration retains high PL efficiency because ethanol results in aggregation of the QDs. Finally, the reduced reaction time during the two-step synthesis plays an important role to retain the initial PL efficiency. For a conventional Stöber synthesis, SiO2 particles were prepared in ethanol solution with a base catalysis and a long reaction period. Because the long reaction period always leads to quenching of the PL of QDs, the PL of inorganic QDs embedded in SiO2 spheres was greatly decreased as revealed in the report from Kotov’s group [12]. In that report, the formation of a silica coating on CdTe QDs results in a drastic reduction of PL efficiency because surface deterioration slowly quench the PL of the QDs during a long reaction time (5 days). However, such surface reactions were not observed in our experiment because of the two-step synthesis with a short preparation period (ca. 7 h). Even though the thickness of the silica shell prepared in the short reaction time in Step 1 is thin, this is enough to prevent the surface from deterioration during Step 2. As we explained in the preparation of glass plates incorporating emitting CdTe QDs [24,25], the movement of molecules and ions in the glass matrix is effectively prohibited by the silica networks. Therefore, the modified Stöber synthesis with a short reaction period is effective to reduce PL degradation of the QDs. A reverse micelle method has been widely used to prepare CdTe–SiO2 beads because it is a convenient and advantageous
Fig. 5. Color images of green- and red-emitting CdTe–SiO2 beads by irradiation of UV light (365 nm). Mean sizes of green- and red-emitting beads are 2 ± 0.8 and 1 ± 0.5 µm, respectively.
Fig. 6. PL spectra of green- and red-emitting CdTe–SiO2 beads with the mean size in micrometer range. Excitation spectrum of the red-emitting beads is also shown (left curve).
technique. However, it is necessary to take a long reaction period because of the very slow hydrolysis of TEOS at water–oil interface. However, the long reaction period results in a relatively low PL efficiency [5,16]. Our current experimental results indicated that the modified Stöber method used for reducing reaction time plays an important role to retain the initial PL efficiency. In Step 1 of our preparation procedure, a thin SiO2 layer was formed on the QDs within a short period via a mod-
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
427
ified Stöber synthesis compared with the reverse micelle route. In this case, the thin SiO2 layer protects the QDs against surface deterioration. Therefore, the CdTe–SiO2 beads retain their high PL efficiency during the subsequent Step 2.
cation of Life Phenomena) of the Japan Science and Technology Agency (JST).
4. Conclusions
[1] W.C.W. Chan, S. Nie, Science 281 (1998) 2016. [2] D.K. Yi, S.T. Selvan, S.S. Lee, G.C. Papaefthymiou, D. Kundaliya, J.Y. Ying, J. Am. Chem. Soc. 127 (2005) 4990. [3] Z. Zhelev, H. Ohba, R. Bakalova, J. Am. Chem. Soc. 128 (2006) 6324. [4] S.T. Selvan, T.T. Tan, J.Y. Ying, Adv. Mater. 17 (2005) 1620. [5] S.T. Selvan, C.L. Li, M. Ando, N. Murase, Chem. Lett. 33 (2004) 434. [6] K.P. Velikov, A. Blaaderen, Langmuir 17 (2001) 4779. [7] T. Nann, P. Mulvaney, Angew. Chem. Int. Ed. 43 (2004) 5393. [8] M. Darbandi, R. Thomann, T. Nann, Chem. Mater. 17 (2005) 5720. [9] V. Salgueiriño-Maceira, M.A. Correa-Duarte, M. Spasova, L.M. LizMarzán, M. Farle, Adv. Funct. Mater. 16 (2006) 509. [10] C.L. Li, N. Murase, Chem. Lett. 34 (2005) 92. [11] W. Stöber, A. Fink, E. Bohn, J. Colloid Interface Sci. 26 (1968) 62. [12] A.L. Rogach, D. Nagesha, J.W. Ostrander, M. Giersig, N.A. Kotov, Chem. Mater. 12 (2000) 2676. [13] Y. Chan, J.P. Zimmer, M. Stroh, J.S. Steckel, R.K. Jain, M.G. Bawendi, Adv. Mater. 16 (2004) 2092. [14] H. Ow, D.R. Larson, M. Srivastava, B.A. Baird, W.W. Webb, U. Wiesner, Nano Lett. 5 (2005) 113. [15] N. Murase, T. Yazawa, Jpn. Patent (Kokai Tokkyo Koho 13-7183), 2001. [16] Y. Yang, M. Gao, Adv. Mater. 17 (2005) 2354. [17] X. Zhao, R.P. Bagwe, W. Tan, Adv. Mater. 16 (2004) 173. [18] N. Murase, C.L. Li, M. Ando, T. Yamashita, K. Takida, A. Tomita, in: Preprints of the Annual Meeting of the Ceramic Society of Japan, Fujisawa, Japan, 2004, p. 296. [19] N.P. Gaponik, D.V. Talapin, A.L. Rogach, K. Hoppe, E.V. Shevchenko, A. Kornowski, A. Eychmüller, H. Weller, J. Phys. Chem. B 106 (2002) 7177. [20] N. Murase, N. Gaponik, H. Weller, Nanoscale Res. Lett. 2 (2007) 230. [21] W. Wang, B. Gu, L. Liang, W. Hamilton, J. Phys. Chem. B 107 (2003) 3400. [22] M. Ando, C.L. Li, N. Murase, Mat. Res. Soc. Symp. Proc. 789 (2004) 123. [23] N. Murase, P. Yang, C.L. Li, J. Phys. Chem. B 109 (2005) 17855. [24] C.L. Li, N. Murase, Langmuir 20 (2004) 1. [25] C.L. Li, M. Ando, N. Murase, J. Non-Cryst. Solids 342 (2004) 32.
