pubs.acs.org/Langmuir © 2009 American Chemical Society
Time-Dependent and Symmetry-Selective Charge-Transfer Contribution to SERS in Gold Nanoparticle Aggregates Jun Hee Yoon, Jung Shin Park, and Sangwoon Yoon* Department of Chemistry, Center for Photofunctional Energy Materials, Dankook University, 126 Jukjeon-dong, Suji-gu, Yongin, Gyeonggi 448-701, Korea Received March 30, 2009. Revised Manuscript Received September 18, 2009 We report the time- and symmetry-dependent surface-enhanced Raman scattering (SERS) of gold nanoparticle (AuNP) aggregates. The addition of p-aminothiophenol (p-ATP) instantly induces the aggregation of AuNPs, confirmed by large absorption in the near-IR region. Dynamic light scattering measurements show that the addition of p-ATP immediately assembles the AuNPs (13 nm) to form aggregates with a mean diameter of ∼200 nm, which then further grow to a size of ∼300 nm. Raman spectra acquired via time lapse show that the a1-symmetry bands of p-ATP are enhanced simultaneously with the formation of the aggregates, indicating that the electromagnetic enhancement largely contributes to the SERS of the AuNP aggregates. In contrast, the enhancement of the b2-symmetry bands occurs ∼10 h after the formation of the aggregates and slowly progresses. The enhancement of the b2 mode is attributed to the charge transfer between AuNPs and adsorbates, rather than the reorientation of the adsorbates because thiophenol and p-methylthiophenol that have surface structures and intermolecular interactions similar to those of p-ATP do not exhibit a symmetry-specific Raman enhancement pattern. To elucidate the disparity in the timescale between the chargetransfer resonance and the formation of the aggregates, we propose two models. A further close approach of the AuNPs constituting the aggregates causes the additional adsorption of the initially adsorbed p-ATP onto neighboring AuNPs, tuning the charge transfer state to be in resonance with the Raman excitation laser. Density functional theory calculations confirm the resonance charge-transfer tunneling through the bridging p-ATP in the AuNP-p-ATP-AuNP structures. Alternatively, the gradual continuing adsorption of p-ATP increases the local Fermi level of AuNPs into the region of resonant charge transfer from the Fermi level to the LUMO of the adsorbates. This model is corroborated by the faster appearance of b2-mode enhancement for the AuNPs with initially higher zeta potentials.
1. Introduction Surface-enhanced Raman scattering (SERS) has attracted a great deal of attention since the discovery of this phenomenon in the 1970s.1-3 Raman scattering from molecules on roughened noble metal surfaces is enormously enhanced. An enhancement factor of up to 1014 has been reported.4,5 Many studies focus on the fundamental understanding of the phenomenon as well as applications that exploit its extremely high sensitivity.6-11 It is generally accepted that SERS occurs via two mechanisms: electromagnetic (EM) enhancement and chemical enhancement. In EM enhancement, the collective oscillation of the conduction electrons in metallic nanoparticles (i.e., surface plasmons) is resonantly excited by the incident laser, producing intensified EM fields around the nanoparticles that enhance the Raman scattering of molecules in the vicinity.12-14 The EM enhancement *Corresponding author. E-mail:
[email protected]. (1) Fleischman, M.; Hendra, P. J.; McQuillan, A. J. Chem. Phys. Lett. 1974, 26, 123. (2) Jeanmaire, D. L.; Van Duyne, R. P. J. Electroanal. Chem. 1977, 84, 1. (3) Albrecht, M. G.; Creighton, J. A. J. Am. Chem. Soc. 1977, 99, 5215. (4) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.; Dasari, R. R.; Feld, M. S. Phys. Rev. Lett. 1997, 78, 1667. (5) Nie, S.; Emory, S. R. Science 1997, 275, 1102. (6) Moskovits, M. Rev. Mod. Phys. 1985, 57, 783. (7) Campion, A.; Kambhampati, P. Chem. Soc. Rev. 1998, 27, 241. (8) Haynes, C. L.; McFarland, A. D.; Van Duyne, R. P. Anal. Chem. 2005, 77, 338 A. (9) Cao, Y. C.; Jin, R.; Mirkin, C. A. Science 2002, 297, 1536. (10) Camden, J. P.; Dieringer, J. A.; Zhao, J.; Van Duyne, R. P. Acc. Chem. Res. 2008, 41, 1653. (11) Brus, L. Acc. Chem. Res. 2008, 41, 1742. (12) Gersten, J.; Nitzan, A. J. Chem. Phys. 1980, 73, 3023. (13) Moskovits, M. J. Chem. Phys. 1982, 77, 4408. (14) Schatz, G. C. Acc. Chem. Res. 1984, 17, 370.
