SERS Enhancement Factors Studies of Silver Nanoprism and

May 7, 2009 - Adsorption and Detection of Sport Doping Drugs on Metallic Plasmonic ... Size Selection and Concentration of Silver Nanoparticles by ...
6 downloads 0 Views 3MB Size
9520

J. Phys. Chem. C 2009, 113, 9520–9525

SERS Enhancement Factors Studies of Silver Nanoprism and Spherical Nanoparticle Colloids in The Presence of Bromide Ions Shih-Hong Ciou,† Yi-Wei Cao,† Huai-Cing Huang,† De-Yan Su,† and Cheng-Liang Huang*,† Department of Applied Chemistry, National Chiayi UniVersity, Chiayi City, Taiwan ReceiVed: October 19, 2008; ReVised Manuscript ReceiVed: March 23, 2009

Surface-enhanced Raman Scattering (SERS) of R6G has been observed in two different silver colloids in the presence of KBr. It was found in spherical nanoparticles with an average diameter of 10 ( 2 nm and in their photoinduced conversion products with sodium light irradiation, silver nanoprisms, with an average edge length of 70 ( 30 nm and thickness of about 10 ( 2 nm. The SERS intensity of R6G (4.5 × 10-9 M) in the silver nanoprism colloid was 20× stronger than in the spherical nanoparticle colloid with the addition of KBr (0.05 M). Comparing the Raman signal intensity of water (55.5 M) and the SERS intensity of R6G (4.5 × 10-9 M) in the same spectrum showed that the analytical enhancement factors (AEF) in the truncated silver nanoprism colloid and spherical nanoparticle colloid are about 1.6 × 105 and 8 × 103, respectively. Monitoring the fluorescent intensity of R6G versus the concentration of R6G in two colloids in the absence of KBr displayed that the nanoprisms had better fluorescence quenching ability than the spherical nanoparticles by a factor of around 3. It implied that light irradiation not only converted the morphology of the spherical nanoparticles to nanoprisms, but also changed the surface character to have a higher affinity to the dye molecule. The surface enhancement factors (SEF) can also be calculated from AEF and the percentage of dye molecules adsorbed on the surface in the presence of KBr. The SEF of the nanoprism and spherical nanoparticle colloids are about 3.2 × 105 and 9 × 103, respectively. The UV-vis-NIR spectra and TEM images show that the spherical nanoparticles remained dispersed, while the nanoprisms were sculptured by the bromide ions after the addition of KBr. 1. Introduction Surface enhanced Raman Scattering (SERS) is one of the major topics of the surface enhanced phenomena, which include surface enhanced absorption, fluorescence, and photochemistry in molecules near a rough surface. They also include enhanced second-harmonic generation, hyper-Raman scattering, and other nonlinear processes.1 SERS has been studied for more than three decades and has become a very useful tool in many fields.2-4 Compared to other kinds of surface enhanced techniques, one of the most important advantages of SERS is to provide a highly discriminated spectrum with strong intensity of the interesting analyte. Similar molecules and even similar microorganisms can be distinguished due to their highly discriminated SERS spectra.5-7 SERS is usually performed on noble metal surfaces and can be recorded in a variety of environments. There are many kinds of active substrates with a surface of fair roughness that were developed for SERS study,8 e.g., colloids,9 colloid in solgel,10 electrochemically roughened electrodes,2,3 vapor-deposited metal island films,11 and lithography-produced nanostructures.12 Among those kinds of substrates, the colloid system is the most easily prepared and has fairly high enhancement factors.9 The silver colloids were the substrate believed to have the largest enhancement factors among noble metal colloids and have been used to perform single molecule SERS experiments.13,14 Many silver colloidal substrates in different shapes, which include spheres, nanorods, triangular plates, hexagonal plates, and nanocubes to polyhedrons have been synthesized.15 In the colloid * To whom all correspondence should be addressed: Fax: +886-52717963. E-mail: [email protected]. † Department of Applied Chemistry, National Chiayi Unversity.

