pubs.acs.

org/NanoLett

Infrared Perfect Absorber and Its Application

As Plasmonic Sensor
Na Liu, Martin Mesch, Thomas Weiss, Mario Hentschel, and Harald Giessen*
4. Physikalisches Institut, Universität Stuttgart, D-70569 Stuttgart, Germany

ABSTRACT We experimentally demonstrate a perfect plasmonic absorber at λ ) 1.6 µm. Its polarization-independent absorbance
is 99% at normal incidence and remains very high over a wide angular range of incidence around (80°. We introduce a novel concept
to utilize this perfect absorber as plasmonic sensor for refractive index sensing. This sensing strategy offers great potential to maintain
the performance of localized surface plasmon sensors even in nonlaboratory environments due to its simple and robust measurement
scheme.

KEYWORDS Perfect absorbers, LSPR resonances, infrared sensors, glucose sensing, plasmonics

I metamaterial losses.18 There are also theoretical proposals

n plasmonics, the optical properties of metallic nano-
particles are the basis of many fascinating applications
such as chemical and biomedical sensing,1 surface-
enhanced spectroscopy,2 and near-field scanning optical
microscopy.3 The collective excitation of conduction elec-
trons in gold or silver nanoparticles can lead to a localized
surface plasmon resonance (LSPR) when interacting with an
incident light field.4 Such LSPR strongly depends on the size,
shape, and surrounding dielectric environment of nano-
structures.5 In particular, the latter dependence opens a
route toward refractive index sensing in which small con-
centrations of target molecules can be detected.5 So far,
different plasmonic nanostructures have been used to op-
timize LSPR sensors with a large spectral shift for a given
change in refractive index, including nanospheres, nanoshells,
nanorice, nanostars, and etc.1,6-10 More complex plasmonic
sensors based on specific physical mechanisms, for ex-
ample, on the classical analog of electromagnetically induced
transparency11-15 have also been explored.
In practical applications, losses are inevitable in plas-
monic metallic nanostructures. Significant effort has been
paid to achieve low-loss devices, for instance by optimizing
structural geometries16 and by using gain materials.17 Re-
cently, the concept of perfect metamaterial absorbers initi-
ated a new research area in which losses are actually put to
the advantage.18-22 Metamaterials are resonant metallic
nanostructures with unit cells much smaller than the operat-
ing wavelength of light.23-26 It has been shown that perfect
absorption is possible with metamaterials by properly en-
gineering of electric and magnetic responses in the gigahertz
frequency regime.18-21 The basic idea is to minimize the
reflectance through impedance matching and simulta-
neously eliminate the transmittance by maximizing the
* To whom correspondence should be addressed. E-mail: giessen@physik.uni-
stuttgart.de.
Received for review: 12/11/2009
Published on Web: 06/18/2010
of wide-angle perfect absorbers in the gigahertz frequency
regime using microstructures in combination with thick
metallic films.27 Alternatively, perfect absorption can also
be achieved with structured metallic surfaces,28 microcavi-
ties,29 and subwavelength hole arrays.30 In general, it is
desirable to have perfect absorbers which are insensitive on
the incident angle and the polarization of light for practical
applications.
In this letter, we introduce a novel plasmonic device
which combines the concepts of a perfect absorber and an
LSPR sensor. We demonstrate experimentally for the first
time a narrow-band perfect absorber working as plasmonic
sensor in the near-infrared regime. We show that this
plasmonic device yields ∼99% absorbance in the experi-
ment and remains highly absorptive over a wide range of
incident angles for both transverse electric (TE) and trans-
verse magnetic (TM) configurations. We also show that
careful consideration of damping in the metal film is of
particular importance for realizing perfect absorbance in
experiment.
Furthermore, we demonstrate that our plasmonic device
can work as LSPR sensor, where mode volumes in the
attoliter range are common. Different from existing LSPR
sensors which measure the spectral shift of a resonance
upon a refractive index change of the surrounding medium,1
our plasmonic absorber sensor detects rather a relative
intensity change dI(λ)/I(λ) at a fixed wavelength λo induced
by a refractive index change dn. A figure of merit FOM*,
which is introduced by J. Becker et al., is defined as FOM*
) max|[dI(λ)/dn(λ)]/(I(λ)|. λo is chosen where FOM* has a
maximum value.31 We have achieved a FOM* around 87 in
the experiment by measuring the intensity changes with
different local dielectric materials (air and water) at the
sample surface. Our FOM* is nearly four times larger than
that of plasmonic gold nanorod sensors.31 Our absorber
sensor concept offers substantial advantages over classical

