Scattering-type scanning near-field optical microscopy (s-SNOM) enables super-resolution spectral imaging of the surfaces of a wide range of materials and nanostructures. However, although s-SNOM offers high lateral spatial resolution, the near-field response characteristics perpendicular to the sample surface have rarely been investigated.
Current s-SNOM methods face challenges in characterizing vertical interactions, which indirectly stems from their large far-field scattering background — that is, reflection or scattering from the tip shaft and cantilever, as well as reflection or scattering caused by sample roughness, in the absence of short-range tip-sample interactions. In s-SNOM, to remove this large background, the probe tip oscillates sinusoidally above the sample in tapping mode, and the near-field signal is extracted at harmonics of the tip oscillation frequency using a lock-in amplifier. However, approach curves do not directly describe the vertical dependence of the near-field signal. They are obtained through a complex signal generation mechanism that depends on the tip oscillation amplitude and the order of lock-in demodulation. As a result, approach curves lack a direct and explicit relationship with the vertical characteristics of the near-field interaction between tip and sample.
The paper introduced in this article proposes an accurate and fast reconstruction method to obtain the vertical characteristics of near-field interactions. The authors also investigated the bound electromagnetic field components of surface phonon polaritons on boron nitride nanotubes, finding that they decay within a range of 20 nm, with significant phase changes in the near-field signal.
Figure 1. a) Schematic of the experimental setup and working principle. The inset on the right shows the lock-in demodulated amplitude of the 18th harmonic obtained from experimental measurements using a platinum-coated probe and a gold substrate.
b) Reconstructed raw s-SNOM waveform.
c) Scattering-type scanning near-field optical microscopy interaction curve. The dashed line shows the linear trend of the background, which is characteristic of the far-field contribution.
Figure 1a shows the experimental setup and working principle, with which the authors obtained the experimentally measured 18th harmonic signal, and then reproduced its signal waveform through a Fourier series (Equation 1), as shown in Figure 1b. Equation 2 describes the positional variation of the probe tip. Figure 1c shows the s-SNOM interaction curve obtained from Equation 3, where the dashed portion represents the fitted far-field contribution. From the figure, it can be seen that when the tip-sample distance is within 50 nm, nonlinear growth of the signal is observed, indicating that the near-field signal dominates. At larger tip-sample distances, the far-field signal is predominant.
For the AC far-field signal, a probe in sinusoidal mechanical oscillation scatters the nanoscale uniform region of the incident infrared field. Therefore, the AC signal from the far-field contribution is sinusoidal and appears only in the first-harmonic demodulation. Thus, the following equation can be obtained:
If the value of η can be obtained in some way, the near-field signal can be fully determined. The authors argue that when the tip-sample distance D is large, the near-field signal is insensitive to changes in D, which can be used to solve for η, corresponding to minimizing the slope of the following formula:
To verify the accuracy of the results, the authors compared the experimentally obtained results with the point-dipole model (Figure 2a). Figure 2b shows the near-field time-domain signal obtained after solving for η. Figure 2c shows the s-SNOM interaction curve obtained from Equation 5, where the dashed line is the point-dipole model interaction curve. Figure 2d shows the normalized approach curves derived from the second to fifth harmonics of the interaction curve; from the figure, it can be seen that the third-harmonic normalization is nearly consistent with the experimentally obtained third-harmonic approach curve.
Figure 2. a) Point-dipole model simulation of the relationship between the effective polarizability amplitude and the tip-sample distance.
b) Reconstructed near-field waveform based on previous data. The contribution of the first harmonic is found to be 0.262.
c) s-SNOM interaction curve. The dashed curve shows the plotted result of the point-dipole model. d) Normalized approach curves derived from the second to fifth harmonics of the interaction curve. The inset shows a comparison between the derived third-harmonic approach curve and the experimentally obtained third-harmonic approach curve.
Finally, the authors applied the interaction solution formula to the observation of the vertical response of surface phonon polaritons on boron nitride nanotubes, as shown in Figure 3.
Figure 3. a) Atomic force microscope topography image of a boron nitride nanotube.
b) Scattering-type near-field optical microscopy image at 1400 cm⁻¹.
c) Scattering-type near-field optical microscopy image at 1420 cm⁻¹.
d) Scattering-type near-field optical microscopy image at 1360 cm⁻¹.
e) Interaction curves at position 1 (red) and position 2 (blue) at 1400 cm⁻¹.
f) Interaction curves at position 3 (red) and position 4 (blue) at 1420 cm⁻¹. The interaction curves at positions 3 and 4 at 1360 cm⁻¹ are shown in the inset.
g) Interaction curve at position 5 at infrared frequencies of 1360, 1400, and 1420 cm⁻¹.
h) 1/e decay length values (black dots) of the interaction curve at position 5 at infrared frequencies from 1350 to 1435 cm⁻¹. The blue curve serves as a reference.
Measurements of the interaction curves show that the near-field signal of polaritonic materials originates from two sources. The first source is the intrinsic near-field scattering, which is based on the modulation of the tip polarizability by the redistribution of charges, carriers, or dipoles in the sample. For boron nitride (BN), at 1360 cm⁻¹, where surface plasmon polaritons do not exist, this contribution dominates. Between 1350 cm⁻¹ and 1360 cm⁻¹, the decay length at an amplitude of 1/e is approximately 40 nm. The second contribution comes from the localized bound electromagnetic field component of the polaritons. The sharp metal tip converts the field component of the surface plasmon polaritons into a detectable scattered light field. The bound field of the surface plasmon polaritons is close to the polariton-supporting surface of boron nitride, so when the tip is very close, it leads to additional enhancement of the near-field scattering. Due to the presence of the short-range surface-bound component, the measured decay length of boron nitride at the surface plasmon polariton-active frequency is shortened compared with the case where polaritons do not exist. The trend of decreasing decay length with increasing frequency (Figure 3f) indicates that at higher infrared frequencies, the field component of the surface plasmon polaritons is more strongly bound to the boron nitride surface.
In summary, the paper introduced in this article developed an experimental method for reconstructing vertical near-field interactions in scattering-type scanning near-field optical microscopy (s-SNOM), which is expected to be used to characterize the vertical field distributions of a variety of nanophotonic materials and structures.
References
Wang, L., Xu, X. Scattering-type scanning near-field optical microscopy with reconstruction of vertical interaction. Nat Commun 6, 8973 (2015). https://doi.org/10.1038/ncomms9973.
Originally published in Chinese on the WeChat official account "见微而知著-觅几科技" (Miji Technology). Original article link: https://mp.weixin.qq.com/s?src=11×tamp=1785059356&ver=6866&signature=ViZ3f6CWwEzsZoMF0*nFqhx-KBdC0tygJFHqt3bT6NmTRp6xdLB27RKc6w4UOnT2OagE1U-rLhG9i-vIfGovn--WU1Q3d7*WaYpms-sLkG06D8rlR1c-wuJ8iV3xTOtg&new=1