Langmuir 2009, 25(21), 12475–12480
is even more pronounced in the sharp curvature of nanoparticles and in small gaps between nanoparticles. In chemical enhancement, the adsorption of molecules onto nanoparticles couples the molecular electronic states with the metallic electronic band to form a new charge-transfer (CT) state. Resonance Raman scattering through the CT state increases the signal.7,15 The ability to control the structure and assembly of nanoparticles in recent years has advanced our understanding of SERS. Many novel nanoarchitectures have been explored, including nanowire bundles,16 nanoprisms,17 nanoshells,18 and nanoparticle/substrate structures spaced by self-assembled monolayers.19 In those structures, SERS originates from the nanogaps, predominantly contributed by the EM enhancement.16,19-22 Gold nanoparticle (AuNP) aggregates in solution also provide an excellent platform for studying SERS.23 The addition of thiol compounds to AuNP solutions induces the aggregation of AuNPs by way of the reduction of electrostatic repulsion. The resulting (15) Lombardi, J. R.; Birke, R. L.; Lu, T. H.; Xu, J. J. Chem. Phys. 1986, 84, 4174. (16) Lee, S. J.; Morrill, A. R.; Moskovits, M. J. Am. Chem. Soc. 2006, 128, 2200. (17) Bae, Y.; Kim, N. H.; Kim, M.; Lee, K. Y.; Han, S. W. J. Am. Chem. Soc. 2008, 130, 5432. (18) Oldenburg, S. J.; Westcott, S. L.; Averitt, R. D.; Halas, N. J. J. Chem. Phys. 1999, 111, 4729. (19) Freeman, R. G.; Grabar, K. C.; Allison, K. J.; Bright, R. M.; Davis, J. A.; Guthrie, A. P.; Hommer, M. B.; Jackson, M. A.; Smith, P. C.; Walter, D. G.; Natan, M. J. Science 1995, 267, 1629. (20) Wiley, B. J.; Im, S. H.; Li, Z.-Y.; McLellan, J.; Siekkinen, A.; Xia, Y. J. Phys. Chem. B 2006, 110, 15666. (21) Zhou, Q.; Fan, Q.; Zhuang, Y.; Li, Y.; Zhao, G.; Zheng, J. J. Phys. Chem. B 2006, 110, 12029. (22) Jackson, J. B.; Halas, N. J. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 17930. (23) Ghosh, S. K.; Pal, T. Chem. Rev. 2007, 107, 4797.
Published on Web 10/09/2009
DOI: 10.1021/la9031865
12475
Article
Yoon et al.
close distances between the AuNPs in the aggregates satisfy the condition for SERS via the EM enhancement, producing SERS for the molecules at the interstitial sites. Furthermore, the SERS properties of the AuNP aggregates can be controlled by modifying the interparticle interactions by varying experimental conditions such as the size and surface functionality of the AuNPs, the pH of the solution, and the solvent.23-25 Despite the extensive studies of SERS from AuNP aggregates, most work has focused on SERS arising from EM enhancement. Here, we report the observation of SERS from AuNP aggregates contributed by the CT enhancement, which unexpectedly occurs far later than SERS via the EM enhancement or the formation of the AuNP aggregates. This slow appearance of the CT provides an important clue to the subtle structural evolution of AuNP aggregates.
