OPEN

SUBJECT AREAS:
SUB-WAVELENGTH
OPTICS
OPTOELECTRONIC DEVICES AND
COMPONENTS
PHOTONIC DEVICES

High-Polarization-Discriminating
Infrared Detection Using a Single
Quantum Well Sandwiched in Plasmonic
Micro-Cavity
Qian Li, ZhiFeng Li, Ning Li, XiaoShuang Chen, PingPing Chen, XueChu Shen & Wei Lu

Received
29 May 2014
Accepted
20 August 2014
Published
11 September 2014

Correspondence and
requests for materials
should be addressed to
W.L. (luwei@mail.sitp.
ac.cn)

National Laboratory for Infrared Physics, Shanghai Institute of Technical Physics, Chinese Academy of Sciences, 500 Yutian Road,
Shanghai 200083, P. R. China and Synergetic Innovation Center of Quantum Information & Quantum Physics, University of Science
and Technology of China, Hefei, Anhui 230026, P. R. China.

Polarimetric imaging has proved its value in medical diagnostics, bionics, remote sensing, astronomy, and
in many other wide fields. Pixel-level solid monolithically integrated polarimetric imaging photo-detectors
are the trend for infrared polarimetric imaging devices. For better polarimetric imaging performance the
high polarization discriminating detectors are very much critical. Here we demonstrate the high infrared
light polarization resolving capabilities of a quantum well (QW) detector in hybrid structure of single QW
and plasmonic micro-cavity that uses QW as an active structure in the near field regime of plasmonic effect
enhanced cavity, in which the first photoelectric conversion in such a plasmonic micro-cavity has been
realized. The detectors extinction ratio reaches 65 at the wavelength of 14.7 mm, about 6 times enhanced in
such a type of pixel-level polarization long wave infrared photodetectors. The enhancement mechanism is
attributed to artificial plasmonic modulation on optical propagation and distribution in the plasmonic
micro-cavities.

ight polarization provides richer sets of descriptive physical constraints for the interpretation of the imaged
scene in addition to intensity and wavelength1. Polarimetric imaging has been intensively developed in
various fields of applications. In modern medicine it could become a fast and accurate optical method for
detecting cancer and determining the stage of the disease2. In bionics it helps making machine vision to detect
motion/collision of moving objects1. In military reconnaissance it is used to single out man-made objects from
camouflages when there is little contrast in intensity imagery3. In remote sensing it can even monitor the change
of the earth because human activities often change the polarization state of the electromagnetic (EM) energy
reflected or emitted from the surface of the earth4. Conventional polarization measurements employ timesequenced or simultaneous measurements by using various stand-alone polarizers and retarders to obtain the
polarization Stokes parameters5. Free standing polarizers extinction ratio can easily reach 100 or even higher,
satisfying the most requirements of static and indoor measurements. For fast moving objects or field measurements, recent developments involve micro-polarizers that are directly integrated into the structure of individual
detecting pixels to form polarization detecting focal plane arrays (FPAs) so that real-time fast polarimetric
imaging are very much facilitated to sustain the applications in, for example, air- or satellite-born remote sensing.
The extinction ratio of micro-polarizers, however, is far less than their stand-alone counterparts. People have
been seeking high extinction ratio wire-grid micro-polarizers to acquire more sensitive all-solid polarimetric
imagers. Once succeeded, especially in long wave infrared (LWIR) with the wavelength longer than about 8 mm
where the polarization signatures are dominated by emission from target and can be very stable in time when
scene temperatures are stable3, application opportunities will be widely open, for example, like live-tumor
diagnosis and any other meticulous target recognitions. In visible range of the spectrum the extinction ratio of
micro-polarizers has reached ,606. In infrared (IR) range, however, the extinction ratio is only 2.3 for IR
detectors7 (expressed as contrast ratio (diattenuation) 5 (Rx-Ry)/(Rx1Ry) . 40% in the referenced paper),
and will saturate to only about 10 (responsivity contrast ,82%) even when the pixel size is enlarged to 100 mm8.
The reason might be attributed to the cross-talks between neighbor pixels or the wave-leakage through pixel edges
SCIENTIFIC REPORTS | 4 : 6332 | DOI: 10.1038/srep06332
SCIENTIFIC REPORTS SREP-14-03585.3d 1/9/14 13:20:44

