Surface-plasmon heating and the study of bio-molecules
Surface-enhanced Raman scattering (SERS) is a method in which analytes are adsorbed onto nanoscopically structured metallic surfaces and excited by strong electric fields that are induced by surface plasmon resonance. It has great potential as an analytical tool. The technique has facilitated the study of molecular structures at dilute concentrations1 and also of the dynamic behavior of single molecules.2 It thus can be employed to investigate ligand-receptor binding—of significance for drug-protein complexes—and heterogeneity in individual molecules, such as catalysis rates of a single enzyme. However, issues concerning the irreproducibility of SERS spectra, which is more pronounced at ultra-low concentrations, have not been fully addressed, and the many confounding factors have not been identified. This hinders the quantitative application of this otherwise-powerful technique. For this reason, we have undertaken studies that detail the nano-environment in which the Raman-active analytes reside.
Much of the theoretical and experimental work with SERS has concentrated on elaborating general mechanisms of the phenomenon3,4 and in optimizing effective substrates. Only recently has the SERS activity of molecules at low concentrations been studied meticulously. However, there is a dearth of detailed knowledge. For example, how would heating generated by localised surface plasmon resonance (SPR) contribute to measured SERS signals or to the large fluctuations observed in previous single-molecule SERS (SM-SERS) experiments? Is the thermal energy associated with SPR-heating comparable to the activation energy barrier between conformation minima in bio-molecules? How would heating influence the binding constant of antigen-antibody complexes? These and similar issues have been largely neglected.
Interpretations of SERS spectra in previous reports are based mainly on the assumption that the analyte molecule is rigid. This is valid for small and simple molecules such Rhodamine R6G and Crystal violet (CV), two widely-used analytes in SERS experiments. However, bio-molecules are flexible and possess many different conformational states separated by energy barriers. Thermal energy generated by SPR-heating, even under a low laser irradiation power on the order of a few hundred W/cm2, may be sufficient to cause them to denature or deviate from native structures. To ensure the correct interpretation of the SERS spectrum, the SPR heat-induced conformational changes must therefore be taken into account. In addition, SPR-heating is also likely to change the inter-particle distances within the SERS-active particle aggregates, which may partially explain the observed dependence of the fluctuation rates of SM-SERS signals on laser power.1,2
In recent work, we studied the effect of SPR- induced heating on SERS signals. To allow for inter-particle movements, a self-assembled mono-layer of SERS-active colloidal gold was prepared by depositing a drop containing an appropriate concentration of Au nano-particles (15nm) onto an electrostatically charged glass slide. Atomic force microscopy (AFM) showed that the monolayer thus formed consisted of dense but physically separated nano-particles. The electrostatic nature of the particle-glass interactions is crucial because it improves particle mobility, and so facilitated observation of SPR-induced movements. The substrate, covered with a layer of CV to be used as a probe for the local plasmon fields, was left to dry overnight in a dessicator. A 2.5mW, 633nm-excitation laser was focused onto the sample via a 20×0.40 NA objective. Enhancement by a factor >10000 can be obtained from the dried sample with virtually no sign of reduction in signal strength under continuous irradiation for 200 seconds. This suggests good thermal stability of individual Au particles, with no heat-induced deformation.
By contrast, SERS intensities derived from CV-coated Au substrates covered with a layer of aqueous solution showed a different behaviour. Not only were overall intensities reduced by 55 times with respect to those measured in a dried sample, but all peak intensities were found to decay exponentially over 200 seconds, beyond which the signals level off to a non-zero value above background noise (see Figure 1). Even with an excitation laser intensity four times lower, signal decay can still be detected. Because the loss of SERS signals was irrecoverable, the possibility of thermal-induced desorption of the adsorbate from the substrate is ruled out. A comparison of the UV-vis data subsequently obtained from the irradiated areas on the Au colloidal substrates, and results of the discrete dipole approximation (DDA) simulation, showed that changes (<10nm) in the interparticle distances within each domain of closely-spaced Au particles were responsible for the observed decay. Isolated Au particles were not perturbed.
Our simple experiment demonstrates the role of a liquid layer in SPR-heating-induced interparticle movements in closely-spaced particles, probably by lowering the activation-energy barrier. Further studies are required to determine the exact mechanisms behind the observed phenomenon. It would also be of interest to investigate whether the heating effect also contributes to the SM-SERS fluctuations commonly observed in Au/Ag aggregates. We plan to study the extent to which the binding constant of a ligand-protein complex adsorbed on a SERS substrate is affected by SPR-heating. Such a study would be imperative to optimal design of a miniaturized label-free SERS-immunoassay device.