system, there are several kinds of experimental parameters that can be optimized to obtain an SERS spectrum with an acceptable S/N value, such as the shapes and sizes of the nanostructures, pH values, pumping laser wavelength, and the addition of different kinds of salts. Many studies have reported that added salts have two major influential effects on the SERS’s spectra quality. The first is that the added salts can induce the aggregation of the metal nanoparticles, and therefore generate a larger local electrical field in the junction between the nanoparticles, resulting in larger SERS enhancement factors.16 The second is that the added salts can help to form an insoluble complex to stabilize the SERS active sites, hence increasing the enhancement factors.17 It is believed that aggregation is a major effect in a colloid system.18 Silver colloid systems had better enhancement factors in the presence of halide ions than those of other kinds of anions.17 Additionally, the presence of a small amount of halide anions in a silver colloid ensures higher reproducibility of the SERS spectra.19 In this study, we use the R6G as the probe molecule to investigate the surface enhancement factors of silver nanoprism and spherical nanoparticle colloids in the presence of KBr. We found that the AEF of the nanoprism colloid system was better than that of the spherical nanoparticle colloid system by a factor of 20 in the presence of 0.05 M KBr. The surface enhancement factors (SEF) can also be calculated from the AEF and the percentage of dye molecules adsorbed on the surface in the presence of KBr. The SEF of the nanoprism colloid and spherical nanoparticle colloid are about 3.2 × 105 and 9 × 103, respectively. The UV-vis-NIR spectrum of the nanoprism colloid shows that the out-of-plane quadrupole surface plasmon resonance (SPR) red-shifted and the in-plane dipole SPR blue-shifted, as the potassium bromide

10.1021/jp809687v CCC: $40.75  2009 American Chemical Society Published on Web 05/07/2009

SERS of Silver Nanoprism and Nanoparticle Colloids

J. Phys. Chem. C, Vol. 113, No. 22, 2009 9521

was added. It implied that the aspect ratios of silver nanoprisms decreased after the addition of KBr. We found that the decrease of aspect ratios comes from the shape transformation of nanoprisms sculptured by bromide ions. 2. Experimental Section 2.1. Materials. Silver nitrate was purchased from Sigma Aldrich. R6G was obtained from Exciton. Sodium citrate, silver nitrate, sodium borohydride, KF, KCl, KBr, and KI were purchased from Sigma Aldrich. All reagents were used without further purification. The Milli-Q grade water (>18 MΩ) was used for all solution preparation throughout all the experiments. 2.2. Instrumentation. Samples for transmission electron microscopy (TEM) images were examined using a Hitachi H-7100 TEM operated at 75 KV. Before analysis by TEM, the silver colloid was placed on carbon-coated copper and air-dried. All optical extinction spectra were recorded at 25 °C on a Hitachi U-2800 UV-vis-NIR spectrophotometer using 1 cm light path quartz cuvette. The fluorescence spectra were recorded with a Hitachi F-2500 fluorescence spectrometer. All surfaceenhanced Raman spectra were excited with a 532 nm laser (Coherent DPSS) and the scattering light was filtered by a notch filter (Kaiser Optical), dispersed by the monochromator (Jobin Yvon, Triax 550), and detected by a liquid nitrogen cooling CCD. 2.3. Colloids Preparation. Silver colloid was prepared according to the procedures in ref 20. A solution of sodium citrate (3.0 × 10-2 M, 1 mL) and a solution of silver nitrate (1.0 × 10-2 M, 1 mL) were added to 97 mL pure water with rapid stirring. Then the solution of sodium borohydride (5.0 × 10-3 M, 1 mL) was added dropwise to the mixture, under vigorous magnetic stirring. The solution immediately changed to a yellow color. The extinction spectrum of this solution displayed a narrow peak at 398 nm. This spectrum indicated that the diameter of the silver nanoparticles was about 10 nm. For convenience, the colloidal solution will be referred to as colloid C1 throughout this work. After being stirred for 30 min, the prepared solution was irradiated with the sodium lamp (Philips 100-W λ ) 589 nm). The typical power of the light on the solution is about 0.22 W/cm2. After 12 h irradiation, the solution changed to a blue color. Similarly, this colloidal solution will be referred to as colloid C2 throughout this work. 2.4. SERS Experiments. A solution of R6G (10-8 M, 0.9 mL) was added to 1 mL colloid C1 and colloid C2. After 30 min, the solution of mixtures was added with the different kinds of salts, KF, KCl, KBr, and KI (1M, 0.1 mL). Prepared samples were placed in a quartz cuvette, and excited by a 30 mW 532 nm laser. The typical acquisition time of a SERS spectrum was 30 s.

