APPLIED PHYSICS LETTERS 98, 151107 共2011兲
A half wave retarder made of bilayer subwavelength metallic apertures
Z. Marcet,1,2 H. B. Chan,2,a兲 D. W. Carr,3 J. E. Bower,4 R. A. Cirelli,4 F. Klemens,4
W. M. Mansfield,4 J. F. Miner,4 C. S. Pai,4 and I. I. Kravchenko5
1
Department of Physics, University of Florida, Gainesville, Florida 32611, USA
Department of Physics and William Mong Institute of Nano Science and Technology, The Hong Kong
University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong
3
Symphony Acoustics, Rio Rancho, New Mexico 87124, USA
4
Bell Laboratories, Alcatel-Lucent, Murray Hill, New Jersey 07974, USA
5
Center for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge,
Tennessee 37830, USA
2
共Received 26 January 2011; accepted 27 March 2011; published online 13 April 2011兲
We demonstrate a half wave plate whose principle of operation is based on the strong evanescent
field coupling between two metal layers with arrays of subwavelength slits. The device is divided
into two kinds of pixels in which the slits are oriented in orthogonal directions. By tuning the phase
delay of the transmitted light through the lateral displacement between the top and bottom layers,
the polarization of linearly polarized light at 1.55 ␮m can be rotated by up to 90°. The polarization
extinction ratio of the transmitted light exceeds 22 dB. © 2011 American Institute of Physics.
关doi:10.1063/1.3579245兴
Nano optical devices exploit the interaction between
near field and far field radiations in subwavelength metallic
and dielectric structures to achieve novel optical capabilities.
For example, the excitation of surface plasmons is responsible for the remarkable transmission1 and beaming
properties2 of periodic subwavelength metallic apertures.
The subwavelength features on the metal surface can be engineered to control the strong evanescent fields associated
with the surface excitations.3 Here, we describe the fabrication of a half wave plate where the retardation in one polarization is achieved through tailoring the evanescent field
coupling between two metal layers with arrays of subwavelength apertures.
Polarization state controllers are essential components
in optical systems. Conventional wave plates rely on birefringent crystals to impart a phase delay to one of the polarizations. The wavelength of operation is determined by the
material properties and therefore is not readily tunable.
Moreover, it is challenging to integrate them with other
optical components due to their bulkiness. Alternative
ways to control the polarization have recently emerged, including liquid-crystals,4–6 waveguides,7–10 micromechanical
devices,11,12 and chiral metamaterials.13 In particular, chiral
metamaterials offer a compact, planar platform to rotate the
polarization of transmitted or reflected light. However, the
difficulty in generating strong optical activity has limited the
polarization azimuth rotation to ⬃30°.14 The device reported
here is a half wave retarder based on the enhanced transmission through bilayer subwavelength metallic slit arrays at
resonance. By utilizing the controllable phase delay of the
transmitted light through the lateral displacement of the two
arrays and the strong dependence of the transmission on the
polarization of the incident light, the polarization of linearly
polarized light at 1.55 ␮m can be rotated by up to 90° in the
zero order transmission. The polarization extinction ratio of
the transmitted light exceeds 22 dB.
a兲
Electronic mail: hochan@ust.hk.
0003-6951/2011/98共15兲/151107/3/$30.00
Our sample is made of two 0.2 ␮m thick aluminum
films separated by 0.34 ␮m, each with subwavelength slit
apertures arranged in specific orientations. The slits in the
aluminum layers are created by reactive ion etching with
a photoresist etch mask defined by deep ultraviolet
lithography.15 As shown in Fig. 1共a兲, the top aluminum layer
is divided into a checkerboard pattern with two kinds of
square pixels. Each pixel 共25⫻ 25 ␮m2兲 consists of an array
of long slits with width of 0.49 ␮m and periodicity of
1 ␮m. The orientations of the slits are perpendicular to each
other in adjacent pixels. For pixels labeled as type 1 in Fig.