To prepare highly luminescent CdTe–SiO2 beads, we have developed a facile two-step synthesis, which involved coating a thin SiO2 layer on the surface of the QDs to get SiO2 -coated CdTe QDs via a modified Stöber method and subsequent encapsulation into SiO2 beads by a reverse micelle route. The reverse micelle route results in the formation of SiO2 beads with controlled size (from 30 nm to 2 µm) by changing W/O ratio, reaction time, the injection speed of the SiO2 -coated CdTe QDs solution, and the amount of TEOS. A red-emitting bead with 30 nm in diameter encapsulated roughly 14 QDs with retaining their initial PL efficiency (65%) in colloidal solution. This PL efficiency is the highest among all other reported values. The modified Stöber method is effective to retain high PL efficiency of the QDs because the thin SiO2 layer was coated and TGA molecules were retained on the surface of the QDs. The high PL efficiency of the CdTe–SiO2 beads opens up new possibilities for various applications including biological labeling and light-emitting devices. Acknowledgments We would like to thank Dr. Chunliang Li for preparing the CdTe nanocrystals. This study was supported in part by the CREST program (Research Area: Novel Measuring and Analytical Technology Contributions to the Elucidation and Appli-
References
Encapsulation of emitting CdTe QDs within silica beads to retain initial photoluminescence efficiency Ping Yang, Masanori Ando, Norio Murase ∗ Photonics Research Institute, National Institute of Advanced Industrial Science and Technology, Midorigaoka, Ikeda-city, Osaka 563-8577, Japan Received 23 March 2007; accepted 23 August 2007 Available online 31 August 2007
Abstract Highly luminescent silica beads (30 nm–2 µm ∅) incorporating CdTe quantum dots (QDs) were prepared via a two-step preparation procedure, namely a modified Stöber synthesis and a subsequent reverse micelle route. In the modified Stöber synthesis, the silica molecules are deposited on the surface of the QDs. After this first step, these coated QDs were incorporated into silica beads via a reverse micelle route. Inductively coupled plasma analysis revealed a red-emitting silica bead of 30 nm in diameter thus prepared encapsulated roughly 14 CdTe QDs. These glass beads (30–40 nm ∅) retained the initial photoluminescence (PL) efficiencies of the colloidal QDs (27 and 65% for the green- and red-emitting beads, respectively). The protection of QDs by a silica layer at the first step, together with the short total reaction time, is the main reason for the retention of the PL efficiency. The size of the glass beads can be easily controlled over the wide range by adjusting the injection speed and the ratio of chemicals used for the reverse micelle preparation. Since the original efficiency was maintained in the beads and is the highest ever reported for QD-containing silica beads, the method presented here is of significant importance for applications of silica beads to biological probes. © 2007 Elsevier Inc. All rights reserved. Keywords: CdTe; Nanocrystals; Silica bead; Luminescence
1. Introduction Highly luminescent semiconductor quantum dots (QDs) have been attracting much interest because they have sizedependent emission wavelength reflecting three-dimensional quantum confinement effect, narrow width of emission spectra, wide excitation wavelength range and fast emission decay that prevents the emission intensity from saturation. These features of the luminescent QDs would lead to various applications including fluorescent biomarkers and informational electronics such as displays and lighting. The development of biocompatible highly luminescent materials for cell staining and visualization has been an interesting objective in cellular biology and ultra-sensitive immunoassay, because luminescent probes for biomolecular labeling and recognition are of great importance in the fields of chemistry, biology, and medical sciences as well as in biotechnology. Semiconductor QDs covalently coupled to * Corresponding author. Fax: +81 72 751 9637.