Figure 1. The UV-vis-NIR extinction spectrum of the silver colloid C1 (a) and colloid C2 (b).

3. Results and Discussion Figure 1a is the UV-vis-NIR extinction spectrum of colloid C1. It displays a narrow peak centered at 398 nm with a fwhm of about 60 nm. Figure 2a shows the TEM images of these silver nanoparticles. Most of the nanoparticles are spherical in shape and about 10 nm in diameter. Colloid C1 changed color from yellow to blue after light irradiation for 12 h. Figure 1b is the UV-vis-NIR extinction spectrum of colloid C2. In the spectrum, there are three distinct peaks at 332, 487, and 642 nm. According to the assignments in previous studies,21,22 the small and sharp peak at 332 nm is attributed to the out-of-plane quadrupole SPR of silver nanoplates with thickness of about 10 nm. The second peak at 487 nm is attributed to the in-plane quadrupole SPR, and the third peak at 642 nm is attributed to

Figure 2. (a) The TEM image of silver nanoparticles in colloid C1. (b) The TEM image of the truncated silver nanoparticles in colloid C2.

the in-plane dipole SPR with the edge length of about 70 nm.22 The peak at about 332 nm is a very distinct character of silver nanoplates. Figure 2b shows the TEM images of colloid C2, demonstrating that only less than 10% spherical nanoparticles survived after irradiation and the remaining majorities are truncated nanoprisms with different edge lengths. The edge length can be adjusted by tuning the excitation wavelength of

9522

J. Phys. Chem. C, Vol. 113, No. 22, 2009

Ciou et al.

Figure 3. SERS spectra of R6G excited by 532 nm laser in silver colloid C1 and C2. Both spectra have the same concentration of R6G (5 × 10-9 M), and the same salt added, 0.05 M KBr. Spectrum A was enhanced with colloid C2 and spectrum B was enhanced with colloid C1. The intensity of spectrum B had been multiplied by 10. The recording time is 10 s, and 3× over average. The laser power is controlled at 30 mW.

the light. Light irradiation can cause the anisotropic growth rate of the nanostructures, and different wavelength light can excite the dipole resonance of the nanoparticles in different sizes. Consequently, this produces the nanoprisms in different edge lengths. It had a relation that the longer the light excited the nanoparticles, the larger the nanoprisms that were generated.22 Figure 3 shows the SERS spectra of 4.5 × 10-9 M R6G with the presence of 0.05 M KBr in colloid C1 and C2. Several kinds of salts, KF, KCl, KBr, and KI were added to the dye silver colloidal solution. After several attempts, it was found that KBr was the best halide salt able to increase the SERS intensities of R6G. In Figure 3, Spectrum A, enhanced with colloid C2, was more than 20× stronger than Spectrum B, enhanced with colloid C1 both in the presence of 0.05 M KBr. In spectra A, there were two broad bands between 3250-3420 cm-1, which corresponded to the symmetry stretching and asymmetry stretching modes of the water molecules in liquid phase. Since the water concentration was fixed at 55.5 M throughout all of the experiments, the Raman signals of water played the role of the internal standard to help to check the relative intensities of SERS signals and further to estimate the enhancement factors of R6G in these two colloid solutions. Figure 4 shows the normal Raman Spectrum of 5 × 10-3 M R6G and water excited by 632.8 nm He-Ne laser. There are seven distinct Raman signals of R6G at the region of 500-1700 cm-1, and two water broad peaks around 3250-3420 cm-1. The Raman spectrum was floated on the all-scale fluorescence background. The peak at 1653 cm-1 was chosen to estimate enhancement factors. After the procedures of deconvolution with the Voigt functions, the ratio of peak area at 1653 cm-1 of R6G compared to the area of water signals at 3380 cm-1 is about 0.2. The differential Raman cross-section of R6G excited at 632.8 nm (σR6G,632.8) compared to that of water (σH2O,632.8) can be estimated by the following equation:

σR6G,632.8 IR6G,632.8 /CR6G ) σH2O,632.8 IH2O,632.8 /CH2O

(1)

where the IR6G,632.8 and IH2O are the differential Raman cross sections of R6G and water excited at 632.8 nm, respectively. CR6G and CH2O are the concentrations of R6G and water, respectively. In Figure 4, CR6G ) 5 × 10-3 M, CH2O ) 55.5 M,

Figure 4. Normal Raman spectrum of 5 × 10-3 M R6G and water excited by 632.8 nm He-Ne laser. The recording time is 30 s, and 3× over average. The laser power is controlled at 15 mW. A peak at 1653 cm-1 (denoted by the star) was chosen to calculate the enhancement factors.

