Gap-mode enhancement on MoS₂ probed by functionalized tip-enhanced Raman spectroscopy

Tip-enhanced Raman scattering (TERS) is a spectroscopic technique that combines surface-enhanced Raman scattering (SERS) with scanning probe microscopy, e.g., atomic force microscopy (AFM) or scanning tunneling microscopy (STM), providing chemical and topographic information with a high spatial resolution.^1–4 The strongest SERS signals are typically attributed to the so-called “hot spots,” i.e., localized electromagnetic fields in nanoscale gap junctions between metallic nanoparticles.^5,6 Similarly, a sharp metallic tip can generate a hot spot when placed in proximity of a metallic surface, resulting in the electromagnetic near-field enhancement at the tip apex.^4,7,8 This gap-mode TERS has been intensively used for nanoscale imaging and single molecule detection.^9,10 Au and Ag are the conventional metallic gap-mode TERS substrates which, in some cases, may impose disadvantages such as spectral background, molecular deformation, and metal-catalyzed side reactions.^11,12 Recent developments used Au nanoparticles on dielectric substrates to circumvent limitations of the metallic substrates,^13,14 some of which cannot be easily overcome due to the metallic nature of the nanoparticles.

Recent studies demonstrated SERS enhancement on semiconducting two-dimensional (2D) materials, which depends on the molecule-substrate coupling via the chemical enhancement mechanism.^11,15 In particular, graphene has been extensively studied due to its intriguing electric and chemical properties, leading to chemical enhancement of Raman signals termed graphene-enhanced Raman scattering (GERS).^16,17 TERS on 2D materials, however, is yet to be demonstrated although the corresponding theoretical analysis has been developed.¹⁸ Here, we demonstrate gap-mode TERS on a bulk molybdenum disulfide (MoS₂) substrate, which is an emerging layered transition metal dichalcogenide material with promising optoelectronic applications.^19–21 

Previous studies used an alternative gap-mode TERS approach by functionalizing the tip with probe molecules instead of depositing them onto metallic substrates.^22,23 This approach may be used to design reference materials for the comparison of different TERS substrates based on the functionalized tip measurements under identical conditions such as molecular concentration and orientation.^24,25 Here, we compare the relative TERS enhancement on Au and MoS₂ substrates using a Au nanosphere tip functionalized with copper phthalocyanine (CuPc) molecules and investigate the efficiency of MoS₂ as a substrate for gap-mode TERS.

Fig. 1(a) shows the sketch of the functionalized TERS process. Tip-enhanced Raman spectra were obtained using an iHR550 spectrometer (Horiba) and a Synapse CCD detector. Briefly, a 532 nm excitation laser (5 W Verdi, Coherent Inc.) was focused by an aspheric lens (N.A. = 0.45) on the xyz-controlled sample stage (5 mW at the sample) in an atomic force microscope (AFM) system (M4000, Nanonics Imaging, Ltd.). A single Au-nanosphere (100 nm radius) affixed to the tip-apex of a SiO₂ cantilever (Nanocs, Inc.) was functionalized with CuPc molecules by scanning the surface of bulk MoS₂ covered by vapor-deposited CuPc (scanning area = 9 μm², thickness of CuPc < 10 nm) using the tapping-mode AFM (tip-oscillation frequency ∼30 kHz, acquisition time = 5 s). TERS measurements were performed by approaching the CuPc-functionalized tip to three different substrates: Au foil (Goodfellow, 99.95%, thickness = 0.05 mm), bulk MoS₂ (SPI Supplies, thickness ∼1 mm), and SiO₂. The average tip-substrate distance was 10 nm ± 2 nm. The average distance is controlled by the oscillating amplitude of the tip used in the tapping-mode AFM and was the same in all measurements. The multilayer MoS₂ substrate was mechanically exfoliated to expose a clean surface before the measurements. Due to the surface inhomogeneity, we performed measurements at several spatial locations on MoS₂, half of which showed strong TERS enhancement results presented below whereas the other half showed smaller enhancements.

Fig. 1(b) shows TERS spectra of CuPc obtained with a Au nanosphere tip in the vicinity of three different substrates: Au, MoS₂, and SiO₂, and in air where the spectrum was measured from the tip itself. All three substrates show the same order of magnitude signals as a result of the relatively large tip-substrate distance. The similar enhancement observed on the MoS₂ and Au substrates indicates the tip-substrate coupling and gap-mode enhancement on MoS₂ in the TERS measurements. We were able to obtain similar results with a different tip shape and size to confirm the results with a different tip geometry. It should be noted that the tip itself can have surface plasmons leading to an enhanced local field in the absence of any substrate. Therefore, TERS signals on SiO₂ are similar to the signals of the tip itself in air in the absence of any substrate, see Fig. 1(b).