E-mail address: [email protected] (N. Murase). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.08.058
biomolecules are expected to be used in ultra-sensitive biological detection because of their intense luminescence, narrow emission spectral width, and higher stability against photobleaching compared with organic dyes [1]. Recently, increasing attention has been directed to the coating of emitting QDs with a layer of transparent silica. Such coating is expected to bring many advantages because the thin silica layer as a protective capping material on the QDs increases the mechanical stability, enables a transfer into various organic and aqueous solvents, and protects QDs against oxidation and agglomeration. Furthermore, the QDs would benefit from having a silica shell to impart biocompatibility, because surface of silica can be easily modified to link bioconjugators [2,3]. There have been several reports to deposit a silica layer on semiconductor QDs [4–9]. To achieve noticeable signal increment in ultra-sensitive bioanalysis, silica spheres encapsulating highly luminescent CdTe QDs at high concentrations are a good candidate. It has recently been reported that thioglycolic acid (TGA)-capped CdTe QDs prepared by an aqueous solution method show very high photoluminescence (PL) efficiency
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
of 65% without post-preparative treatment when reaction conditions are optimized [10]. The Stöber synthesis of colloidal silica, by which silica particles with a diameter less than µm range are obtained, was first described in 1968 [11]. QD–SiO2 nanoparticles having so called raisin-bun-type composite structure were formed by addition of sodium silicate, with either CdTe, CdSe or CdSe/CdS QDs [12]. Bawendi’s group reported on the incorporation of luminescent CdSe/ZnS QDs into a SiO2 shell around SiO2 spheres of several µm in diameter, with PL efficiency of up to 13% [13]. Later, Nann and co-workers indicated that the silicacoating on organic solvent-dispersible QDs results in ordered structures under optimal conditions. They successfully obtained 30–120 nm ∅ silica spheres incorporating single QD in each sphere [7]. However, the formation of a silica shell on the QDs leads to a drastic reduction of PL efficiency compared with their initial value [13]. In contrast, Ow et al. prepared silica– organic dye composite spheres with 20–30 nm in diameter by a Stöber synthesis, and reported that the composite spheres exhibited 20 times brighter PL than that of the organic dye before coating silica layer [14]. These observations indicate that the PL efficiency of such composite spheres is extremely sensitive to incorporation conditions of fluorescent materials in silica spheres. Therefore, it is necessary to develop a novel preparation procedure for retaining the high PL efficiency of QDs in SiO2 beads. A reverse micelle approach, which was proposed 6 years ago by us [15], is a convenient and advantageous technique to coat a silica layer on QDs. Compared with the Stöber method, a reverse micelle route can easily yield more uniform sized silica spheres in the range of 30–150 nm in diameter [3,16] because the size of a water pool, which is a reaction field for the formation of silica spheres, is rather uniform. Furthermore, it is reported that the reverse micelle route normally does not depend very much on reaction conditions, leading to the formation of small silica particles with relatively low density of defects, and tunable size [8]. In this regard, several groups have developed different reverse micelle approaches to form QD–SiO2 composite beads. Recently, single QD encapsulated within a silica bead was prepared by Nann and co-workers [8]. Gao’s group prepared emitting silica beads incorporating CdTe QDs by a reverse micelle route with a PL efficiency of 7% [16]. Ying’s group reported on robust, non-cytotoxic, silica-coated CdSe QDs with 20% PL efficiency [4]. Bakalova’s group prepared single CdSe/ZnS QD with a SiO2 shell by dot-micelles, in which hydrophobic QDs were dispersed in detergent aqueous solution to form micelles. Their silica-coated QDs revealed about 64% PL efficiency [3]. To the best our knowledge, this value (64%) is the highest reported in previous literature. However, their PL efficiency of the silica-coated single QD micelles in aqueous solution is 30–50% lower than that of the initial QDs in chloroform. Therefore, when using a conventional Stöber synthesis or a conventional reverse micelle route, it would be rather difficult to obtain QD-incorporated silica spheres with retention of the initial PL efficiency, because a long reaction period would possibly cause surface deterioration, such as increase of surface defect density that slowly quenches the PL of
421
