Ultrasmall radial polarizer array based on patterned plasmonic nanoslits
Kentaro Iwami, Miho Ishii, Yuzuru Kuramochi, Kenichi Ida, and Norihiro Umeda
Citation: Appl. Phys. Lett. 101, 161119 (2012); doi: 10.1063/1.4761943
View online: http://dx.doi.org/10.1063/1.4761943
View Table of Contents: http://apl.aip.org/resource/1/APPLAB/v101/i16
Published by the American Institute of Physics.
Localized surface plasmon resonances in highly doped semiconductors nanostructures
Appl. Phys. Lett. 101, 161113 (2012)
Polarization dependant scattering as a tool to retrieve the buried phase information of surface plasmon polaritons
Appl. Phys. Lett. 101, 161603 (2012)
Controllable plasmonic antennas with ultra narrow bandwidth based on silver nano-flags
Appl. Phys. Lett. 101, 153118 (2012)
Resonant plasmonic effects in periodic graphene antidot arrays
Appl. Phys. Lett. 101, 151119 (2012)
Silver nanoisland enhanced Raman interaction in graphene
Appl. Phys. Lett. 101, 153113 (2012)
Additional information on Appl. Phys. Lett.
Journal Homepage: http://apl.aip.org/
Journal Information: http://apl.aip.org/about/about_the_journal
Top downloads: http://apl.aip.org/features/most_downloaded
Information for Authors: http://apl.aip.org/authors
Ultrasmall radial polarizer array based on patterned plasmonic nanoslits
Miho Ishii, Yuzuru Kuramochi,
Kenichi Ida, and Norihiro Umeda
Department of Mechanical Systems Engineering, Tokyo University of Agriculture and Technology,
2-24-16 Nakacho, Koganei, Tokyo 184-8588, Japan
(Received 4 September 2012; accepted 8 October 2012; published online 19 October 2012)
We developed an ultrasmall radial or azimuthal polarization converter of size ranged from 10 to
100 lm. The converter consists of a half-wave plate divided into four quadrants, each of which is
made of periodic gold nanoslits at different orientations in steps of 45. The converter and its array
were fabricated and followed by an evaluation with the S
enarmont method and the polarization
microscopy. Due to the giant birefringence of the gold nanoslits, each slit segment achieved 180
retardation at the wavelength of 633 nm, and the conversion from linear to radial polarization was
C2012 American Institute of Physics.[http://dx.doi.org/10.1063/1.4761943]
Radially and azimuthally polarized lights have attracted
much interest because of their unique axially symmetric dis-
tribution of the polarization state in a beam cross section, in
contrast to the homogeneously polarized beams. These polar-
izations have been widely studied and applied to many ﬁelds,
including ﬁne focusing for high resolution optical micros-
and photothermal therapy.
In particular, the generation of a large longitudinal electric
ﬁeld component is a signiﬁcant characteristic of focused
radially polarized light, and it has opened up unique applica-
tions such as particle acceleration.
The generation of these polarizations relies on the local
control of both the polarization orientation and phase. Sev-
eral methods have been studied to accomplish this, including
polarization selection at a laser cavity
or use of a radial
polarization converter (RPC). Of the wide variety of RPCs
that have been studied, examples include orientation-tailored
liquid crystal (LC),
LC-based spatial light modulators
spatially varying retardation (SVR) based on opti-
or photonic crystal,
and glass nanostructuring.
On the other hand, recent advances in microelectrome-
chanical systems (MEMS) have opened up a world of
“parallel optics,” represented by micromirror array or micro-
lens array devices, which have been applied in areas from
basic science to the telecommunications of the future.