In order to examine the nature of the electromagnetic gap-mode enhancement separately from the chemical effects, we analyzed the TERS signals of the strongest CuPc transitions. Fig. 2 shows the relative enhancement on Au (circles) and MoS₂ (triangles), the ratio of the CuPc Raman intensity with the tip in the vicinity of the Au and MoS₂ substrates to that of SiO₂. The low frequency modes at 581 and 668 cm⁻¹ show larger relative enhancement due to the selection rules.^26,27 A detailed discussion of the enhancements of different frequency modes and the corresponding mode assignment was previously reported.^11,28 The difference $Δ$ between the relative enhancement on Au and MoS₂ for different CuPc modes is nearly the same. That is in contrast to the previously reported different enhancement of low and high frequency modes on graphene compared to MoS₂ due to the chemical mechanism in SERS measurements.¹¹ The mode-independent difference in the relative enhancement indicates that the same electromagnetic mechanism dominates on both substrates, i.e., the enhancement due to the chemical effect plays a minor role. Negligible chemical contribution is expected because of the relatively large average tip-substrate distance of ∼10 nm during the tapping-mode AFM. The chemical interactions such as charge transfer are expected for smaller sub-nm gaps.^29,30

In order to compare our experimental results with the well-established theory, we performed calculations based on the electrostatic model where the metallic tip is modeled by a point dipole induced at z = a + d, where a is the tip curvature radius and d is the tip-substrate distance. The interaction of the dipole and its image at z = − a − d results in an enhanced local field in the nanoscale gap junction,^31–33 as shown in Fig. 3(a). The TERS enhancement is determined by calculating the ratio of the local field's magnitude to that of the far field. That ratio denoted by $κ$ is derived with a dipole polarized normal to the surface.³¹ In order to generalize κ for either polarization, we added a constant γ to the denominator of the last term and set it to be 1 when the dipole is normal to the surface, or 1/2 when it is parallel to the surface.³² Hence, the resulting expression is

where $α$ is the dipole polarizability and $β$ is the coefficient associated with the image dipole. Solving the Maxwell's equations for the electric field by applying proper boundary conditions, yields $α=a3(εt−1)/(εt+2)$ and $β=(εs−1)/(εs+1 )$⁠, where $εt$ and $εs$ are the dielectric functions of the tip and the substrate, respectively. We used previously reported values of the dielectric functions of Au and MoS₂.^34,35 Because the SERS enhancement is proportional to the fourth power of the local field,^5,6 the relative enhancement is calculated by taking the fourth power of the field enhancement on Au and MoS₂ to that on SiO₂, i.e., the relative enhancement = $|κi/κSiO2|4$⁠, where i corresponds to Au or MoS₂ substrates. The calculation of the relative enhancement of the functionalized TERS is simplified compared to the conventional TERS due to the cancellation of the illumination area effects, i.e., the same number of molecules contribute to both near field and far field on any substrate. This functionalized tip approach therefore also eliminates the far field background of the conventional TERS.

Fig. 3(b) shows the calculated relative enhancement on Au and MoS₂, for both polarizations vs. the tip-substrate distance. The decrease of the relative enhancement with the distance is slower compared to the conventional TERS enhancement.^33,36 The experimentally observed relative enhancement on Au agrees with the calculated relative enhancement, whereas the measurements on MoS₂ show a range of enhancement values due to the uneven distribution of charge carriers on the surface,^37,38 and the spatial dependence of the tip-sample interactions at the polycrystalline bulk MoS₂ edges and grain boundaries.³⁰ 

In summary, we demonstrated electromagnetic gap-mode enhancement on MoS₂ in TERS measurements with a Au nanosphere tip functionalized with CuPc molecules. This functionalized TERS approach allows for the suppression of the conventional TERS background and direct comparison of the enhancement on various substrates under highly reproducible conditions. Our results show that MoS₂ can generate large relative electromagnetic gap-mode enhancement comparable to that of conventional plasmonic substrates, which could be used as a substrate for sensing and nanoscale bio-imaging applications.

We acknowledge the support of the National Science Foundation Grants Nos. EEC-0540832 (MIRTHE ERC), PHY-1068554, PHY-1241032 (INSPIRE CREATIV), PHY-1307153, and CHE-1609608, the Office of Naval Research, and the Robert A. Welch Foundation (Award Nos. A-1261 and A-1547).