www.nature.com/scientificreports
that contaminates the polarization as for imaging FPAs the detecting
pixel is completely exposed within the light beam and the longer
wavelength in IR range will make the cross-talk/wave-leakage much
easier.
Recent intensively studied plasmonics911 has demonstrated
strong capabilities on manipulating light-matter coupling by modifying the incident optical field and forming resonant condition for
photoelectric conversion in near- and far-field regimes. Micro-cavities have displayed an excellent avenue to greatly enhance the field
intensity by eliminating far-field effects, which has made a direct
observation of individual quantum particles12, leading to a very high
intense nanolaser beams13, etc. In this article we report the achievement on the combination of plasmonics and micro-cavities with the
insertion of an active layer for photo-detections. By manipulating the
localized surface plasmon (LSP) mode and the surface plasmon
polariton (SPP) mode we have fulfilled the modulations of wave
propagation and field distribution of IR photons in all-solid wiregrid polarizer scheme, realizing high extinction ratio of 65 at LWIR of
14.7 mm.

covering area is 230 3 200 mm2. The top and bottom electrodes as
well as the mesa-isolation boundary are shown in Figure 1 b. Sample
preparation processes can be found in Supporting Information (SI)
S1. The small distance, about 900 nm, between the two Au layers
makes them couple with each other and form a plasmonic microcavity, squeezing the incident light into the cavity. When a nonpolarized light illuminates, the transverse magnetic (TM) and
transverse electric (TE) polarized light will have different coupling
properties. At the wavelength of plasmonic resonance for TM light,
most of TM light will couple into the cavity. The light wavevector will
be transformed from z to x direction, generating strong electrical
components along z direction, Ez, which can be absorbed by QW
with inter-subband transition processes14. TE light will have different
resonant wavelength and will be weakly coupled. It will have the
electrical components in y direction, which is forbidden for
the inter-subband transition by the quantum selection rule14. The
inter-subband transition in QW induces photocurrents in the biased
PCQWID, with which we obtain high contrast responses to TM and
TE light.

Results
Device structure of the PCQWID. The configuration and operation
principles of our grating Plasmonic micro-Cavity Quantum Well
Infrared Detector (PCQWID) are shown in Figure 1, together with
the SEM images of the fabricated PCQWID and the cleaved facet. A
single GaAs/Al0.13Ga0.87As QW with the thickness of 207 nm and
the two contact layers of 190 nm and 490 nm are sandwiched
between the upper Au periodic grating and the bottom Au
reflection layer whose thicknesses are both 100 nm. The grating

Polarization discriminating ability of PCQWID. In ideal cases the

TM light can generate photocurrent while the TE light is forbidden,
the non-ideal factors, however, in the structural and fabrications of
the PCQWID will introduce scatterings, creating Ez for TE light and
reducing Ez for TM light. They are non-intrinsic photocurrents in
PCQWID, which will contaminate the polarization. To
quantitatively analyze the polarization discriminating capability,
we introduce Equation 1 to express the extinction ratio at the
cavity resonant wavelength in our hybrid structured PCQWID,
defined as

Figure 1 | Schematic diagraph of the PCQWID. (a) The incident TM and TE polarized lights have different coupling behaviors to form different cavity
modes. The figure shows the electric field distributions of TM LSP (Localized Surface Plasmon) mode at its calculated resonant wavelength of 14.9 mm,
TM SPP (Surface Plasmon Polariton) mode at its calculated resonant wavelength of 12.4 mm, and TE modes at the same wavelength of 14.9 and 12.4 mm,
respectively. The right panel is the schematics of the active layer. The first two energy levels of the single QW are separated of 97 meV, corresponding to a
wavelength of 12.7 mm. (b) SEM image of the cleaved facet of the cavity structure. Geometrical parameters are designed as p 5 9.2,10.6 mm,
s54.9,5.9 mm. Both the top and bottom Au layers are 100 nm thick. (c) SEM image of a fabricated PCQWID, the grating covering area is 230 3
200 mm2.
SCIENTIFIC REPORTS | 4 : 6332 | DOI: 10.1038/srep06332