the QDs. In literature, strong PL was only observed in luminescent SiO2 spheres of several tens of nm in diameter doped with organic dye molecules [17]. They indicated dyes encapsulated within SiO2 spheres exhibited excellent photostability, facile surface modification for bioconjugation, and size uniformity and tunability. Since semiconductor QDs have advantages over organic dyes due to better long-term stability and wavelength tunability, development of highly luminescent QDs–SiO2 beads would provide wide applicability. According to our previous work [5], CdTe–SiO2 composite particles of a few tens of nm in diameter with 5–10% PL efficiency were obtained by a conventional reverse micelle route after reaction for 48 h. Therefore, it is a challenge to develop a novel preparation method with shorter reaction period to retain the initial PL efficiency of the QDs in SiO2 beads. In this paper, we report on successful preparation of highly luminescent, size-controlled (30 nm–2 µm ∅) SiO2 beads incorporating multiple CdTe QDs via a two-step preparation procedure. At the first step, SiO2 -coated CdTe QDs were prepared by a modified Stöber method. Subsequently, the SiO2 -coated CdTe QDs solution was injected into an organic solution to get sizecontrolled CdTe–SiO2 beads by a reverse micelle route. The emitting beads retained the initial PL efficiency of the QDs in colloidal solution. 2. Experimental 2.1. Starting materials All chemicals used were of analytical grade or of the highest purity available. Pure water was of a Milli-Q synthesis grade. TGA-capped CdTe (red- and green-emitting) QDs in aqueous solutions were prepared by a procedure using cadmium perchlorate and hydrogen telluride as described in a previous paper [10]. Green- and red-emitting TGA-capped CdTe QDs with the mean size 2.6 and 3.9 nm in diameter, respectively, were used for further preparation. The PL efficiencies of the green- and red-emitting QDs in aqueous solutions were 27 and 65%, respectively. 2.2. Step 1: Synthesis of SiO2 -coated CdTe QDs A two-step preparation procedure was used to derive the CdTe–SiO2 beads (Scheme 1). In the first step, the SiO2 -coated CdTe QDs were prepared by a modified Stöber technique. Typically, CdTe colloidal solution (1 mL), water (1 mL), diluted ammonia (50 µL, the volume ratio of water to ammonia (25%) is 3:1), and tetraethyl orthosilicate (TEOS, 0.2 mL, [H2 O]/[TEOS] ∼ 120) were mixed in a beaker covered with aluminum foil to reduce the evaporation of ammonia during coating. In this stage, the pH of the precursor is about 10. After continuous stirring for 3 h, the solution of the SiO2 -coated CdTe QDs was obtained. 2.3. Step 2: Preparation of CdTe–SiO2 beads CdTe–SiO2 beads were prepared by a reverse micelle route using the above mentioned SiO2 -coated CdTe QDs as a starting
422
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
Scheme 1. Encapsulation of emitting CdTe QDs within a SiO2 bead via a modified Stöber synthesis (Step 1) and a reverse micelle route (Step 2).
material. A reverse micelle employing Igepal CO-520 as a surfactant, and cyclohexane as a hydrophobic organic solvent was prepared with different water/oil ratios (W/O). Typically, Igepal CO 520 (2.25 g) added in cyclohexane (12.5 g) was stirred until the solution remained clear. Then, the SiO2 -coated QDs solution (pH ∼ 9, 2 mL) obtained in Step 1 was injected into the stock solution dropwise to form the reverse micelles under vigorous stirring. Except for W/O, the size of a water pool in the reverse micelle depended strongly on the injection speed of water under an alkaline condition. To encapsulate the multiple SiO2 -coated CdTe QDs into a SiO2 bead, TEOS (0.3 mL) was injected into the microemulsion. After that, diluted ammonia was injected to adjust pH to ∼10. The microemulsion was then stirred for 4 h. 2.4. Characterization Absorption and PL spectra were taken using conventional spectrometers (Hitachi U-4000 and F-4500). The PL efficiencies of the emitting of CdTe–SiO2 beads and CdTe QDs in solutions were estimated by the method previously reported [18]. Briefly, the absorbance dependence of PL intensity was measured by using standard solution (quinine sulfate in 0.05 M H2 SO4 ). After that, the absorbance and PL of the sample solution in interest were measured. The efficiency was derived by comparing those with the previously measured ones for the standard solution. To measure precisely the absorption and PL spectra of the emitting beads of several tens nm in size, a blank microemulsion without emitting beads was prepared as a reference, and the absorption and PL spectra of the reference were subtracted from those of the sample microemulsion. Transmission electron microscopy (TEM) observation and energy dispersive X-ray (EDX) analysis were performed on a Hitachi H-9000NA (300 kV) microscope and a FEI Technai G2 F20 (200 kV) microscope, respectively. For preparing a TEM specimen, the beads were first precipitated by ace-
tonitrile (CH3 CN) from the microemulsion, and centrifuged at 4000 rpm for 15 min. After that, the powder sample was dried at 40 ◦ C for 1 h. The green-emitting CdTe–QDs contain sulfur ions solely on the surface, whereas the red-emitting ones impregnate them due to hydrolysis of surfactant (TGA) during lengthy reflux upon synthesis [19,20]. Therefore, we used green-emitting QDs to prepare samples for measuring the S/Cd ratio of the samples before and after Step 1. The molar ratio of Cd/Si of red-emitting CdTe–SiO2 beads was measured with a Horiba Jobin Yvon JY-238U Ultima inductively coupled plasma (ICP) spectrometer. The optical images and mean size of the CdTe–SiO2 beads of few micrometers in diameter were obtained with a Nikon Eclipse 80i fluorescent microscope. For the beads with sub-micrometer in diameter, size and its distribution in solution were obtained with a dynamic light scattering particles analyzer (Nanotrac, Nikkiso). Zeta-potential was measured by an electrophoretic light scattering spectrophotometer (ELS-8000, Otsuka Electronic) at pH 10. 