The use of radially or azimuthally polarized light in the ﬁeld
of parallel optics may offer many opportunities, for example,
high-throughput nanoscale surface imaging, sensing, and
modiﬁcation by combining them with MEMS probe array
system. However, such an application has not yet been real-
ized because axially symmetric beams cannot be arranged in
parallel. Therefore, a RPC array is critically needed. Con-
ventional RPCs, however, have disadvantages for this pur-
pose. The main obstacle is the difﬁculty in fabricating
microscale RPCs. The size and resolution of SLMs are not
MEMS compatible, and the low damage threshold of LC-
based devices restricts their application. SVR plates, which
consist of half-wave plate crystal pieces divided into four or
eight segments, are difﬁcult to miniaturize. The sub-
wavelength gratings formed on dielectric substrates have
advantages for micromachining, but the application is lim-
ited to the mid-infrared range because the micromachining
of optically transparent material with the sub-wavelength re-
solution is difﬁcult. Although photonic crystals would satisfy
the above requirements, its fabrication is difﬁcult because
multilayer deposition is essential. Minimum featuring size of
glass-nanostructured RPC is still unclear.
Periodic nanoslit arrays on noble metal thin ﬁlms were
utilized to induce giant birefringence, and more than a 180
phase shift was achieved with nanoscale-thickness.
structure is most compatible with miniaturization because it
can be fabricated through conventional electron beam (EB)
or nano-imprint lithography combined with a lift-off process,
as is expected for future mass-produced wave plates.
Furthermore, by arranging nanoslit arrays into a pattern, the
spatial distribution of polarization can be theoretically tai-
lored. In this paper, we design, develop, and demonstrate an
ultrasmall RPC based on a half-wave plate divided into four
quadrants that are composed of patterned plasmonic nanoslit
Figure 1shows a schematic drawing of a plasmonic
RPC and its cross section. As shown in Fig. 1(a), the RPC
consists of four segments of a nanoslit array of a gold thin
ﬁlm deposited on a transparent substrate. Each array segment
has identical slit width and period. Figure 1(b) shows the ori-
entation of polarization. Since the phase shifts are opposite
between the transverse electric (TE: the electric ﬁeld is par-
allel to the nanoslits) and the transverse magnetic (TM: the
electric ﬁeld is perpendicular to the nanoslits) polarizations,
a giant birefringence can be achieved.
As shown in Fig.
1(a), the four segments, or quadrants, are arranged by chang-
ing each slit orientation in steps of 45. In this structure, the
TE axis (fast axis) is treated as the principal axis of the slit
array because the phase shift for the TE wave is positive, in
contrast to the negative phase shift for the TM wave. There-
fore, the principal axes of the four quadrants from 1 to 4 can
be described as 45,90
, 135, and 0, respectively. If each
quadrant achieves a retardation of 180, the arrays work as a
half-wave plate divided into fourths. Consequently, 45and
Author to whom correspondence should be addressed. Electronic mail:
email@example.com. URL: http://nmems.lab.tuat.ac.jp/en/.
Present address: Department of Precision Engineering, the University of
Tokyo, Tokyo, 113-8656, Japan.
C2012 American Institute of Physics101, 161119-1
APPLIED PHYSICS LETTERS 101, 161119 (2012)
135linear-polarized incident light can be converted into
radially and azimuthally polarized light, respectively.
When developing a plasmonic half-wave plate of metal
nanoslits, design parameters should be optimized; however,
it is difﬁcult to estimate the phase shift accurately on the ba-
sis of analytical waveguide theory. For example, propagation
constants for the TE and TM waves are given by,
0ðp=wef f Þ2
where kTE and kTM are propagation constants for the TE and
TM modes, respectively; and wand wef f are the raw and
effective widths of the slit, respectively; and mis a dielectric
function of the metal. As shown in Eq. (1),kTE has a sharp
cutoff at k¼wef f =2. However, for gold thin ﬁlms, cutoff is
not sharp but gradual around this wavelength.
numerical analysis is essential to the design of a plasmonic
nanoslit half-wave plate.