1共a兲, the slit array is oriented along the y direction. In the
bottom layer, an identical array of slits is present, almost
laterally aligned to the top array 关Fig. 1共b兲兴. For type 2 pixels
with slits along the x direction, the array of slits in the bottom layer is laterally shifted from the top layer by about half
the period 关Fig. 1共c兲兴. The two aluminum layers are completely surrounded by silicon oxide. Fabrication of the metal
layers was performed on a silicon substrate. Completed
structures were then transferred onto a quartz wafer for optical measurements.15
To understand the working principle of the above structure as a half wave retarder, we first consider the transmission properties of a nonpixilated, single-layer structure with
all slits oriented along the y direction. Linearly polarized
light with electric field along the x 共y兲 axis is denoted as Ex
共Ey兲 polarized. We measured a transmission peak of up to
75% at 1.55 ␮m for Ex polarized light. Taking into account
the refractive index of silicon oxide 共n = 1.47兲, the peak
wavelength is slightly higher than the periodicity of the slit
array 共1 ␮m兲. It is well-known that when Ex polarized light
impinges on metallic gratings oriented along the y direction,
the transmission can exceed the diffraction limit at wavelengths comparable to the period of the slits. Such extraordinarily high transmission has been attributed to the excitation
of surface plasmons,1 surface evanescent waves16 and/or
guided modes.17 When two such single layers are placed in
close proximity so that the strong evanescent fields on their
surface couple, both the intensity15 and the phase18 of Ex
98, 151107-1
© 2011 American Institute of Physics
Downloaded 02 Aug 2011 to 143.89.149.207. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
151107-2
Marcet et al.
FIG. 1. 共Color兲 共a兲 Top view optical image of the half-wave retarder. Cross
sectional scanning electron micrograph of 共b兲 aligned and 共c兲 misaligned
double layer slit arrays, in structures similar to the sample. 共d兲 Measured
transmission spectrum of Ex-polarized 共red line兲 and Ey-polarized 共blue line兲
light through the pixilated structure. 共e兲 RCWA calculations of transmission
through both aligned 共red兲 and misaligned 共blue兲 double layer arrays. 共f兲
Phase delay of transmitted light changes by about ␲ radians when the lateral
shift between the top and bottom arrays is increased from zero to half the
period. Field distribution calculated by RCWA for 共g兲 aligned and 共h兲 misaligned double layer arrays. Light at 1.55 ␮m is incident from the bottom
with unit amplitude. The color scale is associated with the magnitude and
direction of the magnetic field that is directed into and out of the page. The
electric field that points along the page is not shown.
polarized light can be controlled by the lateral shift in the
two layers. The transmitted intensity for Ey polarized light, in
contrast, is lower by several orders of magnitude and is practically negligible. Therefore, the pixilated bilayer structure
acts as a polarization beam splitter followed by a polarization
beam combiner, in the sense that Ex polarized light passes
through type 1 pixels and Ey polarized light passes through
type 2 pixels. As we will explain below, the phases acquired
by these two components in this process differ by ␲. Figure
1共d兲 shows the measured transmission through the pixilated
structure for Ex 共red line兲 and Ey 共blue line兲 polarizations at
normal incidence, measured with a Fourier transform infrared specrometer. The transmitted intensity depicted by the
red 共blue兲 line originates from type 1 共2兲 pixels because light
impinging on the other kind of pixels is entirely blocked. For
instance, the maximum transmission of ⬃32% for Ey polarized light in Fig. 1共d兲 originates from ⬃64% transmission
through type 2 pixels. The transmission of both types of
pixels for the corresponding polarization components remains strong due to the efficient coupling of the surface
waves excited on the two layers.15 It should be noted that the
lateral shift s1 between the two slit arrays within pixel 1 is
about 0.1 ␮m instead of exactly perfectly aligned. We
choose a sample with this value of s1 共out of many samples
covering a wide range of s1兲 so that the transmissions of the
Ex polarization through pixel 1 and the Ey polarization
through pixel 2 are identical at our laser wavelength of
Appl. Phys. Lett. 98, 151107 共2011兲
ជ I of the incident light makes an angle
FIG. 2. 共Color兲 共a兲 The electric field E
␾ with the x-axis of the pixilated slit array. 共b兲 The y component of the
electric field 共blue arrow兲 acquires an extra half-wave phase delay relative to
the x component 共red arrow兲, leading to a rotation of the electric field from
ជ I to Eជ T, by an angle of 2␾. 关共c兲–共e兲兴 Transmission measurement 共crosses兲 at
E
a wavelength of 1.55 ␮m as a function of the angle ␣ of the second polarizer 共before the photodetector兲 relative to the first polarizer 共that polarizes
the incident light兲. Full scale transmitted intensity is 16%. The pixilated slit
array is inserted between the two polarizers at ␾ equal to 共c兲 0°, 共d兲 15°, and
共e兲 45°. The purple lines represent the expected cos2共␣ − 2␾兲 dependence of
the transmitted light intensity on ␣.