© 2010 American Chemical Society 2342 DOI: 10.1021/nl9041033 | Nano Lett. 2010, 10, 2342–2348

FIGURE 1. Schematic of the perfect absorber structure and the
incident light polarization configuration. The diameter and thickness
of the gold disks are 352 and 20 nm, respectively. The periods in
both x and y-directions are 600 nm. The thickness of the MgF2 spacer
is 30 nm and the thickness of the gold mirror is 200 nm. The whole
structure resides on a glass substrate.
sensing methods and is of special interest in numerous
applications due to its low-background detection scheme.
Figure 1 illustrates the geometry of the absorber sensor
structure. It consists of two functional layers. The top layer
is a two-dimensional gold disk array and the bottom layer
is a gold mirror. The two layers are separated by an MgF2
dielectric spacer. Because of the presence of the gold mirror,
the transmittance of the structure is totally eliminated across
the entire near-infrared frequency regime (i.e., T ) 0). The
structure is designed to be polarization independent in x-
and y-directions at normal incidence. To investigate the
resonant behavior of the structure, numerical simulations
were performed by employing a Finite Integration Time
Domain algorithm.32 For excitation of the structure, we use
normally incident light with its polarization along the x-
direction as shown in Figure 1. The permittivity of the MgF2
spacer33 is taken as 1.9, and the permittivity of bulk gold in
the near-infrared34 is described by the Drude model with the
plasma frequency ωpl ) 1.37 × 1016 s-1 and the damping
constant ωc ) 4.08 × 1013 s-1. Owing to the surface
scattering and grain boundary effects in thin films, the
damping constant of the gold film in the real system is likely
higher than that of bulk gold.35 To elucidate the influence of
the damping constant of the gold film on the minimum
reflectance value at resonance, Figure 2a presents the
simulated reflectance spectra for damping constants of one,
three, and five times that of bulk gold. As shown in Figure
2a, reflectance dips with different amplitudes are observ-
able. In particular, a strong resonance with nearly zero
reflectance (R ) 0.28%) is achieved for three times damping
constant of bulk gold (see the red curve in Figure 2a).
Consequently, a perfect absorber is obtained (A ) 1 - T -
R). Nevertheless, for a damping constant equal to one time
that of bulk gold, the reflectance minimum is around 30%
and it leads to only 70% absorbance. Notably, this illustrates
© 2010 American Chemical Society
that different from many plasmonic applications where
losses deteriorate the performance of potential devices for
the realization of perfect absorbers the substantial losses in
the metal are desirable and can be exploited as well. In the
case of five times damping constant of bulk gold, the
reflectance minimum slightly deviates from perfect absor-
bance. The losses in gold with a damping constant equal to
three times that of bulk gold are sufficient to yield a strong
narrow-band resonance. In fact, the structural parameters
described in Figure 1 were particularly optimized for the gold
film (i.e., perfectly impedance matched18 to the air on top
of the sample), which has a damping constant three times
that of bulk gold because it provides the best match to our
experimental gold film.
To better understand the nature of our perfect absorber,
the current distribution at resonance was simulated and is
depicted in Figure 2b. It is evident that antiparallel currents
are excited in the gold disk and the bottom gold layer.36,37
Actually, this is often called a magnetic resonance because
the circulating currents result in a magnetic moment which
can strongly interact with the magnetic field of the incident
light.36,37 At resonance, a strong enhancement of the local-
ized electromagnetic field is established between the two
layers. Consequently, electromagnetic energy can be ef-
ficiently confined in the intermediate MgF2 spacer and
therefore no light is reflected back. This gives rise to a
pronounced reflectance dip in the spectrum with nearly zero
intensity, therefore leading to ∼100% absorbance. In fact,
our device can work as a perfect absorber over a wide range
of incident angles. Figure 2c shows the angular dispersions
of the absorbance peak at various angles of incidence for
both TE and TM configurations. For the TM polarization, the
absorbance peak is nearly independent of the incident angle
and it is 96% even at 80°. This is because the direction of
the magnetic field of the incident light remains unchanged
with various incident angles and it can efficiently drive the
circulating currents at all angles of incidence. Conversely,
for the TE polarization, the magnetic field cannot drive the
circulating currents efficiently at large angles.19 Neverthe-
less, the absorbance still remains at 50% at 80°.
The designed structure is compatible with nanofabrica-
tion techniques. The thick gold film and the MgF2 dielectric
layer were subsequently deposited onto a glass substrate
using electron-gun evaporation. The disk array was defined
in positive resist (polymethylmethacrylate, PMMA) using
standard electron beam lithography followed by a gold lift-
off procedure. The sample has a structure area of 100 µm
× 100 µm. The top left picture in Figure 3a shows the normal
view of the sample obtained by a scanning electron micros-
copy. To demonstrate the feasibility of our absorber working
as a LSPR sensor, Figure 3a shows the results of a proof-of-
principle experiment, displaying the measured reflectance
spectra with air (n ) 1) and water (n ) 1.312),38 on the
sample surface. The measurements were performed by a
Fourier-transform infrared spectrometer with electric field
2343 DOI: 10.1021/nl9041033 | Nano Lett. 2010, 10, 2342-–2348
FIGURE 2. (a) Simulated reflectance spectra in dependence on the damping constant of the gold film. Reflectance with zero intensity is achieved
using a damping constant that is equal to three times that of bulk gold. (b) Calculated current distribution at resonance where perfect absorbance
occurs. Antiparallel currents are excited in the gold disk and the gold film. (c) The simulated angular dispersions of the absorbance peak for
TE and TM configurations.