To determine the optimal slit design, a numerical simu-
lation based on commercially available ﬁnite-difference
time-domain (FDTD) software (RSoft, FULLWAVE 8.1) was
performed. As a half-wave plate, identical transmittance
between the TE and the TM waves is important for the plas-
monic nanoslit array, together with a 180phase shift. Figure
2(a) shows contour plots of the phase shifts and the ratio of
amplitude transmittance (ATE=ATM) for a 500-nm-period slit
array under a 633 nm incident wavelength. In Fig. 2(a), the
solid and dashed contours correspond to the phase shifts and
transmittance ratios, respectively. Since the contours of a
180phase shift and an identical transmittance ratio intersect
at a gold ﬁlm thickness of 320 nm and a slit width of 240 nm,
a plasmonic nanoslit half-wave plate can be developed at
this point. Fig. 2(b) shows the regions at which phase shifts
can be achieved between 175and 185together with trans-
mittance ratios between 0.95 and 1.05 for each incident
wavelength on a 500-nm-period gold slit array. As shown in
Fig. 2(b), a plasmonic nanoslit half-wave plate can be
achieved for a broad range in the visible spectrum. Fig. 2(c)
shows amplitude transmittance of above-mentioned regions.
The RPC pattern was deﬁned on a quartz glass substrate
of thickness 525 lm through EB lithography. A 5-nm-thick
chromium layer was deposited by the EB deposition for ad-
hesion, followed by the deposition of a gold layer. Next, the
EB resists were removed by an organic solvent, and a metal
nanoslit pattern was developed (lift-off process). Sizes of
single RPC ranged from 10 to 100 lm. Two types of slit
arrays were fabricated, one with a slit width of 300 nm and
the other, 350 nm with a slit pitch of 525 nm due to lithogra-
phy limitation. In order to achieve a 180phase shift at the
300-nm-wide slit, the gold thin ﬁlm was deposited with a
thickness of 445 nm. Fig. 3(a) shows a scanning electron mi-
croscopy (SEM) image of the fabricated 3 3 array of the
10 lm RPC array with the slit width of 300 nm.
FIG. 1. (a) Schematic drawing of the RPC based on a plasmonic gold nano-
slit pattern. It consists of four quadrants, each having 180retardation. The
RPC is arranged by changing the slit orientation in steps of 45. As the nano-
slit array has its principal axis (fast axis) along a slit, a 45linear-polarized
light can be converted to a radially polarized light. (b) Schematic of a cross
section of each slit array allocated on each quadrant.
FIG. 2. (a) Contour plots of phase shifts (solid curves) and the ratio of am-
plitude transmittance (ATE=ATM, dashed curves) for the 633 nm incident
wavelength. The pitch of the slit array was ﬁxed to 500 nm. (b) Regions to
achieve both 175185phase shift and 0:951:05 amplitude ratio for each
incident wavelength. (c) Amplitude transmittance spectrum of above-
FIG. 3. (a) An SEM image of a fabricated 3 3 RPC array. The size of an
RPC unit is 10 lm. (b) Experimental setup for the retardation measurement
based on the S
enarmont method. (c) Retardation spectra on 525-nm-period
and 300-nm-wide gold nanoslit by S
enarmont method measurement (circle),
FDTD calculation (square), and theoretical estimation (solid line), respec-
tively. Error bars indicate standard deviations over four quadrants (n¼50
for each quadrant).
161119-2 Iwami et al. Appl. Phys. Lett. 101, 161119 (2012)
The phase shifts of each quadrant were measured by the
enarmont method. Figure 3(b) shows the experimental
setup. A He-Ne laser (k¼633 nm) beam passes through a
linear polarizer whose transmission axis is 0(LP0), and is
focused onto each quadrant. The RPC with the size of
100 lm was used. The RPC was rotated with a degree of /
to keep the angle between the fast axis of each quadrant at
45. Therefore, /for the quadrants 1, 2, 3, and 4 were 0,
45,90, and 135, respectively. The transmitted beam
was collimated by an objective lens and passed through a
90Babinet-Soleil compensator as a quarter-wave plate
(QWP90) and a 90linear polarizer as an analyzer. Finally,
the power of the transmitted beam was measured by rotating
the analyzer (LP90þh).