www.nature.com/scientificreports

r~
~

Rx P0,TM (1{R) (1{e{aQW s ) G

~
Ry
P0,TE (1{e{aQW L )
0:35|jEz,TM j2 (1{S) (1{R) (1{e{aQW s ) G


Ey,TE 2 S (1{e{aQW L )

where r is the polarization extinction ratio, Rx and Ry are the detector

responses to the lights polarized along x (TM) and y (TE) directions,
respectively. P0,TM and P0,TE are the effective optical power in TM
and TE light within the cavity, respectively. R is the lateral reflectivity
at the interfaces of Fabry-Perot resonators formed between the
single-metal and double-metal regions15. aQW is the QW
absorption coefficient. s and L are the widths of metal strip and
detector mesa, respectively. G and S are enhancement and
scattering factors that will be described in the later text. jEz,TM j2

2
and Ey,TE  are the simulated average relative intensities of
electric field by taking the input electric field intensity in the
incident optical beam as the unity. 0.35 is the transmissivity of the
polarizer used in the experiment. The detailed explanation and
calculations of Equation 1 are described in SI S5. According to the
selection rule, only Ez component contributes to the photo-electric
conversion. So P0,TM and P0,TE are the effective Ez in TM and TE
light, respectively. In TE light originaly there is no Ez component.
P0,TE is induced by the Ez scattered from Ey. We introduce a
scattering factor S whose value represents the ratio of Ez
component scattering into Ey component for TM light (TM loss)
and Ey scattering into Ez component for TE light (TE leakage),
which are resulted from the extrinsic structural imperfection. As
an approximation, we take the same S for the above two scattering
processes. We describe the detailed defination and calculation of S
and aQW in SI S2S3. The last numerator and denominator represent
the photon numbers absorbed by the QW for TM and TE polarized
light, respectively.
Photo-responsivity spectrum of the PCQWID and analysis. We
investigate the polarization discriminating ability by illuminating the
device with polarized IR light (with a polarizer whose extinction ratio
is over 600 in the studied waveband) in Fourier transform IR
spectrometer and by constructing the polarization azimuth dependence of photocurrent. The typical photocurrent responsivity spectra
for TM and TE light are shown in Figure 2a in red and blue lines. Also
shown is that of a 45 degree edge facet light-coupling device made
from the same QW wafer as the reference in green line, which is the
intrinsic QW absorption curve14. Photocurrent generation in
PCQWID involves three processes including the light coupling
into cavity structure and forming cavity modes, the scatterings of
TM loss and TE leakage, and the photo-electric conversion in the
QW. To understand the optical behaviors in the coupling cavity from
photo-responsivity spectra, we need to extract the cavity characteristics by eliminating the absorption peak of the QW. For TM
and TE light, the QW absorption curves in this cavity structure
can be expressed as 1{e{aQW (l) s and 1{e{aQW (l) L , respectively,
where aQW(l) can be calculated from the photo-responsivity
spectrum (green curve in Figure 2a) and the blackbody response of
the 45 degree facet device. The detailed calculations can also be found
in SI S4. We divide the PCQWID photo-responsivity curves with the
respective absorptions, yielding the measured TM and TE mode
spectra, as presented in the upper panels of Figure 2b and 2c. The
spectrum of TM light demonstrates an obvious cavity mode
behavior. There are two peaks at about 12.4 and 14.7 mm, which
are attributed to SPP and LSP cavity modes16, respectively, in good
agreement with the theoretical simulation of electric field intensity
jEz j2 in the lower panel of Figure 2b that shows the SPP and LSP
peaks at 12.8 and 14.9 mm. We will discuss these two cavity modes in
SCIENTIFIC REPORTS | 4 : 6332 | DOI: 10.1038/srep06332

