Nanoparticle Spectroscopy: Dipole Coupling in Two-Dimensional Arrays of L-Shaped Silver Nanoparticles Jiha Sung, Erin M. Hicks, Richard P. Van Duyne, and Kenneth G. Spears* Chemistry Department, Northwestern UniVersity, EVanston, Illinois 60208-3113 ReceiVed: March 19, 2007; In Final Form: May 10, 2007 The plasmon resonance was measured for two-dimensional arrays of L-shaped Ag nanoparticles fabricated by electron beam lithography. A variety of particle sizes were studied with nominal total edge lengths of ∼150 nm, 63 nm arm widths, and 30 nm height. The single nanoparticle localized surface plasmon resonance (LSPR) of the L particles had two polarized components, which independently coupled in the arrays to create plasmon resonances for the array. The arrays had peak resonance locations and bandwidths that were dependent on grid spacing and particle number in the grid. The array plasmon resonance had a minimum bandwidth of 700-800 cm -1 at a grid spacing ∼75 nm smaller than the grid having the largest red shift of the plasmon resonance. This bandwidth is about half the single nanoparticle resonance bandwidth. For arrays with small numbers of nanoparticles, the resonant wavelength and bandwidth had large deviations from the semi-infinite arrays but approached those results as the number of nanoparticles increased to 25 particles on an edge, which defines the range of effective dipole coupling for a 400 nm grid spacing. This observation is consistent with optical changes observed by scanning across a 300 × 300 μm 2 pad. A solvent effect on these arrays demonstrated a red shift with similar bandwidth effects and some small grating-induced features due to waveguide effects. I. Introduction Nanomaterials are of current interest in a wide variety of fields from medicine to microelectronics. The optical properties of nanoscale-fabricated noble metal nanoparticles have drawn particular interest both experimentally and theoretically because of their impact in technological applications such as bio/ chemosensors, 1-4 optical filters, 5,6 plasmonic waveguides, 7-10 and substrates for surface-enhanced spectroscopies. 11-13 The property behind all of these applications is the localized surface plasmon resonance (LSPR), which is a collective oscillation of the conduction electrons that occurs when light impinges on a nanoparticle at a specific wavelength. The resonance peak position and shape of the LSPR is governed by the nanoparticle shape, size, composition and dielectric environment. The LSPR resonance creates enhanced light scattering, absorption, and local enhancement of the electromagnetic field. For arrays of nano- particles the LSPR optical properties become dependent on the array interactions to create a modified plasmon resonance that is characteristic of the whole array; therefore, one needs to understand nanoparticle coupling versus grid spacing, effects of shape and size, and the nanoparticle dielectric environment. To effectively study a wide variety of arrays, a fabrication method is needed that has precise, user-defined placement of the nanoparticles. Current fabrication techniques include natural lithography, such as nanosphere lithography (NSL) 14,15 and photolithography, 16,17 and direct-write methods, such as electron beam lithography (EBL) 18,19 or dip-pen nanolithography. 20 Each type has its own advantages and disadvantages, but for experi- ments involving the precise placement of nanoparticles on a surface, a direct-write method is an excellent tool. Specifically, EBL can be used to create nanoparticles with different shapes, sizes, spacing, and orientation in two-dimensional (2D) arrays. A partial review of nanoparticle fabrication by EBL is given by Canfield et al. 21 Optical properties of one- or two-dimensional noble metal nanoparticle arrays have been studied experimentally 18,22-27 and theoretically by several groups. Short-range coupling effects in EBL fabricated hexagonal and square arrays of triangular and circular Au and Ag nanoparticles were studied optically by Van Duyne and co-workers. 18 They observed a shift of the LSPR, depending on lattice spacing, and a related theoretical work explained these effects in terms of radiative dipolar coupling between the nanoparticles and retardation effects. 18,28 For one- dimensional (1D) chains, they were able to experimentally find 25 a narrower shoulder on the plasmon resonance due to long- range dipole interactions along the chain, which was predicted by Schatz and co-workers. 29,30 Others have looked in more detail at two-dimensional arrays with large grid spacing. Aussenegg and co-workers have studied extinction spectra of two- dimensional Au nanoparticle arrays with a variety of nanopar- ticle geometries such as cylinders, 23,24,26 nanorods, 23,24 and nanowire gratings. 27 In one study of Au arrays, 26 this group monitored the resonance peak with white light spectroscopy and the plasmon lifetime with time-resolved collinear autocorrelation measurements. They observed a red shift with increasing grid spacing and a dramatic increase in the plasmon damping at a critical grating constant. Their physical explanation for these effects relied on the models of Meier and Wokaun, 31,32 and the experiments showed the importance of coherent scattering into a substrate as the means to enhance dipole coupling and thereby increase the radiative component of bandwidth. In a related recent work, 24 they showed a predicted grating-induced reso- nance 31,32 in an array of Au nanorods having large transition dipoles. Others have looked into different shapes and lattice structures as well. 32-34 10368 J. Phys. Chem. C 2007, 111, 10368-10376 10.1021/jp0721853 CCC: $37.00 © 2007 American Chemical Society Published on Web 06/26/2007