Table Ishows the summary of retardation for RPCs with
slit widths of 350 and 300 nm for 633 nm light. For both slits,
the slit period was 525 nm. As expected, the 300 nm-slit RPC
shows retardation of about 180for every quadrant. This
result conﬁrms that each quadrant of the 300-nm-wide nano-
slit array acts as a half-wave plate. In 350-nm-wide slit, the
large retardation difference between orthogonal arrangement
(quadrants 2 and 4) and diagonal arrangement (quadrants
1 and 3) is observed. It is possibly due to shape error since
variable shaped e-beam was used.
Figure 3(c) shows the retardation spectra of a 300-
nm-wide slit under laser irradiation at four wavelengths
(532, 576, 614, and 633 nm) measured by the S
method (circle), FDTD calculation (square), and theoretical
estimation (solid line), respectively. As shown in Fig. 3(c),
theoretical estimation based on Eqs. (1) and (2) does not
agree with the other methods in the short wavelength
region. In contrast, the FDTD calculation agrees well with
the experimental result.
To determine the principal axes of each quadrant, the
dependence of the rotation angle on transmitted intensity
was measured for the 350-nm-slit RPC, and the results were
evaluated by ﬁtting to the theoretical curve of the S0compo-
nent of Stokes vector. In this setup, we assumed each quad-
rant as an optical retarder with the principal axis of /þ45
and a retardation of D. For example, the S0component of
transmitted light through the quadrant 1 (/¼0)is
expressed as follows:
Fig. 4shows the normalized intensity and the ﬁtted S0
curves for each RPC quadrant. As shown here, all the experi-
mental results agree well with the retarder model, and the
principal axes of each quadrant are conﬁrmed as the TE
direction (parallel to the slit). Therefore, it is conﬁrmed that
the fabricated nanoslit array pattern can act as four-quadrant
half-wave plates, that is, an RPC.
To demonstrate the fabricated RPC visually, a polariza-
tion microscope was used to observe it. Fig. 5shows optical
microscope images under polarized illumination with and
without an optical bandpass ﬁlter at 633 nm. As illustrated in
input linear polarization should be converted
into radial polarization by the RPC. The transmitted light is
imaged by a charge-coupled device (CCD) camera through
an analyzer with the transmission axis at 45(upper images)
and 135(lower images), with and without the 633-nm opti-
cal bandpass ﬁlter. As shown in the images with the 633 nm
ﬁlter, the RPC pattern exhibits the expected contrast; the
quadrants 1 and 3 are bright with the 45analyzer, and the
quadrants 2 and 4 are bright with the 135analyzer. In con-
trast, the white light images do not show clear RPC-like
response because of the wavelength-dependent retardations
from the nanoslit. The differences in brightness between
kitty-cornered quadrants obtained with 633 nm ﬁlters can be
attributed to the differences of transmittance between the TE
and the TM polarizations resulting from the fabrication ﬂuc-
tuation in slit dimensions. Intra-quadrant inhomogeneity of
transmittance is also obtained, the reason can be due to par-
tial peeling or sticking occurred during the development pro-
cedure of EB resist.
TABLE I. Retardation (deg) for 633-nm irradiation of each nanoslit array
quadrant with slit widths of 350 and 300 nm. The period and thickness of
both slit arrays are 525 and 445 nm, respectively.
Quadrant 350 nm 300 nm
1 116.3 175.7
2 71.65 184.9
3 147.9 184.7
4 70.24 181.1
FIG. 4. Normalized intensity of transmitted light through each quadrant of
RPC and ﬁtting with S0component of Stokes vector.
FIG. 5. Polarization microscope images of the 300-nm-wide slit RPC illumi-
nated by white light with and without a 633 nm bandpath ﬁlter. Dashed lines
at RPC symbol indicate fast axes of quadrants (slit direction). Symbols
Q1-Q4 indicate the number of each quadrant.
161119-3 Iwami et al. Appl. Phys. Lett. 101, 161119 (2012)
In conclusion, a radial polarization converter that made
use of a gold nanoslit pattern was designed and fabricated on
a glass substrate. The size of the RPC ranged from 10 to
100 lm in length and was 445 nm in thickness. With proper
slit width and gold thickness design, the RPC will be useful in
a large portion of the visible spectrum ranged 442 to 780 nm.