the later text in details. As for TE light shown in the upper panel of
Figure 2c, the very weak response (three orders of magnitude smaller
than TM, near extinction condition in experiment) causes large
noises, especially in the wavelength range larger than 15 mm, as
indicated by the error bars. The overall trend shows a peak-like
 2
shape, similar with the calculated Ey  as shown in the lower
panel of Figure 2c, although the peak position shifts by 1 mm.
Using Finite Difference Time Domain (FDTD) method and taking
the input electric field intensity in the incident optical beam as the
unity, the calculated distribution of the square of relative electric field
components of TM light, jEz,TMj2, are shown in Figure 2d and 2e for
both LSP and SPP modes. In which the average value of jEz,TMj2 is:
1
jEz,TM j ~
A
2

397nm s=2

jEz (x,z)j2 dx dz

190 nm {s=2

where A is the QW absorption area, A55.5 3 0.2 mm2 as indicated by

the black boxes in Figure 2d and 2e. The calculated results show that
jEz,TM j2 5 14.7 in LSP mode. The ratio of jEz,TM j2 between LSP and
SPP is 14.7/3.354.5, consistent with the experimental intensity ratio
value of 4 that can be obtained from Figure 2b. As one can also see
that in LSP mode the electrical field is localized under the metal strip
to form Fabry-Perot resonances except those strongest parts condensed at the edges (Figure 2d) while in SPP mode it is weak and
dispersive within the cavity with its strongest parts at the surface of
the dielectric media (Figure 2e). The strong localization effect in LSP
mode makes much higher jEz,TMj2, whose highest value can over 30,
about 6 times higher than the maximum value of 5 in SPP mode
inside the cavity. Similarly, the average relative values of electric field

2
component for TE light, Ey,TE  , can also be calculated, which is 0.3
and 1.0 at the calculated TM LSP and SPP mode peak wavelengths of
14.9 and 12.8 mm, respectively. It is interesting to see that at the
wavelength of TM LSP mode the average relative intensity of TM
light is much higher than that of TE light (14.7/0.3), while at the
wavelength of TM SPP mode the intensity ratio is only 3.3/1.0.
According to Equation 1, the higher TM/TE ratio will yield higher
extinction. This is one of the reasons we can obtain high extinction
ratio in such a plasmonic micro-cavity structure, especially at the
peak wavelength of LSP mode.
Experimental polarization characteristics of PCQWID. The
experimental polarization characteristics of our PCQWID are
shown in Figure 3a in black solid dots. The data are taken from the
average photocurrent values at the wavelength range between 14.2
14.9 mm as indicated by the black dashed lines in Figure 3b, owing to
the large fluctuations in the extinction ratio values (refer to Figure 3c
for the measured extinction ratio values). Figure 3b is the
photocurrent curves changing with the azimuth positions of the
polarizer. The largest extinction ratio does not occur at the peak
wavelength of 13.5 mm, at which the photocurrent is the largest.
This is because that the photocurrent curves are modulated by
both cavity modes and the QW absorption peak. In our
experiment the QW absorption peak is at about 12.5 mm (green
line in Figure 2a) and the cavity mode peak is at 14.7 mm. The
averaged experiment values in solid dots agree with the normalized
square of sine function curve in red line, which is a standard
characteristics of a polarizer, indicating a perfect polarization
discriminating behavior of the device. Figure 3c is the measured
and simulated extinction ratio dependence on the wavelength. The
measured curve in blue line is obtained by dividing the two
photocurrent response spectra at polarizer angles of 90 and 0
degrees in Figure 3b. The maximum extinction ratio appears at the
wavelength around 14.7 mm, close to the simulated TM LSP peak
wavelength. The simulated curve is from the calculation of the
3

www.nature.com/scientificreports

Figure 2 | Photo responding properties and cavity mode behaviors of PCQWID. (a) Typical photocurrent spectra for TM (red) and TE (blue) light,
together with that of 45 degree edge facet device (green). (b) Experimental (upper, blue) and simulated (lower, red) cavity mode dependence on
wavelength for TM light. The experimental curve is the photo-responsivity spectrum divided by the quantum well absorption curve (1{e{aQW (l) s ). (c)
Similar curves as (b) for TE light. The error bars show the large noise at wavelength larger than 15 mm in experiment. (d) Calculated intensity distribution
of the square of electric field component in z direction of LSP cavity mode for TM light, | Ez,TM | 2, within the cavity. (e) Similar distribution as (d) for SPP
cavity mode.