polarization as illustrated in Figure 1 at normal incidence.
The sensing principle relies on the fact that zero reflectance
(i.e., perfect impedance matching) occurs only for a certain
refractive index of the surrounding medium. The variation
of the refractive index of the surrounding medium gives rise
to nonzero reflectance (i.e., nonperfect absorbance) and
therefore allows for the extremely sensitive detection of the
intensity change in reflectance at a fixed frequency. As
characterized by the black curve in Figure 3a, the experi-
mental reflectance reaches a minimum of 1% at 185.6 THz
(1.6 µm) in air, which corresponds to an experimental
absorbance of 99%. When water (red curve) is applied onto
the sample surface, a clear increase of the reflectance
intensity from 1% to 28.7% at 185.6 THz is visible, resulting
from the refractive index change of the local dielectric
environment. The correspondingly simulated reflectance

© 2010 American Chemical Society 2344 DOI: 10.1021/nl9041033 | Nano Lett. 2010, 10, 2342-–2348
FIGURE 3. (a) Top: Experimental demonstration of perfect absorbance in air. Experimental tuning of the reflectance and absorbance spectra by changing
the dielectric environment which is adjacent to the gold disks from air to water is shown. The SEM image of the sample is presented in the left column.
Bottom: The correspondingly simulated reflectance and absorbance spectra with different dielectric materials (air and water) on the structure surface. In
simulations, a damping constant equal to three times that of bulk gold was utilized. The agreement between experimental and simulated results is nearly
perfect. (b) Experimental FOM* as a function of frequency. The highest value of FOM* is 87, which is reached at a frequency slightly away from the minimum
reflectance. In the discrete calculation of the experimental FOM*, the reflectance difference between water (Iwater (λ)) and air (Iair (λ)) was taken for the
derivative in eq (1). Iair (λ) was taken for the denominator I (λ) in eq 1. The experimental reflectance spectrum with air on the structure surface is replotted
with a black curve.
© 2010 American Chemical Society 2345 DOI: 10.1021/nl9041033 | Nano Lett. 2010, 10, 2342-–2348
and absorbance spectra characterized by dashed lines are
also shown in Figure 3a. In simulations, a damping constant
equal to three times that of bulk gold for the gold film was
used. The agreement between experimental and simulated
results is nearly perfect.
Subsequently, the FOM* can be calculated according to

FOM* ) max | dI(λ)/dn(λ)

I(λ) | (1)

where dI(λ)/I(λ) is the relative intensity change at a fixed

wavelength induced by a refractive index change dn.31 I (λο)
corresponds to the intensity where FOM* reaches a maxi-
mum value. Figure 3b presents the experimental FOM* for
water (with a dn of 0.312) as a function of frequency. The
maximum value of FOM* is around 87 and it occurs at the
position where the slope of dI(λ)/I(λ) is highest right next to
the perfect absorption. Compared to the experimental FOM*
of plasmonic gold nanorods31 which is around 24, our FOM*
is much higher. The fundamental advantage of our absorber
sensor lies in the fact that it allows detection of photons that
