Single-Molecule and Single-Particle-Based Correlation Studies

Nov 20, 2013 - Individual EF values and their fractional populations are directly obtainable in single molecule-SERS (sm-SERS), and combining SMS with...
1 downloads 4 Views 2MB Size
Letter pubs.acs.org/NanoLett

Single-Molecule and Single-Particle-Based Correlation Studies between Localized Surface Plasmons of Dimeric Nanostructures with ∼1 nm Gap and Surface-Enhanced Raman Scattering Haemi Lee,∥,† Jung-Hoon Lee,∥,§ Seung Min Jin,† Yung Doug Suh,*,†,‡ and Jwa-Min Nam*,§ †

Laboratory for Advanced Molecular Probing (LAMP), Research Center for Convergence Nanotechnology, Korea Research Institute of Chemical Technology, Daejeon 305-600, South Korea ‡ School of Chemical Engineering, Sungkyunkwan University, Suwon 440-746, Korea § Department of Chemistry, Seoul National University, Seoul 151-747, South Korea S Supporting Information *

ABSTRACT: Understanding the detailed electromagnetic field distribution inside a plasmonically coupled nanostructure, especially for structures with ∼1 nm plasmonic gap, is the fundamental basis for the control and use of the strong optical properties of plasmonic nanostructures. Using a multistep AFM tipmatching strategy that enables us to gain the optical spectra with the optimal signal-to-noise ratio as well as high reliability in correlation measurement between localized surface plasmon (LSP) and surface-enhanced Raman scattering (SERS), the coupled longitudinal dipolar and high-order multipolar LSPs were detected within a dimeric structure, where a single Raman dye is located via a single-DNA hybridization between two differently sized Au−Ag core−shell particles. On the basis of the characterization of each LSP component, the distinct phase differences, attributed to different quantities of the excited quadrupolar LSPs, between the transverse and longitudinal regimes were observed for the first time. By assessing the relative ratio of dipolar and quadrupolar LSPs, we found that these LSPs of the dimer with ∼1 nm gap were simultaneously excited, and large longitudinal bonding dipolar LSP/longitudinal bonding quadrupolar LSP value is required to generate high SERS signal intensity. Interestingly, a minor population of the examined dimers exhibited strong SERS intensities along not only the dimer axis but also the direction that arises from the interaction between the coupled transverse dipolar and longitudinal bonding quadrupolar LSPs. Overall, our high-precision correlation measurement strategy with a plasmonic heterodimer with ∼1 nm gap allows for the observation of the characteristic spectral features with the optimal signal-to-noise ratio and the subpopulation of plasmonic dimers with a distinct SERS behavior, hidden by a majority of dimer population, and the method and results can be useful in understanding the whole distribution of SERS enhancement factor values and designing plasmonic nanoantenna structures. KEYWORDS: Plasmonic nanogap, single-nanoparticle spectroscopy, single-molecule SERS, plasmonic heterodimer, localized surface plasmon, single-particle spectroscopy

D

that the SERS enhancement factors (EFs) at the varying hot spots exhibit a broad distribution with a long population tail, a phenomenon that could be largely attributable to subtle change in the plasmonic nanostructures and the random orientation and position of an analyte molecule within the plasmonic hot spot.8 For example, a broad distribution ranging from ∼105 to ∼1012 can result in a considerable miscalculation of the measured SERS EF values, and it is almost impossible to quantitatively and reliably measure the SERS intensity from nanostructures. Therefore, it is of paramount importance to fully understand the relationship between nanostructures, Raman dyes, and optical signals.

esigning, synthesizing, manipulating, and understanding plamonically coupled multimeric structures including dimers and trimers have been important and widely studied topics due to their usefulness and potential applications in optics, materials science, nanoscience, energy, biotechnology, and medicine. Among them, plasmonic dimeric nanostructures have been the most heavily studied structures because they generate a very strong plasmonic coupling, their plamonic properties are relatively well understood, and the structural reproducibility for these structures is better than more complicated structures.1−7 However, it is still not possible to control their optical properties to obtain reliable and quantifiable Rayleigh scattering and SERS signals mainly because tuning subnanometer structural details and magnitude of the electromagnetic field within a single structure is highly challenging. This is a critical issue because it has been reported © 2013 American Chemical Society

Received: September 13, 2013 Revised: November 14, 2013 Published: November 20, 2013 6113

dx.doi.org/10.1021/nl4034297 | Nano Lett. 2013, 13, 6113−6121

Nano Letters

Letter

heterodimeric structures are the simplest and reproducibly synthesizable coupled system with a strong electromagnetic field within the interparticle nanogap and there is a chance to find the abnormal or interesting optical properties with these structures due to the anisotropy with respect to the direction of incident polarization. The excitation of coupled longitudinal dipolar LSP along the dimer axis is believed to be mainly responsible for the strongly enhanced field. However, it has been reported that higher multipolar LSP excitation cannot be ignored when particle size increases.15−17 Accordingly, in this study we thoroughly investigated the contributions from both the dipolar and quadrupolar LSPs that generate sm-SERS signals. Scattered light was identified on a video camera with a micrometer-level accuracy and sent to a spectrograph through a diffraction limit-sized pinhole in a conventional correlation measurement approach. In our approach, we facilitated precise measurements with an accuracy of a few nanometers by employing an AFM-based multistep tip-matching strategy, as schematically shown in Figure 1A. The multistep tip-matching strategy was systematically carried out through the following steps. First, the simultaneous visualization of the AFM cantilever and a focused laser spot on a video camera with ∼100 × 100 μm2 field of view was performed. Next, the AFM tip apex was located at the center of the focused laser spot on the video image. We then found the position of the tip apex by finding Rayleigh scattering image with the highest scattering intensity. Finally, the position of the AFM tip apex was finely adjusted at the nanometer level by observing the intensity of a representative Raman peak at 520 cm−1 of the AFM tip. To quantitatively assess the uncertainty in lateral displacement between the AFM tip apex and the laser spot at the objective center, we used the AFM and Rayleigh scattering images simultaneously.3 As a result, correlation accuracy was estimated to be ∼2.8 nm when a chosen nanostructure was illuminated with an incident light with a 250 nm diffraction limit, as shown by the light blue line and the FWHM of the blue-dotted curve in Figure 1A. Owing to such a small lateral uncertainty, high reliability in correlating a structure with an optical signal was achieved. Such a precision provides higher incident photon flux, efficiently suppressing unwanted backgrounds as well. Thus, high correlation accuracy increases both the signal-to-noise ratio and signal-to-background ratio when obtaining optical signals from a single nanostructure (or molecule). As shown in Figure 1A, the experimentally obtained normalized signal-tonoise ratio varies considerably with change in the relative displacement between the AFM tip apex and the objective center. In a typical experiment, first, plasmonic nanodumbbell-type dimers with a single Raman dye (Cy3 in this case) were prepared. Using transmission electron microscopy (TEM) measurements with 70 dimers, the diameters of two Au core nanoparticles were determined to be 40 and 50 nm, respectively, as shown in Figure 1B. The Ag shell thickness and average interparticle nanogap distance were also estimated to be ∼5 and 1.05 ± 0.51 nm, respectively. This DNAengineered dimeric nanostructure with two differently sized Au cores was designed to optimally control both the nanogap distance and the position of the Cy3 in the gap with subnanometer accuracy.1−3 Although plasmonic coupling gets significantly stronger and more quantifiable for 1 nm or smaller plasmonic gap, plasmonic coupling signal can be significantly smaller when interparticle gap is too small (e.g.,