coupled TM and TE optical power inside the cavity and their

absorptions (see SI S5). It agrees well with the experiments,
reproducing the two peaks of LSP and SPP at the respective
wavelength and their relative values, verifying the analysis that we
proposed for describing the polarization extinction ratio dependence
in PCQWID, which provides guidance for further optimization of
the extinction ratio.
The above cavity and polarization results quantitatively demonstrate the cavity-mode effects in PCQWID and strongly suggest
that for a high extinction ratio one should choose the PCQWID
responding peak wavelength to be at that of the LSP mode. Moving
the TE peak to shorter wavelength by decreasing the grating period
while retaining the metal strip width will decrease the TE intensity at
the LSP wavelength and will further increase the extinction ratio. In
addition, tuning the QW absorption peak to the LSP resonant wavelength by adjusting the intersubband transition energy is also important so as to increase the QW absorption, according to Equation 1.
Cavity mode of LSP and SPP. To clarify the cavity mode behavior of
LSP and SPP, we have studied the mode variations against the grating
parameters. LSP resonance is resulted from local electron density
SCIENTIFIC REPORTS | 4 : 6332 | DOI: 10.1038/srep06332

oscillation around a single metal strip or sphere17, and is strongly

dependent on metal size16. It couples the incident light from free
space, transferring the light energy into the cavity, supporting the
propagating TM wave to form Fabry-Perot resonantors18. This kind
of surface plasmons, similar with those in metal holes and spherical
particles16, does not require the wave vector matched excitation and
is independent to periodicities. SPP resonance is propagating EM
waves coupled to the electron plasma of a conductor at a dielectric
interface18. In our one-dimensional grating structure, the periodic
array provides the necessary plane wavevector to excite SPP mode.
From FDTD analysis, the LSP resonant wavelength increases with
the increasing width s and almost unchanged with period p. While
the SPP resonant wavelength is almost independent on s and
increases with the increasing p. These properties have been
reproduced in series of PCQWID devices as shown in Figure 4a
and 4b, where the blue spheres are the experimental peak
wavelengths of LSP and SPP modes taken from the mode spectra
(One of the measured mode spectra for p510.6 mm and s55.5 mm is
shown in the upper panel of Figure 2(b). The colored maps are the
calculations of the optical power in the cavity, which indicates
obvious concentrations at the resonant wavelengths of LSP and
4

www.nature.com/scientificreports

Figure 3 | Polarization characteristics and extinction ratio of PCQWID. (a) Measured photocurrent average intensity at l514.2,14.9 mm against
incident light polarization angle. The red line shows normalized calculation of the square sine curve, showing a typical polarizer behavior. (b) the original
photocurrent spectra vs. wavelength at different polarization angles. The photocurrent peak does not corresponding to the highest extinction ratio, as
explained in the context. (c) Measured and simulated extinction ratio spectra for the PCQWID, indicating the maximum extinction ratio occurs at
around the peak wavelength of LSP mode.