are reflected by the nonperfect absorber upon refractive
index change versus a nearly dark reference measurement
where only few photons are reflected from the perfect
absorber. For optimum operation, specific care has to be
given to reduce the background intensity due to scattering
as well as due to detector noise. Our experimentally deter-
mined sensitivity in terms of wavelength shift per refractive
index unit is around 400 nm/RIU, which is compatible with
excellent current LSPR sensor data.15 It is also noteworthy
that our sensor design is highly scalable and can be tuned
to the visible or gigahertz frequency regimes by changing
the sizes appropriately. In addition, for specific applications
perfect absorbance can also be designed for surrounding
dielectrics other than air, such as water or glucose solution
with a certain refractive index. As an example, Figure 4a
presents the simulated spectral tuning of an absorber sensor
which exhibits perfect absorbance for water as the sur-
rounding medium. The reflectance spectrum, which is char- FIGURE 4. (a) Simulated reflectance spectra of an absorber sensor
acterized by the black curve in Figure 4a, reaches an designed for water as reference medium. Spectral tuning occurs with
extremely low intensity at 0.04% at 175.5 THz in water. For glucose solutions varying from 0 to 25% that have different refractive
indices. The resonance position of the perfect absorber in an aqueous
clarity, the resonance position is highlighted by a pink environment at 175.5 THz is highlighted by a dashed pink line. The
dashed line as shown in Figure 4a. It is evident that at 175.5 diameter and thickness of the gold disks are 330 and 19 nm, respec-
THz the reflectance increases with increasing the refractive tively. The periods in both x- and y-directions are 600 nm. The
thickness of the MgF2 spacer is 24 nm and the thickness of the gold
index of the glucose solution, which is applied on the mirror is 200 nm. The whole structure resides on a glass substrate. In
structure surface. The perfect absorber can therefore be a simulations, a damping constant equal to three times that of bulk gold
highly promising device for detecting refractive index changes was utilized. (b) Calculated FOM* as a function of frequency: In the
calculation of the numerical FOM*, we used a finer differential quotient
of a sensing agent. FOM* calculations based on eq 1 by with n ) 1.322 and 1.302. The average intensity value of the cases n )
extracting the data from Figure 4a show that the highest 1.322 and 1.302 was then taken for the denominator I (λ) in eq 1. The
value of FOM* is achieved not exactly at the frequency of maximum value is 94, and it is reached slightly away from the
reflectance minimum. The simulated reflectance spectrum with water
perfect absorption, but rather slightly detuned (see Figure on the structure surface is replotted with a black curve.
4b). Such detuning is also demonstrated by J. Becker et al.
using their nanorod plasmonic sensor. Subsequently, the ing drastically low reflectance near resonance is the key
careful optimization of the structural parameters for achiev- issue to obtain a very high FOM*.