3. Results and discussion 3.1. Formation of SiO2 -coated QDs and CdTe–SiO2 beads The formation of sol–gel silica is based on complicated hydrolysis–condensation reactions of silicon alkoxide (Si(OR)4 ) in alcoholic solutions. The polymerization of silica sol in an alkaline solution occurs by internal condensation and crosslinking to give dense silica particles. We prepared the SiO2 -coated CdTe QDs in aqueous solution with ammonia as a catalyst and without addition of ethanol via a modified Stöber method (Step 1). Since ethanol is not added, TEOS is not mixed homogeneously with the aqueous solution of CdTe QDs, and therefore the hydrolysis of TEOS takes place slowly at the interface of TEOS and water. In addition, it is known that ammonia catalyst accelerates the hydrolysis of TEOS [8]. In the case of the present experiment, the hydrolysis of TEOS was slow because of a relatively low concentration of ammonia (0.04–0.09 M) and without ethanol added. This slow hydrolysis speed resulted in low silica sol concentration. This low concentration prevents silica from their own nucleation. Consequently, silica was deposited onto CdTe QDs to form the SiO2 -coated CdTe QDs. The QDs seem to act as nuclei for deposition of a silica layer. Similar mechanism was reported on the formation of ZnS–SiO2 composite particles (∼50 nm ∅) [6]. The authors reported that at low ammonia concentration (0.1 M), any nucleation of pure SiO2 particles did not occur during silica coating. In addition, Gao’s group reported on SiO2 -coated QD using TGA and thioglycerol as capping agents via a reverse micelle method without any further surface modification on the QDs [16]. Previously, a mechanism with ligands exchange was proposed to explain the formation of a SiO2 layer on the surface of QDs. Nann and co-workers prepared QD–SiO2 beads [8], in which homogeneous silica layer was coated on the QDs by the exchange between TEOS and hydrophobic ligands. However, as describe below, we propose a mechanism without ligands exchange to explain the formation of SiO2 -coated CdTe QDs. For
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
423
Table 1 Effect of injection speed of H2 O phase, W/O and O/S ratios on mean size of glass beads with reaction time for 4 h and 0.06 M TEOS added at pH 9 during Step 2
Scheme 2. Formation mechanism of SiO2 -coated CdTe QDs during Step 1.
TGA-capped CdTe QDs, the sulfur in mercapto group linked to the Cd2+ on the surface of the QDs. The QDs act as nuclei for deposition of a silica layer. TGA was maintained on the surface of the QDs after Step 1. A proposed mechanism of a thin silica layer on the surface of the QDs was shown in Scheme 2. The ≡Si–O–from the hydrolysis of TEOS linked to the surface of the QDs because hydrolysis products are negatively charged [6, 16] and there are Cd2+ ions on the surface of the QDs. A similar mechanism was proposed by Bakalova and co-workers [3] to indicate the formation of single CdSe/ZnS QD with a SiO2 shell. The measurements of S/Cd ratio and zeta-potential at pH 10 supported our arguments. The EDX analysis revealed the molar ratio of S/Cd of the green-emitting QDs and the green-emitting SiO2 -coated CdTe QDs are 0.65± 0.27 and 0.70± 0.36, respectively. These values are very closed to the value (0.63 ± 0.01) derived from bulk powder analysis [20]. This means S/Cd ratio was maintained before and after Step 1. The zeta-potential of the SiO2 -coated CdTe QDs solution (−41 mV) is similar with that of the QDs colloidal solution (−40 mV). Furthermore, the zeta-potential of the SiO2 -coated CdTe QDs solution is changed into −53 mV after the TGA in the solution was removed, while the zeta-potential of a blank sample (SiO2 nanoparticles in an aqueous solution without the QDs prepared by a same preparation procedure) is −55 mV. This small increase of zeta-potential is ascribed to TGA molecules retained on the SiO2 shell. The SiO2 nanoparticles are much more negatively charged compared with TGA in alkaline solution because base-catalyzed hydrolysis and condensation of TEOS monomer provides the resulting particles with negative surface charge in water due to the ionization of the surface hydroxyl groups [21]. In addition, because of the slow hydrolysis and condensation speed of TEOS and the short reaction time during Step 1, there are many –OH and –OC2 H5 groups retained within the thin silica layer. Therefore, the glass network of thin silica layer is not developed very well after Step 1. Subsequently, the SiO2 -coated CdTe QDs were used to grow a SiO2 shell by a reverse micelle route (Step 2). In a typical preparation, cyclohexane served as a continuous oil phase where TEOS and Igepal CO-520 were dissolved. The SiO2 coated CdTe QDs solution and ammonia were then added. The hydrolysis of TEOS takes place at the interface of water and oil
Mean size
W/O (molar ratio)a
O/S (molar ratio)b
Injection speed (mL/min)
500 nm 1 µm 2 µm
0.51 0.51 0.51
37 37 37
0.09 0.30 1.40
30 nm 40 nm 1 µm 2 µm
0.50 0.40 0.56 0.75
30 35 30 30
0.05 0.10 0.40 0.50
a Molar ratio of water to oil in microemulsion. b Molar ratio of oil to surfactant in microemulsion.