IR6G,632.8/IH2O,632.8 ) 0.2. σR6G,632.8/σH2O,632.8 can be calculated to be about 2 × 103. Finding the Raman cross-section of water excited at 632.8 nm is not readily available. Nevertheless, it can be found that the Raman cross-section of water excited at 488 nm is 8 × 10-30 cm2/molecule sr in the literature.23 Since 488 and 632.8 nm are both too far away from the resonance wavelength to excite a water molecule to a real state, the ratio of the Raman cross sections excited at two wavelengths is simply proportional to (ν0 - νmn)4:24

σλ1 σλ2

)

(ν1 - νmn)4 (ν2 - νmn)4

(2)

where σλ1 and σλ2 are the normal Raman cross sections excited at the wavelength λ1 and λ2, which are both far away from the resonance absorption wavelength. Additionally, ν1 and ν2 are the frequencies corresponding to λ1 and λ2, and νmn is the Raman scattering frequency of the specific vibrational mode. Then the Raman cross-section of the peak at 3380 cm-1 of water excited at 632.8 nm, σH2O,632.8, can be calculated and equal to 2.2 × 10-30 cm2/molecule sr. σR6G,632.8 of the peak at 1653 cm-1 is equal to 4.4 × 10-27 cm2/molecule sr. This cross-section is much larger than the value which Nie et al. had estimated,13 but similar to the experimental results in the literature.25 The cause of the large Raman cross section of R6G at 632.8 nm is believed to be the preresonance Raman effect, since the excitation wavelength is close to the resonance wavelength.25 The differential cross-section of the 1653 cm-1 mode of R6G at resonance without surface enhancement, σR6G,RR, was recently predicted by the theoretical calculation to be on the order of 10-25 cm2/ molecule sr.26 The analytical enhancement factors (AEF) of R6G in the bulk silver-salts colloidal solution can be estimated by the following equation:25

AEF )

σcolloid,bulk σR6G,RR

(3)

Spectrum A in Figure 3 shows that the area of 4.5 × 10-9 M R6G SERS signal peaked at 1651 cm-1 compared to the water Raman signals peaked at 3380 cm-1 was about 1/4, and the concentration ratio is about 8 × 10-11. Raman cross-section of water at 532 nm, σH2O,532, can be calculated by using eq 2 and is about 5.2 × 10-30 cm2/molecule sr. σcolloid,c2 is equal to 1.6 × 10-20 cm2/molecule sr. By using eq 3, the analytical

SERS of Silver Nanoprism and Nanoparticle Colloids

Figure 5. (a) The fluorescence spectra of R6G excited by 532 nm laser. Spectrum R (dash), B1 (short dash), and B2 (solid) show 5 × 10-8 M R6G, 5 × 10-8 M R6G in colloid C1, and 5 × 10-8 M R6G in colloid C2, respectively. For convenience, the three spectra were normalized, and the maximum signal was set equal to 100. (b) The fluorescence spectra of R6G excited by 532 nm laser. Spectrum R (dash), B1 (short dash), and B2 (solid) show the supernatants of 5 × 10-8 M R6G, 5 × 10-8 M R6G in colloid C1, and 5 × 10-8 M R6G in colloid C2, respectively. For convenience, the three spectra were normalized, and the maximum signal of Spectrum R was set equal to 100.