© 2010 American Chemical Society 2346 DOI: 10.1021/nl9041033 | Nano Lett. 2010, 10, 2342-–2348
 In essence, our sensor scheme suggests potential refrac-
tive index sensor platforms in which LSPR sensing is based
on straightforward reflectance measurements simply using
a single-wavelength light source and the exact incidence
angle does not matter as much. Such advantage will benefit
the development of simple and cheap sensors for chemical
sensing and biomedical diagnostics. So far, the best plas-
monic sensors have been mostly synthesized by chemical
methods1 because gold or silver nanostructures obtained
from chemical synthesis can be single crystalline. The
introduction of the perfect absorber sensors discards this
restriction since the intrinsic losses in the metal are essential
to achieve perfect absorption. As a result, many nanofabri-
cation technologies such as electron beam lithography, mask
colloidal lithography,10 interference lithography,39 nano-
sphere lithography,40 nanoimprint lithography,41 and focused-
ion beam writing15 can be widely applied to manufacture
ultrasensitive plasmonic sensors.
In conclusion, we have experimentally demonstrated a
plasmonic perfect absorber in the near-infrared, which is
independent of polarization. Specifically, this wide-angle
absorber can work as an LSPR sensor with high FOM* close
to triple digit values. The results agree well with theoretical
predictions. Also, our absorber sensor is very flexible and
can be designed for achieving perfect absorbance in specific
surrounding materials by impedance matching, even when
lossy metals are being employed in the nanostructure. The
combination of basic sensing principles and particular physi-
cal phenomena opens up a new pathway for plasmonic
sensing and it will facilitate a new class of nanoscopic bio/
chemosensors. The presented sensing strategy has general
applicability for a large variety of nanofabrication technolo-
gies and it has profound significance for developing cheap
sensing equipment in which only simple measurement
technologies are needed.
Acknowledgment. We thank Professor M. Dressel for
useful discussions and comments. We acknowledge S. Hein
for his metamaterial visualizations. We thank Patrick Mai
and Andreas Tittl for extensive help. We acknowledge H.
Gräbeldinger and M. Ubl for technical assistance. This work
was financially supported by Deutsche Forschungsgemein-
schaft (SPP1391 and FOR557), by Landesstiftung BW, and
by BMBF (13N9048 and 13N10146).
REFERENCES AND NOTES
(1) Lal, S.; Link, S.; Halas, N. J. Nat. Photonics 2007, 1, 641–648.
(2) Moskovits, M. Rev. Mod. Phys. 1985, 57, 783–826.