Improvement of the low transmittance of nanoslit is the fur-
ther challenge to a practical device, which will be achieved by
optimizing slit pitch, modifying pattern from slit to for exam-
or possibly adopting a technique of electro-
magnetically induced transparency.
We are expecting to
broaden the usefulness to the ultraviolet region by changing
the plasmonic material from gold to silver or aluminum.
The authors would like to thank the Grant-in-Aid for
Young Researchers (A) No. 23686016 from the Japanese So-
ciety for the Promotion of Science (JSPS) for ﬁnancially sup-
porting this research. A part of this research was supported
by the “Nanotechnology Network” of the Ministry of Educa-
tion, Culture, Sports, Science and Technology (MEXT), Ja-
pan for electron beam lithography at Toyota Technological
Institute and VLSI Design and Education Center (VDEC),
the University of Tokyo.
T. Wilson, F. Massoumian, and R. Juskaitis, Opt. Eng. 42, 3088 (2003).
V. G. Niziev and A. V. Nesterov, J. Phys. D: Appl. Phys. 32, 1455 (1999).
Q. Zhan, Opt. Express 12, 3377 (2004).
W. Chen, D. C. Abeysinghe, R. L. Nelson, and Q. Zhan, Nano Lett. 9,
Y. Saito, N. Hayazawa, H. Kataura, T. Murakami, K. Tsukagoshi, Y.
Inouye, and S. Kawata, Chem. Phys. Lett. 410, 136 (2005).
H. Kang, B. Jia, J. Li, D. Morrish, and M. Gu, Appl. Phys. Lett. 96,
Y. Liu, D. Cline, and P. He, Nucl. Instrum. Methods Phys. Res. A 424,
A. Nesterov and V. Niziev, J. Phys. D: Appl. Phys. 32, 2871 (1999).
L. Cai, J. Zhang, W. Bai, Q. Wang, X. Wei, and G. Song, Appl. Phys.
Lett. 97, 201101 (2010).
H. Ren, Y.-H. Lin, and S.-T. Wu, Appl. Phys. Lett. 89, 051114 (2006).
I. Nishiyama, N. Yoshida, Y. Otani, and N. Umeda, Meas. Sci. Technol.
18, 1673 (2007).
G. Machavariani, Y. Lumer, I. Moshe, A. Meir, and S. Jackel, Opt. Com-
mun. 281, 732 (2008).
H. Kawauchi, Y. Kozawa, S. Sato, T. Sato, and S. Kawakami, Opt. Lett.
33, 399 (2008).
Z. Bomzon, G. Biener, V. Kleiner, and E. Hasman, Opt. Lett. 27, 285
M. Beresna, M. GeceviCius, P. G. Kazansky, and T. Gertus, Appl. Phys.
Lett. 98, 201101 (2011).
S. Singh-Gasson, R. D. Green, Y. Yue, C. Nelson, F. Blattner, M. R. Suss-
man, and F. Cerrina, Nat. Biotechnol. 17, 974 (1999).
L. Schares, D. M. Kuchta, and A. F. Benner, in 2010 18th IEEE Sym-
posium on High Performance Interconnects (IEEE, 2010), pp. 104–
S.-Y. Hsu, K.-L. Lee, E.-H. Lin, M.-C. Lee, and P.-K. Wei, Appl. Phys.
Lett. 95, 013105 (2009).
P. F. Chimento, N. V. Kuzmin, J. Bosman, P. F. A. Alkemade, G. W.’t
Hooft, and E. R. Eliel, Opt. Express 19, 24219 (2011).
M. Shcherbakov, M. Dobynde, T. Dolgova, D.-P. Tsai, and A. Fedyanin,
Phys. Rev. B 82, 193402 (2010).
H. Li, X. Luo, Z. Guo, K. Wen, and L. Yan, Mod. Phys. Lett. B 26,
N. Liu, L. Langguth, T. Weiss, J. K€
astel, M. Fleischhauer, T. Pfau, and H.
Giessen, Nature Mater. 8, 758 (2009).
161119-4 Iwami et al. Appl. Phys. Lett. 101, 161119 (2012)