2. Experimental Section We employed the Turkevich method to synthesize AuNPs.26 Adding trisodium citrate (50 mL, 3.4 10-2 M) to a heated HAuCl4 solution (950 mL, 2.7 10-4 M) with vigorous stirring yields AuNPs with a diameter of 13 nm, determined by highresolution transmission electron microscopy (HR-TEM, JEM3010, JEOL). The AuNPs show a characteristic surface plasmon band at 521 nm. The aggregation of AuNPs is induced by the addition of 2 mL of 2.6 μM p-aminothiophenol (p-ATP) to 2 mL of aqueous AuNP solution. We determined the added amount of p-ATP for stable AuNP aggregates by performing a series of concentration-dependence experiments (Supporting Information). The final concentration of 1.3 μM of p-ATP corresponds to the 30% surface coverage of individual AuNPs, assuming perfect adsorption probability and an occupation area of 0.22 nm2 for p-ATP.27 Under this condition, AuNP aggregates remain stable without precipitation for at least 15 days. The average particle sizes are measured by dynamic light scattering (ELS-Z, Otsuka Electronics). The surface properties of AuNPs are characterized by the zeta potential. Centrifugation of AuNP solutions at 15 000 rpm for 10 min partially removes the citrate anions surrounding the nanoparticles, increasing the zeta potential from -49 to -30 mV. SERS spectra of the AuNP aggregates are measured using a Raman microscope (RamanRxn1 Microprobe, Kaiser Optical Systems). A diode laser at 785 nm is focused on the cell through an objective with 10 magnification. Raman scattering is collected at 180° by the same objective and delivered to a holographic spectrometer (f/1.8) by optical fibers. All spectra presented are the averages of three spectra acquired at an exposure time of 20 s. The spectral resolution is 4 cm-1. For density functional theory calculations, we used the Gaussian 03 program.28 (24) Kim, T.; Lee, K.; Gong, M.-s.; Joo, S.-W. Langmuir 2005, 21, 9524. (25) Zhong, Z.; Patskovskyy, S.; Bouvrette, P.; Luong, J. H. T.; Gedanken, A. J. Phys. Chem. B 2004, 108, 4046. (26) Turkevich, J.; Stevenson, P. C.; Hillier, J. Discuss. Faraday Soc. 1951, 11, 55. (27) Mohri, N.; Inoue, M.; Ara, Y.; Yoshikawa, K. Langmuir 1995, 11, 1612. (28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03; Gaussian, Inc.: Wallingford, CT, 2004.
12476 DOI: 10.1021/la9031865
Figure 1. Evolution of extinction spectra of AuNP solutions upon addition of p-ATP (1.3 μM). The selected spectra shown are acquired 5 min, 15 min, 45 min, 1.5 h, 3 h, 6 h, 18 h, 24 h, 50 h, 74 h, 100 h, and 125 h after the addition, as indicated. The spectrum of pure AuNP solutions (---) is presented for comparison. The arrows are a visual guide to show the direction of spectral changes with time.
3. Results and Discussion Formation of AuNP Aggregates. We observed the formation of AuNP aggregates upon addition of p-ATP. We acquired extinction spectra of the AuNP solution as a function of time as shown in Figure 1. The spectrum represented by the dashed line shows the surface plasmon absorption of dispersed AuNPs before aggregation occurs. Immediately after the addition of p-ATP, the surface plasmon band at 521 nm shifts to 525 nm with its intensity decreased and a new broad absorption band appears at 660 nm. As time passes, the 525 nm band further decreases and the absorption in the long-wavelength region broadens and red shifts. Such changes in absorption are also visible to the naked eye. The color of the solution changes from red to dark red to purple. The new absorption band in the long-wavelength region is attributed to the dipole plasmon resonance excitation of coupled AuNPs.29-31 The position and width of the band are sensitive to the degree of interaction between the dipolar surface plasmons of AuNPs. For example, as the distance between the nanoparticles decreases, the dipolar interaction becomes stronger, resulting in the red shift of the extinction spectrum.25,29 Therefore, the instant appearance of the absorption band at long wavelengths, followed by the gradual red shift, indicates the immediate formation of AuNP aggregates upon addition of p-ATP and the subsequent slow evolution of the aggregates. According to the DLVO theory, AuNPs are kept apart by electrostatic repulsion provided by the surrounding capping materials, overcoming the van der Waals attraction between particles.32 In our experiments, the added p-ATP quickly displaces the citrate anions on the surface of AuNPs because of the high affinity of the thiol group for Au. Consequent reduction of the surface charges of the AuNPs causes immediate aggregation of the nanoparticles, as indicated by the immediate appearance of the absorption band in the long-wavelength region. As more p-ATP molecules adsorb, it is anticipated that more AuNPs are assembled, producing larger AuNP aggregates. The interparticle distance within the aggregates also decreases because of further neutralization of surface charges. Larger aggregates and stronger interparticle interactions induce the broadening and red shift of the absorption band, consistent with our observation.25,29 (29) Jensen, T.; Kelly, L.; Lazarides, A.; Schatz, G. C. J. Cluster Sci. 1999, 10, 295. (30) Lazarides, A. A.; Kelly, K. L.; Jensen, T. R.; Schatz, G. C. J. Mol. Struct.- Theochem 2000, 529, 59. (31) Lazarides, A. A.; Schatz, G. C. J. Phys. Chem. B 2000, 104, 460. (32) Cao, G. Nanostructures and Nanomaterials: Synthesis, Properties, and Applications; Imperial College Press: London, 2004.
