Self-Assembly of Silver Nanoparticles - American Chemical Society

Self-Assembly of Silver Nanoparticles: Synthesis, Stabilization, Optical Properties, and. Application in Surface-Enhanced Raman Scattering. Sudipa Pan...
0 downloads 0 Views 486KB Size
13436

J. Phys. Chem. B 2006, 110, 13436-13444

Self-Assembly of Silver Nanoparticles: Synthesis, Stabilization, Optical Properties, and Application in Surface-Enhanced Raman Scattering Sudipa Panigrahi,† Snigdhamayee Praharaj,† Soumen Basu,† Sujit Kumar Ghosh,† Subhra Jana,† Surojit Pande,† Tuan Vo-Dinh,‡ Hongjin Jiang,§ and Tarasankar Pal*,† Department of Chemistry, Indian Institute of Technology, Kharagpur-721302, India, Center for AdVanced Biomedical Photonics, Oak Ridge National Laboratory, Oak Ridge, Tennessee, and Department of Material Science and Engineering, Georgia Institute of Technology, Atlanta, Georgia ReceiVed: April 5, 2006; In Final Form: May 14, 2006

Silver nanoparticle aggregates were synthesized in large scale using resorcinol under alkaline condition to obtain an assembly of silver clusters. Stable dispersion of the cluster in aqueous medium has been examined out of resorcinol-capped silver nanoparticle assemblies. The UV-vis spectroscopy during the particle evolution has been studied in detail. From the high-resolution TEM (HRTEM) image and XRD pattern it was confirmed that the particles are made of pure silver only. The capping action of resorcinol has been authenticated from the FTIR spectra. UV-vis spectroscopy and TEM images reveal that the temperature, effect of vibrational energy, heat shock, and time-dependent particle evolution have unique bearing on the stability and surface properties of the clusters. The concentrations of silver nitrate, resorcinol, and NaOH have important influence on the particle evolution and its size. TEM images incite us to examine the aggregates to capitulate surfaceenhanced Raman scattering (SERS) to the single molecular level using crystal violet (CV) and cresyl fast violet (CFV) as molecular probes. The SERS intensity of CV increases with increasing the size of the silver aggregate.

1. Introduction Particles in the nanometer size range have attracted increasing attention with the growing interest in nanotechnological disciplines. The intrinsic properties of metal nanoparticles are mainly governed by their size, shape, composition, crystallinity, and structure. Colloidal inorganic nanoparticles exhibit size- and shapedependent phenomena that are expected to lead to superstructures with a range of practical applications.1,2 During the past few years, several techniques have been developed for directing the self-assembly of nanocrystals into ordered aggregates.3-5 Self-assembly of nanoparticles has been of tremendous interest to the science and technology since it provides effective building blocks for physical, chemical, and biological applications. Recently, there has been much experimental interest in the optical properties of noble metal nanoparticle aggregates.6-10 This interest has arisen in part because of new methods for linking nanoparticles into clusters and complicated aggregates11 and in part because of advances in the technology for characterizing the nanoparticles using AFM, STM,12,13 surface-enhanced Raman scattering (SERS),12-15 optical measurements,16 and near-field methods17 that illuminate only a few nanoparticles. Self-assembly based on selective control of noncovalent interactions18-20 provides a powerful tool for the creation of structured systems at a molecular level, and application of this methodology to macromolecular systems provides a means of extending such structures to the macroscopic length scale.21 * To whom correspondence should be addressed. E-mail: tpal@ chem.iitkgp.ernet.in. † Indian Institute of Technology. ‡ Oak Ridge National Laboratory. § Georgia Institute of Technology.

