High-Resolution Distance Dependence Interrogation of Scanning Ion-Conductance Microscopic Tip-Enhanced Raman Spectroscopy Enabled by Two-Dimensional Molybdenum Disulfide Substrates

Abstract

Scanning ion conductance microscopy (SICM) is a powerful surface imaging tool operated in the electrolytic environment. Tip-enhanced Raman spectroscopy (TERS) can give more information in addition to the morphology provided by the SICM by utilizing the label-free Raman spectroscopy aided by the localized plasmonic enhancement from the metal-coated probes. In this study, the integration of SICM with TERS is demonstrated through employing a silver-coated plasmonic nanopipette. Leveraging a two-dimensional (2D) molybdenum disulfide (MoS₂) as a model system, the SICM-TERS enhancement factor was estimated to be ~10⁵, supported by Finite-Difference Time-Domain (FDTD) simulation. Moreover, the sub-nm distance dependence SICM-TERS study reveals the tensile stress and structural changes caused by the nanopipette. These findings illustrate the potential of SICM-TERS for providing comprehensive morphological and chemical insights in electrolytic environments, paving the way for future investigation of electrocatalytic and biological systems.

Graphical Abstract

Scanning probe microscopy (SPM) is known for its well-established capability of imaging and manipulating surfaces with unprecedented spatial resolution. In the family of SPMs, only two are suitable for imaging soft samples, such as biological systems, under controlled liquid environments: atomic force microscopy (AFM) with ultrasoft probes¹ and scanning ion-conductance microscopy (SICM).² While the external force native to AFM can be used to provide mechanical stimuli and induce biological processes such as ion-channel regulation³ and bioluminescence⁴, the conformation of small surface molecules is likely to be altered, resulting in a diminished spatial resolution for live-sample imaging.⁵ In SICM, a glass nanopipette is employed as the scanning tip, enabling non-distorted, non-destructive nanoscale imaging of the biological surfaces in liquid environments via changes in the ionic current between the nanopipette and a reference electrode. Thus far, SICM has been applied for direct observation of single active ion channels,⁶ tracking small-molecule migration,⁷ detecting single-stranded DNA,⁸ nanoscale biopsies,⁹ and patch-clamp electrophysiology measurements.¹⁰ This technique provides a unique opportunity to study interfacial phenomena under biologically relevant environments with <5 nm spatial resolution.¹¹

However, similar to other SPMs, SICM primarily provides topological information about the samples, and ionic current cannot be used to investigate the chemical specificity of surface features. For example, surface proteins appear as bright protrusions in SICM images, but specific proteins cannot be chemically identified using this technique. On the contrary, rich and detailed chemical information about the system of interest can be obtained via Raman spectroscopy, a label-free vibrational spectroscopic technique. Despite the many benefits Raman spectroscopy has for evaluating and probing biological systems (label-free, no water background in the fingerprint regions, sensitive to structural and environmental changes, etc.), low signal-to-noise (S/N) ratio and diffraction-limited resolution have hindered its application in biological research.

Surface-enhanced Raman spectroscopy (SERS) overcomes the low S/N ratio challenge by amplifying and localizing electromagnetic fields of plasmonic nanostructures. The scanning probes are naturally in the nanometer regime and can provide the needed electromagnetic enhancement if fabricated with plasmonic metals. This joint SPM and SERS approach is termed tip-enhanced Raman spectroscopy (TERS), which combines the atomic spatial resolution of SPMs with the single-molecule chemical sensitivities of SERS.¹² Under ultrahigh vacuum (UHV) conditions, UHV-TERS has demonstrated Ångström-scale spatial resolution^13-16 and has gained new insights on surface molecule orientation¹³ and molecule-molecule interaction¹⁷. Recently, TERS has been extended into the solution phase with AFM^18,19 and electrochemical AFM/STM.^20-24 If this unprecedented spatial resolution can be realized in the liquid environment, it is an ideal identification and spatial mapping technique in three dimensions of soft materials.

