5 Near-Field Scanning Optical Microscope Combined with Digital …
127
function of the emission angle in polar coordinates for the three cases studied previously: tip in free space (Fig. 5.11a), in contact with the glass substrate (Fig. 5.11b),
and in contact with the gold film on a glass substrate (Fig. 5.11c). The plotted experimental data points (in blue) are in spherical coordinates and are selected in a way
to obey two selection rules. First, they are chosen to lie in the x–z plane of the
NSOM tip perpendicular to the surface of the sample. Second, we made sure that
each angle θ has only one attributed intensity value. To this end, we select the data
that lie only in a spherical shell of thickness ρ = 30 nm centered on the aperture
tip. The corresponding FDTD simulation results of the far-field radiation patterns
of a superposition of a magnetic and electric dipole are plotted in black. Very good
agreement is obtained in all the three cases. The angular radiation patterns confirm
the observations presented in Sect. 5.3.2 (Figs. 5.7, 5.8 and 5.9). In fact, we observe
that the intensity of the scattered light from the tip in free space decays exponentially
around a maximum value centered at θ = 0° (Fig. 5.11d). We also quantitatively
verify that for the tip placed in contact with the glass sample (Fig. 5.11e), most of
the light is scattered exactly at the critical angle ±|θ c,glass | = 41.8° in addition to a
broad distribution around θ = 0°. As for the plasmonic sample, we verify that the
generated leaky plasmons are transmitted into the substrate at an angle of ±|θ c,gold | =
43.3°, which is higher than the critical angle of the medium. This highly directional
scattering, narrower than in the previous cases, again reinforces our previous observation of the two lobes with a preferential direction caused by the surface plasmons.
Surface plasmons are indeed highly selective in angle, both for their excitation and,
as shown here, their leakage.
Next, we aim at comparing the behavior of the NSOM probe in the case where it is
placed in the far-field region above the sample to the case where it is placed in contact
with the sample in the near-field region. To do that, we place the tip at a distance
of 3 μm above the sample surface, and repeat the same characterization procedure.
The angular radiation pattern is plotted in Fig. 5.12. By comparing Fig. 5.12b to
Fig. 5.11e, we notice the very high resemblance between the two plots, which allows
us to conclude that the gold film has no effect on the transmitted light through the
substrate when the incident light source is placed in the far field. This is because
in this case, only small wavevectors are created by the tip, and therefore surface
plasmons cannot be generated. The angular emission thus corresponds to that of a
glass substrate, with maxima occurring at an angle of ± 40.5° (below θ c,glass ). We
verified this result by performing the same experiment with the tip placed at a 3 μm
height above a glass substrate, and the same exact emission angle was observed. In
addition, due to the weak transmission of the gold film, the maximum intensity at ρ
= 10 μm is found to decrease by a factor of 2.5 when the tip is placed in the far field.
These results agree with those obtained by Hecht et al. [17] using back-focal plane
and real space imaging. The complex EM field represented by the product ρ|A(x,
y, z)|cos(φ(x, y, z)) is shown in Fig. 5.12c, where we can also clearly see that there
are no leaky surface plasmons generated as opposed to the case where the tip is in
contact with the film (Fig. 5.9c).
127
function of the emission angle in polar coordinates for the three cases studied previously: tip in free space (Fig. 5.11a), in contact with the glass substrate (Fig. 5.11b),
and in contact with the gold film on a glass substrate (Fig. 5.11c). The plotted experimental data points (in blue) are in spherical coordinates and are selected in a way
to obey two selection rules. First, they are chosen to lie in the x–z plane of the
NSOM tip perpendicular to the surface of the sample. Second, we made sure that
each angle θ has only one attributed intensity value. To this end, we select the data
that lie only in a spherical shell of thickness ρ = 30 nm centered on the aperture
tip. The corresponding FDTD simulation results of the far-field radiation patterns
of a superposition of a magnetic and electric dipole are plotted in black. Very good
agreement is obtained in all the three cases. The angular radiation patterns confirm
the observations presented in Sect. 5.3.2 (Figs. 5.7, 5.8 and 5.9). In fact, we observe
that the intensity of the scattered light from the tip in free space decays exponentially
around a maximum value centered at θ = 0° (Fig. 5.11d). We also quantitatively
verify that for the tip placed in contact with the glass sample (Fig. 5.11e), most of
the light is scattered exactly at the critical angle ±|θ c,glass | = 41.8° in addition to a
broad distribution around θ = 0°. As for the plasmonic sample, we verify that the
generated leaky plasmons are transmitted into the substrate at an angle of ±|θ c,gold | =
43.3°, which is higher than the critical angle of the medium. This highly directional
scattering, narrower than in the previous cases, again reinforces our previous observation of the two lobes with a preferential direction caused by the surface plasmons.
Surface plasmons are indeed highly selective in angle, both for their excitation and,
as shown here, their leakage.
Next, we aim at comparing the behavior of the NSOM probe in the case where it is
placed in the far-field region above the sample to the case where it is placed in contact
with the sample in the near-field region. To do that, we place the tip at a distance
of 3 μm above the sample surface, and repeat the same characterization procedure.
The angular radiation pattern is plotted in Fig. 5.12. By comparing Fig. 5.12b to
Fig. 5.11e, we notice the very high resemblance between the two plots, which allows
us to conclude that the gold film has no effect on the transmitted light through the
substrate when the incident light source is placed in the far field. This is because
in this case, only small wavevectors are created by the tip, and therefore surface
plasmons cannot be generated. The angular emission thus corresponds to that of a
glass substrate, with maxima occurring at an angle of ± 40.5° (below θ c,glass ). We
verified this result by performing the same experiment with the tip placed at a 3 μm
height above a glass substrate, and the same exact emission angle was observed. In
addition, due to the weak transmission of the gold film, the maximum intensity at ρ
= 10 μm is found to decrease by a factor of 2.5 when the tip is placed in the far field.
These results agree with those obtained by Hecht et al. [17] using back-focal plane
and real space imaging. The complex EM field represented by the product ρ|A(x,
y, z)|cos(φ(x, y, z)) is shown in Fig. 5.12c, where we can also clearly see that there
are no leaky surface plasmons generated as opposed to the case where the tip is in
contact with the film (Fig. 5.9c).