Aggregation processes in colloids have been the subjects of numerous experimental, theoretical, and computational studies.22-24 The fabrication of assemblies of perfect nanometerscale crystallites (quantum crystal) identically replicated in unlimited quantities in such a state that can be manipulated and understood as pure macromolecular substances is an ultimate challenge in modern materials research with outstanding fundamental and potential technological consequences. Despite numerous reports, the experimental investigation of nanoparticles in liquid is limited, primarily due to two factors: the small size prohibits the application of the most commonly used imaging techniques; it is more difficult to access the properties of individual nanoparticles than those of aggregates or clusters. Kimura et al. have reported the formation of a threedimensional superlattice of gold nanoparticles.25 Mirkin and coworkers26 and Alivisatos et al.27 have demonstrated the formation of aggregated metal clusters, using DNA as the recognition element. Near-field optical absorbance of gold nanoparticle aggregates has been observed and studied in detail.28 Some of the amino acids can induce aggregation for gold nanoparticles.29 Recently, Monti et al.30 have reported variable single aggregate of luminescence spectrum. The aggregation of the silver colloids leads to a less stable system with limited use in time. A possible solution to such a problem is the immobilization of the nanoparticles onto substrates. This can be accomplished by a direct procedure31 or by a previous functionalization of the surface.32 Our group has reported the reversible formation of gold aggregates by UV photoactivation.33 Mandal et al. have reported the aggregation of cysteine-capped silver particles.34 In the literature, there are several reports on gold nanoparticle aggregates, but syntheses of silver aggregates have met with limited success. On the other hand, the physical chemistry underpinning these interactions and the degree of influence of

10.1021/jp062119l CCC: $33.50 © 2006 American Chemical Society Published on Web 06/21/2006

Self-Assembly of Silver Nanoparticles environmental factors that remain for silver nanoparticles are rarely investigated due its oxidation problem. In this article, we address the wet chemical synthesis of silver nanoparticle aggregates using resorcinol as a reducing agent under alkaline medium. The process leads to the formation of silver nanoparticle aggregates in the gram scale, which remain well dispersed in water. The synthesized particles are stable for months together and show no sign of oxidation or dissociation. The size of the particles in the aggregate can be tailored by controlling the concentration of resorcinol, silver nitrate, and sodium hydroxide and by adjusting the time of heating. The aggregates of silver particle have been characterized by UVvisible spectroscopy, transmission electron microscopy, and X-ray diffraction spectroscopic studies. We also focus our study on the effect of interparticle coupling on the particle plasmon resonances, especially the shift of the plasmon resonance wavelength as a function of particle separation. As expected, the experimental results indicate that the resonance wavelength of two coupled particles in close proximity is significantly redshifted from that of the individual particles. FTIR spectroscopic study revealed the binding of the resorcinol molecules to the surface of silver particles. From the XRD pattern and HRTEM images we confirmed that the aggregates are made of pure silver particles only. The effect of thermal energy and vibrational energy on the stability of the aggregates was studied, and the results present the understanding to manipulate reversible aggregation and dispersion of the particles. Lastly, the aggregates of silver particles were employed successfully to study the surface-enhanced Raman scattering (SERS) using cresyl fast violet (CFV) and crystal violet (CV) as molecular probes. From the SERS study using the aggregates, detection of molecular probes almost to the single molecular level is possible. SERS intensity was found to be dependent on the state of aggregation of silver particles. 2. Experimental Section 2.1. Reagents and Instruments. All the reagents used were of AR grade. Resorcinol was purchased from Merck and was recrystallized from hot water. Silver nitrate was obtained from Aldrich, and a 0.1 M aqueous solution was used as stock solution. Sodium hydroxide was obtained from Merck, India. Mili-Q water was used throughout the experiment. Absorption spectra were recorded in a Spectrascan UV 2600 spectrophotometer (Chemito, India), and the solutions were taken in a 1 cm well-stoppered quartz cuvette. Transmission electron microscopic (TEM) analyses were performed in a Hitachi H-9000 NAR instrument on samples prepared by placing a drop of fresh silver solution on Cu grids precoated with carbon films, followed by solvent evaporation under vacuum. Fourier transform infrared (FTIR) spectral characteristics of the samples were collected in reflectance mode with Nexus 870 ThermoNicolet instrument coupled with a Thermo-Nicolet continuum FTIR microscope. One drop of the test solution was placed on a KBr pellet and was dried under vacuum for 6 h before analysis. The FTIR spectrum was recorded over 32 scans of each sample, and the background spectrum was automatically subtracted. The X-ray diffraction (XRD) pattern was recorded in an X′pert pro diffractometer with Co (KR ) 1.788 91 Å) radiation. Surfaceenhanced Raman scattering studies were performed using a Renishaw Raman spectrometer. Two different types of lasers [He-Ne (632.8 nm) and Ar+ ion (514.5 nm) lasers] were used, and the laser power at the sample was 2 mW. An integration time of 10 ms was used for each measurement. 2.2. Synthesis of the Silver Nanoparticle Assembly. In a beaker, 10 mL of resorcinol solution in water was taken. Then