Herein, SICM-TERS on a two-dimensional (2D) molybdenum disulfide (MoS₂) was demonstrated in an ionic environment using a silver-coated nanopipette. Inspired by the previously demonstrated AFM-TERS and SICM setups,^25,26 a custom-built SICM is mounted on an inverted optical microscope to enable Raman spectroscopic detection at the SICM probe apex in a back-scattering geometry (Figure 1). A three-dimensional inchworm motorized stage controls the nanopipette position and aligns it to the laser focus. A piezo-flexure nanostage is used to achieve the precise and fast response required for SICM imaging and feedback.

Figure 1.

Schematic (left) and Photograph (right) of the SICM-TERS setup. The monolayer MoS₂ is placed at the bottom of a petri dish filled with 0.1 M KCl solution. A silver-coated nanopipette is immersed into the solution from the top. A 532 nm continuous laser is focused from the bottom onto the tip apex through a 40× long working distance objective lens. The TERS signal is collected with the same objective lens in a back-scattering geometry.

A TERS enhancement factor of ~10⁵ has been observed in the non-gap-mode geometry when the tip is in close proximity to the MoS₂ surface. Additionally, Raman mode shifting and splitting have also been observed when the nanopipette moved toward the 2D MoS₂ film and eventually inserted into the film. The highly structurally sensitive TERS signals enable the characterization of SICM working distance and the interaction between the probe and the surface at the sub-nm scale.

In a TERS experiment, the plasmonic properties of the SPM probes are important. For the presented SICM-TERS, the nanopipette was pulled by a CO₂ laser puller (P-2000, Sutter Instrument) using a quartz capillary with an outer diameter of 1.0 mm and inner diameter of 0.7 mm (Sutter Instrument). After the nanopipette was made, a high vacuum sputter coater (VTC-16-3HD, MTI corporation, modified with a Varian TPS-compact turbo pumping station, base pressure is ~5 × 10⁻⁵ torr) was used to sputter the pipette with 25 nm of silver on one side in an Argon environment at 4.5 × 10⁻³ torr held at 120 °C. The SEM images were acquired with a Phenom pharos G2 Desktop FEG-SEM (ThermoFisher Scientific). Figures 2a and b show representative scanning electron microscopy (SEM) cross-section images of a silver-sputtered SICM-TERS nanopipette at 45° and 90° angles, respectively. The opening of the nanopipette tip is around 50 nm in radius, estimated by the SEM images. The SEM image represents a direct view of the nanopipette, but it needs to be broken for mounting. Since the nanopipette opening directly correlates with the total resistance,²⁷ a current-voltage (I-V) curve has been performed for the conductance measurement (Figure S1 in Supplementary Information). The calculation based on an established model^28,29 shows the inner pipette radius ~45 nm, which agrees with the SEM image estimation.

Figure 2.

(a) and (b) SEM images of the nanopipette tip at 45° (a) and 90° (b) cross-section angles. (c) FDTD simulation of the electric field power density around the nanopipette: The glass part of the pipette is shown in light gray, while the silver coating is shown in dark gray. On the end of the pipette, there are two hemispheres to mimic the silver extension part, as shown in panel (a).

The custom-built SICM-TERS performed in this study uses transparent and insulating samples mounted on the inverted microscope. The longitudinal mode from a tightly focused linearly polarized laser excitation enables the excitation of the plasmon at the tip apex.^30-32 Therefore, neither gap-mode plasmon couplings nor quantum tunneling is expected. The near-field responses of the silver-coated plasmonic nanopipettes can be simulated with computational classical electrodynamics.^33,34 The simulated nanopipette geometry is based on the parameters extracted from the SEM images. The computation region is filled with water (refractive index 1.31) to account for the plasmon redshifts in liquids.³⁵ Furthermore, since the majority of the enhanced Raman signal was attributed to thehot spot with the strongest enhancement,³⁶ two hemispheres of silver were added on the tip apex to mimic the Ag protrusions seen in the SEM image (Figure 2a). The simulated electric field power density (${\mid E\mid }^2$) show that the electric field is primarily enhanced near the silver hemispheres (Figure 2c). The enhancement of ${\mid E\mid }^2$ is around 10³ by comparing the tip apex part with the far away region, which gives ~10⁶ TERS enhancement factor considering the Raman signal enhancement is proportional to ${\mid E\mid }^4$.³⁷