SPP. The experimental data are almost overlapped over the

calculated peaks, confirming our attributions of the LSP and SPP
modes in the plasmonic micro-cavity. Therefore it is clear that we
have observed the co-existence of LSP and SPP in experiment. The
reason would be attributed to the relatively thick of the cavity
thickness. For our PCQWID rather than plasmonic micro-cavity
absorbers19 such a co-existence is very important for sustaining
quantum efficiency of the detectors, here we call it a two-mode
manipulation, which enables the simultaneous tailoring on the two
modes by adjusting the parameters of the cavity. LSP resonance
yields high extinction ratio, the importance of SPP mode, however,
should not be neglected. Comparing with the ultra-thin absorbers19,
the cavity thickness of our device cannot be reduced to several
nanometers so as to contain the photo-electric conversion active
layers and allow the wave propagation to retain quantum efficiency
of photo-electric conversion. The enhancement factor can be as high
as 1.25 (see SI S4.3). Also notice that the QW absorption is weak

because the increment of QW absorption will need to increase the

doping level in the well and will cause heightening of the dark
current. The assistance of SPP mode helps to keep good quantum
efficiency in photoelectric conversion, making it to be ,6% with only
single QW, similar with that of ,5% in conventional 50 periods
QWIPs.

Discussion
We have demonstrated the high infrared light polarization discriminating capabilities of metal-QW-metal plasmonic micro-cavities.
The photocurrent response spectra directly and effectively disclose
the behavior of the optical waves inside the cavity. Benefited from
double-selection-rules of the plasmonic cavity and the QW, we
observe well-enhanced polarization extinction ratio of about 65 at
LWIR of 14.7 mm. The maximum extinction ratio appears at the
wavelength close to the LSP resonant mode, in good agreement with
the theoretical calculations. Future studies will include tailoring the

Figure 4 | Cavity mode dependence on the grating parameters of PCQWID. Blue spheres are experimental data taken from the peaks in LSP and SPP
mode spectra, as the one shown in the upper panel of Figure 2b. (a) With different periods at Au strip width s55.5 mm. (b) With different Au strip widths
at period p510.0 mm.
SCIENTIFIC REPORTS | 4 : 6332 | DOI: 10.1038/srep06332

www.nature.com/scientificreports
micro-cavity to enable circular polarization discriminating by chiral
structures. It is reasonable to expect that our PCQWID structure can
be employed for FPA device, in which the constraint of cavity resonance would reduce the side-edged light leakage and maintain high
pixel-level extinction ratio, to achieve high quality polarimetric
image in LWIR range.

Method
Fabrication of plasmonic micro-cavity quantum well infrared detector
(PCQWID). The fabrication starts on GaAs/AlGaAs single quantum well (QW)
epitaxial layers substrated on a GaAs wafer. It was defined into 250 3 350 mm2 mesas
by standard photolithography and chemical etching. The mesa depth was controlled
to reach the bottom contact layer. A GeAu(100 nm)/Ni(20 nm)/Au(400 nm) layer
was deposited by electron-beam evaporation and patterned on the mesas as the top
electrode and on the side bottom contact layer as the bottom common electrode. Then
a rapid thermal annealing was processed to form Ohmic contact. Another Ti(20 nm)/
Au(80 nm) layer was deposited and lift-offed to form the surface plasmon grating.
The patterned surface was then bonded on a GaAs polishing wafer. Following that, a
series of mechanism polishing and selective wet etching using the mixture of citric
acid (C6H8O7) and hydrogen peroxide (H2O2) were performed to remove the
600 mm thick GaAs substrate and reaching the Al0.55Ga0.45As etch stop layer. Then
a reflection Ti(20 nm)/Au(80 nm) layer was deposited as the bottom reflection layer
of the cavity. This bottom side was fixed on a sapphire wafer with UV curing adhesive.
The preparation finished with a de-bonding process from the GaAs polishing wafer.
The flowchart for preparing the PCQWID is shown in Figure S1.