(3) Paesler, M. A.; Moyer, P. J. Near-field Optics: TheoryInstrumenta-
tion, and Applications; Wiley: New York, 1996.
(4) Kreibig, U.; Vollmer, M. Optical Properties of Metal Clusters;
Springer: Berlin, 1995.
(5) Stewart, M. E.; Anderton, C. R.; Thompson, L. B.; Maria, J.; Gray,
S. K.; Rogers, J. A.; Nuzzo, R. G. Chem. Rev. 2008, 108, 494–521.
(6) Soennichsen, C.; Alivisatos, A. P. Nano Lett. 2005, 5, 301–304.
(7) Park, T. H.; Mirin, N.; Lassiter, J. B.; Nehl, C. L.; Halas, N.;
Nordlander, P. ACS Nano 2008, 2, 25–32.
© 2010 American Chemical Society
(8) Lassiter, J. B.; Aizpurua, J.; Hernandez, L. I.; Brandl, D. W.;
Romero, L.; Lal, S.; Hafner, J. H.; Nordlander, P.; Halas, N. J. Nano
Lett. 2008, 8, 1212–1218.
(9) Dmitriev, A.; Haegglund, C.; Chen, S.; Fredriksson, H.; Pakizeh,
T.; Kaell, M.; Sutherland, D. S. Nano Lett. 2008, 8, 3893–3898.
(10) Larsson, E. M.; Alegret, J.; Kaell, M.; Sutherland, D. S. Nano Lett.
2007, 7, 1256–1263.
(11) Hao, F.; Nordlander, P.; Sonnerfraud, Y.; Van Dorpe, P.; Maier,
S. A. ACS Nano 2009, 3, 643–652.
(12) Hao, F.; Sonnefraud, Y.; Van Dorpe, P.; Maier, S. A.; Halas, N. J.;
Nordlander, P. Nano Lett. 2008, 8, 3983–3988.
(13) Zhang, S.; Genov, D.; Wang, Y.; Liu, M.; Zhang, X. Phys. Rev. Lett.
2008, 101, No. 047401.
(14) Liu, N.; Langguth, L.; Weiss, T.; Kästel, J.; Fleischhauer, M.; Pfau,
T.; Giessen, H. Nat. Mater. 2009, 8, 758–762.
(15) Liu, N.; Weiss, T.; Mesch, M.; Langguth, L.; Eigenthaler, U.;
Hirschner, M.; Sönnichsen, C.; Giessen, H. Nano Lett. 2009, 10,
1103–1107.
(16) Zhou, J.; Koschny, T.; Soukoulis, C. M. Opt. Express 2008, 16,
11147–11152.
(17) Plum, E.; Fedotov, V. A.; Kuo, P.; Tsai, D. P.; Zheludev, N. I. Opt.
Express 2009, 17, 8548–8551.
(18) Landy, N. I.; Sajuyigbe, S.; Mock, J. J.; Smith, D. R.; Padilla, W. J.
Phys. Rev. Lett. 2008, 100, 207402.
(19) Tao, H.; Bingham, C. M.; Strikwerda, A. C.; Pilon, D.; Shreken-
hamer, D.; Landy, N. I.; Fan, K.; Zhang, X.; Padilla, W. J.; Averitt,
R. D. Phys. Rev. B 2008, 78, 241103.
(20) Avitzour, Y.; Urzhumov, Y. A.; Shvets, G. Phys. Rev. B 2009, 79,
No. 045131. Mosallaei, H.; Sarabandi, K. A one-layer ultra-thin
meta-surface absorber. In Antennas Propag. Soc. Int. Symp., 2005
IEEE; IEEE: Los Alamitos, CA, 2005; Vol. 1B, pp 615–618. Wu,
C.; Avitzour, Y.; Shvets, G. Ultra-thin, wide-angle perfect absorber
for infrared frequencies. In Proc. SPIE, Proceedings of Metamateri-
als: Fundamentals and Applications, San Diego, CA, August 10–14,
2008; Noginov, M. A., Zheludev, N. I., Boardman, A. D., Engheta,
N., Eds.; SPIE: Bellingham WA; 2008; Vol. 7029, p 70290W.
(21) Wang, B. N.; Koschny, T.; Soukoulis, C. M. Phys. Rev. B 2009, 80,
No. 033108.