enhancement factor (AEF) in colloid C2 is about 1.6 × 105, and the AEF in colloid C1 is about 8 × 103. The AEF in colloid C2 is similar to the value performed in the silver colloids prepared by the standard citrate reduction of Lee and Meisel, while the AEF in colloid C1 is much less than that in the referenced literature.25 There are three possibilities to explain why colloid C2 has a much better enhancement factor than colloid C1. One of these is the different adsorption abilities of these two kinds of colloids. The absorption ability of R6G/silver colloid system can be obtained from the fluorescence quenching experiments.27 Figure 5, parts a and b, shows the fluorescence spectra of 5 × 10-8 M R6G excited at 532 nm under several different conditions without the addition of KBr. The damping features on the fluorescence spectra do not come from the vibrational progression of R6G but from the specific behavior of our CCD detector when the signals of the neighbor pixels are very similar. The damping features would not be observed in the spectra with sharp peaks, such as the SERS spectra in Figure 3. Spectra in Figure 5, parts a and b, show that the sols and their corresponding supernatants after the centrifugation process exhibited almost the same fluorescence intensities. Therefore, it could be proposed that only the free dye molecules in the colloid solution can

J. Phys. Chem. C, Vol. 113, No. 22, 2009 9523 contribute to the fluorescence intensity and those molecules which were adsorbed on the surface of nanoparticles could not fluoresce.27 Consequently, the adsorption of R6G on the surface can be easily estimated by the reduction of fluorescence intensity in the silver colloids. The adsorption of R6G (5 × 10-8 M) in colloid C1 and C2 in the absence of bromide ions are about 10% and 35%, respectively. It shows that colloid C2 had a stronger affinity to the R6G molecules in the absence of KBr. Since the percentage of surface atoms of nanoprisms with 70 nm edge length and 10 nm thickness is about 5%, and the percentage of surface atoms on spherical nanoparticles with diameter of 10 nm is about 8%, the very different adsorption percentages on the two colloids would not come from the surface area but likely come from another surface property. It is probable that the process of photoinduced shape conversion not only changed the morphology of the nanostructures,28 but also changed the chemical or physical characteristics of the particle surfaces. The higher adsorption ability of colloid C2 in the absence KBr probably comes from the fact that R6G molecules favor to bind on {111} facet, which is the basal plane of silver nanoprisms, lowering the energy. The addition of salt into R6G/silver sols would cause the change of the ionic strength and the structure of electrical double layer; hence the adsorption ability of NP’s surface to dye molecules might be altered. Figure 6a shows the fluorescence spectra accompanied with SERS signals of 5 × 10-8 M R6G in two colloids in the presence of 0.05 M KBr. Fluorescence reduction of R6G in colloid C1 and C2 silver-halide systems were about 90% and 50%, respectively. The increased adsorption ability in both colloidal systems probably comes from the coadsorption of bromide ions and R6G molecules on the surface of nanoparticles. Nie et al. had demonstrated that silver cations would surround the surface of the nanoparticles, and play a very important role in the mechanism of chemical enhancement.17 Since R6G is a positive ion in a water solution, the bromide ion possibly functions as an electrostatic bridge to connect the surface and R6G cations, and therefore increases the adsorption. Figure 5a shows that the SERS signals of R6G could not be observed in both colloids in the absence of bromide ions, while they could be observed after the addition of KBr, shown in Figure 6a. The SERS signals had to increase to several tens of times their original strength to emerge from the large fluorescence background. Since the estimated values of adsorptions only increased by a few percent in colloid C2 and about 9-fold in colloid C1 after the addition of KBr, it is believed that the addition of bromide ions had caused the larger electromagnetic or chemical enhancement in both colloid systems. Previous studies have revealed that halide ions can activate chemical enhancement to increase the SERS signals 100 to 1000×.17 Figure 6b shows the fluorescence spectra of 5 × 10-8 M R6G in the supernatants of the dye-bromide-silver colloidal systems after the centrifugation process. Spectra T1 and T2 are correlated to the fluorescence spectra of the supernatants of colloids C1 and C2 after the centrifugation process, respectively. The SERS signals totally disappeared in these two colloidal systems because the silver NPs were moved out. The fluorescence intensities are almost the same, about 38% and 41% for colloids C1 and C2, respectively, relative to R6G solution. Comparing the fluorescence spectra of T1 (supernatants) and S1 in Figure 6, the dramatic increase of fluorescence intensity of colloid C1 sol after the centrifugation process probably comes from the fact that a portion of R6G molecules that were loosely adsorbed on the surface would desorb from the surface due to the inertial effect in such a centrifugation process.