phases. Hydrolyzed products are subsequently added to form a dense silica shell around the SiO2 -coated CdTe QDs, providing a shielding. Because a water pool in a reverse micelle leads to the formation of a bead, the number of the QDs in a SiO2 bead depends strongly on the size of a water pool and the concentration of the QDs. It is found that the size is governed by the ratio of oil to surfactant (O/S), the ratio of water to oil (W/O), the injection speed of water phase under alkaline condition, the amount of TEOS, and reaction time. Especially for the large beads, their sizes depend strongly on the injection speed. Our experiment demonstrated the mean sizes of emitting beads are 2 µm, 1 µm, and 500 nm when the injection speeds of a water phase are 1.40, 0.30, and 0.09 mL/min, in which the molar ratios of W/O and O/S were kept at 0.51 and 37 (pH ∼ 9). Table 1 summarized these situations together with other typical results. The size distributions of samples measured by a dynamic light scattering method are shown in Fig. 1. The size of the green-emitting QDs shown in Fig. 1a is 2.6 ± 0.4 nm. This result is almost same as that derived from the first absorption peak of a colloidal solution (2.6 nm). The size of the green-emitting SiO2 -coated CdTe QDs is 3.8 ± 0.4 nm (Fig. 1b) because a thin silica layer (∼0.6 nm) was grown on the surface of the QDs after Step 1. EDX analysis of these specimens showed the ratio of S/Cd remained un-change during Step 1 as explained before. The thickness of the SiO2 layer depended strongly on the amount of TEOS added and reaction time in Step 1. After Step 2, the size of green-emitting CdTe–SiO2 beads in microemulsion is 31 ± 7 nm (Fig. 1c). Similarly, the TEM photograph and size distribution of red-emitting beads are shown in Fig. 2. The results from TEM observation (31 ± 8 nm) and the measurement of size distribution (30 ± 7 nm) are almost same. Furthermore, the TEM pictures of red-emitting SiO2 beads with different size (100 nm–1 µm) were shown in Fig. 3. The QDs were observed in the bead in Fig. 3a. Fusion of sol–gel derived silica at edge (right) of the bead was generated because of the focused electron beam (300 keV, ca. 8 µA) during observation. We can identify at least 40 QDs having the mean size of 4 nm in the bead. The size is almost the same as that derived from their first absorption peak (3.9 nm).
424
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
Fig. 1. Size distributions of green-emitting CdTe QDs (a), SiO2 -coated CdTe QDs (b), and CdTe–SiO2 beads (c) measured by a dynamic light scattering method.
Fig. 2. TEM photograph (a) and size distribution (b) of red-emitting CdTe–SiO2 beads.
3.2. Intense PL of CdTe–SiO2 beads Fig. 4 shows PL and absorption spectra of the red- and greenemitting CdTe–SiO2 beads with about 30 nm in diameter. The first absorption and PL peaks of the red-emitting beads did not change comparing with those of the as-prepared QDs in aqueous solution. However, a little blue shift was observed for the green-emitting beads. This phenomenon is associated with the different surface state of the red- and green-emitting QDs [20,23]. Because the mean size of the emitting beads is about 30 nm, a scattering effect was observed in their absorption spectra. The PL efficiencies of the prepared green- and red-emitting beads in microemulsion were estimated to be as high as 27 and 65%, respectively, when the mean size of these beads is 30– 40 nm. These values are same with those of initial colloidal so-
lutions. To the best of our knowledge, this PL efficiency (65%) is the highest compared with those reported in previous literature. Gao’s group indicated the PL efficiency of SiO2 -coated CdTe QDs redispersed in water is 7%. The PL efficiency of the CdTe QDs as-prepared in aqueous solution is 20% [16]. Bakalova and co-workers reported SiO2 -coated single QD micelles in water reveal 64% PL efficiency, even though this value is 30–50% lower than the initial QDs in chloroform [3]. An ICP analysis demonstrated the molar ratio of Cd/Si of red-emitting beads is 2.4 ± 0.1%. This corresponds to the volume ratio of CdTe QDs in the silica beads of approximately 3.5%. Therefore, a red-emitting bead with 30 nm ∅ encapsulates roughly 14 QDs. The present preparation method provides an effective way to enhance PL intensity from a SiO2 bead by encapsulating multiple QDs within it.
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
425
Fig. 3. TEM photographs of red-emitting SiO2 beads with different size. In case of (a), originally spherical shape of the beads was changed during the observation of QDs inside of the beads due to the irradiation of focused electron beam.