J. Phys. Chem. B, Vol. 110, No. 27, 2006 13437 0.5 mL of a 0.1 M AgNO3 solution was added and the solution was mixed thoroughly. Finally, 10 mL of 0.5 M NaOH was added to it. The final concentration of silver nitrate was 5 mM, and that of resorcinol was 0.5 M. The mixture was kept on a temperature-controlled hot water bath at a temperature of ∼80 °C. After the addition of the AgNO3 solution the mixture turned yellow within a few second indicating the onset of the evolution of silver particles. After about 20 min of heating, the solution turned red. On further heating, the solution changes its color from red to gray to blackish gray. After the solution was heated on a water bath for 90 min, a green tint appeared in the gray colored solution indicating the formation of aggregates of silver nanoparticles. After that, the sol showed no further change in color implying the completion of the reaction. The final solution was cooled and centrifuged at 2000 rpm for 5 min. The liquid at the top of the centrifuge tube was yellow and was discarded. The remaining solid mass was redispersed in distilled water and further centrifuged. After the second centrifugation, the liquid in the tube became colorless. The centrifugation was repeated for 7-8 cycles to remove the excess reagents. The solid mass at the bottom of the tube was redispersed in water and used to characterize the sample. The method is reproducible, and the aggregates of silver nanoparticles prepared by this method remain stable for 3 months when stored under ambient conditions. 3. Results and Discussion 3.1. Evolution and Characterization of Silver Nanoparticle Aggregates. 3.1.1. Absorption Spectra and TEM Study. The aggregates of silver nanoparticles have been synthesized using silver nitrate as the precursor salt and resorcinol as the reducing agent in alkaline medium. The successive changes of the absorption spectra during the evolution of silver nanoparticle aggregates are shown in Figure 1. After the addition of AgNO3 to the alkaline solution of resorcinol, the color of the solution changed from colorless to light yellow indicating the nucleation of the silver particles at their infancy. Absorption measurement of the light yellow solution showed an extinction spectrum with a maximum at 410 nm (trace a, Figure 1). Transmission electron microscopic (TEM) measurement of the solution showed the particles have an average diameter of ∼7 ( 0.5 nm (Figure 2A). On further heating, the solution turned deep yellow. Timedependent absorbance measurement during the turnover of light yellow solution to deep yellow solution showed that the λmax does not change effectively but the absorbance increases. After 20 min of the reaction, the solution changed the color from deep yellow to red with the appearance of a single hump near 520 nm (trace b, Figure 1) and the 410 nm peak red shifted to 415

Figure 1. Successive changes in the absorption spectra of silver nanoparticle aggregates during heating at different time intervals (a) 0 min, (b) 20 min, (c) 40 min, and (d) 90 min.

