11 Plasmonic Functionalities Based on Detuned Electrical Dipoles
425
nanoantenna, respectively. Interestingly, simulations show that the induced current
at ϕ = 1520 nm (Fig. 11.11f) makes a full rotation within one cycle; a property
related to scattered light with elliptical/circular polarization. In order to get circular
polarized light at ϕ ∼ 1520 nm, we should require |E sc,x | = |E sc,y | that can be
satisfied at Ω = 56 ◦ (Fig. 11.11d). Note that in the simplified case of Eq. (11.1), we
would expect circular polarization at Ω = 45 ◦ , but a lower dipole strength of the
short nanoantenna requires Ω > 45 ◦ . Also importantly, the configuration has the
wavelength range 1475–1560 nm where the phase difference variation is less than
2 %, implying a relatively broadband operation of the proposed wave plate.
As an example of versatility of our approach, we construct a quarter-wave plate
from a gold nanobrick of 50 nm thickness (Fig. 11.12a) exhibiting two orthogonal
plasmonic resonances [see the scattering cross section in Fig. 11.12b]. This time,
however, we choose the short axis to be L 1 = 90 nm satisfying the criteria for a
quarter-wave plate at ϕ ∼ 770 nm when L 2 = 140 nm and Ω = 52 ◦ (not shown).
The wavelength range in which the phase difference deviates with less than 2 % is
748–796 nm. For further details on the nanobrick configuration, we refer to [15].
Using electron-beam lithography and lift-off technique, a 400 nm-periodic (50 ×
50 µm 2 ) array of the nanobrick configuration has been fabricated on a silica substrate
[see inset in Fig. 11.12c] and subsequently covered with ∼15 µm PMMA to create
(almost) homogeneous surroundings. Reflection spectra are measured using a broadband halogen light source with a fiber output whose radiation is directed through a
Glan-Thompson polarizer and focused onto the fabricated array by a x60-objective
(NA = 0.85). The reflected light is collected by the same objective and fed into an
optical fiber connected to a spectrometer. The reflection spectra (Fig. 11.12c) show
that for Ω = 0 ◦ and Ω = 90 ◦ the individual resonances along the brick long and
short axis, respectively, are being excited, whereas both resonances are excited when
using an intermediate Ω.
In order to verify experimentally that the reflected light from the fabricated array is
circularly polarized, Ω is fixed to an angle of 50 ◦ and the radiation from a Ti:sapphire
laser at 780 nm is focused onto the array with a x10-objective (NA = 0.25), collected by the same objective and sent through an achromatic quarter-wave plate
(690–1200 nm) followed by a Glan-Thompson analyzer with varying analyzer angle
with respect to (w.r.t.) the long axis of the brick (Fig. 11.12d). If the reflected light
from the fabricated array is circularly polarized it must be converted into linearly
polarized light with an angle of 45 ◦ w.r.t. the fast axis (FA) of the quarter-wave plate.
When fixing the FA w.r.t. the long axis of the brick at 0 ◦ and 90 ◦ , respectively, the
output of the Glan-Thompson analyzer reveals linear polarization at angles −45 ◦ and
45 ◦ , respectively. Therefore, in this case, the criteria of equal strengths and a phase
difference of ∼ η/2 between the two resonances of the brick are approximately met,
and the reflected light is circularly polarized. Note that the minimum reflection is
∼5 % due to the presence of linearly polarized light reflected at the glass/PMMA
interface. On the other hand, when Ω = 0 ◦ or Ω = 90 ◦ and the FA is fixed at 45 ◦ w.r.t.
to the long axis of the brick, the individual resonances are excited, and the commercial quarter-wave plate converts the reflected (linearly polarized) light into circularly
polarized light as evidenced by the analyzer output when varying the analyzer angle.
Précédent

- 433/581

Suivant