Fig. 5 shows the color images of the green- and red-emitting CdTe–SiO2 beads in micrometer range. Their emission spectra are shown in Fig. 6 together with the excitation spectrum of the red-emitting beads. The broad excitation band and intense PL were observed. 3.3. Reason for retaining high PL efficiency In general, the decreased PL efficiency of QDs is attributed to the surface deterioration that slowly quenches the PL during incorporation. Solutions of hydrophobic or hydrophilic QDs have been reported to be used for the preparation of QD–SiO2 beads [8,16]. In any case, exchange of ligands and surface deterioration occurred during the incorporation. Because the cap-exchange usually leads to diminish PL efficiency, low PL efficiency was observed in the previous works. For example, Bawendi and co-workers indicated that the original surface ligands were changed into 5-amino-1-pentanol and 3-aminopropyltrimethoxysilane. Their demonstrated PL efficiency of QDs within SiO2 beads was 13% whereas initial core-shell CdS/ZnS QDs before the incorporation had a PL efficiency of 38% [13]. In contrast, we observed PL efficiency as high as 65% by retaining the initial PL efficiency even in SiO2 beads. The possible reasons for the efficiency retention are as follows.
One reason is the proper pH control during incorporation of the QDs. To obtain high PL efficiency of QDs in SiO2 beads, it is important to retain the stability of the QDs during incorporation. We have indicated that TGA-capped CdTe QDs in aqueous solution exhibits high PL efficiency and good stability in the pH range of 6–10 because carboxyl group in TGA molecule has a dissociated form in such alkaline to weak acidic conditions reflecting the dissociation constant (pKa = 3.67) of TGA [22]. When the pH is less than 5, the carboxyl groups in TGA are changed from the dissociation form with a negative charge to a non-dissociation form without a charge, which makes the QDs unstable and promotes the agglomeration. In the preparation of CdTe–SiO2 beads, we used alkaline solution with pH 8–10 during the Stöber synthesis (Step 1) and the reverse micelle route (Step 2). The second reason is that a thin silica layer was coated on the surface of the QD in Step 1 without removing ligands (TGA). This thin layer plays an important role to retain high PL efficiency and increase the concentration of the QDs in the SiO2 bead. Surface deterioration, such as exchanging or removing of capping ligands, quenched slowly the PL. Such surface reaction was almost negligible in the subsequent preparation procedure because a thin SiO2 shell was coated on the surface of the QDs during Step 1. In addition, no addition of ethanol during incor-
426
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
Fig. 4. Absorption (a) and PL (b) spectra of CdTe–SiO2 beads in microemulsions and CdTe QDs in aqueous solutions. Black, red, green, and blue curves are from green-emitting beads, green-emitting QDs, red-emitting beads, and red-emitting QDs, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
poration retains high PL efficiency because ethanol results in aggregation of the QDs. Finally, the reduced reaction time during the two-step synthesis plays an important role to retain the initial PL efficiency. For a conventional Stöber synthesis, SiO2 particles were prepared in ethanol solution with a base catalysis and a long reaction period. Because the long reaction period always leads to quenching of the PL of QDs, the PL of inorganic QDs embedded in SiO2 spheres was greatly decreased as revealed in the report from Kotov’s group [12]. In that report, the formation of a silica coating on CdTe QDs results in a drastic reduction of PL efficiency because surface deterioration slowly quench the PL of the QDs during a long reaction time (5 days). However, such surface reactions were not observed in our experiment because of the two-step synthesis with a short preparation period (ca. 7 h). Even though the thickness of the silica shell prepared in the short reaction time in Step 1 is thin, this is enough to prevent the surface from deterioration during Step 2. As we explained in the preparation of glass plates incorporating emitting CdTe QDs [24,25], the movement of molecules and ions in the glass matrix is effectively prohibited by the silica networks. Therefore, the modified Stöber synthesis with a short reaction period is effective to reduce PL degradation of the QDs. A reverse micelle method has been widely used to prepare CdTe–SiO2 beads because it is a convenient and advantageous
Fig. 5. Color images of green- and red-emitting CdTe–SiO2 beads by irradiation of UV light (365 nm). Mean sizes of green- and red-emitting beads are 2 ± 0.8 and 1 ± 0.5 µm, respectively.
Fig. 6. PL spectra of green- and red-emitting CdTe–SiO2 beads with the mean size in micrometer range. Excitation spectrum of the red-emitting beads is also shown (left curve).
technique. However, it is necessary to take a long reaction period because of the very slow hydrolysis of TEOS at water–oil interface. However, the long reaction period results in a relatively low PL efficiency [5,16]. Our current experimental results indicated that the modified Stöber method used for reducing reaction time plays an important role to retain the initial PL efficiency. In Step 1 of our preparation procedure, a thin SiO2 layer was formed on the QDs within a short period via a mod-
P. Yang et al. / Journal of Colloid and Interface Science 316 (2007) 420–427
427
ified Stöber synthesis compared with the reverse micelle route. In this case, the thin SiO2 layer protects the QDs against surface deterioration. Therefore, the CdTe–SiO2 beads retain their high PL efficiency during the subsequent Step 2.
cation of Life Phenomena) of the Japan Science and Technology Agency (JST).
