Anomalous 2D Band Decay in TERS Reveals Liquid Meniscus Formation

Scientists are pushing the boundaries of nanoscale material analysis with Tip-Enhanced Raman Spectroscopy (TERS), a powerful technique that bypasses conventional limits of optical resolution. André G. Pereira (Universidade Federal de Minas Gerais and FabNS), Raul Corrêa, and Bianca Carneiro, alongside colleagues including Thiago L. Vasconcelos (Instituto Nacional de Metrologia, Qualidade e Tecnologia) and Vitor Monken, have uncovered a surprising phenomenon during TERS analysis of graphene in ambient conditions , an unusual relationship between tip-sample distance and signal intensity. Their research, detailing an anomalous decay profile, suggests the formation of liquid menisci due to capillary forces subtly alters the effective distance between the TERS tip and the graphene surface, offering crucial insight into the complexities of nanoscale measurements and paving the way for more accurate 2D material characterisation.

The team achieved high TERS efficiency and accuracy in both tip-approach and tip-retract procedures, enabling the detection of subtle effects previously obscured by limitations in spatial resolution. This work establishes that the observed anomalies can be accurately described by incorporating a deformation term into the effective tip-sample distance, a deformation attributed to the formation of a liquid meniscus driven by capillary forces.

The study unveils a significant discrepancy between experimental TERS data and current theoretical predictions, specifically concerning the near-field intensity decay as the tip retracts from a graphene surface. Researchers performed TERS experiments on graphene mechanically exfoliated from a National Graphite source, utilising an AFM-TERS setup equipped with an oil immersion objective lens (NA = 1.4) and a radially polarized HeNe laser (λ = 632.8nm) with power levels of 200 or 700 μW. Data processing was carried out using the FabNS PortoFlow software (v1.21), allowing for precise analysis of the Raman signal. The team focused on the prominent 2D band of graphene, observing its behaviour at varying tip-sample distances, and meticulously normalised the TERS near-field intensity to highlight the anomalous decay.
Experiments involved precisely controlling the tip-sample separation using a quartz tuning fork in constant drive and shear force configuration, monitoring frequency shifts to maintain a setpoint of 0.9Hz. The researchers conducted 54 tip-retraction experiments with different Plasmon-Tunable Tip Pyramids (PTTP) and sample types across two independent laboratories, with air humidity maintained between 40% and 55%. A remarkable 44 of these experiments exhibited a clear discontinuity in the expected decay profile, occurring within the 8-18nm range, indicating a systematic deviation from the standard TERS model. Detailed statistical analysis, available in supplementary materials, confirms the robustness of these findings and supports the hypothesis of a meniscus-induced deformation.

This breakthrough reveals that the standard TERS model, which assumes a simple geometric relationship between tip-sample distance and signal intensity, is insufficient to accurately describe the observed behaviour in ambient conditions. The research establishes the necessity of incorporating the effect of a liquid meniscus, a thin film of water formed by capillary action, to achieve a satisfactory fit to the experimental data. By analysing the tuning fork oscillation amplitude alongside the TERS signal, the team demonstrated a clear correlation between the meniscus formation and the observed anomalies, providing compelling evidence for this novel explanation. This work pioneers the addition of a deformation term to existing TERS intensity theoretical models, accurately describing experimental data and rationalizing observations as resulting from a liquid meniscus formed by capillary forces. Graphene flakes, mechanically exfoliated from National Graphite, were deposited onto Knittel coverslips using a pickup transfer method for the TERS experiments conducted in ambient conditions.

Although airborne contaminants render the graphene hydrophobic, the study acknowledges the persistence of a layered water molecule structure on the graphene surface, as reported in existing literature. Researchers anticipated meniscus formation when the tip approached the sample, building on prior work demonstrating this effect at close proximity. Measurements were performed using a custom-built AFM-TERS system, equipped with an oil immersion objective lens boasting a numerical aperture of 1.4. The system harnessed a radially polarized HeNe laser (λ = 632.8nm) at powers of 200 or 700 μW for excitation, and the back-scattered signal was directed to an Andor Shamrock SR-303i spectrometer for analysis.

Plasmon-Tunable Tip Pyramids (PTTP) were mounted on a quartz tuning fork, facilitating near-field enhancement. Data processing utilized FabNS’s PortoFlow software (v1.21) to extract meaningful information from the complex spectra. The team engineered tip-approach cycles until a frequency shift of 0.9Hz was reached, initiating Raman signal collection with 700 μW laser power. Focusing on the prominent 2D band around 2650cm−1, due to its higher intensity compared to the G band at ∼1580cm−1, the study normalized the TERS near-field intensity, INF(z) = ITERS(z)−ITERS(z = 30), where ITERS(z = 30) represents the far-field intensity.

A comparison of the experimental data with both undeformed and deformed TERS models revealed the inadequacy of the former, highlighting the necessity of the deformation correction for accurate fitting. A comprehensive dataset comprising 54 tip-retraction experiments, conducted across two laboratories with humidity between 40% and 55%, demonstrated a significant discrepancy between the undeformed model and experimental curves in 44 cases, typically marked by a discontinuity between z = 8 −18nm. Experiments revealed that the observed anomaly can be accurately described by incorporating an additional deformation term into the effective tip-sample distance, suggesting the presence of a liquid meniscus formed by capillary forces. The team measured the Raman spectrum of graphene, focusing on the prominent 2D band observed around 2650cm−1, at varying tip-sample distances during retraction experiments.

At a closest distance of z0 = 5nm, the Raman signal was collected using a HeNe laser with a power of 700 μW, and the resulting spectra exhibited a clear 2D band peak. Data processing, conducted using PortoFlow software (v1.21), allowed for the normalization of the TERS near-field intensity, defined as INF(z) = ITERS(z)−ITERS(z = 30), where ITERS(z = 30) represents the far-field intensity. Measurements confirm that the standard, undeformed TERS model failed to accurately fit the experimental data, highlighting the need for a correction factor to account for the observed discrepancies. Specifically, the normalized TERS near-field intensity decay, INF(z), showed a significant deviation from the predicted curve, with a discontinuity appearing in the range of z = 8 −18nm in 44 out of 54 tip-retraction experiments conducted across two laboratories with humidity levels between 40% and 55%.

This discrepancy, detailed in supplementary material, prompted the researchers to propose a model incorporating the effect of a liquid meniscus on the graphene surface. The team’s analysis of tuning fork oscillation amplitude, monitored in constant drive mode, further supports the hypothesis of meniscus formation, as variations in amplitude reflect changes in the tip-sample interaction. Tests prove that incorporating this meniscus effect into the TERS model allows for accurate fitting of the experimental data, demonstrating the validity of the proposed explanation. The frequency shift during tip approach was measured at 0.9Hz, establishing the initial tip-sample interaction point. The. While no Raman peaks associated with water were detected, potentially due to the nanoscale volume of the meniscus, the authors acknowledge this as a point for further investigation. The authors note a limitation in definitively confirming the true tip-sample distance, as the meniscus’s influence on tip position remains unclear. Future research directions include conducting experiments in a vacuum environment to further elucidate these findings and potentially detect Raman signals from trace surface contaminants, as observed in similar studies with molybdenum diselenide. This work contributes to a more nuanced understanding of TERS, particularly in ambient conditions, and highlights the importance of considering capillary forces and meniscus formation in accurate nanoscale optical measurements.