1.55 ␮m 关Fig. 1共d兲兴. As described later, equal transmission
intensity of the Ex and Ey polarizations is a crucial factor in
determining the performance of this device as a half wave
retarder. Figure 1共e兲 shows the transmitted intensity calculated with rigorous coupled wave analysis 共RCWA兲 共Ref. 19兲
using the slit array dimensions of the two kinds of pixels and
tabulated optical properties of aluminum.20 The calculated
transmission is in good qualitative agreement with measurement.
The different lateral shifts in the orthogonal pixels produce a phase delay that is close to half-wave in the transmission of the two polarizations. Figure 1共f兲 shows that the calculated phase of the transmitted light varies by almost ␲ as
the top array is laterally shifted from the bottom array between zero and half the period. To achieve a phase shift
between light transmitted through pixels 1 and 2 that is
closer to ␲, the sample was tilted about the y-axis by a small
angle of 2.5°. The transmission intensity remains largely unchanged. Figures 1共g兲 and 1共h兲 compare the spatial distribution of the magnetic fields for the two pixels, showing the
same field directions on the incident side 共bottom兲 but opposite field directions on the transmitted side 共top兲. Such phase
delay can be understood in terms of the surface electromagnetic fields on the two individual layers coupling differently
depending on whether the fields are parallel or opposite to
each other, as explained in Ref. 18. The half-period lateral
shift in the type 2 pixels reverses the direction of the electromagnetic field for the transmitted Ex polarization relative
to the Ey polarization that was transmitted through the type 1
pixels.
We use a 1.55 ␮m laser to demonstrate the capability of
the pixilated structure to rotate the polarization. Through a
shadow mask, the linearly polarized, collimated laser beam
illuminated a sample area of 400⫻ 400 ␮m2 共about 256 pixels兲. As shown in Fig. 2共a兲, the x-axis of the sample was
oriented at an angle ␾ relative to the electric field of the
linearly polarized incident light. A second linear polarizer in
Downloaded 02 Aug 2011 to 143.89.149.207. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
151107-3
Appl. Phys. Lett. 98, 151107 共2011兲
Marcet et al.
front of a photodetector serves as an analyzer to measure the
polarization of light transmitted through the pixilated sample
relative to the polarization of the incident light. Depending
on ␾, the electric field of the incident light decomposes into
different x and y components along the sample axis 关Fig.
2共b兲兴. For ␾ = 0°, all the light is transmitted through the type
1 pixels. The polarization of the transmitted light remains
unchanged, as shown in Fig. 2共c兲 by the square cosine dependence of the measured intensity at the photodetector on
the angle ␣ of the second polarizer with respect to the first
one. When ␾ is changed to 45°, half of the incident light is
transmitted through the type 1 pixels. The other half is transmitted through the type 2 pixels with the direction of the
electromagnetic field reversed, as illustrated in Fig. 2共b兲.
Consequently, the transmitted light remains linearly polarized with polarization rotated by 90°, as shown in Fig. 2共e兲.
The structure therefore acts as a half wave plate that rotates
the polarization of the incident linearly polarized light by 2␾.
Figure 2共d兲 shows measured polarization rotation of 30° for
␾ = 15°. For all our measurements at different ␾, the polarization extinction ratio of the transmitted light is found to
exceed 22 dB. It is important to note that since the maximum
intensities in Figs. 2共c兲–2共e兲 are largely independent of ␾,
the pixilated bilayer structure is not a simple polarizer that
projects the polarization component along a certain direction,
but is a polarization controller that imparts a polarization
rotation to the transmitted light.
Unlike conventional wave plates, the half-wave retarder
described here is not based on birefringence of the material.