9524

J. Phys. Chem. C, Vol. 113, No. 22, 2009

Ciou et al.

SEF AEF SEF AEF ) or ) Nsurface Nbulk Csurface Cbulk

Figure 6. (a) The fluorescence and Raman spectra of R6G excited by 532 nm laser. Spectrum R (dash), S1 (short dash), and S2 (solid) show 5 × 10-8 M R6G, 5 × 10-8 M R6G in colloid C1, and 5 × 10-8 M R6G in colloid C2, respectively. All solutions were in the presence of 0.05 M KBr. For convenience, the three spectra were normalized, and the maximum signal of Spectrum R was set equal to 100. (b) The fluorescence spectra of R6G excited by 532 nm laser. Spectrum R (dash), B1 (short dash), and B2 (solid) show the supernatants of 5 × 10-8 M R6G, 5 × 10-8 M R6G in colloid C1, and 5 × 10-8 M R6G in colloid C2, respectively. 0.05 M KBr were added to all the solutions for 30 min prior to centrifugation process. For convenience, the three spectra were normalized, and the maximum signal of spectrum R was set equal to 100.

In contrast with colloid C1, the fluorescence intensity of R6G in colloid C2 with the addition of KBr decreased a little after centrifugation process. This is probably due to the fact that a fraction of R6G molecules, which were coadsorbed with bromide ions on the surface of silver nanoprisms, might have fluorescence. The additional fluorescence contributed from the coadsorbed molecules possibly resulted from the fact that the excited fluorophores were not quenched completely by the surface or surface enhanced fluorescence occurred accompanying the SERS signals. However, it remained an unsolved problem for us to distinguish the mechanisms of the fluorescence from the molecules on the surface. Some further sophisticated works, such as fluorescence profile or lifetime measurements,29 might be helpful to understand this phenomenon. Since not all the R6G molecules were adsorbed on the surface of nanoparticles, the surface enhancement factor (SEF) can be estimated by SERS intensities and the ratio of molecules which were adsorbed on the surface. The SEF can be calculated from the analytical enhancement factors (AEF) by the following equation:

(4)

Nsurface and Nbulk are the number of molecules on the surface and in the bulk solution, respectively. From the fluorescence intensities of the spectra in Figure 6a, it can be estimated that the percentages of R6G molecule adsorbed on the surface of the nanoparticles in colloids C1 and C2 in the presence of bromide ions are about 90% and 50%, respectively. The SEFcolloid,C1 and SEFcolloid,C2 are 9 × 103 and 3.2 × 105, respectively. After considering the different adsorption abilities in the presence of bromide ions, the surface enhancement factor of colloid C2 is about 30× larger than that of colloid C1. From the discussion above, the fact that SEF of colloid C2 is much larger than that of colloid C1 in the presence of KBr could possibly come from the fact that the R6G molecules that were bound tightly on {111} facet of the basal plane of nanoprisms might have more chemical enhancement than those that were loosely adsorbed on the surface of spherical NPs. The second possibility is that colloid C2 possibly generates a larger local electrical field because the SPR bands of nanoprisms were better at overlapping with the excitation laser wavelength than that of spherical nanoparticles, as shown in Figure 1. Moreover, the diameter of the spherical nanoparticles in colloid C1 is only about 10 nm, much smaller than the silver substrates, which were about 100 nm, that were believed to have the better enhancement factors in the previous studies.13,18 The third possibility is that more nanostructures with SERS active sites were formed in colloid C2 than in colloid C1 after the addition of bromide ions. The UV-vis-NIR extinction spectra of the colloid systems would be altered by the addition of salts due to the formation of aggregation or disaggregation of NPs.19 Figure 7, parts a and b, shows the UV-vis-NIR extinction spectra of colloid C1 and colloid C2 respectively, in the presence of 0.05 M KBr. There is almost no difference between Figure 7a and b. This means that the added salt almost did not cause aggregation in colloid C1. However, Figure 7b shows a very different spectrum compared to Figure 1b. The out-of-plane quadrupole SPR band red-shifted from 332 to 360 nm and in-plane dipole SPR band blue-shifted from 642 to 440 nm after the addition of KBr. Both shifts of SPR bands indicated the decrease of aspect ratios of the two-dimensional anisotropic nanostructures.30 Figure 8 is the TEM image, which corresponds to colloid C2 with the addition of 10-5 M KBr for 5 min, then dripped onto the carbon-coated copper grids for air drying. In

Figure 7. (a) The UV-vis-NIR extinction spectrum (dash) of the silver colloid C1 in the presence of 0.05 M KBr. (b) The UV-vis-NIR extinction spectrum of the silver colloid C2 in the presence of 0.05 M KBr.

