Angularly Resolved Tip-Enhanced Raman Spectroscopy

Abstract

Despite intensive research in tip-enhanced Raman spectroscopy (TERS), the angular distribution of Raman scattering in the TERS gap remains experimentally unreported leaving its relevance to the TERS signal formation to be seldomly discussed. Here, we investigate the angular distribution of the tip-enhanced Raman signal in the Fourier plane using a model system composed of flat-lying cobalt (II) hexadecafluoro-phthalocyanine (CoPcF₁₆) molecules physically adsorbed on a smooth gold surface. Both in-plane and out-of-plane vibrational modes are observed, where the out-of-plane Raman modes at about 678 and 740 cm^-1 have different angular intensity distributions than those of in-plane Raman modes at 1309 and 1373 cm^-1. We interpret the angular spectrum of the TERS signal considering the molecular vibrational modes computed with density functional theory (DFT) for the free and gold-deposited molecule, and the directed Raman scattering by the gap-mode predicted by finite-difference time-domain (FDTD) simulations. We contend that the TERS gap directs the Raman vibrational modes differently, leading to distinct angularly distributed Raman scattering intensities. These findings emphasize the nonnegligible role of the TERS detection scheme in understanding spectral features, such as the relative peak intensity ratio variations for studying molecular orientations, or for monitoring chemical reactions.

Keywords: Angular‐resolved emission; Antenna directivity; Back focal plane imaging; Energy momentum spectroscopy; Tip‐enhanced Raman spectroscopy.

Figure 1

Experimental setting. a) Scheme of the angular resolved radiation intensity from flat‐lying CoPcF₁₆ molecules in a TERS gap with the in‐plane and out‐of‐plane Raman dipoles. b) Top‐illumination TERS using a parabolic mirror to focus the RPDM beam onto the tip‐sample gap.

Figure 2

TERS spectra and calculated Raman vibrational motions. The excitation wavelength is λ = 633 nm, the acquisition time for the Stokes spectrum is 30 s, and for the anti‐Stokes spectrum is 180 s. a) Raman spectra without tip engagement (dotted lines) and with tip engagement (solid lines) of CoPcF₁₆ on a smooth gold film with different excitation polarizations. Inset: Chemical structure of CoPcF₁₆. b) Nuclear displacements of vibrational modes calculated for $\stackrel{\sim }{\nu }$ = 678, 740 (out‐of‐plane), and 1373 cm⁻¹ (in‐plane) using DFT. c)–e) Anti‐Stokes‐, Stokes‐, and overtone Raman scattering for small, medium, and large gap distance, respectively.

Figure 3

Calculated gap‐distance and wavelength‐dependent far‐field emission intensity |E²|. a) Definition of simulation parameters: θ is the polar angle versus the optical axis (z), and φ is the azimuthal angle in the x‐y‐sample plane. b) Simulated plasmon resonances for the tip alone, and the tip‐sample gap (d = 2 nm). c) and d) Radiation patterns for λ _em = 640 nm at varying gap‐distances simulated for an emitter dipole c) vertical, and d) parallel to the sample plane, respectively. Right: Normalized polar plots in the x‐z‐plane for varying gap‐distances, sectioned from φ = 90° to φ = 270° in the respective polar plots. e) and f) Radiation patterns for different emission wavelengths (λ = 640, 660, 680, 700, 720 nm) calculated for a dipole e) vertical, and f) parallel to the sample plane in a 2 nm tip‐molecule gap, respectively. Right: x‐z‐sections of normalized emission intensity for varying wavelengths, sectioned from φ = 90° to φ = 270° in the respective polar plots.

Figure 4

Angularly resolved TERS. The excitation wavelength is λ = 633 nm; the acquisition time is 90 s for all Raman spectra and 180 s for back‐focal plane images. a) Back‐focal plane and energy momentum TERS setup scheme. The collected k‐space image is limited by the numerical aperture of the parabolic mirror. The angular range marked as black cannot be detected due to the hole in the parabolic mirror. Spatial slicing is used to select TERS signal from a defined angular range. b) Back‐focal plane image recorded at a tip‐sample gap of about 2 nm. c) Radiation intensity at varying θ angle in a 2 nm tip‐sample gap. d) and e) Raman spectra collected in Slice 1 (S1) and Slice 3 (S3) at large gap (LG) and small gap (SG), respectively. The rest of the subtracted spectra is shown in blue. The red dashed line indicates the spectral background fit. f) TERS spectra collected in five angular ranges, and the reference TERS without angular selection. The red dash line indicates the gap plasmon resonance calculated in Figure  3b in the spectral range of 640 and 710 nm. g) Experimental and h) simulated far‐field emission intensity from Slice 1 to Slice 5 for dipoles vertical or parallel to the sample plane, respectively. i) and j) The same as g) and h) but for Slices 6 to 10.