The wavelength of operation can be chosen by creating slit
arrays with different periodicity to tune the surface wave
resonance. Using similar principles, quarter wave plates can
also be created to generate elliptically polarized light from
linearly polarized incident light. Most importantly, it offers
the possibility of dynamical control of polarization in future
designs if one of the two metal plates can be suspended to
allow controlled motion between the two layers. One such
configuration involves a lower metal layer fixed to the substrate and a movable upper layer supported by springs. When
the top layer moves in the x direction by half the array period, the top and bottom slit arrays become aligned in both
types of pixels. As a result, the polarization rotating effect
can be completely turned off. Such capability of dynamic
polarization control by nanomechanical motion could prove
to be useful in optical systems, telecommunication networks
and interchip data communications.
This work was supported by the National Science Foundation under Grant No. ECS-0621944. Z. Marcet acknowledges support from South East Alliance for Graduate Education and the Professoriate. A portion of this research was
conducted at the Center for Nanophase Material Sciences,
which is sponsored at Oak Ridge National Laboratory by the
Division of Scientific User Facilities, U.S. Department of
Energy.
1
T. W. Ebbesen, H. J. Lezec, H. F. Ghaemi, T. Thio, and P. A. Wolff, Nature
共London兲 391, 667 共1998兲.
2
H. J. Lezec, A. Degiron, E. Devaux, R. A. Linke, L. Martin-Moreno, F. J.
Garcia-Vidal, and T. W. Ebbesen, Science 297, 820 共2002兲.
3
W. L. Barnes, A. Dereux, and T. W. Ebbesen, Nature 共London兲 424, 824
共2003兲.
4
I. Moreno, J. L. Martinez, and J. A. Davis, Appl. Opt. 46, 881 共2007兲.
5
H. Ren and S.-T. Wua, Appl. Phys. Lett. 90, 121123 共2007兲.
6
P. Wang and A. Asundi, Rev. Sci. Instrum. 79, 063105 共2008兲.
7
N. R. Erickson and R. M. Groslein, IEEE Trans. Microwave Theory Tech.
55, 2495 共2007兲.
8
N.-N. Feng, R. Sun, J. Michel, and L. C. Kimerling, Opt. Lett. 32, 2131
共2007兲.
9
H. Fukuda, K. Yamada, T. Tsuchizawa, T. Watanabe, H. Shinojima, and S.
Itabashi, Opt. Express 16, 2628 共2008兲.
10
T. Mangeat, L. Escoubas, F. Flory, L. Roussel, M. D. Micheli, and P.
Coudray, Opt. Express 15, 12436 共2007兲.
11
C. Pu, L. Y. Lin, E. L. Goldstein, N. J. Frigo, and R. W. Tkach, IEEE
Photonics Technol. Lett. 12, 1358 共2000兲.
12
S. Sumriddetchkajorn and N. A. Riza, Appl. Opt. 41, 3506 共2002兲.
13
M. Kuwata-Gonokami, N. Saito, Y. Ino, M. Kauranen, K. Jefimovs, T.
Vallius, J. Turunen, and Y. Svirko, Phys. Rev. Lett. 95, 227401 共2005兲.
14
A. Papakostas, A. Potts, D. Bagnall, S. L. Prosvirnin, H. J. Coles, and N.
I. Zheludev, Phys. Rev. Lett. 90, 107404 共2003兲.
15
H. B. Chan, Z. Marcet, K. Woo, D. B. Tanner, D. W. Carr, J. E. Bower, R.
A. Cirelli, E. Ferry, F. Klemens, J. Miner, C. S. Pai, and J. A. Taylor, Opt.
Lett. 31, 516 共2006兲.
16
H. J. Lezec and T. Thio, Opt. Express 12, 3629 共2004兲.
17
J. A. Porto, F. J. Garcia-Vidal, and J. B. Pendry, Phys. Rev. Lett. 83, 2845
共1999兲.
18
Z. Marcet, J. W. Paster, D. W. Carr, J. E. Bower, R. A. Cirelli, F. Klemens,
W. M. Mansfield, J. F. Miner, C. S. Pai, and H. B. Chan, Opt. Lett. 33,
1410 共2008兲.
19
M. G. Moharam, D. A. Pommet, E. B. Grann, and T. K. Gaylord, J. Opt.
Soc. Am. A 12, 1077 共1995兲.
20
Handbook of Optical Constants of Solids, edited by E. D. Palik 共McGrawHill, New York, 1950兲.
Downloaded 02 Aug 2011 to 143.89.149.207. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions