Figure 1

Schematic representation of the experiment for imaging standing surface plasmons in the Kretschmann configuration using a photon scanning tunneling microscope. The left (L) and right (R) portions of the incident beam stimulate the same region of the film to excite two counterpropagating surface plasmons which will interfere to set up standing surface plasmons.Reuse & Permissions

Figure 2

Schematic representation of the probe-substrate geometry for imaging standing surface plasmons. The involved material domains are characterized by ${ϵ}_t$ for the probe (here taken as quartz), $ϵ\left(\omega \right)$ for the metal (here gold), and ${ϵ}_s$ for the quartz substrate. The angle ${\theta }_t$ is the angle between the symmetry axis $z$ and an asymptote to the probe. The upper surface of the film is defined by the plane $z=0$, while $z=-d$ defines the bottom plane. The counterpropagating plasmon wave vectors are denoted by $\pm \kappa$, while the excitation wave vector $k=2\pi ∕\lambda$.Reuse & Permissions

Figure 3

The simulated behavior of the influence function, Eq. (4). (a) The higher the magnitude of ${\mathcal{T}}_q^m\left({\mu }_t\right)$ for a probe defined by ${\mu }_t=\cos {\theta }_t$, the higher the extent to which the surface modes of the configuration in Fig. 2 deviate from those of an unperturbed foil. Noting that for a particular mode of the influence function represented by the pair $(m,q)$, the closer to the surface and blunter the probe, the higher the magnitude above zero as expected. In the absence of the probe—that is, when ${ϵ}_t=0$—then ${\mathcal{T}}_q^m\left({\mu }_t\right)=0$ for all $(m,q)$. This is particularly shown in (b) for the strongest modes [corresponding to $q=0$; see (a)], the effect being largest in the metallic tip limit $ϵ\to \infty$.Reuse & Permissions

Figure 4

The $m=1$ modification of the dispersion relation of a thin metal film on a dielectric half space corresponding to the symmetric (a) and antisymmetric (b) modes of the surface charge distribution. As can be seen for a fixed film thickness $d$, only a small momentum region experiences a negligible alteration, a behavior observed in all the roots of Eq. (3). The visible range $400–700\mathrm{nm}$ corresponds to the $\kappa d$ range (for a $45\text{−}\mathrm{nm}$-thick metal film) of 0.71–0.40.Reuse & Permissions

Figure 5

Standing surface plasmons excited at $\lambda =632.8\mathrm{nm}$ on a $55\text{−}\mathrm{nm}$ gold foil is sensed by the probe of a photon scanning tunneling microscope. When imaging the field of the plasmons, the presence of the dielectric probe has negligible effect on the modes of the metal film and plasmon interference can be sustained in arbitrarily large regions in comparison with the excitation wavelength. The formation of the interference pattern due to the standing surface plasmons in the metal film is further supported by a decrease in the intensity of the fringes upon (1) a continuous variation of the polarization state of the incident field from $p\to s$, (2) an inhibition of either L or R beam, and (3) a shift away from resonance of the angular or spectral position of the excitation beam. No image processing was performed on the presented data.Reuse & Permissions

Figure 6

Photon scanning tunneling microscope images of standing surface plasmons over a $4\lambda \times 4\lambda$ region, excited, respectively, at (a) $\lambda =442\mathrm{nm}$ using a $25\text{−}\mathrm{nm}$ gold film and (b) $\lambda =1550\mathrm{nm}$ using a $29.5\text{−}\mathrm{nm}$ gold film. The scale bars indicate the relative range of the amplitude of in the detected signal.Reuse & Permissions

Figure 7

(a) A line profile (symbols) in the direction parallel to the plasmon wave vector ${\kappa }_x$. The fitted function $u\left(x\right)=2.1-18.9\cos (20.4x+0.3)$ (solid curve) yields a period of $0.3\mu \mathrm{m}$. We note that this tunneling signal, detected by the probe and photomultiplier combination, is a direct consequence of the plasmon field amplitudes. In particular, we note the difference of this signal from the traditional PSTM signal, where the tunneling photons are provided by the evanescent field produced by total internal reflection at the sample/substrate interface. (b) Power spectral density along a line perpendicular to the fringes. The inset shows the two-dimensional FFT spectrum of the image in Fig. 5. The off-vertical axis placement of the spikes is a result of the small angle of the fringes with respect to the horizontal axis in image of Fig. 5.Reuse & Permissions