Sep 2013. http://www.nobelprize.org/nobel_prizes/physics/laureates/2012/

haroche-lecture.html.
13. Yang, X. et al. Experimental realization of three-dimensional indefinite cavities at
the nanoscale with anomalous scaling laws. Nature Photon. 6, 450454 (2012).
14. Levine, B. F. Quantum-well infrared photodetectors. J. Appl. Phys 74, R1R81
(1993).
15. Todorov, Y. et al. Strong light-matter coupling in subwavelength metal-dielectric
microcavities at terahertz frequencies. Phys. Rev. Lett. 102, 186402 (2009).
16. Law, S. et al. Towards nano-scale photonics with micro-scale photons: the
opportunities and challenges of mid-infrared plasmonics. Nanophotonics 2,
103130 (2013).
17. Stegeman, G. I. et al. Excitation of surface polaritons by end-fire coupling. Opt.
Lett. 8, 386388 (1983).
18. Maier, S. A. Plasmonics: Fundamentals and Applications (Springer
Science1Business Media LLC, New York, 2007).
19. Alaee, R. et al. Deep-Subwavelength Plasmonic Nanoresonators Exploiting
Extreme Coupling. Nano Lett. 13, 34823486 (2013).

Acknowledgments
We gratefully acknowledge the financial support from National Basic Research Program of
China under Grant Nos. 2011CB925604 and 2013CB632705 and NNSFC Nos. 61376053,
10990103, 91021015 and 91121009. We acknowledge the contribution in device process
from Dr. WangPing Wang.

Author contributions
1. Sarkar, M. et al. Biologically Inspired CMOS Image Sensor for Fast Motion and
Polarization Detection. IEEE Sensors J. 13, 10651073 (2013).
2. Novikova, T. et al. Polarimetric imaging for cancer diagnosis and staging. Opt.
Photon. News 23, 2633 (2012).
3. Tyo, J. S. et al. Review of passive imaging polarimetry for remote sensing
applications. Appl. Optics, 45, 54535469 (2006).
4. Schott, J. R. Fundamentals of polarimetric remote sensing (SPIE Press, 2009).
5. Ramkhalawon, R. D. et al. Imaging the polarization of a light field. Opt. Express 21,
41064115 (2013).
6. Gruev, V. et al. CCD polarization imaging sensor with aluminum nanowire
optical filters. Opt. Exp. 18, 1908719094 (2010).
7. Nedelcu, A. et al. Multispectral and polarimetric imaging in the LWIR:
Intersubband detectors as a versatile solution. Infrared Phys. Techn. 59, 125132
(2013).
8. Antoni, T. et al. High contrast polarization sensitive quantum well infrared
photodetectors. Appl. Phys. Lett. 90, 201107 (2007).
9. Genet, C. et al. Light in tiny holes. Nature 445, 3946 (2007).
10. Chang, C. C. et al. A Surface Plasmon Enhanced Infrared Photodetector Based on
InAs Quantum Dots. Nano Lett. 10, 17041709 (2010).
11. Berry, C. W. et al. Significant performance enhancement in photoconductive
terahertz optoelectronics by incorporating plasmonic contact electrodes. Nature
Commun. 4, 1622 (2013).
12. Haroche, S. Nobel Lecture: Controlling Photons in a Box and Exploring the
Quantum to Classical Boundary. Nobelprize.org. Nobel Media AB 2013. Web. 4

SCIENTIFIC REPORTS | 4 : 6332 | DOI: 10.1038/srep06332

Q.L. performed FDTD calculation and designed and fabricated the plasmonic cavity
quantum well infrared detector prototypes and conducted all the measurements. Z.F.L. and
N.L. supervised the fabrications and measurements. X.S.C. assisted in the design and
numerical modelling of PCQWID. P.P.C. grew the quantum well epitaxy layers. W.L.
proposed the physical idea and organized the research. X.C.S. supervised the project. All
authors discussed the results and commented on the manuscript.

Additional information
Supplementary information accompanies this paper at http://www.nature.com/
scientificreports
Competing financial interests: The authors declare no competing financial interests.
How to cite this article: Li, Q. et al. High-Polarization-Discriminating Infrared Detection
Using a Single Quantum Well Sandwiched in Plasmonic Micro-Cavity. Sci. Rep. 4, 6332;
DOI:10.1038/srep06332 (2014).
This work is licensed under a Creative Commons Attribution 4.0 International
License. The images or other third party material in this article are included in the
articles Creative Commons license, unless indicated otherwise in the credit line; if
the material is not included under the Creative Commons license, users will need
to obtain permission from the license holder in order to reproduce the material. To
view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/