Monolayer MoS₂, a versatile material for electronics, photonics, and electrocatalysis,^38-40 is used as a proof-of-concept system for SICM-TERS. The bandgap of the 2D MoS₂ matches the green spectral region; therefore, the material has a large resonance Raman cross-section when excited with a green laser. The growth of monolayer MoS₂ films on sapphire substrates by the conventional chemical vapor deposition method and the PMMA-transfer method were described in previous studies.^39,41 Figure 3a illustrates the two major MoS₂ Raman modes: the ${\mathrm{E}}_{2\mathrm{g}}^1$ (385 cm⁻¹) and ${\mathrm{A}}_{1\mathrm{g}}$ (402 cm⁻¹), representing the in-plane vibration of both molybdenum and sulfur atoms and the out-of-plane vibration of sulfur atoms, respectively.^42-44 The normal Raman spectrum of MoS₂ was measured (blue solid line in Figure 3b) when the pipette tip was far from the surface. As the Ag-coated nanopipette approaches the MoS₂ sample, the Raman spectrum is enhanced, as shown by the red solid line in Figure 3b.

Figure 3.

(a) Illustration of two Raman modes of MoS₂ monolayer (the in-plane vibration $E_{2g}^1$ and the out-of-plane motion $A_{1g}$). (b) The normal Raman (blue solid line) and SICM-TERS Raman (red solid line) spectra. (c) The Raman modes intensity as a function of the distance between the tip and the MoS₂ monolayer. The zero position of Z is set to the distance at which the intensity change begins.

The Raman signal enhancement happens at such a small distance that the changes show a sudden jump even with the 0.1 nm step size of the sample movement (Figure 3b). Moreover, the in- and out-of-plane Raman modes behave slightly differently. The intensity enhancement of the out-of-plane mode (${\mathrm{A}}_{1\mathrm{g}}$) appears to be higher than that of the in-plane mode (${\mathrm{E}}_{2\mathrm{g}}^1$), which can be attributed to the electric field around the silver-coated tip that is vertically closer to the conductor surface, as it satisfies the boundary conditions of Maxwell’s equations. By the surface selection rule of TERS,⁴⁵, the out-of-plane modes align better with the plasmonic dipole at the tip-sample junction. Therefore, the ${\mathrm{A}}_{1\mathrm{g}}$ is preferably enhanced over the in-plane ${\mathrm{E}}_{2\mathrm{g}}^1$ mode. The same experiment was performed on an uncoated glass nanopipette, showing no Raman signal enhancement (Figure S4 in Supplementary Information).

Furthermore, the rich chemical information provided by Raman spectroscopy can be exploited to elucidate the effects that cause structural changes. Specifically, the two MoS₂ Raman modes are sensitive to local strain introduced onto the materials, resulting in mode shifting and splitting.^46,47 Figure 4a shows a waterfall plot of the SICM-TERS spectra when the surface approaches the tip with a step-size of 0.1 nm. The distance-dependent SICM-TERS data show a clear signal enhancement as the tip-sample gap decreases. Moreover, a small Raman mode shift was observed in the first 0.5 nm (Figure 4b). Based on previously reported strain-induced shift values of MoS₂ Raman modes,⁴⁷ the strain caused by the nanopipette in the proximity of the materials is estimated to be ~0.2% - 0.4%, observing the shifts of ${\mathrm{E}}_{2\mathrm{g}}^1$ and ${\mathrm{A}}_{1\mathrm{g}}$ modes, respectively. Even though SICM was known to cause less deformation than AFM,⁵ the MoS₂ SICM-TERS spectra indicate the Ag nanopipette still introduces a small amount of pressure within the extent of the plasmonic near-field.

Figure 4.

(a) SICM-TERS spectral evaluation of the MoS₂ monolayer while the nanopipette tip approaches the sample. The step size between the TERS spectra is 0.1 nm (excluding the bottom black far-field Raman spectrum). (b) MoS₂ Raman peak shifts as a function of tip-sample distance extracted from (a). Inset: optical microscope image of MoS₂ film after the experiment.