Langmuir 2009, 25(21), 12475–12480
Yoon et al.
Article
Figure 2. Structural evolution of AuNP aggregates. The average size of the AuNP aggregates was measured using dynamic light scattering as time elapsed after the addition of p-ATP to the AuNP (13 nm) solution.
The structural evolution of the AuNP aggregates was measured using dynamic light scattering (DLS).33 The average size of the AuNP aggregates was measured as time elapsed after the addition of p-ATP. Figure 2 shows that the addition of p-ATP instantly assembles the individual AuNPs (13 nm) to produce AuNP aggregates with a mean diameter of ∼200 nm. The AuNP aggregates then grow to a size of ∼300 nm over 10 h, after which they remain stable in suspension. The structural evolution measured by DLS is consistent with our interpretation of the UV-visible absorption spectra discussed above. SERS of AuNP Aggregates. To study the SERS behavior of the AuNP aggregates, we acquired Raman spectra as time elapsed after the addition of p-ATP to the AuNP solution to a final concentration of 1.3 μM. Figure 3 presents the selected Raman spectra obtained at the time delays indicated. We observe strongly enhanced Raman spectra of p-ATP whereas no Raman signal is detected for the same concentration of p-ATP (1.3 μM) in ethanol without AuNPs. From the comparison of the enhanced Raman spectrum of p-ATP from the AuNP aggregates to that of 0.5 M p-ATP in ethanol (Figure 3b), we determine the enhancement factor, 5 106. The assignment of the Raman spectra of p-ATP has been discussed in previous publications.34,35 In addition to the large enhancement, in Figure 3 the modespecific changes in Raman intensity with time are noteworthy. Whereas most Raman peaks are enhanced instantly and remain largely unchanged, three vibrational peaks at 1137, 1384, and 1426 cm-1, marked by the asterisks, appear at 13 h and slowly rise. We note that these three vibrational modes are of the b2-symmetry species whereas all of the other peaks in the spectrum correspond to either a1 or b1 symmetry.34,35 Dissimilar evolution of the Raman peaks between the two symmetry modes is clearly demonstrated by plotting the intensities of the 1078 cm-1 (C-S stretch, a1) and the 1137 cm-1 (C-H bend, b2) bands as a function of time, as shown in Figure 4. Also included in Figure 4 for comparison is the change in absorbance at 785 nm indicative of the formation of AuNP aggregates. Figure 4 shows that the rise in the Raman intensity of the 1078 cm-1 band is on a similar time scale to that of the absorbance at 785 nm, suggesting that the Raman enhancement of the 1078 cm-1 band is concurrent with the formation of the AuNP aggregates. In contrast, the 1137 cm-1 band of b2 symmetry does not even appear until 10 h has passed and slowly rises over a long period of time. Although (33) It is difficult to measure the morphology of aggregates using transmission electron microscopy (TEM) because, depending on the number of particles sampled on the carbon-coated copper grid, the TEM images look different and can be misleading. (34) Osawa, M.; Matsuda, N.; Yoshii, K.; Uchida, I. J. Phys. Chem. 1994, 98, 12702. (35) Kim, K.; Yoon, J. K. J. Phys. Chem. B 2005, 109, 20731.
Langmuir 2009, 25(21), 12475–12480
Figure 3. (a) Raman spectra from AuNP solutions acquired at each indicated time delay after the addition of p-ATP (1.3 μM). (b) Normal Raman spectrum of 0.5 M p-ATP in ethanol in the absence of AuNPs. All spectra were obtained under the same conditions. The asterisks indicate the b2-symmetry vibrational modes that show a different intensity evolution from that of other modes. Solvent spectra have been subtracted.
Figure 4. Change in the (a) absorbance at 785 nm and Raman intensity at (b) 1078 and (c) 1137 cm-1 as a function of elapsed time after the addition of p-ATP to AuNPs.
the intensity is weak, the 1384 cm-1 (δCHþνCC, b2) and 1426 cm-1 (νCCþδCH, b2) bands show similar behavior. Symmetry- and Time-Dependent Changes in SERS Intensities. The aggregation of AuNPs produces nanogaps between the AuNPs where surface plasmon resonance excitation generates intense electromagnetic fields. Such an EM enhancement makes a major contribution to the observed SERS spectra of p-ATP. As soon as the addition of p-ATP induces the aggregation, the a1-symmetry-type Raman bands are enhanced via the EM enhancement. However, significantly slow enhancement of the b2-symmetry bands suggests that some other mechanisms come DOI: 10.1021/la9031865
12477
Article
Yoon et al.
Figure 5. Reorientation model for selective enhancement of the b2-symmetry bands of p-ATP. As more p-ATP molecules adsorb onto the surface of AuNPs, the adsorbates reorient because of the intermolecular interactions. The symmetry-dependent enhancement from the surface selection rule is given for each orientation.
into play in the increase in those bands. Here we present two possible origins of the enhancement of the b2-symmetry bands. One possibility is the reorientation of p-ATP adsorbed on the surfaces of AuNPs. As shown in Figure 5, the adsorbed p-ATP may undergo orientation changes because of increased intermolecular interactions caused by the increase in the surface density. The surface selection rule of the EM enhancement model tells us that the intensities of vibrational modes polarized perpendicular to the surface should be preferentially enhanced.13,36-38 Under the assumption that p-ATP belongs to the C2v point group, the b2-symmetry vibrational modes (e.g., in-plane C-H bend) must be enhanced as the molecule reorients from its flat-lying position to a standing-up position. Despite apparent agreement of the model with the experimental results, we rule out this possibility on the basis of the following two reasons: First, we added p-ATP in the amount corresponding to the submonolayer coverage of AuNPs (0.3 ML). Thus, it is unlikely that the adsorbed p-ATP changes its orientation by crowding from the increased surface coverage. Second, similar symmetry-dependent Raman intensity changes are not observed for thiophenol (TP) and p-methylthiophenol (p-MTP). The adsorption rate of TP and p-MTP onto AuNPs must be the same as that of p-ATP because all three bind to AuNPs through a Au-S bond. Furthermore, the benzene rings that they have in common provide similar intermolecular interactions. Therefore, if indeed the enhancement of the b2 mode arises from the geometrical changes in adsorbates, then the SERS spectra of TP and p-MTP should show a similar pattern to that of p-ATP. However, Figure 6 shows that no symmetry-dependent Raman intensity changes are observed for TP and p-MTP contrary to p-ATP. Therefore, we believe that SERS of the b2 mode is associated with the chemical properties of molecules rather than the orientation effects. The other possibility for the selective enhancement of the b2-symmetry bands is the resonance charge transfer between AuNPs and p-ATP. When the charge-transfer transition from the metal to the adsorbates or vice versa is in resonance with the Raman excitation energy, then resonance-Raman-like enhancement occurs. Preferential enhancement of the b2 mode of p-ATP due to the CT has been reported for other systems. Osawa and coworkers found that the SERS spectrum of p-ATP adsorbed onto the Ag electrode is different from the normal Raman spectrum of free p-ATP in that the former is dominated by the b2-symmetry bands and the latter exhibits only a1-symmetry bands.34 From the dependence of the intensities of the b2-symmetry bands on (36) Hexter, R. M.; Albrecht, M. G. Spectrochim. Acta, Part A 1979, 35, 233. (37) Nichols, H.; Hexter, R. M. J. Chem. Phys. 1981, 74, 3126. (38) Richardson, N. v.; Sass, J. K. Chem. Phys. Lett. 1979, 62, 267. (39) Carron, K. T.; Hurley, L. G. J. Phys. Chem. 1991, 95, 9979. (40) Ohta, N.; Yagi, I. J. Phys. Chem. C 2008, 112, 17603. (41) Maitani, M. M.; Ohlberg, D. A. A.; Li, Z.; Allara, D. L.; Stewart, D. R.; Williams, R. S. J. Am. Chem. Soc. 2009, 131, 6310.
12478 DOI: 10.1021/la9031865
Figure 6. Time-dependent SERS of the AuNP aggregates induced by the addition of (a) p-ATP, (b) TP, and (c) p-MTP. Each Raman spectrum was acquired at the indicated time delay on the left after the addition of molecules. To compare the symmetry-dependent Raman intensity changes with time, we mark the a1-symmetry and b2-symmetry bands of each molecule by open circles and filled triangles, respectively. The assignments of the TP and p-MTP spectra are from refs 39-41.
the applied electrochemical potential and Raman excitation wavelength, they suggested that the enhancement in the b2-symmetry bands is largely a contribution from the chargetransfer resonance transition between silver and the adsorbates. Zheng and co-workers noticed a significant enhancement of the b2 modes of p-ATP interconnecting silver nanoparticles in the Ag-p-ATP-Ag assembly.42,43 They ascribed the enhancement of the b2 modes to the charge transfer between the silver particles tunneling through the bridging p-ATP molecules. Theoretical studies reveal that the intensity borrowing of the CT transition from allowed molecular transitions (the Herzberg-Teller effect) is associated with the preferential enhancement of modes that are not totally symmetric.15,44,45 In particular, Lombardi and Birke presented a unified expression for SERS intensities that includes all of the contributions from surface plasmon resonance (i.e., the EM enhancement), CT resonance, and molecular resonance and showed that one can calculate the relative contributions of each resonance.44 By employing the equation proposed by Lombardi and Birke, we found that the degree of the CT contribution to the overall SERS enhancement increases from 0 to 0.11 for the 1137 cm-1 band, with 0 denoting no CT contribution and 1 denoting a predominant CT contribution. Other theoretical calculations also verify that the b2 mode increases via CT. The time-dependent density functional theory calculations for metal-molecule complexes and metal-molecule-metal junction architectures showed that resonant charge-transfer tunneling through the molecular junction is responsible for the selective enhancement of b2 modes in Raman spectra.46,47 On the basis of these experimental and theoretical studies, we suggest that the enhancement of the b2 modes of p-ATP from the AuNP aggregates is attributed to CT, (42) Zhou, Q.; Li, X.; Fan, Q.; Zhang, X.; Zheng, J. Angew. Chem., Int. Ed. 2006, 45, 3970. (43) Zhou, Q.; Zhao, G.; Chao, Y.; Li, Y.; Wu, Y.; Zheng, J. J. Phys. Chem. C 2007, 111, 1951. (44) Lombardi, J. R.; Birke, R. L. J. Phys. Chem. C 2008, 112, 5605. (45) Lombardi, J. R.; Birke, R. L. Acc. Chem. Res. 2009, 42, 734. (46) Liu, S.; Zhao, X.; Li, Y.; Zhao, X.; Chena, M. J. Chem. Phys. 2009, 130, 234509. (47) Sun, M.; Xu, H. ChemPhysChem 2009, 10, 392.
Langmuir 2009, 25(21), 12475–12480
Yoon et al.
Figure 7. Possible mechanisms for the slow CT resonance in AuNP aggregates. (a) Alteration of a CT state by the formation of AuNP-p-ATP-AuNP interlinked structures. A CT state formed by the adsorption of p-ATP on an AuNP is not in resonance with the incident Raman excitation laser. Upon aging of AuNPs in the aggregates, the additional adsorption occurs through the N-Au bonding, altering the CT state to allow the resonant CT transition. (b) Change in the local Fermi level (EF) of AuNPs by the adsorption of p-ATP. Even after the aggregates are formed, as more p-ATPs are adsorbed, the surface potential of the AuNPs constituting the aggregates increases, enabling resonant CT from the Fermi level of AuNP to the LUMO of p-ATP.