13438 J. Phys. Chem. B, Vol. 110, No. 27, 2006

Panigrahi et al. nm. TEM images of the red solution showed the particles in the size range of 10-12 nm (Figure 2B). At this stage the particles were almost monodispersed with tight size distribution. As the heating of the solution was continued, the solution changed its color from red to gray to blackish gray that was reflected by the successive red shifting of the 520 nm peak in the absorption spectra while the 415 nm peak also slightly red shifted. The surface plasmon absorption of the gray solution exhibited two peaks with maxima at 425 and 665 nm (trace c, Figure 1). TEM image (Figure 2C) of the gray colored solution showed decreased interparticle separation with an increase of individual particle diameter of 15 nm. After 90 min of heating, there was an appearance of green tint in the solution. The solution was cooled and centrifuged several times. At this stage the first peak at 425 nm was constant but the second peak further shifted to the 738 nm. (trace d, Figure 1). The transmission electron micrograph of the finally formed colloid showed aggregates of silver particles with an average particle size of 20 nm diameter in the aggregate (Figure 2D). After that, the solution on further heating produced a black precipitate. At this stage, the particle assembly could not maintain the colloidal dimension, and hence, the aggregates were precipitated as black mass. Metal colloids show pronounced surface plasmon (SP) resonances, whose spectral position can be varied over a wide range by a suitable choice of the metal, the dielectric, and the geometry of the colloid.35 The gradual development of the surface plasmon band during the evolution of silver nanoparticle aggregate (Figure 1) can be explained as follows. An isolated pseudospherical silver nanoparticle in aqueous solution is characterized by a single intense near-UV band centered at ∼410 nm, which arises due to the excitation of the dipolar surface plasmon of the silver nanoparticles.36 The UV-visible absorption spectra of a fairly dilute dispersion of colloidal particles can be calculated from the “Mie theory”.37 The absorbance, A, for a dispersion of N particles/unit volume is given by38

A)

CNl 2.303

(1)

where C and l are the absorption cross section and the optical path length, respectively. In this case optical path length l is 1 cm and it is fixed. Thus, the increase in absorbance of the solution at 410 nm indicates the increase in particle concentration. In the quasi-static limit, i.e., 2πR < λ (where R is the radius of the particles and λ the wavelength of light in the media), only the electric dipole term, developed in Mie’s theory, is significant. Then the cross section can be expressed as

C)

(18πV2(ω)m3/2) λ[(1(ω) + 2m)2 + 2(ω)2]

(2)

where V and λ are the volume of the spherical particle and the incident wavelength, corresponding to a frequency ω, respectively. The complex relative permittivity of the metal, (ω), is expressed as

(ω) ) 1(ω) + i2(ω)

Figure 2. TEM images of the aggregates during their evolution: (A) 0 min; (B) 20 min; (C) 40 min; (D) 90 min.

(3)

where 1 is the real part and 2 is the imaginary part of . When 2 is small or does not change so much around the band, the position of maximum absorption occurs for all 1(ω). From the above equations, it is evident that the absorption spectrum markedly depends on the dielectric constant of the surrounding media, m. Furthermore, the plasmon peak width is related to

Self-Assembly of Silver Nanoparticles

J. Phys. Chem. B, Vol. 110, No. 27, 2006 13439

the dielectric constant of the surrounding environment, m.1 When the particles are in close packed assembly, then the electromagnetic field of one particle influences its neighbor. It has been shown theoretically and experimentally that aggregation of silver nanoparticles leads to another plasmon absorption at longer wavelengths when the individual nanoparticles are electronically coupled to each other. The oscillating electrons in one particle feel the electric field due to the oscillation of the second particle, which can lead to a collective plasmon oscillation of the aggregated system. The frequency and intensity of the latter depend on the degree of aggregation as well as orientation of the individual particles within the aggregate. As the dipole plasmon red shifts, a quadrupole resonance grows in. Because of the rapid variation of the real part of  with wavelength, the quadrupole resonance also starts at about 410 nm for small particles. The quadrupole resonance also red shifts as the particle size increases, but the effect is much smaller than for the dipole resonance. But still the dipole resonance is more intense than the quadrupole resonance. This clearly demonstrates that the peak near 520 nm arises from the dipolar term. Taking into account the variation of the plasmon peak with the dielectric constant, it is reported that the shift toward lower energy was observed due to an increase in the dielectric constant of the composite medium of the particles, which is the superposition of several factors,39 such as the spherical particles, particle-particle interactions, and the nature of the surrounding medium. Metal sols normally aggregate through two global (nanoparticle-concentration-dependent) mechanisms called clusterparticle aggregation and cluster-cluster aggregation.23,24 Electromagnetic coupling of clusters is effective for cluster-cluster distances smaller than 5 times the cluster radius and may lead to complicated extinction spectra depending on size and shape of the formed cluster aggregate by a splitting of single cluster resonance.40 In this case, it was observed that the effect of polarization has a great influence on the aggregate formation since larger particles are more polarized than the smaller one. For spherical particles with polarization P