SERS of Silver Nanoprism and Nanoparticle Colloids

J. Phys. Chem. C, Vol. 113, No. 22, 2009 9525 The latter colloid had 20× better analytical enhancement factors than the former colloid. There are three possibilities to explain these results. The first is that the silver nanoprisms have a better ability to adsorb the R6G molecules in the absence of bromide ions than spherical nanoparticles. The second is that the SPR bands of nanoprisms resonate better than that of spherical nanoparticles with the excitation laser. The third is that some small atomic silver clusters with high surface enhancement effect were probably formed during the nanoprism’s sculpturing process in the presence of KBr. Acknowledgment. The authors are thankful for the financial support of the National Science Foundation of Taiwan (NSC 96-2113-M-415-007-). C.L. is grateful to Profs. Yuan T. Lee, Chi-Kung Ni, Jon T. Hougen, I-Chia Chen, and Wenlung Chen for their long-time support and encouragement. References and Notes

Figure 8. The TEM image of the colloid C2. The colloid was added with 0.01 M KBr for 5 min prior to drip a drop on to the TEM grid. The arrows point at the shadows of triangular shape.

Figure 8, most nanoparticles are circular in shape; however, three nanostructures, indicated by the arrows, showed that the darker circular shapes were surrounded by lighter triangular shadows. It meant that the silver nanoprisms were sculptured by bromide ions and had been transformed into nanodisks. The triangular shadows possibly corresponded to the residues of sculptured nanoprisms, which remained around the nanodisks because the diffusion rate was limited on the TEM grid. Xu et al. reported that the chloride ions could selectively sculpture the {110} facets of silver nanoprisms and transform their shapes into planar disk-like sheets.31 They also observed the SERS signals of 4-MPY in the colloid with the addition of chloride ions were 10× stronger than that without the addition of chloride ions. Consequently, they proposed that the increase of enhancement factors would probably come from the existence of small clusters, which were formed after sculpturing, and the adsorption of halide ions on the surface.31 Bromide ion, similar to chloride ion, can also function as an etching agent to sculpture the corners of silver nanoprisms into nanodisks. The dissolved silver atoms or ions would possibly form small atomic silver clusters,31 such as Ag4+, which were proposed to have larger surface enhancements.32,33 Since the sculpturing process only exists in colloid C2 but not in colloid C1 with the addition of bromide ions, the formation of small atomic clusters seems to be one of the reasons to explain the larger SEF of colloid C2 sol system. However, the fact that the SERS signals of colloid C2 system completely disappeared after centrifugation, shown in Figure 6b, indicated that the materials corresponding to surface enhancement were totally moved out. To our knowledge, the free atomic cluster is so light that it would not precipitate in such a centrifugation process. Therefore, it seems more probable that the atomic clusters can help to increase the surface enhancement effect only when they are bound or coadsorbed with the bromide ions on the surface of NPs. This is also a possible explanation for why colloid C2 had a better SERS enhancement factor than colloid C1 in the presence of KBr. 4. Conclusions The SERS spectra of R6G were performed on silver spherical nanoparticle and nanoprism colloids in the presence of KBr.