which is further corroborated by the investigations presented below. Once we accept that the major contribution to the enhancement of the b2 modes is the resonant CT transition in the AuNP aggregates, the question arises as to why CT occurs on a different time scale from the formation of the AuNP aggregates. The p-ATP molecules quickly adsorb onto the AuNPs through a strong Au-S bond, displacing citrate anions on the surfaces. The reduced surface charges bring the AuNPs together immediately to form aggregates with a size of ∼200 nm (Figure 2). The resulting close distances between the AuNPs in the aggregates provide an environment for EM enhancement that causes the instantaneous enhancement of the a1 modes of adsorbed p-ATP. In contrast, despite the instant formation of the AuNP aggregates, it takes ∼10 h for the Raman enhancement by CT to take place. Here we propose two models for the slow appearance of the CT resonance, both of which are illustrated in Figure 7. The first possibility is the alteration of a CT state by the formation of AuNP-p-ATP-AuNP interlinked structures. The adsorption of molecules onto metal surfaces creates a CT state by coupling the molecular electronic states with the electronic band of the metal.15 Resonance Raman scattering through the CT state enhances the Raman signal. It is possible that the CT state created by the initial adsorption of p-ATP onto AuNPs through a Au-S bond is not quite in resonance with the incident Raman excitation wavelength (Figure 7a). As the aging of AuNP aggregates occurs, a further close approach of AuNPs constituting the aggregates causes the additional adsorption of the initially adsorbed p-ATP onto neighboring AuNPs through the nonbonding electrons of Langmuir 2009, 25(21), 12475–12480
Article
Figure 8. Optimized geometry, HOMO, and LUMO of (a) pATP, (b) p-ATP bound to Au on the thiol side, and (c) p-ATP bound to Au’s on both the thiol and amine sides. The lower diagram presents the calculated molecular orbital energies in electronvolts. The HOMO-LUMO energy gap of each system is denoted in nanometers for comparison with the Raman excitation laser wavelength (785 nm) for the resonance transition.
amine, tuning the CT state to be in resonance with the Raman excitation laser.48 To determine if this model is feasible, we carried out density functional theory calculations. We constructed three molecular structures: free p-ATP, p-ATP bound to Au on the thiol side, and p-ATP bound to Au on both the thiol and amine sides. After the optimization of the structures at the B3LYP level of theory using LANL2DZ basis sets, we calculated the molecular orbital energy levels.28 Figure 8 presents optimized geometries, lowest unoccupied molecular orbitals (LUMO), highest occupied molecular orbitals (HOMO), and corresponding molecular orbital energy levels of each system. The energy gap between the HOMO and LUMO decreases as Au binds to thiol and further to both thiol and amine, finally matching the wavelength of incident light (785 nm). An inspection of the molecular orbitals reveals that the charges are indeed transferred through the p-ATP bridge from one Au atom to the other as the transition occurs from HOMO to LUMO, as shown in Figure 8c. In stark contrast, similar calculations for TP and p-MTP show that the HOMO-LUMO energy gap of the metal-molecule-metal systems is far from being in resonance with the Raman excitation energy (Supporting Information), consistent with the experimental results shown in Figure 6 where no symmetry-specific CT enhancement is observed for those molecules. Although our calculations have limited applications because we have used isolated Au atoms rather than AuNPs, they qualitatively explain the lowering of the (48) A further close approach of the AuNPs by the additional adsorption of the amine group onto neighboring AuNPs should also result in an increase in the a1symmetry band because of the electromagnetic enhancement. However, it seems that the change in the interparticle distance caused by the additional adsorption of p-ATP is very small probably because the second adsorption occurs within the already-formed aggregates. No further spectral red shift of the SPR band in the long-wavelength region while the charge transfer is occurring (>10 h) also indicates that the change in the interparticle distance is not significant enough to cause an increase in the electromagnetic enhancement effect.
DOI: 10.1021/la9031865
12479
Article
Yoon et al.
Despite our experimental and theoretical attempts to elucidate the time-dependent CT phenomena in the AuNP aggregates, at this stage it is difficult to resolve unequivocally which model is more appropriate. It is also possible that both models are indistinguishable and work cooperatively. Further investigations including the modification of AuNP surfaces with other materials and varying Raman excitation wavelengths will provide better insight into the contribution of the CT enhancement to SERS in the AuNP aggregates.
4. Conclusions
Figure 9. Time-dependent SERS spectra of AuNP aggregates formed by AuNPs with (a) lower (ζ = -49 mV) and (b) higher (ζ = -30 mV) zeta potentials. The arrows mark the b2-symmetry vibrational band of p-ATP at 1137 cm-1, which is largely enhanced by the charge transfer.