P ) R[Eloc + Erad]

(4)

where the radiative correction field, Erad, is

Erad ) (2/3)ik3P + k2/aP

(5)

where “a” is the particle radius. The first term in this expression describes radiative damping. It arises from spontaneous emission of radiation by the induced dipole. This emission grows rapidly with particle size, eventually reducing the size of the induced dipole and increasing the plasmon line width. The second term comes from depolarization of the radiation across the particle surface due to the finite ratio of particle size to wavelength. This dynamic depolarization term causes red shifting of the plasmon resonance as the particle size is increased. An equivalent theory for other particle shapes has been described by Zeman and Schatz.41 The net effect of both these terms is to produce a modified polarization P of the particle multiplied by the following correction factor:

F ) (1 - (2/3)ik3R - k2/aR)-1

(6)

It is to be noted that the radiative damping contribution to the correction factor is proportional to the product of the polarizability.42 It was observed from the TEM images that when

Figure 3. High-resolution transmission electron microscopy (HRTEM) image of the aggregate.

Figure 4. X-ray diffraction pattern of the silver aggregate.

particles were of 15 nm in size, they induce aggregation, whereas the smaller particles do not form aggregates. For spheres that are not too large, this dependence is already known and is described in terms of radiation damping and dynamic depolarization effects. 3.1.2. HRTEM Image. A TEM image at high magnification of these aggregates confirms that they are made of silver nanoparticles. High resolution shows a 4-fold symmetry and hexagonal network. The HRTEM picture of silver nanoparticles is shown in Figure 3. The lattice spacing of the nanoparticle is 0.25 nm, which corresponds to the lattice spacing of {100} plane of pure silver. Hence, the solid mass is composed of pure silver nanoparticles and no sign of oxidation was observed. The HRTEM image indicates that the sample is single crystalline in nature. The presence of 4-fold symmetry clearly indicates formation of fcc structure. Such packing of silver nanoparticles has been demonstrated in the literature previously.43 3.1.3. X-ray Diffraction Pattern. The XRD pattern of the silver aggregate is shown in Figure 4. The reflection peaks above 30° are indexed as the (111), (200), (220), (311), and (222) plane of cubic silver. As shown in Figure 4, the XRD pattern of silver nanoparticles shows that the peak positions are consistent with those of bulk silver.44 It is well-known that silver colloid is very unstable, and the instability is believed to be due to the random aggregation or surface oxidation of the silver nanoparticles. In our case the synthesized particles are stable enough. This is presumably due to the fact that the capping action of the resorcinol monolayer on the surface of the particles can prevent the nanoparticle aggregates from the surface oxidation or agglomeration in the solvent. 3.1.4. FTIR Spectra. Figure 5 shows the FTIR spectra of the pure resorcinol (trace a Figure 5) and resorcinol-capped silver nanoparticles (trace b, Figure 5). The FTIR spectra of the pure resorcinol and synthesized silver aggregates revealed that the solid silver particles are capped by the resorcinol moiety. The peak in the 3500 cm-1 region red shifted, and also the peak became sharp after capping to the silver surface that confirmed the surface binding of resorcinol to the silver surface. The red shifting was also observed for the other peak (1650 cm-1). The

13440 J. Phys. Chem. B, Vol. 110, No. 27, 2006

Figure 5. FTIR spectra of pure resorcinol and silver aggregate: (a) resorcinol; (b) resorcinol-capped silver aggregate.