(1) Moskovits, M. ReV. Mod. Phys. 1985, 57, 783. (2) Fleischmann, M.; Hendra, P. J.; McQuillan, A. J. Chem. Phys. Lett. 1974, 26, 163. (3) Jeanmaire, D. L.; Van Duyne, R. P. J. Electroanal. Chem. 1977, 84, 1. (4) Albrecht, M. G.; Creighton, J. A. J. Am. Chem. Soc. 1977, 99, 5215. (5) Premasiri, W. R.; Moir, D. T.; Klempner, M. S.; Krieger, N.; Jones, G., II; Ziegler, L. D. J. Phys. Chem. B 2005, 109, 312. (6) Jarvis, R. M.; Brooker, A.; Goodacre, R. Faraday Discuss. 2006, 132, 281. (7) Wood, B. R.; Heraud, P.; Stojkovic, S.; Morrison, D.; Beardall, J.; McNaughton, D. Anal. Chem. 2005, 77, 4955. (8) Zhang, X.; Yonzon, C. R.; Young, M. A.; Stuart, D. A.; Van Duyne, R. P. IEE Proc.-Nanobiotechnol. 2005, 152, 195. (9) Lee, P. C.; Meisel, D. J. Phys. Chem. 1982, 86, 3391. (10) Volkan, M.; Stokes, D. L.; Vo-Dinh, T. J. Raman Spectrosc. 1999, 30, 1057. (11) Zeisel, D.; Deckert, V.; Zenobi, R.; Vo-Dinh, T. Chem. Phys. Lett. 1998, 283, 381. (12) Dick, L. A.; McFarland, A. D.; Haynes, C. L.; Van Duyne, R. P. J. Phys. Chem. B 2002, 106, 853. (13) Nie, S.; Emory, S. R. Science 1997, 275, 1102. (14) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.; Dasari, R. R.; Feld, M. S. Phys. ReV. Lett. 1997, 78, 1667. (15) Yang, Y.; Matsubara, S.; Xiong, L.; Hayakawa, T.; Nogami, M. J. Phys. Chem. C 2007, 111, 9095. (16) Jiang, J.; Bosnick, K.; Maillard, M.; Brus, L. J. Phys. Chem. B 2003, 107, 9964. (17) Doering, W. E.; Nie, S. J. Phys. Chem. B 2002, 106, 311. (18) Campion, A.; Kambhampati, P. Chem. Soc. ReV. 1998, 27, 241. (19) Muniz-Miranda, M.; Sbrana, G. J. Raman Spectrosc. 1996, 27, 105. (20) Jia, H.; Xua, W.; Ana, J.; Li, D.; Zhao, B. Spectrochim. Acta, Part A 2006, 64, 956. (21) Jin, R.; Cao, Y. W.; Mirkin, C. A.; Kelly, K. L.; Schatz, G. C.; Zheng, J. G. Science 2001, 294, 1901. (22) Jin, R.; Cao, Y. C.; Hao, E.; Me´ traux, G. S.; ; Schatz, G. C.; Mirkin, C. A. Nature 2003, 425, 487. (23) Faris, G. W.; Copeland, R. A. Appl. Opt. 1997, 36, 2686. (24) Albercht, A. C. J. Chem. Phys. 1961, 34, 1476. (25) Ru, E. C. Le.; Blackie, E.; Meyer, M.; Etchegoin, P. G. J. Phys. Chem. C 2007, 111, 13794. (26) Jensen, L.; Schatz, G. C. J. Phys. Chem. A 2006, 110, 5973. (27) Hildebrandt, P.; Stockburger, M. J. Phys. Chem. 1984, 88, 5935. (28) Xue, C.; Me´ traux, G. S.; Millstone, J. E.; Mirkin, C. A. J. Am. Chem. Soc. 2008, 130, 8337. (29) Ru, E. C. L.; Etchegoin, P. G.; Grand, J.; Fe´lidj, N.; Aubard, J.; Le´vi, G. J. Phys. Chem. C 2007, 111, 16076. (30) (a) Kelly, K. L.; Coronado, E.; Zhao, L. L.; Schatz, G. C. J. Phys. Chem. B 2003, 107, 668. (b) Bohren, C. F.; Huffman, D. R. Absorption and Scattering of Light by Small Particles; Wiley Interscience: New York, 1983. (31) An, J.; Tang, B.; Zheng, X.; Zhou, J.; Dong, F.; Xu, S.; Wang, Y.; Zhao, B.; Xu, W. J. Phys. Chem. C 2008, 112, 15176. (32) Otto, A.; Mrozek, I.; Grabhorn, H.; Akemann, W. J. Phys.: Condens. Matter 1992, 4, 1143–1212. (33) Roy, D.; Furtak, T. E. Chem. Phys. Lett. 1986, 124, 299.

JP809687V