Single-molecule (SM) surface-enhanced Raman spectroscopy (SERS) and tip-enhanced Raman spectroscopy (TERS) have emerged as analytical techniques for characterizing molecular systems in nanoscale environments. SERS and TERS use plasmonically enhanced Raman scattering to characterize the chemical information on single molecules. Additionally, TERS can image single molecules with subnanometer spatial resolution. In this review, we cover the development and history of SERS and TERS, including the concept of SERS hot spots and the plasmonic nanostructures necessary for SM detection, the past and current methodologies for verifying SMSERS, and investigations into understanding the signal heterogeneities observed with SMSERS. Moving on to TERS, we cover tip fabrication and the physical origins of the subnanometer spatial resolution. Then, we highlight recent advances of SMSERS and TERS in fields such as electrochemistry, catalysis, and SM electronics, which all benefit from the vibrational characterization of single molecules. SMSERS and TERS provide new insights on molecular behavior that would otherwise be obscured in an ensemble-averaged measurement.

Surface- and tip-enhanced Raman and LSPR spectroscopies have developed over the past 15 years as unique tools for uncovering the properties of single particles and single molecules that are unobservable in ensemble measurements. Measurements of individual events provide insight into the distribution of molecular properties that are averaged over in ensemble experiments. Raman and LSPR spectroscopy can provide detailed information on the identity of molecular species and changes in the local environment, respectively. In this review a detailed discussion is presented on single-molecule and single-particle Raman and LSPR spectroscopy focusing on the major developments in the fields and applications of the techniques.

Verification of single-molecule (SM) detection for surface-enhanced Raman spectroscopy (SERS) requires the use of two analytes via either the bianalyte or isotopologue approach. For both approaches, the preferential observation of the individual analytes over a combination of both analytes is used to conclude that SM detection has been achieved. Isotopologues are preferred because they have identical surface binding affinities and Raman cross sections, whereas bianalyte pairs typically do not. We conducted multianalyte SERS studies to investigate the limitations of the bianalyte approach. The bianalyte partners, Rhodamine 6G (R6G-d 0) and crystal violet (CV-d 0), were directly compared, while SM detection was verified (or disproved) using their corresponding isotopologues (R6G-d 4, CV-d 12). We found that the significant difference in counts between R6G and CV can provide misleading evidence for SMSERS. We then rationalized these results using a joint Poisson-binomial model with unequal detection probabilities and adjusted the relative concentrations of R6G and CV to achieve a comparable distribution of SMSERS counts. Using this information, we outlined the necessary considerations, such as accounting for the differences in molecular properties, for reliable SMSERS proofs. Moreover, we showed that multianalyte experiments at the SM level are achievable, opening the opportunity for new types of SM studies.

We provide conclusive proof of single molecule (SM) detection by surface-enhanced Raman spectroscopy (SERS) for discrete Ag triangular nanopyramids prepared via nanosphere lithography (NSL). While the observation of SMSERS has been well-demonstrated using various chemically synthesized nanoparticle substrates, they have a high degree of polydispersity in shape, size, and aggregation state resulting in an interest to develop more reproducible and uniform nanoparticles. Here triangular-based nanopyramids were characterized by scanning electron microscopy to confirm their geometry and interparticle spacing. Then the isotopologue approach with Rhodamine 6G was used to conclusively prove SM sensitivity for the individual nanoparticles. NSL’s straightforward, simple fabrication procedure produces large active arrays. Furthermore, the tunable localized surface plasmon resonance makes NSL prepared substrates effective for the detection of resonant molecules by SMSERS.

Plasmonic near fields, wherein light is magnified and focused within nanoscale volumes, are utilized in a broad array of technologies including optoelectronics, catalysis, and sensing. Within these nanoscale cavities, increases in temperature are expected and indeed have been demonstrated. Heat generation can be beneficial or detrimental for a given system or technique, but in either case it is useful to have knowledge of local temperatures. Surface-enhanced Raman spectroscopy (SERS), potentially down to the limit of single-molecule (SM) detection, has been suggested as a viable route for measuring nanoscale temperatures through simultaneous collection of Stokes and anti-Stokes SER scattering, as the ratio of their intensities is related to the Boltzmann distribution. We have rigorously verified SM detection in anti-Stokes SERS of rhodamine 6G on aggregated Ag nanoparticles using the isotopologue method. We observe a broad distribution in the ratio of anti-Stokes and Stokes signal intensities among SM events. An equivalent distribution in high-coverage, single-aggregate SERS suggests that the observed variance is not a SM phenomenon. We find that the variance is instead caused by a combination of local heating differences among hot spots as well as variations in the near-field strength as a function of frequency, effectively causing nonequivalent enhancement factors (EFs) for anti-Stokes and Stokes scattering. Additionally, we demonstrate that dark-field scattering cannot account for the frequency dependence of the optical near field. Finite-difference time-domain simulations for nanoparticle aggregates predict a significant wavelength dependence to the ratio of anti-Stokes/Stokes EFs, confirming that the observed variation in this ratio has strong nonthermal contributions. Finally, we outline the considerations that must be addressed in order to accurately evaluate local temperatures using SERS.