peak in the 2100 cm-1 is clearly seen in the resorcinol molecule but completely vanishes in the aggregated particles due to coordination of resorcinol to the silver surface.34 3.2. Stabilization of the Aggregate. Metal sols are often stabilized in solution by the presence of a charged double layer surrounding each colloidal nanoparticle that produces a coulomb barrier to aggregation.40 The nanoparticles can be induced to form an aggregate by replacing the charged surface species by either uncharged adsorbate or by neutralizing the surface charge as evident from the DLVO theory.40,45 The stability of the aggregates, in the present experiment, may be due to the hydrogen-bonding capability of the resorcinol molecule or its oxidation product, tetrahydroxybiphenyl, which stabilizes the particles. In fact, a higher ξ (zeta) potential and thickness of the electrical bilayer (Stern layer) exist on the silver nanoparticles because of the higher negative electric charge provided by the alkaline solution of resorcinol and its oxidation product. Moreover, the existence of a high negative electric charge in silver nanoparticles may have an important influence on the nucleation and growing mechanism involved in the formation of these particles, being responsible for the final observed morphologies. 3.3. Effect of NaOH. The aggregates of silver nanoparticles have been synthesized by the wet chemical reduction of silver nitrate using resorcinol in alkaline medium. For the synthesis of silver colloids, resorcinol molecules serve as the reducing agents as well as the capping agent. When the concentration of NaOH was 0.5 mM, the red shifted peak was obtained. Since the aqueous solution of resorcinol is feebly acidic (pka1 ) 9.15 and pka2 ) 11.06),46 after the addition of resorcinol solution and AgNO3 to this system the pH of the solution turned to 6.2 ( 0.2. But this pH went down further to 5.9 at the end of the reaction. The pH of the reaction was found to play an important role in assembling the nanoparticles. At a high concentration of NaOH (10 mM), the black precipitate was obtained due to the formation of Ag2O. However, at an intermediate concentration of NaOH (1 mM), only one broad peak peaking at 532 nm was observed (trace a, Figure 6). When the concentration of NaOH was 10 µM, we could see the light yellow solution indicating the formation of spherical silver nanoparticles. This solution has absorption maxima at ∼405 nm (trace b, Figure 6). It is assumed that, at the prescribed concentration of NaOH (0.5 mM), resorcinol molecules brought the particles together presumably through a hydrogen-bonding interaction to form a close-pack assembly. This confirms that the suppressing effect on aggregation previously seen is indeed linked to the deprotonation of the hydroxyl group, which does not occur at lower concentration of NaOH. The reaction does not take place at all in absence of NaOH, highlighting the importance of the species

Panigrahi et al.

Figure 6. UV-vis spectra of the aggregate at different concentrations of NaOH: (a) [NaOH] ) 1 mM; (b) [NaOH] ) 10 µM.

Figure 7. UV-vis spectra of the aggregate at different concentrations of resorcinol: (a) 10-1 M; (b) 10-2 M; (c) 10-3 M; (d) 10-4 M.

for particle generation.47 It seems intuitively obvious that only dibasic resorcinol can act as cross-linking agents for pairs of silver particles, thereby inducing a color change, as was observed. 3.4. Effect of Resorcinol and Silver Nitrate. The effect of concentration of resorcinol and silver nitrate was studied in detail. The optimum concentration of resorcinol was found to be 0.5 M for the stable silver aggregate formation. When the concentration of resorcinol was 10 mM, then one peak with a λmax at 411 nm was observed (trace a, Figure 7). Increased concentration of resorcinol (100 mM) produced a single hump at about 497 nm along with a peak at 414 nm. But a further increase in concentration (1.0 M) of resorcinol indicated a broad peak. At this high concentration (1.0 M) of resorcinol the silver particles were precipitated. On the other hand, when we decreased the concentration of resorcinol to ca. 1 mM, the peak red shifted to 429 nm, and a still lower concentration (0.1 mM) indicated a plasmon peak at 449 nm. Further lowering of resorcinol concentration (