326
I. I. Smolyaninov and V. N. Smolyaninova
Fig. 13.9 a This image obtained using a conventional optical microscope presents the results of
two imaging experiments (top portion of the image) performed simultaneously with four control
experiments seen at the bottom of the same image. The rows of PMMA dots shown in the inset
AFM image were fabricated near the two top and two bottom hyperlenses. No such pattern was
made near the two hyperlenses visible in the center of the image. Upon illumination with an external
laser, the two rows of PMMA dots separated by a 130 nm gap gave rise to two divergent plasmon
rays shown by the arrows, which are clearly visible in the top portion of the image. The four control
hyperlenses visible at the bottom do not produce such rays, because there is no sample to image for
the two hyperlenses in the center, and the two bottom hyperlenses are inverted. b The same pattern
produced on an ITO instead of gold film demonstrates a pattern of ordinary light scattering by the
structure without any hyperlens imaging effects
dot structure was designed for phase-matched plasmon generation in the upward
direction, as seen in the image. That is why no plasmon rays are visible when the
hyperlens structures are inverted, as seen in the bottom of Fig. 13.9a. When the gold
film was replaced with an ITO film in another control experiment performed using
the same experimental geometry, no hyperlens imaging occurred since no surface
plasmons are generated on ITO surface (see Fig. 13.9b). These experiments clearly
verify the imaging mechanism and increased spatial resolution of the 2D plasmonic
hyperlens.
13.4 Super-Resolution Microscopy Based on Photon
Tunneling
Most of the linear super-resolution microscopy techniques rely on some form of
conversion of “fast” normally evanescent Fourier components of the object electromagnetic field into the propagating far field. For example, in scanning photon
tunneling microscopy (PSTM) [1] evanescent light tunnels into a small tapered fiber
probe, which is scanned over the sample surface. The probe converts evanescent light
into photons which propagate inside the fiber. On the other hand, in the 2D and 3D
hyperlens geometries [10–13] a similar conversion is achieved using a relatively small
volume of hyperbolic material, which typically consists of closely spaced concentric
metal-dielectric interfaces, which support surface plasmon polaritons. Therefore, a
hyperlens may be understood as a form of photon tunneling device, which relies on
radial plasmon tunneling between the adjacent interfaces. Comparison of PSTM and
hyperlens techniques indicate that a simple non-scanning super-resolution imaging
device may be built based on radial photon tunneling (Fig. 13.10). No metamaterial
is necessary for its operation. Similar to our recent cloaking [20] and “trapped rainbow” [21] demonstrations, the “metamaterial properties” needed to observe these
effects are emulated by tapered optical waveguides. Similar to PSTM, high spatial
resolution of this technique is guaranteed by very strong exponential dependence
of tunneling probability on the tunneling distance. This is illustrated in Fig. 13.11,
which demonstrates results of numerical simulation of light propagation inside our
I. I. Smolyaninov and V. N. Smolyaninova
Fig. 13.9 a This image obtained using a conventional optical microscope presents the results of
two imaging experiments (top portion of the image) performed simultaneously with four control
experiments seen at the bottom of the same image. The rows of PMMA dots shown in the inset
AFM image were fabricated near the two top and two bottom hyperlenses. No such pattern was
made near the two hyperlenses visible in the center of the image. Upon illumination with an external
laser, the two rows of PMMA dots separated by a 130 nm gap gave rise to two divergent plasmon
rays shown by the arrows, which are clearly visible in the top portion of the image. The four control
hyperlenses visible at the bottom do not produce such rays, because there is no sample to image for
the two hyperlenses in the center, and the two bottom hyperlenses are inverted. b The same pattern
produced on an ITO instead of gold film demonstrates a pattern of ordinary light scattering by the
structure without any hyperlens imaging effects
dot structure was designed for phase-matched plasmon generation in the upward
direction, as seen in the image. That is why no plasmon rays are visible when the
hyperlens structures are inverted, as seen in the bottom of Fig. 13.9a. When the gold
film was replaced with an ITO film in another control experiment performed using
the same experimental geometry, no hyperlens imaging occurred since no surface
plasmons are generated on ITO surface (see Fig. 13.9b). These experiments clearly
verify the imaging mechanism and increased spatial resolution of the 2D plasmonic
hyperlens.
13.4 Super-Resolution Microscopy Based on Photon
Tunneling
Most of the linear super-resolution microscopy techniques rely on some form of
conversion of “fast” normally evanescent Fourier components of the object electromagnetic field into the propagating far field. For example, in scanning photon
tunneling microscopy (PSTM) [1] evanescent light tunnels into a small tapered fiber
probe, which is scanned over the sample surface. The probe converts evanescent light
into photons which propagate inside the fiber. On the other hand, in the 2D and 3D
hyperlens geometries [10–13] a similar conversion is achieved using a relatively small
volume of hyperbolic material, which typically consists of closely spaced concentric
metal-dielectric interfaces, which support surface plasmon polaritons. Therefore, a
hyperlens may be understood as a form of photon tunneling device, which relies on
radial plasmon tunneling between the adjacent interfaces. Comparison of PSTM and
hyperlens techniques indicate that a simple non-scanning super-resolution imaging
device may be built based on radial photon tunneling (Fig. 13.10). No metamaterial
is necessary for its operation. Similar to our recent cloaking [20] and “trapped rainbow” [21] demonstrations, the “metamaterial properties” needed to observe these
effects are emulated by tapered optical waveguides. Similar to PSTM, high spatial
resolution of this technique is guaranteed by very strong exponential dependence
of tunneling probability on the tunneling distance. This is illustrated in Fig. 13.11,
which demonstrates results of numerical simulation of light propagation inside our