4. Conclusions
[1] W.C.W. Chan, S. Nie, Science 281 (1998) 2016. [2] D.K. Yi, S.T. Selvan, S.S. Lee, G.C. Papaefthymiou, D. Kundaliya, J.Y. Ying, J. Am. Chem. Soc. 127 (2005) 4990. [3] Z. Zhelev, H. Ohba, R. Bakalova, J. Am. Chem. Soc. 128 (2006) 6324. [4] S.T. Selvan, T.T. Tan, J.Y. Ying, Adv. Mater. 17 (2005) 1620. [5] S.T. Selvan, C.L. Li, M. Ando, N. Murase, Chem. Lett. 33 (2004) 434. [6] K.P. Velikov, A. Blaaderen, Langmuir 17 (2001) 4779. [7] T. Nann, P. Mulvaney, Angew. Chem. Int. Ed. 43 (2004) 5393. [8] M. Darbandi, R. Thomann, T. Nann, Chem. Mater. 17 (2005) 5720. [9] V. Salgueiriño-Maceira, M.A. Correa-Duarte, M. Spasova, L.M. LizMarzán, M. Farle, Adv. Funct. Mater. 16 (2006) 509. [10] C.L. Li, N. Murase, Chem. Lett. 34 (2005) 92. [11] W. Stöber, A. Fink, E. Bohn, J. Colloid Interface Sci. 26 (1968) 62. [12] A.L. Rogach, D. Nagesha, J.W. Ostrander, M. Giersig, N.A. Kotov, Chem. Mater. 12 (2000) 2676. [13] Y. Chan, J.P. Zimmer, M. Stroh, J.S. Steckel, R.K. Jain, M.G. Bawendi, Adv. Mater. 16 (2004) 2092. [14] H. Ow, D.R. Larson, M. Srivastava, B.A. Baird, W.W. Webb, U. Wiesner, Nano Lett. 5 (2005) 113. [15] N. Murase, T. Yazawa, Jpn. Patent (Kokai Tokkyo Koho 13-7183), 2001. [16] Y. Yang, M. Gao, Adv. Mater. 17 (2005) 2354. [17] X. Zhao, R.P. Bagwe, W. Tan, Adv. Mater. 16 (2004) 173. [18] N. Murase, C.L. Li, M. Ando, T. Yamashita, K. Takida, A. Tomita, in: Preprints of the Annual Meeting of the Ceramic Society of Japan, Fujisawa, Japan, 2004, p. 296. [19] N.P. Gaponik, D.V. Talapin, A.L. Rogach, K. Hoppe, E.V. Shevchenko, A. Kornowski, A. Eychmüller, H. Weller, J. Phys. Chem. B 106 (2002) 7177. [20] N. Murase, N. Gaponik, H. Weller, Nanoscale Res. Lett. 2 (2007) 230. [21] W. Wang, B. Gu, L. Liang, W. Hamilton, J. Phys. Chem. B 107 (2003) 3400. [22] M. Ando, C.L. Li, N. Murase, Mat. Res. Soc. Symp. Proc. 789 (2004) 123. [23] N. Murase, P. Yang, C.L. Li, J. Phys. Chem. B 109 (2005) 17855. [24] C.L. Li, N. Murase, Langmuir 20 (2004) 1. [25] C.L. Li, M. Ando, N. Murase, J. Non-Cryst. Solids 342 (2004) 32.
To prepare highly luminescent CdTe–SiO2 beads, we have developed a facile two-step synthesis, which involved coating a thin SiO2 layer on the surface of the QDs to get SiO2 -coated CdTe QDs via a modified Stöber method and subsequent encapsulation into SiO2 beads by a reverse micelle route. The reverse micelle route results in the formation of SiO2 beads with controlled size (from 30 nm to 2 µm) by changing W/O ratio, reaction time, the injection speed of the SiO2 -coated CdTe QDs solution, and the amount of TEOS. A red-emitting bead with 30 nm in diameter encapsulated roughly 14 QDs with retaining their initial PL efficiency (65%) in colloidal solution. This PL efficiency is the highest among all other reported values. The modified Stöber method is effective to retain high PL efficiency of the QDs because the thin SiO2 layer was coated and TGA molecules were retained on the surface of the QDs. The high PL efficiency of the CdTe–SiO2 beads opens up new possibilities for various applications including biological labeling and light-emitting devices. Acknowledgments We would like to thank Dr. Chunliang Li for preparing the CdTe nanocrystals. This study was supported in part by the CREST program (Research Area: Novel Measuring and Analytical Technology Contributions to the Elucidation and Appli-
References