(22) Teperik, T. V.; Garcia de Abajo, F. J.; Borisov, A. G.; Abdelsalam,
M.; Bartlett, P. N.; Sugawara, Y.; Baumberg, J. J. Nat. Photonics
2008, 2, 299–301.
(23) Smith, D. R.; Pendry, J. B.; Wiltshire, M. C. K. Science 2004, 305,
788–792.
(24) Shalaev, M. Nat. Photonics 2007, 1, 41–48.
(25) Liu, N.; Guo, H. C.; Fu, L. W.; Kaiser, S.; Schweizer, H.; Giessen,
H. Nat. Mater. 2008, 7, 31–37.
(26) Liu, N.; Liu, H.; Zhu, S. N.; Giessen, H. Nat. Photonics 2009, 3,
157–162.
(27) Diem, M.; Koschny, T.; Soukoulis, C. M. Phys. Rev. B 2009, 79,
No. 033101.
(28) Laroche, M.; Carminati, R.; Greffet, J. J. Phys. Rev. Lett. 2006, 96,
123903.
(29) Celanovic, I.; Perreault, D.; Kassakian, J. Phys. Rev. B 2005, 72,
No. 075127.
(30) Hu, C. G.; Liu, L. Y.; Chen, X. N.; Luo, X. G. Opt. Express 2009,
17, 16745.
(31) Becker, J.; Truegler, A.; Jakab, A.; Hohenester, U.; Soennichsen,
C. Plasmonics 2010, 5, 161.
(32) All the simulations in this paper are performed using the following
software package: CST Microwave Studio; CST GmbH: Germany, 2010.
(33) Dodge, M. J. Appl. Opt. 1984, 23, 1980–1985.
(34) Ordal, M. A.; Long, L. L.; Bell, R. J.; Bell, S. E.; Bell, R. R.; Alexander,
R. W., Jr.; Ward, C. A. Appl. Opt. 1983, 22, 1099–1119.
(35) Zhang, S.; Fan, W.; Malloy, K. J.; Brueck, S. R. J.; Panoiu, N. C.;
Osgood, R. M. J. Opt. Soc. Am. B 2006, 23, 434–438.
(36) Liu, N.; Guo, H. C.; Fu, L. W.; Kaiser, S.; Schweizer, H.; Giessen,
H. Adv. Mater. 2007, 19, 3628–3632. Cai, W. S.; Chettiar, U. K.;
Yuan, H. K.; de Silva, V. C.; Kildishev, A. V.; Drachev, V. P.;
Shalaev, V. M. Opt. Express 2007, 15, 3333–3341.
(37) Previous work on magnetic mirrors has demonstrated the ability
of magnetic resonances in metamaterials for loss enhancement,
see e.g., Fedotov, V. A.; Mladyonov, P. L.; Prosvirnin, S. L.;
Zheludev, N. I. Phys. Rev. E 2005, 72, 056613. Schwanecke, A. S.;
2347 DOI: 10.1021/nl9041033 | Nano Lett. 2010, 10, 2342-–2348
 Fedotov, V. A.; Khardikov, V. V.; Prosvirnin, S. L.; Chen, Y.;
Zheludev, N. I. J. Opt. A: Pure Appl. Opt. 2007, 9, L1–L2.
(38) Fernandez-Prini, R.; Harvey, A. H.,; Palmer, D. A. Aqueous Systems at
Elevated Temperatures and pressures; Elsevier: New York, 2004.
(39) Feth, N.; Enkrich, C.; Wegener, M.; Linden, S. Opt. Express 2007,
15, 501–507.
© 2010 American Chemical Society
(40) Gwinner, M.; Koroknay, E.; Fu, L.; Patoka, P.; Kandulski, W.;
Giersig, M.; Giessen, H. Small 2009, 5, 400–406.
(41) Wu, W.; Kim, E.; Ponizovskaya, E.; Liu, Y. M.; Yu, Z. N.; Fang,
N.; Shen, Y. R.; Bratkovsky, A. M.; Tong, W.; Sun, C.;
Zhang, X.; Wang, S.; Williams, R. S. Appl. Phys. A 2007, 87,
143–150.
2348 DOI: 10.1021/nl9041033 | Nano Lett. 2010, 10, 2342-–2348