pubs.acs.org/NanoLett
Charge Transfer Enhancement in the SERS of a Single Molecule Won-Hwa Park, and Zee Hwan Kim* Department of Chemistry and BK21 Division of Chemistry, Korea University, Seoul 136-701, Korea ABSTRACT We measured the surface-enhanced Raman scattering (SERS) of individual gold nanoparticle-4-aminobenzenethiol (ABT)gold film junctions to investigate the charge-transfer (CT) enhancement of the SERS signals. Despite the mild electromagnetic field enhancement (∼105) and high surface density of the ABT-molecules (∼240 molecules/hotspot) at the junctions, we observed the clear spectral and temporal signatures of CT-enhanced single-molecule SERS (SM-SERS). The result reveals that only a small fraction of the molecules at the junction has a significant CT-enhancement of 101∼103, whereas the rest of the molecules are nearly CTinactive. Furthermore, the result also proves that overall (charge-transfer and electromagnetic) enhancement of 106∼108 is sufficient to observe the SM-SERS of an electronically off-resonant molecule, which disproves the widespread belief that a minimum enhancement of ∼1014 is required for SM-SERS. KEYWORDS Surface plasmon, single-molecule, charge-transfer, surface-enhanced Raman scattering
S
urface-enhanced Raman scattering (SERS)1 re-emerged as one of the most promising chemical and biochemical assay techniques soon after the reports of the single-molecule SERS (SM-SERS) of electronically resonant dyes.2,3 In these reports and the many others that followed, enhancement factors (EFs) as large as 1014 were claimed. The enhancement was attributed to the strong electromagnetic (EM) field enhancement that occurs between closely spaced nanoparticles, and the enhancement associated with electronic resonances of dye molecules.4,5 Although these pioneering reports opened up new possibilities for the vibrational spectroscopy of individual molecules, they also gave the general impression that SM-SERS is possible only for electronically resonant dyes under an extreme local field. Recently, however, a few researchers6-9 have argued that the EF required for SM-SERS and the role of EM-enhancement may have been significantly overestimated. In particular, Etchegoin and co-workers7,8 employed bianalyte technique to show that the minimum EF for SM-SERS must be significantly smaller than the originally proposed values. Ward et al.,6 Ichimura et al.,10 and Fromm et al.11 observed fluctuations in the SERS and tip-enhanced Raman scattering (TERS) spectra of nonresonant organic molecules, which were interpreted as SM-SERS. Zuloaga et al.9 reported that the classical electrodynamic simulation tends to overestimate the local-field at the narrow metallic gap. These reports signify that the SM-SERS of an electronically off-resonant molecule may be possible using only moderate local field intensities. In addition, they also suggest that chemical mechanism may play an important role in SM-SERS.
The photoinduced metal-molecule charge-transfer (hereafter called charge-transfer for short) is believed to provide an extra chemical enhancement in SERS,12 yet its microscopic mechanism and relative importance are still hotly debated. For example, the estimates of the CT-enhancement factors11,13-16 range from a factor of 1013,14 to as large as 107-108.11,15 In addition to the vibronic charge-transfer coupling model12 that describes the basic selection rule and energetic requirements for CT process, microscopic models that emphasize the role of surface defects17-19 and tunneling in the metal-molecule-metal bridge20-22 have been proposed. The important questions regarding the CT-enhancement include the following: (1) How to quantify CT-enhancement factors?11,16 (2) Which of the vibrational bands are CT-active?23,24 and (3) What are the influences of molecular structures and metal-molecule contacts on the CT-enhancement?20-22,24 Despite extensive experimental and theoretical efforts, many of these questions remain unanswered thus far. The difficulty mostly originates from the problem of disentangling the EM- and CT-contributions to the SERS signals from nanostructures with many uncontrollable variables, such as the precise geometries of the junctions and the molecular orientations. Herein, we investigate the contribution of the CT-enhancement in the SERS signal of a single 4-aminobenzenethiol (ABT). The main purpose of the current work is two-fold. First, we present spectral and temporal proofs that the SM-SERS of an electronically off-resonant molecule can be achieved with a moderate signal enhancement (both EM and CT) of 106∼108, which disproves a widespread belief that the SM-SERS requires EFs larger than 1014. Second, we analyze the mode-specific on/off blinking ratios in the SMSERS spectra to quantify the CT-enhancement factors of each vibrational mode of a single ABT and to examine the fraction of ABTs that actually contributes to the observed CT-
* To whom correspondence should be addressed. E-mail: (Z.H.K.) zhkim@ korea.ac.kr. Telephone: +82-2-3290-3142. Fax: +82-2-3290-3121. Received for review: 06/7/2010 Published on Web: 09/21/2010 © 2010 American Chemical Society
4040
DOI: 10.1021/nl102026p | Nano Lett. 2010, 10, 4040–4048
FIGURE 1. (a) Schematic diagram of the sample and measurement scheme. The AFM tip (not shown) scans above the sample, and the Raman excitation (λ ) 633 nm) and detection are carried out through the thin-film side of the sample. (b) A representative SERS spectrum (integration time of 50 s) of ABT obtained from an Au NP-ABT-TF junction, along with the peak assignment.21,31 (c) AFM topography. (d,e) Confocal SERS images of 7a (d) and 9b (e) vibrational modes of ABT. (f) The |E/E0|4 distribution (E ) local field, E0 ) incident field) near a Au NP (diameter of 200 nm)-TF (thickness of 10 nm) junction simulated by 3D-FDTD method. The upper and lower panels of (f) show |E/E0|4-distributions sampled along the planes horizontal (xy) and vertical (xz) to the sample plane (xy), respectively.
enhancement. Unlike in the majority of reports on SM-SERS2,3,8,10,16,25,26 in which random aggregates of nanoparticles serve as hotspots, we employ well-defined Au nanoparticle (NP)-molecule-Au thin film (TF) junctions27-30 to better distinguish the EM- and CT-contributions in the SERS signals. Figure 1a sketches the structure of the sample and the measurement scheme. The AuNP (diameter of 200 nm)-ABT self-assembled monolayer-AuTF (thickness of 10 nm) junction structures are fabricated on a glass substrate by the method described in Supporting Information-A. The SERS signals were measured with an epi-confocal Raman/atomic force microscope (AFM) equipped with a high-NA objective lens (Plan Apochromat, 60×, 1.45 NA, Olympus), a CCD-camera based spectrometer (500im, focal length ) 50 cm, Chromex, and DU401, Andor Tech; overall spectral resolution is ∼2 cm-1), and a HeNe laser as the excitation light source (λex ) 633 nm, Melles Griot,