HOMO-LUMO energy gap by the additional binding of amine to Au and the nature of the charge transfer upon transition. More advanced calculations for similar systems show that the additional adsorption of the Ag2 cluster onto the amine of p-ATP that is bound to the Au2 cluster decreases the transition energy from 543.44 to 853.73 nm, in qualitative agreement with our results.46 Therefore, it is plausible that the contact of the amine group of the adsorbed p-ATP with neighboring AuNPs tunes the CT state into a resonance region matching the Raman excitation energy, resulting in the enhancement of the b2 modes. Another possible cause of the difference in the charge-transfer timescale from that of the formation of aggregates is the gradual change in the surface potential of AuNPs, as illustrated in Figure 7b. Even after the AuNP aggregates are formed, as more p-ATP molecules adsorb onto the AuNPs, the surface potential of the AuNPs changes, leading to the alteration of local Fermi energy levels of the metal nanoparticles. The substitution of neutral p-ATP for citrate anions on the AuNPs increases the surface potential, better matching the CT transition from the Fermi level of AuNPs to the LUMO of the adsorbates with the Raman excitation energy. We tested this model by examining the effect of the surface potential of AuNPs on the b2-mode enhancement. Figure 9 presents the time-dependent SERS spectra of the AuNP aggregates induced by adding p-ATP to the AuNPs with two different surface potentials (ζ). For the nanoparticles with higher zeta potential (ζ = -30 mV) shown in Figure 9b, the increase in the 1137 cm-1 b2-mode peak, marked by an arrow, is faster than when the nanoparticles have a lower surface potential (-49 mV). When the initial zeta potential of AuNPs is higher, it should take less time to increase the surface potential to the region of a CT resonance by the adsorption of the same amount of p-ATP. Therefore, the enhancement of the b2 modes by the CT resonance should appear faster for higher zeta potential, agreeing with the experimental results.
12480 DOI: 10.1021/la9031865
We observed that the intensities of the SERS spectra of p-ATP from the AuNP aggregates evolve differently with time, depending on the vibrational symmetry of each peak. Most vibrational peaks of the a1- or b1-symmetry species were instantaneously enhanced as the AuNP aggregates were formed. In contrast, the enhancement of the b2-mode peaks (1137, 1384, and 1426 cm-1) were not observed until 10 h had passed and slowly increased with time. We attributed the enhancement of the b2 modes to the charge-transfer resonance rather than to the reorientation effect because (a) we used a small amount of p-ATP to induce the aggregation, corresponding to 0.3 ML of AuNP surfaces assuming the perfect adsorption probability, and (b) TP and p-MTP, which have similar structures on the AuNP surfaces and similar intermolecular interactions to those of p-ATP, did not exhibit any symmetrydependent SERS behavior. We proposed two models for the time-dependent charge-transfer resonance in the AuNP aggregates. As the AnNPs consisting of the aggregates approach each other closely, the adsorption of the amine group of p-ATP onto the neighboring AuNPs alters the CT state and causes it to be in resonance with the Raman excitation wavelength, producing SERS. Density functional theory calculations as well as the slow red-shift of the UV-visible absorption spectra support this model. However, it is also possible that the slow adsorption of p-ATP onto the AuNPs after the formation of the aggregates gradually increases the local Fermi energy level of the AuNPs into the region of the resonant charge transfer from the Fermi level to the LUMO of the adsorbates. This model was corroborated by the faster appearance of the b2-mode enhancement for the AuNPs with initially higher zeta potentials. Further investigations including the modification of the AuNP surfaces with other materials and varying Raman excitation wavelengths will provide better insight into the contribution of CT enhancement of SERS in the AuNP aggregates. Acknowledgment. We gratefully acknowledge the support of the Korea Science and Engineering Foundation (R01-2007-00011332-0) and the Korea Research Foundation (KRF-2006-331C00141). This work was also supported by the GRRC program of Gyeonggi province (GRRC Dankook 2009-B03). Supporting Information Available: Experimental method and density functional theory calculations of TP and p-MTP. This material is available free of charge via the Internet at http://pubs.acs.org.
Langmuir 2009, 25(21), 12475–12480