However, when the sample moved even closer toward the tip (−0.6 nm to −1.1 nm), the splitting of the ${\mathrm{E}}_{2\mathrm{g}}^1$ peak has also been observed besides the intensity enhancement. This result indicates the lifting of the degenerated $E$ mode due to an increase of the tensile stress (>0.8%)^48-50 of the film, which is also observed with the AFM-TERS spectra of MoS₂.^51,52 At this short distance, the plasmonic nanopipette introduces a non-negligible strain to the materials. A direct comparison with force-controlled EC-AFM-TERS would be ideal to further elucidate the strain effect of MoS2 under electrolytic environments.

Possible tip insertion was observed by the SICM-TERS spectra as the tip-sample gap further shrinks below −1.2 nm. The ${\mathrm{E}}_{2\mathrm{g}}^1$ mode further splits, and the signal of the ${\mathrm{A}}_{1\mathrm{g}}$ decreases. The reduction of ${\mathrm{A}}_{1\mathrm{g}}$ suggests that the tip has been inserted into the MoS_2, and the most enhancing region has moved over passed materials. The silver island film on the shaft of the Ag-coated nanopipette enabled the detection of the disrupted materials in a tip-SERS fashion.⁵³ Moreover, the further split of ${\mathrm{E}}_{2\mathrm{g}}^1$ modes resemble the edge structure of MoS₂, suggesting the disruption of the materials caused by the down-moving nanopipette. The process of the tip insertion lasts less than 1 nm, according to Figure 4a, which is consistent with the thickness of monolayer MoS₂ (around 0.7 nm).⁵⁴ An optical micrograph shown in the inset of Figure 4b clearly shows a hole in the MoS₂ sample after the experiment.

The SICM-TERS enhancement factor ($\text{EF}$) can be evaluated based on the spectra shown in Figure 3 and 4, which have an apparent signal enhancement $\text{EF}$ of 1.7 and 7.0, respectively. To estimate the SICM-TERS enhancement factor (${\text{EF}}_{\mathrm{o}}$), the tip and the laser spot sizes must be considered. The laser beam size was measured to be about 1.7 μm of full width at half maxima (FWHM) with the 40× objective lens (NA 0.6). With the pipette opening size of 100 nm in diameter and the coating tip about 25 nm in diameter, the $\text{EF}$ has a relation with the SICM-TERS ${\text{EF}}_0$ as shown below:

$$EF=(1-P)+E{F}_0\times P$$ (1)

where $P$ is the portion of the tip-enhanced laser compared to the whole laser beam, as calculated in the Supplementary Information. Using Equation 1, the SICM-TERS ${\text{EF}}_0$ is estimated to be ~10⁴ to ~10⁵, which suggests that the plasmonic resonances were the dominating enhancement mechanism rather than the lightning rod effect. This estimation does not take into account atomistic features that might be present on the tip apex, and it is likely to be underestimated as the fine feature can further contribute to the electromagnetic field confinement. In principle, SICM-TERS resolution could be defined by the atomistic plasmonic nanostructures^55,56 and therefore, it could exceed the SICM imaging resolution that depends on the nanopipette opening size. Moreover, many other factors might affect the enhanced electromagnetic field, such as absorbed ions between the tip, and sample and electric double layers if a third electrode is connected to the sample. SICM-TERS can potentially elucidate these effects, complementing the observations from other EC-TERS measurements.

In conclusion, the capability of TERS is extended to a liquid environment within a SICM framework, marking a pioneering endeavor in the field. The experimentally observed SICM-TERS enhancement factor (~10⁵) is evaluated by leveraging a 2D MoS₂ sample as a proof-of-concept model and supported by FDTD simulations. This simulation, even limited to classical theory, is sufficient to rule out other non-plasmonic mechanisms, but a more sophisticated high-level theoretical approach is needed as the liquid-phase TERS field progresses. Moreover, the strain-induced MoS₂ vibrational mode shifting and splitting reveals that the optimal working distance for SICM-TERS is between 0.4 – 1 nm above the surface without causing significant material deformation. This working distance of TERS is consistent with the recently reported values for UHV-TERS,⁵⁷ and the short-range distance dependence of SERS.⁵⁸ The SICM-TERS spectral resolution achieved highlights its ability to provide detailed morphological and crucial chemical insights into the materials and methodology. Overall, this work not only underscores the synergistic potential of combining SICM with TERS but also sets the stage for its application in soft materials, electrocatalysis, and biological research exploiting solid-liquid interfaces.